Add mac clang 3.6 prebuilts for LLVM rebase to r222494.

Change-Id: I21acdfbba4dcd1782b215635dcc4c2b09bf5e3a0

65 files changed, 96998 insertions, 0 deletions

diff --git a/bin/clang b/bin/clang
new file mode 100755
index 0000000..c1565e4
--- /dev/null
+++ b/bin/clang
diff --git a/bin/clang++ b/bin/clang++
new file mode 120000
index 0000000..060d289
--- /dev/null
+++ b/bin/clang++
@@ -0,0 +1 @@
+clang
\ No newline at end of file
diff --git a/bin/llvm-as b/bin/llvm-as
new file mode 100755
index 0000000..c3c70f5
--- /dev/null
+++ b/bin/llvm-as
diff --git a/bin/llvm-link b/bin/llvm-link
new file mode 100755
index 0000000..85eea9d
--- /dev/null
+++ b/bin/llvm-link
diff --git a/lib/clang/3.6/include/Intrin.h b/lib/clang/3.6/include/Intrin.h
new file mode 100644
index 0000000..016c480
--- /dev/null
+++ b/lib/clang/3.6/include/Intrin.h
@@ -0,0 +1,950 @@
+/* ===-------- Intrin.h ---------------------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+/* Only include this if we're compiling for the windows platform. */
+#ifndef _MSC_VER
+#include_next <Intrin.h>
+#else
+
+#ifndef __INTRIN_H
+#define __INTRIN_H
+
+/* First include the standard intrinsics. */
+#if defined(__i386__) || defined(__x86_64__)
+#include <x86intrin.h>
+#endif
+
+/* For the definition of jmp_buf. */
+#if __STDC_HOSTED__
+#include <setjmp.h>
+#endif
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#if defined(__MMX__)
+/* And the random ones that aren't in those files. */
+__m64 _m_from_float(float);
+__m64 _m_from_int(int _l);
+void _m_prefetch(void *);
+float _m_to_float(__m64);
+int _m_to_int(__m64 _M);
+#endif
+
+/* Other assorted instruction intrinsics. */
+void __addfsbyte(unsigned long, unsigned char);
+void __addfsdword(unsigned long, unsigned long);
+void __addfsword(unsigned long, unsigned short);
+void __code_seg(const char *);
+static __inline__
+void __cpuid(int[4], int);
+static __inline__
+void __cpuidex(int[4], int, int);
+void __debugbreak(void);
+__int64 __emul(int, int);
+unsigned __int64 __emulu(unsigned int, unsigned int);
+void __cdecl __fastfail(unsigned int);
+unsigned int __getcallerseflags(void);
+static __inline__
+void __halt(void);
+unsigned char __inbyte(unsigned short);
+void __inbytestring(unsigned short, unsigned char *, unsigned long);
+void __incfsbyte(unsigned long);
+void __incfsdword(unsigned long);
+void __incfsword(unsigned long);
+unsigned long __indword(unsigned short);
+void __indwordstring(unsigned short, unsigned long *, unsigned long);
+void __int2c(void);
+void __invlpg(void *);
+unsigned short __inword(unsigned short);
+void __inwordstring(unsigned short, unsigned short *, unsigned long);
+void __lidt(void *);
+unsigned __int64 __ll_lshift(unsigned __int64, int);
+__int64 __ll_rshift(__int64, int);
+void __llwpcb(void *);
+unsigned char __lwpins32(unsigned int, unsigned int, unsigned int);
+void __lwpval32(unsigned int, unsigned int, unsigned int);
+unsigned int __lzcnt(unsigned int);
+unsigned short __lzcnt16(unsigned short);
+static __inline__
+void __movsb(unsigned char *, unsigned char const *, size_t);
+static __inline__
+void __movsd(unsigned long *, unsigned long const *, size_t);
+static __inline__
+void __movsw(unsigned short *, unsigned short const *, size_t);
+void __nop(void);
+void __nvreg_restore_fence(void);
+void __nvreg_save_fence(void);
+void __outbyte(unsigned short, unsigned char);
+void __outbytestring(unsigned short, unsigned char *, unsigned long);
+void __outdword(unsigned short, unsigned long);
+void __outdwordstring(unsigned short, unsigned long *, unsigned long);
+void __outword(unsigned short, unsigned short);
+void __outwordstring(unsigned short, unsigned short *, unsigned long);
+static __inline__
+unsigned int __popcnt(unsigned int);
+static __inline__
+unsigned short __popcnt16(unsigned short);
+unsigned long __readcr0(void);
+unsigned long __readcr2(void);
+static __inline__
+unsigned long __readcr3(void);
+unsigned long __readcr4(void);
+unsigned long __readcr8(void);
+unsigned int __readdr(unsigned int);
+#ifdef __i386__
+static __inline__
+unsigned char __readfsbyte(unsigned long);
+static __inline__
+unsigned long __readfsdword(unsigned long);
+static __inline__
+unsigned __int64 __readfsqword(unsigned long);
+static __inline__
+unsigned short __readfsword(unsigned long);
+#endif
+static __inline__
+unsigned __int64 __readmsr(unsigned long);
+unsigned __int64 __readpmc(unsigned long);
+unsigned long __segmentlimit(unsigned long);
+void __sidt(void *);
+void *__slwpcb(void);
+static __inline__
+void __stosb(unsigned char *, unsigned char, size_t);
+static __inline__
+void __stosd(unsigned long *, unsigned long, size_t);
+static __inline__
+void __stosw(unsigned short *, unsigned short, size_t);
+void __svm_clgi(void);
+void __svm_invlpga(void *, int);
+void __svm_skinit(int);
+void __svm_stgi(void);
+void __svm_vmload(size_t);
+void __svm_vmrun(size_t);
+void __svm_vmsave(size_t);
+void __ud2(void);
+unsigned __int64 __ull_rshift(unsigned __int64, int);
+void __vmx_off(void);
+void __vmx_vmptrst(unsigned __int64 *);
+void __wbinvd(void);
+void __writecr0(unsigned int);
+static __inline__
+void __writecr3(unsigned int);
+void __writecr4(unsigned int);
+void __writecr8(unsigned int);
+void __writedr(unsigned int, unsigned int);
+void __writefsbyte(unsigned long, unsigned char);
+void __writefsdword(unsigned long, unsigned long);
+void __writefsqword(unsigned long, unsigned __int64);
+void __writefsword(unsigned long, unsigned short);
+void __writemsr(unsigned long, unsigned __int64);
+static __inline__
+void *_AddressOfReturnAddress(void);
+static __inline__
+unsigned char _BitScanForward(unsigned long *_Index, unsigned long _Mask);
+static __inline__
+unsigned char _BitScanReverse(unsigned long *_Index, unsigned long _Mask);
+static __inline__
+unsigned char _bittest(long const *, long);
+static __inline__
+unsigned char _bittestandcomplement(long *, long);
+static __inline__
+unsigned char _bittestandreset(long *, long);
+static __inline__
+unsigned char _bittestandset(long *, long);
+unsigned __int64 __cdecl _byteswap_uint64(unsigned __int64);
+unsigned long __cdecl _byteswap_ulong(unsigned long);
+unsigned short __cdecl _byteswap_ushort(unsigned short);
+void __cdecl _disable(void);
+void __cdecl _enable(void);
+void __cdecl _fxrstor(void const *);
+void __cdecl _fxsave(void *);
+long _InterlockedAddLargeStatistic(__int64 volatile *_Addend, long _Value);
+static __inline__
+long _InterlockedAnd(long volatile *_Value, long _Mask);
+static __inline__
+short _InterlockedAnd16(short volatile *_Value, short _Mask);
+static __inline__
+char _InterlockedAnd8(char volatile *_Value, char _Mask);
+unsigned char _interlockedbittestandreset(long volatile *, long);
+static __inline__
+unsigned char _interlockedbittestandset(long volatile *, long);
+static __inline__
+long __cdecl _InterlockedCompareExchange(long volatile *_Destination,
+                                         long _Exchange, long _Comparand);
+long _InterlockedCompareExchange_HLEAcquire(long volatile *, long, long);
+long _InterlockedCompareExchange_HLERelease(long volatile *, long, long);
+static __inline__
+short _InterlockedCompareExchange16(short volatile *_Destination,
+                                    short _Exchange, short _Comparand);
+static __inline__
+__int64 _InterlockedCompareExchange64(__int64 volatile *_Destination,
+                                      __int64 _Exchange, __int64 _Comparand);
+__int64 _InterlockedcompareExchange64_HLEAcquire(__int64 volatile *, __int64,
+                                                 __int64);
+__int64 _InterlockedCompareExchange64_HLERelease(__int64 volatile *, __int64,
+                                                 __int64);
+static __inline__
+char _InterlockedCompareExchange8(char volatile *_Destination, char _Exchange,
+                                  char _Comparand);
+void *_InterlockedCompareExchangePointer_HLEAcquire(void *volatile *, void *,
+                                                    void *);
+void *_InterlockedCompareExchangePointer_HLERelease(void *volatile *, void *,
+                                                    void *);
+static __inline__
+long __cdecl _InterlockedDecrement(long volatile *_Addend);
+static __inline__
+short _InterlockedDecrement16(short volatile *_Addend);
+long _InterlockedExchange(long volatile *_Target, long _Value);
+static __inline__
+short _InterlockedExchange16(short volatile *_Target, short _Value);
+static __inline__
+char _InterlockedExchange8(char volatile *_Target, char _Value);
+static __inline__
+long __cdecl _InterlockedExchangeAdd(long volatile *_Addend, long _Value);
+long _InterlockedExchangeAdd_HLEAcquire(long volatile *, long);
+long _InterlockedExchangeAdd_HLERelease(long volatile *, long);
+static __inline__
+short _InterlockedExchangeAdd16(short volatile *_Addend, short _Value);
+__int64 _InterlockedExchangeAdd64_HLEAcquire(__int64 volatile *, __int64);
+__int64 _InterlockedExchangeAdd64_HLERelease(__int64 volatile *, __int64);
+static __inline__
+char _InterlockedExchangeAdd8(char volatile *_Addend, char _Value);
+static __inline__
+long __cdecl _InterlockedIncrement(long volatile *_Addend);
+static __inline__
+short _InterlockedIncrement16(short volatile *_Addend);
+static __inline__
+long _InterlockedOr(long volatile *_Value, long _Mask);
+static __inline__
+short _InterlockedOr16(short volatile *_Value, short _Mask);
+static __inline__
+char _InterlockedOr8(char volatile *_Value, char _Mask);
+static __inline__
+long _InterlockedXor(long volatile *_Value, long _Mask);
+static __inline__
+short _InterlockedXor16(short volatile *_Value, short _Mask);
+static __inline__
+char _InterlockedXor8(char volatile *_Value, char _Mask);
+void __cdecl _invpcid(unsigned int, void *);
+static __inline__
+unsigned long __cdecl _lrotl(unsigned long, int);
+static __inline__
+unsigned long __cdecl _lrotr(unsigned long, int);
+static __inline__
+static __inline__
+void _ReadBarrier(void);
+static __inline__
+void _ReadWriteBarrier(void);
+static __inline__
+void *_ReturnAddress(void);
+unsigned int _rorx_u32(unsigned int, const unsigned int);
+static __inline__
+unsigned int __cdecl _rotl(unsigned int _Value, int _Shift);
+static __inline__
+unsigned short _rotl16(unsigned short _Value, unsigned char _Shift);
+static __inline__
+unsigned __int64 __cdecl _rotl64(unsigned __int64 _Value, int _Shift);
+static __inline__
+unsigned char _rotl8(unsigned char _Value, unsigned char _Shift);
+static __inline__
+unsigned int __cdecl _rotr(unsigned int _Value, int _Shift);
+static __inline__
+unsigned short _rotr16(unsigned short _Value, unsigned char _Shift);
+static __inline__
+unsigned __int64 __cdecl _rotr64(unsigned __int64 _Value, int _Shift);
+static __inline__
+unsigned char _rotr8(unsigned char _Value, unsigned char _Shift);
+int _sarx_i32(int, unsigned int);
+#if __STDC_HOSTED__
+int __cdecl _setjmp(jmp_buf);
+#endif
+unsigned int _shlx_u32(unsigned int, unsigned int);
+unsigned int _shrx_u32(unsigned int, unsigned int);
+void _Store_HLERelease(long volatile *, long);
+void _Store64_HLERelease(__int64 volatile *, __int64);
+void _StorePointer_HLERelease(void *volatile *, void *);
+static __inline__
+void _WriteBarrier(void);
+unsigned __int32 xbegin(void);
+void _xend(void);
+static __inline__
+unsigned __int64 __cdecl _xgetbv(unsigned int);
+void __cdecl _xrstor(void const *, unsigned __int64);
+void __cdecl _xsave(void *, unsigned __int64);
+void __cdecl _xsaveopt(void *, unsigned __int64);
+void __cdecl _xsetbv(unsigned int, unsigned __int64);
+
+/* These additional intrinsics are turned on in x64/amd64/x86_64 mode. */
+#ifdef __x86_64__
+void __addgsbyte(unsigned long, unsigned char);
+void __addgsdword(unsigned long, unsigned long);
+void __addgsqword(unsigned long, unsigned __int64);
+void __addgsword(unsigned long, unsigned short);
+static __inline__
+void __faststorefence(void);
+void __incgsbyte(unsigned long);
+void __incgsdword(unsigned long);
+void __incgsqword(unsigned long);
+void __incgsword(unsigned long);
+unsigned char __lwpins64(unsigned __int64, unsigned int, unsigned int);
+void __lwpval64(unsigned __int64, unsigned int, unsigned int);
+unsigned __int64 __lzcnt64(unsigned __int64);
+static __inline__
+void __movsq(unsigned long long *, unsigned long long const *, size_t);
+__int64 __mulh(__int64, __int64);
+static __inline__
+unsigned __int64 __popcnt64(unsigned __int64);
+static __inline__
+unsigned char __readgsbyte(unsigned long);
+static __inline__
+unsigned long __readgsdword(unsigned long);
+static __inline__
+unsigned __int64 __readgsqword(unsigned long);
+unsigned short __readgsword(unsigned long);
+unsigned __int64 __shiftleft128(unsigned __int64 _LowPart,
+                                unsigned __int64 _HighPart,
+                                unsigned char _Shift);
+unsigned __int64 __shiftright128(unsigned __int64 _LowPart,
+                                 unsigned __int64 _HighPart,
+                                 unsigned char _Shift);
+static __inline__
+void __stosq(unsigned __int64 *, unsigned __int64, size_t);
+unsigned __int64 __umulh(unsigned __int64, unsigned __int64);
+unsigned char __vmx_on(unsigned __int64 *);
+unsigned char __vmx_vmclear(unsigned __int64 *);
+unsigned char __vmx_vmlaunch(void);
+unsigned char __vmx_vmptrld(unsigned __int64 *);
+unsigned char __vmx_vmread(size_t, size_t *);
+unsigned char __vmx_vmresume(void);
+unsigned char __vmx_vmwrite(size_t, size_t);
+void __writegsbyte(unsigned long, unsigned char);
+void __writegsdword(unsigned long, unsigned long);
+void __writegsqword(unsigned long, unsigned __int64);
+void __writegsword(unsigned long, unsigned short);
+static __inline__
+unsigned char _BitScanForward64(unsigned long *_Index, unsigned __int64 _Mask);
+static __inline__
+unsigned char _BitScanReverse64(unsigned long *_Index, unsigned __int64 _Mask);
+static __inline__
+unsigned char _bittest64(__int64 const *, __int64);
+static __inline__
+unsigned char _bittestandcomplement64(__int64 *, __int64);
+static __inline__
+unsigned char _bittestandreset64(__int64 *, __int64);
+static __inline__
+unsigned char _bittestandset64(__int64 *, __int64);
+unsigned __int64 __cdecl _byteswap_uint64(unsigned __int64);
+void __cdecl _fxrstor64(void const *);
+void __cdecl _fxsave64(void *);
+long _InterlockedAnd_np(long volatile *_Value, long _Mask);
+short _InterlockedAnd16_np(short volatile *_Value, short _Mask);
+__int64 _InterlockedAnd64_np(__int64 volatile *_Value, __int64 _Mask);
+char _InterlockedAnd8_np(char volatile *_Value, char _Mask);
+unsigned char _interlockedbittestandreset64(__int64 volatile *, __int64);
+static __inline__
+unsigned char _interlockedbittestandset64(__int64 volatile *, __int64);
+long _InterlockedCompareExchange_np(long volatile *_Destination, long _Exchange,
+                                    long _Comparand);
+unsigned char _InterlockedCompareExchange128(__int64 volatile *_Destination,
+                                             __int64 _ExchangeHigh,
+                                             __int64 _ExchangeLow,
+                                             __int64 *_CompareandResult);
+unsigned char _InterlockedCompareExchange128_np(__int64 volatile *_Destination,
+                                                __int64 _ExchangeHigh,
+                                                __int64 _ExchangeLow,
+                                                __int64 *_ComparandResult);
+short _InterlockedCompareExchange16_np(short volatile *_Destination,
+                                       short _Exchange, short _Comparand);
+__int64 _InterlockedCompareExchange64_HLEAcquire(__int64 volatile *, __int64,
+                                                 __int64);
+__int64 _InterlockedCompareExchange64_HLERelease(__int64 volatile *, __int64,
+                                                 __int64);
+__int64 _InterlockedCompareExchange64_np(__int64 volatile *_Destination,
+                                         __int64 _Exchange, __int64 _Comparand);
+void *_InterlockedCompareExchangePointer(void *volatile *_Destination,
+                                         void *_Exchange, void *_Comparand);
+void *_InterlockedCompareExchangePointer_np(void *volatile *_Destination,
+                                            void *_Exchange, void *_Comparand);
+static __inline__
+__int64 _InterlockedDecrement64(__int64 volatile *_Addend);
+static __inline__
+__int64 _InterlockedExchange64(__int64 volatile *_Target, __int64 _Value);
+static __inline__
+__int64 _InterlockedExchangeAdd64(__int64 volatile *_Addend, __int64 _Value);
+void *_InterlockedExchangePointer(void *volatile *_Target, void *_Value);
+static __inline__
+__int64 _InterlockedIncrement64(__int64 volatile *_Addend);
+long _InterlockedOr_np(long volatile *_Value, long _Mask);
+short _InterlockedOr16_np(short volatile *_Value, short _Mask);
+static __inline__
+__int64 _InterlockedOr64(__int64 volatile *_Value, __int64 _Mask);
+__int64 _InterlockedOr64_np(__int64 volatile *_Value, __int64 _Mask);
+char _InterlockedOr8_np(char volatile *_Value, char _Mask);
+long _InterlockedXor_np(long volatile *_Value, long _Mask);
+short _InterlockedXor16_np(short volatile *_Value, short _Mask);
+static __inline__
+__int64 _InterlockedXor64(__int64 volatile *_Value, __int64 _Mask);
+__int64 _InterlockedXor64_np(__int64 volatile *_Value, __int64 _Mask);
+char _InterlockedXor8_np(char volatile *_Value, char _Mask);
+static __inline__
+__int64 _mul128(__int64 _Multiplier, __int64 _Multiplicand,
+                __int64 *_HighProduct);
+unsigned __int64 _rorx_u64(unsigned __int64, const unsigned int);
+__int64 _sarx_i64(__int64, unsigned int);
+#if __STDC_HOSTED__
+int __cdecl _setjmpex(jmp_buf);
+#endif
+unsigned __int64 _shlx_u64(unsigned __int64, unsigned int);
+unsigned __int64 shrx_u64(unsigned __int64, unsigned int);
+unsigned __int64 _umul128(unsigned __int64 _Multiplier,
+                          unsigned __int64 _Multiplicand,
+                          unsigned __int64 *_HighProduct);
+void __cdecl _xrstor64(void const *, unsigned __int64);
+void __cdecl _xsave64(void *, unsigned __int64);
+void __cdecl _xsaveopt64(void *, unsigned __int64);
+
+#endif /* __x86_64__ */
+
+/*----------------------------------------------------------------------------*\
+|* Bit Twiddling
+\*----------------------------------------------------------------------------*/
+static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
+_rotl8(unsigned char _Value, unsigned char _Shift) {
+  _Shift &= 0x7;
+  return _Shift ? (_Value << _Shift) | (_Value >> (8 - _Shift)) : _Value;
+}
+static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
+_rotr8(unsigned char _Value, unsigned char _Shift) {
+  _Shift &= 0x7;
+  return _Shift ? (_Value >> _Shift) | (_Value << (8 - _Shift)) : _Value;
+}
+static __inline__ unsigned short __attribute__((__always_inline__, __nodebug__))
+_rotl16(unsigned short _Value, unsigned char _Shift) {
+  _Shift &= 0xf;
+  return _Shift ? (_Value << _Shift) | (_Value >> (16 - _Shift)) : _Value;
+}
+static __inline__ unsigned short __attribute__((__always_inline__, __nodebug__))
+_rotr16(unsigned short _Value, unsigned char _Shift) {
+  _Shift &= 0xf;
+  return _Shift ? (_Value >> _Shift) | (_Value << (16 - _Shift)) : _Value;
+}
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_rotl(unsigned int _Value, int _Shift) {
+  _Shift &= 0x1f;
+  return _Shift ? (_Value << _Shift) | (_Value >> (32 - _Shift)) : _Value;
+}
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_rotr(unsigned int _Value, int _Shift) {
+  _Shift &= 0x1f;
+  return _Shift ? (_Value >> _Shift) | (_Value << (32 - _Shift)) : _Value;
+}
+static __inline__ unsigned long __attribute__((__always_inline__, __nodebug__))
+_lrotl(unsigned long _Value, int _Shift) {
+  _Shift &= 0x1f;
+  return _Shift ? (_Value << _Shift) | (_Value >> (32 - _Shift)) : _Value;
+}
+static __inline__ unsigned long __attribute__((__always_inline__, __nodebug__))
+_lrotr(unsigned long _Value, int _Shift) {
+  _Shift &= 0x1f;
+  return _Shift ? (_Value >> _Shift) | (_Value << (32 - _Shift)) : _Value;
+}
+static
+__inline__ unsigned __int64 __attribute__((__always_inline__, __nodebug__))
+_rotl64(unsigned __int64 _Value, int _Shift) {
+  _Shift &= 0x3f;
+  return _Shift ? (_Value << _Shift) | (_Value >> (64 - _Shift)) : _Value;
+}
+static
+__inline__ unsigned __int64 __attribute__((__always_inline__, __nodebug__))
+_rotr64(unsigned __int64 _Value, int _Shift) {
+  _Shift &= 0x3f;
+  return _Shift ? (_Value >> _Shift) | (_Value << (64 - _Shift)) : _Value;
+}
+/*----------------------------------------------------------------------------*\
+|* Bit Counting and Testing
+\*----------------------------------------------------------------------------*/
+static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
+_BitScanForward(unsigned long *_Index, unsigned long _Mask) {
+  if (!_Mask)
+    return 0;
+  *_Index = __builtin_ctzl(_Mask);
+  return 1;
+}
+static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
+_BitScanReverse(unsigned long *_Index, unsigned long _Mask) {
+  if (!_Mask)
+    return 0;
+  *_Index = 31 - __builtin_clzl(_Mask);
+  return 1;
+}
+static __inline__ unsigned short __attribute__((__always_inline__, __nodebug__))
+__popcnt16(unsigned short value) {
+  return __builtin_popcount((int)value);
+}
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__popcnt(unsigned int value) {
+  return __builtin_popcount(value);
+}
+static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
+_bittest(long const *a, long b) {
+  return (*a >> b) & 1;
+}
+static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
+_bittestandcomplement(long *a, long b) {
+  unsigned char x = (*a >> b) & 1;
+  *a = *a ^ (1 << b);
+  return x;
+}
+static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
+_bittestandreset(long *a, long b) {
+  unsigned char x = (*a >> b) & 1;
+  *a = *a & ~(1 << b);
+  return x;
+}
+static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
+_bittestandset(long *a, long b) {
+  unsigned char x = (*a >> b) & 1;
+  *a = *a | (1 << b);
+  return x;
+}
+#if defined(__i386__) || defined(__x86_64__)
+static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
+_interlockedbittestandset(long volatile *__BitBase, long __BitPos) {
+  unsigned char __Res;
+  __asm__ ("xor %0, %0\n"
+           "lock bts %2, %1\n"
+           "setc %0\n"
+           : "=r" (__Res), "+m"(*__BitBase)
+           : "Ir"(__BitPos));
+  return __Res;
+}
+#endif
+#ifdef __x86_64__
+static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
+_BitScanForward64(unsigned long *_Index, unsigned __int64 _Mask) {
+  if (!_Mask)
+    return 0;
+  *_Index = __builtin_ctzll(_Mask);
+  return 1;
+}
+static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
+_BitScanReverse64(unsigned long *_Index, unsigned __int64 _Mask) {
+  if (!_Mask)
+    return 0;
+  *_Index = 63 - __builtin_clzll(_Mask);
+  return 1;
+}
+static __inline__
+unsigned __int64 __attribute__((__always_inline__, __nodebug__))
+ __popcnt64(unsigned __int64 value) {
+  return __builtin_popcountll(value);
+}
+static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
+_bittest64(__int64 const *a, __int64 b) {
+  return (*a >> b) & 1;
+}
+static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
+_bittestandcomplement64(__int64 *a, __int64 b) {
+  unsigned char x = (*a >> b) & 1;
+  *a = *a ^ (1ll << b);
+  return x;
+}
+static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
+_bittestandreset64(__int64 *a, __int64 b) {
+  unsigned char x = (*a >> b) & 1;
+  *a = *a & ~(1ll << b);
+  return x;
+}
+static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
+_bittestandset64(__int64 *a, __int64 b) {
+  unsigned char x = (*a >> b) & 1;
+  *a = *a | (1ll << b);
+  return x;
+}
+static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
+_interlockedbittestandset64(__int64 volatile *__BitBase, __int64 __BitPos) {
+  unsigned char __Res;
+  __asm__ ("xor %0, %0\n"
+           "lock bts %2, %1\n"
+           "setc %0\n"
+           : "=r" (__Res), "+m"(*__BitBase)
+           : "Ir"(__BitPos));
+  return __Res;
+}
+#endif
+/*----------------------------------------------------------------------------*\
+|* Interlocked Exchange Add
+\*----------------------------------------------------------------------------*/
+static __inline__ char __attribute__((__always_inline__, __nodebug__))
+_InterlockedExchangeAdd8(char volatile *_Addend, char _Value) {
+  return __atomic_add_fetch(_Addend, _Value, 0) - _Value;
+}
+static __inline__ short __attribute__((__always_inline__, __nodebug__))
+_InterlockedExchangeAdd16(short volatile *_Addend, short _Value) {
+  return __atomic_add_fetch(_Addend, _Value, 0) - _Value;
+}
+#ifdef __x86_64__
+static __inline__ __int64 __attribute__((__always_inline__, __nodebug__))
+_InterlockedExchangeAdd64(__int64 volatile *_Addend, __int64 _Value) {
+  return __atomic_add_fetch(_Addend, _Value, 0) - _Value;
+}
+#endif
+/*----------------------------------------------------------------------------*\
+|* Interlocked Exchange Sub
+\*----------------------------------------------------------------------------*/
+static __inline__ char __attribute__((__always_inline__, __nodebug__))
+_InterlockedExchangeSub8(char volatile *_Subend, char _Value) {
+  return __atomic_sub_fetch(_Subend, _Value, 0) + _Value;
+}
+static __inline__ short __attribute__((__always_inline__, __nodebug__))
+_InterlockedExchangeSub16(short volatile *_Subend, short _Value) {
+  return __atomic_sub_fetch(_Subend, _Value, 0) + _Value;
+}
+static __inline__ long __attribute__((__always_inline__, __nodebug__))
+_InterlockedExchangeSub(long volatile *_Subend, long _Value) {
+  return __atomic_sub_fetch(_Subend, _Value, 0) + _Value;
+}
+#ifdef __x86_64__
+static __inline__ __int64 __attribute__((__always_inline__, __nodebug__))
+_InterlockedExchangeSub64(__int64 volatile *_Subend, __int64 _Value) {
+  return __atomic_sub_fetch(_Subend, _Value, 0) + _Value;
+}
+#endif
+/*----------------------------------------------------------------------------*\
+|* Interlocked Increment
+\*----------------------------------------------------------------------------*/
+static __inline__ short __attribute__((__always_inline__, __nodebug__))
+_InterlockedIncrement16(short volatile *_Value) {
+  return __atomic_add_fetch(_Value, 1, 0);
+}
+#ifdef __x86_64__
+static __inline__ __int64 __attribute__((__always_inline__, __nodebug__))
+_InterlockedIncrement64(__int64 volatile *_Value) {
+  return __atomic_add_fetch(_Value, 1, 0);
+}
+#endif
+/*----------------------------------------------------------------------------*\
+|* Interlocked Decrement
+\*----------------------------------------------------------------------------*/
+static __inline__ short __attribute__((__always_inline__, __nodebug__))
+_InterlockedDecrement16(short volatile *_Value) {
+  return __atomic_sub_fetch(_Value, 1, 0);
+}
+#ifdef __x86_64__
+static __inline__ __int64 __attribute__((__always_inline__, __nodebug__))
+_InterlockedDecrement64(__int64 volatile *_Value) {
+  return __atomic_sub_fetch(_Value, 1, 0);
+}
+#endif
+/*----------------------------------------------------------------------------*\
+|* Interlocked And
+\*----------------------------------------------------------------------------*/
+static __inline__ char __attribute__((__always_inline__, __nodebug__))
+_InterlockedAnd8(char volatile *_Value, char _Mask) {
+  return __atomic_and_fetch(_Value, _Mask, 0);
+}
+static __inline__ short __attribute__((__always_inline__, __nodebug__))
+_InterlockedAnd16(short volatile *_Value, short _Mask) {
+  return __atomic_and_fetch(_Value, _Mask, 0);
+}
+static __inline__ long __attribute__((__always_inline__, __nodebug__))
+_InterlockedAnd(long volatile *_Value, long _Mask) {
+  return __atomic_and_fetch(_Value, _Mask, 0);
+}
+#ifdef __x86_64__
+static __inline__ __int64 __attribute__((__always_inline__, __nodebug__))
+_InterlockedAnd64(__int64 volatile *_Value, __int64 _Mask) {
+  return __atomic_and_fetch(_Value, _Mask, 0);
+}
+#endif
+/*----------------------------------------------------------------------------*\
+|* Interlocked Or
+\*----------------------------------------------------------------------------*/
+static __inline__ char __attribute__((__always_inline__, __nodebug__))
+_InterlockedOr8(char volatile *_Value, char _Mask) {
+  return __atomic_or_fetch(_Value, _Mask, 0);
+}
+static __inline__ short __attribute__((__always_inline__, __nodebug__))
+_InterlockedOr16(short volatile *_Value, short _Mask) {
+  return __atomic_or_fetch(_Value, _Mask, 0);
+}
+static __inline__ long __attribute__((__always_inline__, __nodebug__))
+_InterlockedOr(long volatile *_Value, long _Mask) {
+  return __atomic_or_fetch(_Value, _Mask, 0);
+}
+#ifdef __x86_64__
+static __inline__ __int64 __attribute__((__always_inline__, __nodebug__))
+_InterlockedOr64(__int64 volatile *_Value, __int64 _Mask) {
+  return __atomic_or_fetch(_Value, _Mask, 0);
+}
+#endif
+/*----------------------------------------------------------------------------*\
+|* Interlocked Xor
+\*----------------------------------------------------------------------------*/
+static __inline__ char __attribute__((__always_inline__, __nodebug__))
+_InterlockedXor8(char volatile *_Value, char _Mask) {
+  return __atomic_xor_fetch(_Value, _Mask, 0);
+}
+static __inline__ short __attribute__((__always_inline__, __nodebug__))
+_InterlockedXor16(short volatile *_Value, short _Mask) {
+  return __atomic_xor_fetch(_Value, _Mask, 0);
+}
+static __inline__ long __attribute__((__always_inline__, __nodebug__))
+_InterlockedXor(long volatile *_Value, long _Mask) {
+  return __atomic_xor_fetch(_Value, _Mask, 0);
+}
+#ifdef __x86_64__
+static __inline__ __int64 __attribute__((__always_inline__, __nodebug__))
+_InterlockedXor64(__int64 volatile *_Value, __int64 _Mask) {
+  return __atomic_xor_fetch(_Value, _Mask, 0);
+}
+#endif
+/*----------------------------------------------------------------------------*\
+|* Interlocked Exchange
+\*----------------------------------------------------------------------------*/
+static __inline__ char __attribute__((__always_inline__, __nodebug__))
+_InterlockedExchange8(char volatile *_Target, char _Value) {
+  __atomic_exchange(_Target, &_Value, &_Value, 0);
+  return _Value;
+}
+static __inline__ short __attribute__((__always_inline__, __nodebug__))
+_InterlockedExchange16(short volatile *_Target, short _Value) {
+  __atomic_exchange(_Target, &_Value, &_Value, 0);
+  return _Value;
+}
+#ifdef __x86_64__
+static __inline__ __int64 __attribute__((__always_inline__, __nodebug__))
+_InterlockedExchange64(__int64 volatile *_Target, __int64 _Value) {
+  __atomic_exchange(_Target, &_Value, &_Value, 0);
+  return _Value;
+}
+#endif
+/*----------------------------------------------------------------------------*\
+|* Interlocked Compare Exchange
+\*----------------------------------------------------------------------------*/
+static __inline__ char __attribute__((__always_inline__, __nodebug__))
+_InterlockedCompareExchange8(char volatile *_Destination,
+                             char _Exchange, char _Comparand) {
+  __atomic_compare_exchange(_Destination, &_Comparand, &_Exchange, 0, 0, 0);
+  return _Comparand;
+}
+static __inline__ short __attribute__((__always_inline__, __nodebug__))
+_InterlockedCompareExchange16(short volatile *_Destination,
+                              short _Exchange, short _Comparand) {
+  __atomic_compare_exchange(_Destination, &_Comparand, &_Exchange, 0, 0, 0);
+  return _Comparand;
+}
+static __inline__ __int64 __attribute__((__always_inline__, __nodebug__))
+_InterlockedCompareExchange64(__int64 volatile *_Destination,
+                              __int64 _Exchange, __int64 _Comparand) {
+  __atomic_compare_exchange(_Destination, &_Comparand, &_Exchange, 0, 0, 0);
+  return _Comparand;
+}
+/*----------------------------------------------------------------------------*\
+|* Barriers
+\*----------------------------------------------------------------------------*/
+#if defined(__i386__) || defined(__x86_64__)
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+__attribute__((deprecated("use other intrinsics or C++11 atomics instead")))
+_ReadWriteBarrier(void) {
+  __asm__ volatile ("" : : : "memory");
+}
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+__attribute__((deprecated("use other intrinsics or C++11 atomics instead")))
+_ReadBarrier(void) {
+  __asm__ volatile ("" : : : "memory");
+}
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+__attribute__((deprecated("use other intrinsics or C++11 atomics instead")))
+_WriteBarrier(void) {
+  __asm__ volatile ("" : : : "memory");
+}
+#endif
+#ifdef __x86_64__
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+__faststorefence(void) {
+  __asm__ volatile("lock orq $0, (%%rsp)" : : : "memory");
+}
+#endif
+/*----------------------------------------------------------------------------*\
+|* readfs, readgs
+|* (Pointers in address space #256 and #257 are relative to the GS and FS
+|* segment registers, respectively.)
+\*----------------------------------------------------------------------------*/
+#define __ptr_to_addr_space(__addr_space_nbr, __type, __offset)              \
+    ((volatile __type __attribute__((__address_space__(__addr_space_nbr)))*) \
+    (__offset))
+
+#ifdef __i386__
+static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
+__readfsbyte(unsigned long __offset) {
+  return *__ptr_to_addr_space(257, unsigned char, __offset);
+}
+static __inline__ unsigned __int64 __attribute__((__always_inline__, __nodebug__))
+__readfsqword(unsigned long __offset) {
+  return *__ptr_to_addr_space(257, unsigned __int64, __offset);
+}
+static __inline__ unsigned short __attribute__((__always_inline__, __nodebug__))
+__readfsword(unsigned long __offset) {
+  return *__ptr_to_addr_space(257, unsigned short, __offset);
+}
+#endif
+#ifdef __x86_64__
+static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
+__readgsbyte(unsigned long __offset) {
+  return *__ptr_to_addr_space(256, unsigned char, __offset);
+}
+static __inline__ unsigned long __attribute__((__always_inline__, __nodebug__))
+__readgsdword(unsigned long __offset) {
+  return *__ptr_to_addr_space(256, unsigned long, __offset);
+}
+static __inline__ unsigned __int64 __attribute__((__always_inline__, __nodebug__))
+__readgsqword(unsigned long __offset) {
+  return *__ptr_to_addr_space(256, unsigned __int64, __offset);
+}
+static __inline__ unsigned short __attribute__((__always_inline__, __nodebug__))
+__readgsword(unsigned long __offset) {
+  return *__ptr_to_addr_space(256, unsigned short, __offset);
+}
+#endif
+#undef __ptr_to_addr_space
+/*----------------------------------------------------------------------------*\
+|* movs, stos
+\*----------------------------------------------------------------------------*/
+#if defined(__i386__) || defined(__x86_64__)
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+__movsb(unsigned char *__dst, unsigned char const *__src, size_t __n) {
+  __asm__("rep movsb" : : "D"(__dst), "S"(__src), "c"(__n)
+                        : "%edi", "%esi", "%ecx");
+}
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+__movsd(unsigned long *__dst, unsigned long const *__src, size_t __n) {
+  __asm__("rep movsl" : : "D"(__dst), "S"(__src), "c"(__n)
+                        : "%edi", "%esi", "%ecx");
+}
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+__movsw(unsigned short *__dst, unsigned short const *__src, size_t __n) {
+  __asm__("rep movsh" : : "D"(__dst), "S"(__src), "c"(__n)
+                        : "%edi", "%esi", "%ecx");
+}
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+__stosb(unsigned char *__dst, unsigned char __x, size_t __n) {
+  __asm__("rep stosb" : : "D"(__dst), "a"(__x), "c"(__n)
+                        : "%edi", "%ecx");
+}
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+__stosd(unsigned long *__dst, unsigned long __x, size_t __n) {
+  __asm__("rep stosl" : : "D"(__dst), "a"(__x), "c"(__n)
+                        : "%edi", "%ecx");
+}
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+__stosw(unsigned short *__dst, unsigned short __x, size_t __n) {
+  __asm__("rep stosh" : : "D"(__dst), "a"(__x), "c"(__n)
+                        : "%edi", "%ecx");
+}
+#endif
+#ifdef __x86_64__
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+__movsq(unsigned long long *__dst, unsigned long long const *__src, size_t __n) {
+  __asm__("rep movsq" : : "D"(__dst), "S"(__src), "c"(__n)
+                        : "%edi", "%esi", "%ecx");
+}
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+__stosq(unsigned __int64 *__dst, unsigned __int64 __x, size_t __n) {
+  __asm__("rep stosq" : : "D"(__dst), "a"(__x), "c"(__n)
+                        : "%edi", "%ecx");
+}
+#endif
+
+/*----------------------------------------------------------------------------*\
+|* Misc
+\*----------------------------------------------------------------------------*/
+static __inline__ void * __attribute__((__always_inline__, __nodebug__))
+_AddressOfReturnAddress(void) {
+  return (void*)((char*)__builtin_frame_address(0) + sizeof(void*));
+}
+static __inline__ void * __attribute__((__always_inline__, __nodebug__))
+_ReturnAddress(void) {
+  return __builtin_return_address(0);
+}
+#if defined(__i386__) || defined(__x86_64__)
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+__cpuid(int __info[4], int __level) {
+  __asm__ ("cpuid" : "=a"(__info[0]), "=b" (__info[1]), "=c"(__info[2]), "=d"(__info[3])
+                   : "a"(__level));
+}
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+__cpuidex(int __info[4], int __level, int __ecx) {
+  __asm__ ("cpuid" : "=a"(__info[0]), "=b" (__info[1]), "=c"(__info[2]), "=d"(__info[3])
+                   : "a"(__level), "c"(__ecx));
+}
+static __inline__ unsigned __int64 __cdecl __attribute__((__always_inline__, __nodebug__))
+_xgetbv(unsigned int __xcr_no) {
+  unsigned int __eax, __edx;
+  __asm__ ("xgetbv" : "=a" (__eax), "=d" (__edx) : "c" (__xcr_no));
+  return ((unsigned __int64)__edx << 32) | __eax;
+}
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+__halt(void) {
+  __asm__ volatile ("hlt");
+}
+#endif
+
+/*----------------------------------------------------------------------------*\
+|* Privileged intrinsics
+\*----------------------------------------------------------------------------*/
+#if defined(__i386__) || defined(__x86_64__)
+static __inline__ unsigned __int64 __attribute__((__always_inline__, __nodebug__))
+__readmsr(unsigned long __register) {
+  // Loads the contents of a 64-bit model specific register (MSR) specified in
+  // the ECX register into registers EDX:EAX. The EDX register is loaded with
+  // the high-order 32 bits of the MSR and the EAX register is loaded with the
+  // low-order 32 bits. If less than 64 bits are implemented in the MSR being
+  // read, the values returned to EDX:EAX in unimplemented bit locations are
+  // undefined.
+  unsigned long __edx;
+  unsigned long __eax;
+  __asm__ ("rdmsr" : "=d"(__edx), "=a"(__eax) : "c"(__register));
+  return (((unsigned __int64)__edx) << 32) | (unsigned __int64)__eax;
+}
+
+static __inline__ unsigned long __attribute__((always_inline, __nodebug__))
+__readcr3(void) {
+  unsigned long __cr3_val;
+  __asm__ __volatile__ ("mov %%cr3, %0" : "=q"(__cr3_val) : : "memory");
+  return __cr3_val;
+}
+
+static __inline__ void __attribute__((always_inline, __nodebug__))
+__writecr3(unsigned int __cr3_val) {
+  __asm__ ("mov %0, %%cr3" : : "q"(__cr3_val) : "memory");
+}
+#endif
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* __INTRIN_H */
+#endif /* _MSC_VER */
diff --git a/lib/clang/3.6/include/__stddef_max_align_t.h b/lib/clang/3.6/include/__stddef_max_align_t.h
new file mode 100644
index 0000000..a06f412
--- /dev/null
+++ b/lib/clang/3.6/include/__stddef_max_align_t.h
@@ -0,0 +1,40 @@
+/*===---- __stddef_max_align_t.h - Definition of max_align_t for modules ---===
+ *
+ * Copyright (c) 2014 Chandler Carruth
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __CLANG_MAX_ALIGN_T_DEFINED
+#define __CLANG_MAX_ALIGN_T_DEFINED
+
+#ifndef _MSC_VER
+typedef struct {
+  long long __clang_max_align_nonce1
+      __attribute__((__aligned__(__alignof__(long long))));
+  long double __clang_max_align_nonce2
+      __attribute__((__aligned__(__alignof__(long double))));
+} max_align_t;
+#else
+typedef double max_align_t;
+#endif
+
+#endif
diff --git a/lib/clang/3.6/include/__wmmintrin_aes.h b/lib/clang/3.6/include/__wmmintrin_aes.h
new file mode 100644
index 0000000..2bfa027
--- /dev/null
+++ b/lib/clang/3.6/include/__wmmintrin_aes.h
@@ -0,0 +1,67 @@
+/*===---- __wmmintrin_aes.h - AES intrinsics -------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+#ifndef _WMMINTRIN_AES_H
+#define _WMMINTRIN_AES_H
+
+#include <emmintrin.h>
+
+#if !defined (__AES__)
+#  error "AES instructions not enabled"
+#else
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_aesenc_si128(__m128i __V, __m128i __R)
+{
+  return (__m128i)__builtin_ia32_aesenc128(__V, __R);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_aesenclast_si128(__m128i __V, __m128i __R)
+{
+  return (__m128i)__builtin_ia32_aesenclast128(__V, __R);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_aesdec_si128(__m128i __V, __m128i __R)
+{
+  return (__m128i)__builtin_ia32_aesdec128(__V, __R);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_aesdeclast_si128(__m128i __V, __m128i __R)
+{
+  return (__m128i)__builtin_ia32_aesdeclast128(__V, __R);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_aesimc_si128(__m128i __V)
+{
+  return (__m128i)__builtin_ia32_aesimc128(__V);
+}
+
+#define _mm_aeskeygenassist_si128(C, R) \
+  __builtin_ia32_aeskeygenassist128((C), (R))
+
+#endif
+
+#endif  /* _WMMINTRIN_AES_H */
diff --git a/lib/clang/3.6/include/__wmmintrin_pclmul.h b/lib/clang/3.6/include/__wmmintrin_pclmul.h
new file mode 100644
index 0000000..8d1f1b7
--- /dev/null
+++ b/lib/clang/3.6/include/__wmmintrin_pclmul.h
@@ -0,0 +1,34 @@
+/*===---- __wmmintrin_pclmul.h - AES intrinsics ----------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+#ifndef _WMMINTRIN_PCLMUL_H
+#define _WMMINTRIN_PCLMUL_H
+
+#if !defined (__PCLMUL__)
+# error "PCLMUL instruction is not enabled"
+#else
+#define _mm_clmulepi64_si128(__X, __Y, __I) \
+  ((__m128i)__builtin_ia32_pclmulqdq128((__v2di)(__m128i)(__X), \
+                                        (__v2di)(__m128i)(__Y), (char)(__I)))
+#endif
+
+#endif /* _WMMINTRIN_PCLMUL_H */
diff --git a/lib/clang/3.6/include/adxintrin.h b/lib/clang/3.6/include/adxintrin.h
new file mode 100644
index 0000000..9db8bcb
--- /dev/null
+++ b/lib/clang/3.6/include/adxintrin.h
@@ -0,0 +1,83 @@
+/*===---- adxintrin.h - ADX intrinsics -------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __IMMINTRIN_H
+#error "Never use <adxintrin.h> directly; include <immintrin.h> instead."
+#endif
+
+#ifndef __ADXINTRIN_H
+#define __ADXINTRIN_H
+
+/* Intrinsics that are available only if __ADX__ defined */
+#ifdef __ADX__
+static __inline unsigned char __attribute__((__always_inline__, __nodebug__))
+_addcarryx_u32(unsigned char __cf, unsigned int __x, unsigned int __y,
+               unsigned int *__p)
+{
+  return __builtin_ia32_addcarryx_u32(__cf, __x, __y, __p);
+}
+
+#ifdef __x86_64__
+static __inline unsigned char __attribute__((__always_inline__, __nodebug__))
+_addcarryx_u64(unsigned char __cf, unsigned long long __x,
+               unsigned long long __y, unsigned long long  *__p)
+{
+  return __builtin_ia32_addcarryx_u64(__cf, __x, __y, __p);
+}
+#endif
+#endif
+
+/* Intrinsics that are also available if __ADX__ undefined */
+static __inline unsigned char __attribute__((__always_inline__, __nodebug__))
+_addcarry_u32(unsigned char __cf, unsigned int __x, unsigned int __y,
+              unsigned int *__p)
+{
+  return __builtin_ia32_addcarry_u32(__cf, __x, __y, __p);
+}
+
+#ifdef __x86_64__
+static __inline unsigned char __attribute__((__always_inline__, __nodebug__))
+_addcarry_u64(unsigned char __cf, unsigned long long __x,
+              unsigned long long __y, unsigned long long  *__p)
+{
+  return __builtin_ia32_addcarry_u64(__cf, __x, __y, __p);
+}
+#endif
+
+static __inline unsigned char __attribute__((__always_inline__, __nodebug__))
+_subborrow_u32(unsigned char __cf, unsigned int __x, unsigned int __y,
+              unsigned int *__p)
+{
+  return __builtin_ia32_subborrow_u32(__cf, __x, __y, __p);
+}
+
+#ifdef __x86_64__
+static __inline unsigned char __attribute__((__always_inline__, __nodebug__))
+_subborrow_u64(unsigned char __cf, unsigned long long __x,
+               unsigned long long __y, unsigned long long  *__p)
+{
+  return __builtin_ia32_subborrow_u64(__cf, __x, __y, __p);
+}
+#endif
+
+#endif /* __ADXINTRIN_H */
diff --git a/lib/clang/3.6/include/altivec.h b/lib/clang/3.6/include/altivec.h
new file mode 100644
index 0000000..eded7b2
--- /dev/null
+++ b/lib/clang/3.6/include/altivec.h
@@ -0,0 +1,12666 @@
+/*===---- altivec.h - Standard header for type generic math ---------------===*\
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+\*===----------------------------------------------------------------------===*/
+
+#ifndef __ALTIVEC_H
+#define __ALTIVEC_H
+
+#ifndef __ALTIVEC__
+#error "AltiVec support not enabled"
+#endif
+
+/* constants for mapping CR6 bits to predicate result. */
+
+#define __CR6_EQ     0
+#define __CR6_EQ_REV 1
+#define __CR6_LT     2
+#define __CR6_LT_REV 3
+
+#define __ATTRS_o_ai __attribute__((__overloadable__, __always_inline__))
+
+static vector signed char __ATTRS_o_ai
+vec_perm(vector signed char __a, vector signed char __b, vector unsigned char __c);
+
+static vector unsigned char __ATTRS_o_ai
+vec_perm(vector unsigned char __a,
+         vector unsigned char __b,
+         vector unsigned char __c);
+
+static vector bool char __ATTRS_o_ai
+vec_perm(vector bool char __a, vector bool char __b, vector unsigned char __c);
+
+static vector short __ATTRS_o_ai
+vec_perm(vector short __a, vector short __b, vector unsigned char __c);
+
+static vector unsigned short __ATTRS_o_ai
+vec_perm(vector unsigned short __a,
+         vector unsigned short __b,
+         vector unsigned char __c);
+
+static vector bool short __ATTRS_o_ai
+vec_perm(vector bool short __a, vector bool short __b, vector unsigned char __c);
+
+static vector pixel __ATTRS_o_ai
+vec_perm(vector pixel __a, vector pixel __b, vector unsigned char __c);
+
+static vector int __ATTRS_o_ai
+vec_perm(vector int __a, vector int __b, vector unsigned char __c);
+
+static vector unsigned int __ATTRS_o_ai
+vec_perm(vector unsigned int __a, vector unsigned int __b, vector unsigned char __c);
+
+static vector bool int __ATTRS_o_ai
+vec_perm(vector bool int __a, vector bool int __b, vector unsigned char __c);
+
+static vector float __ATTRS_o_ai
+vec_perm(vector float __a, vector float __b, vector unsigned char __c);
+
+static vector unsigned char __ATTRS_o_ai
+vec_xor(vector unsigned char __a, vector unsigned char __b);
+
+/* vec_abs */
+
+#define __builtin_altivec_abs_v16qi vec_abs
+#define __builtin_altivec_abs_v8hi  vec_abs
+#define __builtin_altivec_abs_v4si  vec_abs
+
+static vector signed char __ATTRS_o_ai
+vec_abs(vector signed char __a)
+{
+  return __builtin_altivec_vmaxsb(__a, -__a);
+}
+
+static vector signed short __ATTRS_o_ai
+vec_abs(vector signed short __a)
+{
+  return __builtin_altivec_vmaxsh(__a, -__a);
+}
+
+static vector signed int __ATTRS_o_ai
+vec_abs(vector signed int __a)
+{
+  return __builtin_altivec_vmaxsw(__a, -__a);
+}
+
+static vector float __ATTRS_o_ai
+vec_abs(vector float __a)
+{
+  vector unsigned int __res = (vector unsigned int)__a
+                            & (vector unsigned int)(0x7FFFFFFF);
+  return (vector float)__res;
+}
+
+/* vec_abss */
+
+#define __builtin_altivec_abss_v16qi vec_abss
+#define __builtin_altivec_abss_v8hi  vec_abss
+#define __builtin_altivec_abss_v4si  vec_abss
+
+static vector signed char __ATTRS_o_ai
+vec_abss(vector signed char __a)
+{
+  return __builtin_altivec_vmaxsb
+           (__a, __builtin_altivec_vsubsbs((vector signed char)(0), __a));
+}
+
+static vector signed short __ATTRS_o_ai
+vec_abss(vector signed short __a)
+{
+  return __builtin_altivec_vmaxsh
+           (__a, __builtin_altivec_vsubshs((vector signed short)(0), __a));
+}
+
+static vector signed int __ATTRS_o_ai
+vec_abss(vector signed int __a)
+{
+  return __builtin_altivec_vmaxsw
+           (__a, __builtin_altivec_vsubsws((vector signed int)(0), __a));
+}
+
+/* vec_add */
+
+static vector signed char __ATTRS_o_ai
+vec_add(vector signed char __a, vector signed char __b)
+{
+  return __a + __b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_add(vector bool char __a, vector signed char __b)
+{
+  return (vector signed char)__a + __b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_add(vector signed char __a, vector bool char __b)
+{
+  return __a + (vector signed char)__b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_add(vector unsigned char __a, vector unsigned char __b)
+{
+  return __a + __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_add(vector bool char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)__a + __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_add(vector unsigned char __a, vector bool char __b)
+{
+  return __a + (vector unsigned char)__b;
+}
+
+static vector short __ATTRS_o_ai
+vec_add(vector short __a, vector short __b)
+{
+  return __a + __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_add(vector bool short __a, vector short __b)
+{
+  return (vector short)__a + __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_add(vector short __a, vector bool short __b)
+{
+  return __a + (vector short)__b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_add(vector unsigned short __a, vector unsigned short __b)
+{
+  return __a + __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_add(vector bool short __a, vector unsigned short __b)
+{
+  return (vector unsigned short)__a + __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_add(vector unsigned short __a, vector bool short __b)
+{
+  return __a + (vector unsigned short)__b;
+}
+
+static vector int __ATTRS_o_ai
+vec_add(vector int __a, vector int __b)
+{
+  return __a + __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_add(vector bool int __a, vector int __b)
+{
+  return (vector int)__a + __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_add(vector int __a, vector bool int __b)
+{
+  return __a + (vector int)__b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_add(vector unsigned int __a, vector unsigned int __b)
+{
+  return __a + __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_add(vector bool int __a, vector unsigned int __b)
+{
+  return (vector unsigned int)__a + __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_add(vector unsigned int __a, vector bool int __b)
+{
+  return __a + (vector unsigned int)__b;
+}
+
+static vector float __ATTRS_o_ai
+vec_add(vector float __a, vector float __b)
+{
+  return __a + __b;
+}
+
+/* vec_vaddubm */
+
+#define __builtin_altivec_vaddubm vec_vaddubm
+
+static vector signed char __ATTRS_o_ai
+vec_vaddubm(vector signed char __a, vector signed char __b)
+{
+  return __a + __b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vaddubm(vector bool char __a, vector signed char __b)
+{
+  return (vector signed char)__a + __b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vaddubm(vector signed char __a, vector bool char __b)
+{
+  return __a + (vector signed char)__b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vaddubm(vector unsigned char __a, vector unsigned char __b)
+{
+  return __a + __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vaddubm(vector bool char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)__a + __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vaddubm(vector unsigned char __a, vector bool char __b)
+{
+  return __a + (vector unsigned char)__b;
+}
+
+/* vec_vadduhm */
+
+#define __builtin_altivec_vadduhm vec_vadduhm
+
+static vector short __ATTRS_o_ai
+vec_vadduhm(vector short __a, vector short __b)
+{
+  return __a + __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vadduhm(vector bool short __a, vector short __b)
+{
+  return (vector short)__a + __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vadduhm(vector short __a, vector bool short __b)
+{
+  return __a + (vector short)__b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vadduhm(vector unsigned short __a, vector unsigned short __b)
+{
+  return __a + __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vadduhm(vector bool short __a, vector unsigned short __b)
+{
+  return (vector unsigned short)__a + __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vadduhm(vector unsigned short __a, vector bool short __b)
+{
+  return __a + (vector unsigned short)__b;
+}
+
+/* vec_vadduwm */
+
+#define __builtin_altivec_vadduwm vec_vadduwm
+
+static vector int __ATTRS_o_ai
+vec_vadduwm(vector int __a, vector int __b)
+{
+  return __a + __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vadduwm(vector bool int __a, vector int __b)
+{
+  return (vector int)__a + __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vadduwm(vector int __a, vector bool int __b)
+{
+  return __a + (vector int)__b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vadduwm(vector unsigned int __a, vector unsigned int __b)
+{
+  return __a + __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vadduwm(vector bool int __a, vector unsigned int __b)
+{
+  return (vector unsigned int)__a + __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vadduwm(vector unsigned int __a, vector bool int __b)
+{
+  return __a + (vector unsigned int)__b;
+}
+
+/* vec_vaddfp */
+
+#define __builtin_altivec_vaddfp  vec_vaddfp
+
+static vector float __attribute__((__always_inline__))
+vec_vaddfp(vector float __a, vector float __b)
+{
+  return __a + __b;
+}
+
+/* vec_addc */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_addc(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vaddcuw(__a, __b);
+}
+
+/* vec_vaddcuw */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_vaddcuw(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vaddcuw(__a, __b);
+}
+
+/* vec_adds */
+
+static vector signed char __ATTRS_o_ai
+vec_adds(vector signed char __a, vector signed char __b)
+{
+  return __builtin_altivec_vaddsbs(__a, __b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_adds(vector bool char __a, vector signed char __b)
+{
+  return __builtin_altivec_vaddsbs((vector signed char)__a, __b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_adds(vector signed char __a, vector bool char __b)
+{
+  return __builtin_altivec_vaddsbs(__a, (vector signed char)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_adds(vector unsigned char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vaddubs(__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_adds(vector bool char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vaddubs((vector unsigned char)__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_adds(vector unsigned char __a, vector bool char __b)
+{
+  return __builtin_altivec_vaddubs(__a, (vector unsigned char)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_adds(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vaddshs(__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_adds(vector bool short __a, vector short __b)
+{
+  return __builtin_altivec_vaddshs((vector short)__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_adds(vector short __a, vector bool short __b)
+{
+  return __builtin_altivec_vaddshs(__a, (vector short)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_adds(vector unsigned short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vadduhs(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_adds(vector bool short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vadduhs((vector unsigned short)__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_adds(vector unsigned short __a, vector bool short __b)
+{
+  return __builtin_altivec_vadduhs(__a, (vector unsigned short)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_adds(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vaddsws(__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_adds(vector bool int __a, vector int __b)
+{
+  return __builtin_altivec_vaddsws((vector int)__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_adds(vector int __a, vector bool int __b)
+{
+  return __builtin_altivec_vaddsws(__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_adds(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vadduws(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_adds(vector bool int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vadduws((vector unsigned int)__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_adds(vector unsigned int __a, vector bool int __b)
+{
+  return __builtin_altivec_vadduws(__a, (vector unsigned int)__b);
+}
+
+/* vec_vaddsbs */
+
+static vector signed char __ATTRS_o_ai
+vec_vaddsbs(vector signed char __a, vector signed char __b)
+{
+  return __builtin_altivec_vaddsbs(__a, __b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vaddsbs(vector bool char __a, vector signed char __b)
+{
+  return __builtin_altivec_vaddsbs((vector signed char)__a, __b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vaddsbs(vector signed char __a, vector bool char __b)
+{
+  return __builtin_altivec_vaddsbs(__a, (vector signed char)__b);
+}
+
+/* vec_vaddubs */
+
+static vector unsigned char __ATTRS_o_ai
+vec_vaddubs(vector unsigned char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vaddubs(__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vaddubs(vector bool char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vaddubs((vector unsigned char)__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vaddubs(vector unsigned char __a, vector bool char __b)
+{
+  return __builtin_altivec_vaddubs(__a, (vector unsigned char)__b);
+}
+
+/* vec_vaddshs */
+
+static vector short __ATTRS_o_ai
+vec_vaddshs(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vaddshs(__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vaddshs(vector bool short __a, vector short __b)
+{
+  return __builtin_altivec_vaddshs((vector short)__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vaddshs(vector short __a, vector bool short __b)
+{
+  return __builtin_altivec_vaddshs(__a, (vector short)__b);
+}
+
+/* vec_vadduhs */
+
+static vector unsigned short __ATTRS_o_ai
+vec_vadduhs(vector unsigned short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vadduhs(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vadduhs(vector bool short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vadduhs((vector unsigned short)__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vadduhs(vector unsigned short __a, vector bool short __b)
+{
+  return __builtin_altivec_vadduhs(__a, (vector unsigned short)__b);
+}
+
+/* vec_vaddsws */
+
+static vector int __ATTRS_o_ai
+vec_vaddsws(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vaddsws(__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vaddsws(vector bool int __a, vector int __b)
+{
+  return __builtin_altivec_vaddsws((vector int)__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vaddsws(vector int __a, vector bool int __b)
+{
+  return __builtin_altivec_vaddsws(__a, (vector int)__b);
+}
+
+/* vec_vadduws */
+
+static vector unsigned int __ATTRS_o_ai
+vec_vadduws(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vadduws(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vadduws(vector bool int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vadduws((vector unsigned int)__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vadduws(vector unsigned int __a, vector bool int __b)
+{
+  return __builtin_altivec_vadduws(__a, (vector unsigned int)__b);
+}
+
+/* vec_and */
+
+#define __builtin_altivec_vand vec_and
+
+static vector signed char __ATTRS_o_ai
+vec_and(vector signed char __a, vector signed char __b)
+{
+  return __a & __b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_and(vector bool char __a, vector signed char __b)
+{
+  return (vector signed char)__a & __b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_and(vector signed char __a, vector bool char __b)
+{
+  return __a & (vector signed char)__b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_and(vector unsigned char __a, vector unsigned char __b)
+{
+  return __a & __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_and(vector bool char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)__a & __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_and(vector unsigned char __a, vector bool char __b)
+{
+  return __a & (vector unsigned char)__b;
+}
+
+static vector bool char __ATTRS_o_ai
+vec_and(vector bool char __a, vector bool char __b)
+{
+  return __a & __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_and(vector short __a, vector short __b)
+{
+  return __a & __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_and(vector bool short __a, vector short __b)
+{
+  return (vector short)__a & __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_and(vector short __a, vector bool short __b)
+{
+  return __a & (vector short)__b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_and(vector unsigned short __a, vector unsigned short __b)
+{
+  return __a & __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_and(vector bool short __a, vector unsigned short __b)
+{
+  return (vector unsigned short)__a & __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_and(vector unsigned short __a, vector bool short __b)
+{
+  return __a & (vector unsigned short)__b;
+}
+
+static vector bool short __ATTRS_o_ai
+vec_and(vector bool short __a, vector bool short __b)
+{
+  return __a & __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_and(vector int __a, vector int __b)
+{
+  return __a & __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_and(vector bool int __a, vector int __b)
+{
+  return (vector int)__a & __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_and(vector int __a, vector bool int __b)
+{
+  return __a & (vector int)__b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_and(vector unsigned int __a, vector unsigned int __b)
+{
+  return __a & __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_and(vector bool int __a, vector unsigned int __b)
+{
+  return (vector unsigned int)__a & __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_and(vector unsigned int __a, vector bool int __b)
+{
+  return __a & (vector unsigned int)__b;
+}
+
+static vector bool int __ATTRS_o_ai
+vec_and(vector bool int __a, vector bool int __b)
+{
+  return __a & __b;
+}
+
+static vector float __ATTRS_o_ai
+vec_and(vector float __a, vector float __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a & (vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+static vector float __ATTRS_o_ai
+vec_and(vector bool int __a, vector float __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a & (vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+static vector float __ATTRS_o_ai
+vec_and(vector float __a, vector bool int __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a & (vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+/* vec_vand */
+
+static vector signed char __ATTRS_o_ai
+vec_vand(vector signed char __a, vector signed char __b)
+{
+  return __a & __b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vand(vector bool char __a, vector signed char __b)
+{
+  return (vector signed char)__a & __b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vand(vector signed char __a, vector bool char __b)
+{
+  return __a & (vector signed char)__b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vand(vector unsigned char __a, vector unsigned char __b)
+{
+  return __a & __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vand(vector bool char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)__a & __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vand(vector unsigned char __a, vector bool char __b)
+{
+  return __a & (vector unsigned char)__b;
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vand(vector bool char __a, vector bool char __b)
+{
+  return __a & __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vand(vector short __a, vector short __b)
+{
+  return __a & __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vand(vector bool short __a, vector short __b)
+{
+  return (vector short)__a & __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vand(vector short __a, vector bool short __b)
+{
+  return __a & (vector short)__b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vand(vector unsigned short __a, vector unsigned short __b)
+{
+  return __a & __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vand(vector bool short __a, vector unsigned short __b)
+{
+  return (vector unsigned short)__a & __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vand(vector unsigned short __a, vector bool short __b)
+{
+  return __a & (vector unsigned short)__b;
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vand(vector bool short __a, vector bool short __b)
+{
+  return __a & __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vand(vector int __a, vector int __b)
+{
+  return __a & __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vand(vector bool int __a, vector int __b)
+{
+  return (vector int)__a & __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vand(vector int __a, vector bool int __b)
+{
+  return __a & (vector int)__b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vand(vector unsigned int __a, vector unsigned int __b)
+{
+  return __a & __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vand(vector bool int __a, vector unsigned int __b)
+{
+  return (vector unsigned int)__a & __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vand(vector unsigned int __a, vector bool int __b)
+{
+  return __a & (vector unsigned int)__b;
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vand(vector bool int __a, vector bool int __b)
+{
+  return __a & __b;
+}
+
+static vector float __ATTRS_o_ai
+vec_vand(vector float __a, vector float __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a & (vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+static vector float __ATTRS_o_ai
+vec_vand(vector bool int __a, vector float __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a & (vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+static vector float __ATTRS_o_ai
+vec_vand(vector float __a, vector bool int __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a & (vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+/* vec_andc */
+
+#define __builtin_altivec_vandc vec_andc
+
+static vector signed char __ATTRS_o_ai
+vec_andc(vector signed char __a, vector signed char __b)
+{
+  return __a & ~__b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_andc(vector bool char __a, vector signed char __b)
+{
+  return (vector signed char)__a & ~__b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_andc(vector signed char __a, vector bool char __b)
+{
+  return __a & ~(vector signed char)__b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_andc(vector unsigned char __a, vector unsigned char __b)
+{
+  return __a & ~__b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_andc(vector bool char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)__a & ~__b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_andc(vector unsigned char __a, vector bool char __b)
+{
+  return __a & ~(vector unsigned char)__b;
+}
+
+static vector bool char __ATTRS_o_ai
+vec_andc(vector bool char __a, vector bool char __b)
+{
+  return __a & ~__b;
+}
+
+static vector short __ATTRS_o_ai
+vec_andc(vector short __a, vector short __b)
+{
+  return __a & ~__b;
+}
+
+static vector short __ATTRS_o_ai
+vec_andc(vector bool short __a, vector short __b)
+{
+  return (vector short)__a & ~__b;
+}
+
+static vector short __ATTRS_o_ai
+vec_andc(vector short __a, vector bool short __b)
+{
+  return __a & ~(vector short)__b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_andc(vector unsigned short __a, vector unsigned short __b)
+{
+  return __a & ~__b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_andc(vector bool short __a, vector unsigned short __b)
+{
+  return (vector unsigned short)__a & ~__b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_andc(vector unsigned short __a, vector bool short __b)
+{
+  return __a & ~(vector unsigned short)__b;
+}
+
+static vector bool short __ATTRS_o_ai
+vec_andc(vector bool short __a, vector bool short __b)
+{
+  return __a & ~__b;
+}
+
+static vector int __ATTRS_o_ai
+vec_andc(vector int __a, vector int __b)
+{
+  return __a & ~__b;
+}
+
+static vector int __ATTRS_o_ai
+vec_andc(vector bool int __a, vector int __b)
+{
+  return (vector int)__a & ~__b;
+}
+
+static vector int __ATTRS_o_ai
+vec_andc(vector int __a, vector bool int __b)
+{
+  return __a & ~(vector int)__b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_andc(vector unsigned int __a, vector unsigned int __b)
+{
+  return __a & ~__b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_andc(vector bool int __a, vector unsigned int __b)
+{
+  return (vector unsigned int)__a & ~__b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_andc(vector unsigned int __a, vector bool int __b)
+{
+  return __a & ~(vector unsigned int)__b;
+}
+
+static vector bool int __ATTRS_o_ai
+vec_andc(vector bool int __a, vector bool int __b)
+{
+  return __a & ~__b;
+}
+
+static vector float __ATTRS_o_ai
+vec_andc(vector float __a, vector float __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a & ~(vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+static vector float __ATTRS_o_ai
+vec_andc(vector bool int __a, vector float __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a & ~(vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+static vector float __ATTRS_o_ai
+vec_andc(vector float __a, vector bool int __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a & ~(vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+/* vec_vandc */
+
+static vector signed char __ATTRS_o_ai
+vec_vandc(vector signed char __a, vector signed char __b)
+{
+  return __a & ~__b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vandc(vector bool char __a, vector signed char __b)
+{
+  return (vector signed char)__a & ~__b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vandc(vector signed char __a, vector bool char __b)
+{
+  return __a & ~(vector signed char)__b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vandc(vector unsigned char __a, vector unsigned char __b)
+{
+  return __a & ~__b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vandc(vector bool char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)__a & ~__b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vandc(vector unsigned char __a, vector bool char __b)
+{
+  return __a & ~(vector unsigned char)__b;
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vandc(vector bool char __a, vector bool char __b)
+{
+  return __a & ~__b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vandc(vector short __a, vector short __b)
+{
+  return __a & ~__b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vandc(vector bool short __a, vector short __b)
+{
+  return (vector short)__a & ~__b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vandc(vector short __a, vector bool short __b)
+{
+  return __a & ~(vector short)__b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vandc(vector unsigned short __a, vector unsigned short __b)
+{
+  return __a & ~__b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vandc(vector bool short __a, vector unsigned short __b)
+{
+  return (vector unsigned short)__a & ~__b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vandc(vector unsigned short __a, vector bool short __b)
+{
+  return __a & ~(vector unsigned short)__b;
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vandc(vector bool short __a, vector bool short __b)
+{
+  return __a & ~__b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vandc(vector int __a, vector int __b)
+{
+  return __a & ~__b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vandc(vector bool int __a, vector int __b)
+{
+  return (vector int)__a & ~__b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vandc(vector int __a, vector bool int __b)
+{
+  return __a & ~(vector int)__b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vandc(vector unsigned int __a, vector unsigned int __b)
+{
+  return __a & ~__b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vandc(vector bool int __a, vector unsigned int __b)
+{
+  return (vector unsigned int)__a & ~__b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vandc(vector unsigned int __a, vector bool int __b)
+{
+  return __a & ~(vector unsigned int)__b;
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vandc(vector bool int __a, vector bool int __b)
+{
+  return __a & ~__b;
+}
+
+static vector float __ATTRS_o_ai
+vec_vandc(vector float __a, vector float __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a & ~(vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+static vector float __ATTRS_o_ai
+vec_vandc(vector bool int __a, vector float __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a & ~(vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+static vector float __ATTRS_o_ai
+vec_vandc(vector float __a, vector bool int __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a & ~(vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+/* vec_avg */
+
+static vector signed char __ATTRS_o_ai
+vec_avg(vector signed char __a, vector signed char __b)
+{
+  return __builtin_altivec_vavgsb(__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_avg(vector unsigned char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vavgub(__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_avg(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vavgsh(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_avg(vector unsigned short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vavguh(__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_avg(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vavgsw(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_avg(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vavguw(__a, __b);
+}
+
+/* vec_vavgsb */
+
+static vector signed char __attribute__((__always_inline__))
+vec_vavgsb(vector signed char __a, vector signed char __b)
+{
+  return __builtin_altivec_vavgsb(__a, __b);
+}
+
+/* vec_vavgub */
+
+static vector unsigned char __attribute__((__always_inline__))
+vec_vavgub(vector unsigned char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vavgub(__a, __b);
+}
+
+/* vec_vavgsh */
+
+static vector short __attribute__((__always_inline__))
+vec_vavgsh(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vavgsh(__a, __b);
+}
+
+/* vec_vavguh */
+
+static vector unsigned short __attribute__((__always_inline__))
+vec_vavguh(vector unsigned short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vavguh(__a, __b);
+}
+
+/* vec_vavgsw */
+
+static vector int __attribute__((__always_inline__))
+vec_vavgsw(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vavgsw(__a, __b);
+}
+
+/* vec_vavguw */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_vavguw(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vavguw(__a, __b);
+}
+
+/* vec_ceil */
+
+static vector float __attribute__((__always_inline__))
+vec_ceil(vector float __a)
+{
+  return __builtin_altivec_vrfip(__a);
+}
+
+/* vec_vrfip */
+
+static vector float __attribute__((__always_inline__))
+vec_vrfip(vector float __a)
+{
+  return __builtin_altivec_vrfip(__a);
+}
+
+/* vec_cmpb */
+
+static vector int __attribute__((__always_inline__))
+vec_cmpb(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpbfp(__a, __b);
+}
+
+/* vec_vcmpbfp */
+
+static vector int __attribute__((__always_inline__))
+vec_vcmpbfp(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpbfp(__a, __b);
+}
+
+/* vec_cmpeq */
+
+static vector bool char __ATTRS_o_ai
+vec_cmpeq(vector signed char __a, vector signed char __b)
+{
+  return (vector bool char)
+    __builtin_altivec_vcmpequb((vector char)__a, (vector char)__b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_cmpeq(vector unsigned char __a, vector unsigned char __b)
+{
+  return (vector bool char)
+    __builtin_altivec_vcmpequb((vector char)__a, (vector char)__b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_cmpeq(vector short __a, vector short __b)
+{
+  return (vector bool short)__builtin_altivec_vcmpequh(__a, __b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_cmpeq(vector unsigned short __a, vector unsigned short __b)
+{
+  return (vector bool short)
+    __builtin_altivec_vcmpequh((vector short)__a, (vector short)__b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_cmpeq(vector int __a, vector int __b)
+{
+  return (vector bool int)__builtin_altivec_vcmpequw(__a, __b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_cmpeq(vector unsigned int __a, vector unsigned int __b)
+{
+  return (vector bool int)
+    __builtin_altivec_vcmpequw((vector int)__a, (vector int)__b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_cmpeq(vector float __a, vector float __b)
+{
+  return (vector bool int)__builtin_altivec_vcmpeqfp(__a, __b);
+}
+
+/* vec_cmpge */
+
+static vector bool int __attribute__((__always_inline__))
+vec_cmpge(vector float __a, vector float __b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgefp(__a, __b);
+}
+
+/* vec_vcmpgefp */
+
+static vector bool int __attribute__((__always_inline__))
+vec_vcmpgefp(vector float __a, vector float __b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgefp(__a, __b);
+}
+
+/* vec_cmpgt */
+
+static vector bool char __ATTRS_o_ai
+vec_cmpgt(vector signed char __a, vector signed char __b)
+{
+  return (vector bool char)__builtin_altivec_vcmpgtsb(__a, __b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_cmpgt(vector unsigned char __a, vector unsigned char __b)
+{
+  return (vector bool char)__builtin_altivec_vcmpgtub(__a, __b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_cmpgt(vector short __a, vector short __b)
+{
+  return (vector bool short)__builtin_altivec_vcmpgtsh(__a, __b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_cmpgt(vector unsigned short __a, vector unsigned short __b)
+{
+  return (vector bool short)__builtin_altivec_vcmpgtuh(__a, __b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_cmpgt(vector int __a, vector int __b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgtsw(__a, __b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_cmpgt(vector unsigned int __a, vector unsigned int __b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgtuw(__a, __b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_cmpgt(vector float __a, vector float __b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgtfp(__a, __b);
+}
+
+/* vec_vcmpgtsb */
+
+static vector bool char __attribute__((__always_inline__))
+vec_vcmpgtsb(vector signed char __a, vector signed char __b)
+{
+  return (vector bool char)__builtin_altivec_vcmpgtsb(__a, __b);
+}
+
+/* vec_vcmpgtub */
+
+static vector bool char __attribute__((__always_inline__))
+vec_vcmpgtub(vector unsigned char __a, vector unsigned char __b)
+{
+  return (vector bool char)__builtin_altivec_vcmpgtub(__a, __b);
+}
+
+/* vec_vcmpgtsh */
+
+static vector bool short __attribute__((__always_inline__))
+vec_vcmpgtsh(vector short __a, vector short __b)
+{
+  return (vector bool short)__builtin_altivec_vcmpgtsh(__a, __b);
+}
+
+/* vec_vcmpgtuh */
+
+static vector bool short __attribute__((__always_inline__))
+vec_vcmpgtuh(vector unsigned short __a, vector unsigned short __b)
+{
+  return (vector bool short)__builtin_altivec_vcmpgtuh(__a, __b);
+}
+
+/* vec_vcmpgtsw */
+
+static vector bool int __attribute__((__always_inline__))
+vec_vcmpgtsw(vector int __a, vector int __b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgtsw(__a, __b);
+}
+
+/* vec_vcmpgtuw */
+
+static vector bool int __attribute__((__always_inline__))
+vec_vcmpgtuw(vector unsigned int __a, vector unsigned int __b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgtuw(__a, __b);
+}
+
+/* vec_vcmpgtfp */
+
+static vector bool int __attribute__((__always_inline__))
+vec_vcmpgtfp(vector float __a, vector float __b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgtfp(__a, __b);
+}
+
+/* vec_cmple */
+
+static vector bool int __attribute__((__always_inline__))
+vec_cmple(vector float __a, vector float __b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgefp(__b, __a);
+}
+
+/* vec_cmplt */
+
+static vector bool char __ATTRS_o_ai
+vec_cmplt(vector signed char __a, vector signed char __b)
+{
+  return (vector bool char)__builtin_altivec_vcmpgtsb(__b, __a);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_cmplt(vector unsigned char __a, vector unsigned char __b)
+{
+  return (vector bool char)__builtin_altivec_vcmpgtub(__b, __a);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_cmplt(vector short __a, vector short __b)
+{
+  return (vector bool short)__builtin_altivec_vcmpgtsh(__b, __a);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_cmplt(vector unsigned short __a, vector unsigned short __b)
+{
+  return (vector bool short)__builtin_altivec_vcmpgtuh(__b, __a);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_cmplt(vector int __a, vector int __b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgtsw(__b, __a);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_cmplt(vector unsigned int __a, vector unsigned int __b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgtuw(__b, __a);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_cmplt(vector float __a, vector float __b)
+{
+  return (vector bool int)__builtin_altivec_vcmpgtfp(__b, __a);
+}
+
+/* vec_ctf */
+
+static vector float __ATTRS_o_ai
+vec_ctf(vector int __a, int __b)
+{
+  return __builtin_altivec_vcfsx(__a, __b);
+}
+
+static vector float __ATTRS_o_ai
+vec_ctf(vector unsigned int __a, int __b)
+{
+  return __builtin_altivec_vcfux((vector int)__a, __b);
+}
+
+/* vec_vcfsx */
+
+static vector float __attribute__((__always_inline__))
+vec_vcfsx(vector int __a, int __b)
+{
+  return __builtin_altivec_vcfsx(__a, __b);
+}
+
+/* vec_vcfux */
+
+static vector float __attribute__((__always_inline__))
+vec_vcfux(vector unsigned int __a, int __b)
+{
+  return __builtin_altivec_vcfux((vector int)__a, __b);
+}
+
+/* vec_cts */
+
+static vector int __attribute__((__always_inline__))
+vec_cts(vector float __a, int __b)
+{
+  return __builtin_altivec_vctsxs(__a, __b);
+}
+
+/* vec_vctsxs */
+
+static vector int __attribute__((__always_inline__))
+vec_vctsxs(vector float __a, int __b)
+{
+  return __builtin_altivec_vctsxs(__a, __b);
+}
+
+/* vec_ctu */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_ctu(vector float __a, int __b)
+{
+  return __builtin_altivec_vctuxs(__a, __b);
+}
+
+/* vec_vctuxs */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_vctuxs(vector float __a, int __b)
+{
+  return __builtin_altivec_vctuxs(__a, __b);
+}
+
+/* vec_div */
+#ifdef __VSX__
+static vector float __ATTRS_o_ai
+vec_div(vector float __a, vector float __b)
+{
+  return __builtin_vsx_xvdivsp(__a, __b);
+}
+
+static vector double __ATTRS_o_ai
+vec_div(vector double __a, vector double __b)
+{
+  return __builtin_vsx_xvdivdp(__a, __b);
+}
+#endif
+
+/* vec_dss */
+
+static void __attribute__((__always_inline__))
+vec_dss(int __a)
+{
+  __builtin_altivec_dss(__a);
+}
+
+/* vec_dssall */
+
+static void __attribute__((__always_inline__))
+vec_dssall(void)
+{
+  __builtin_altivec_dssall();
+}
+
+/* vec_dst */
+
+static void __attribute__((__always_inline__))
+vec_dst(const void *__a, int __b, int __c)
+{
+  __builtin_altivec_dst(__a, __b, __c);
+}
+
+/* vec_dstst */
+
+static void __attribute__((__always_inline__))
+vec_dstst(const void *__a, int __b, int __c)
+{
+  __builtin_altivec_dstst(__a, __b, __c);
+}
+
+/* vec_dststt */
+
+static void __attribute__((__always_inline__))
+vec_dststt(const void *__a, int __b, int __c)
+{
+  __builtin_altivec_dststt(__a, __b, __c);
+}
+
+/* vec_dstt */
+
+static void __attribute__((__always_inline__))
+vec_dstt(const void *__a, int __b, int __c)
+{
+  __builtin_altivec_dstt(__a, __b, __c);
+}
+
+/* vec_expte */
+
+static vector float __attribute__((__always_inline__))
+vec_expte(vector float __a)
+{
+  return __builtin_altivec_vexptefp(__a);
+}
+
+/* vec_vexptefp */
+
+static vector float __attribute__((__always_inline__))
+vec_vexptefp(vector float __a)
+{
+  return __builtin_altivec_vexptefp(__a);
+}
+
+/* vec_floor */
+
+static vector float __attribute__((__always_inline__))
+vec_floor(vector float __a)
+{
+  return __builtin_altivec_vrfim(__a);
+}
+
+/* vec_vrfim */
+
+static vector float __attribute__((__always_inline__))
+vec_vrfim(vector float __a)
+{
+  return __builtin_altivec_vrfim(__a);
+}
+
+/* vec_ld */
+
+static vector signed char __ATTRS_o_ai
+vec_ld(int __a, const vector signed char *__b)
+{
+  return (vector signed char)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_ld(int __a, const signed char *__b)
+{
+  return (vector signed char)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_ld(int __a, const vector unsigned char *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_ld(int __a, const unsigned char *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_ld(int __a, const vector bool char *__b)
+{
+  return (vector bool char)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_ld(int __a, const vector short *__b)
+{
+  return (vector short)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_ld(int __a, const short *__b)
+{
+  return (vector short)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_ld(int __a, const vector unsigned short *__b)
+{
+  return (vector unsigned short)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_ld(int __a, const unsigned short *__b)
+{
+  return (vector unsigned short)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_ld(int __a, const vector bool short *__b)
+{
+  return (vector bool short)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_ld(int __a, const vector pixel *__b)
+{
+  return (vector pixel)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_ld(int __a, const vector int *__b)
+{
+  return (vector int)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_ld(int __a, const int *__b)
+{
+  return (vector int)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_ld(int __a, const vector unsigned int *__b)
+{
+  return (vector unsigned int)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_ld(int __a, const unsigned int *__b)
+{
+  return (vector unsigned int)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_ld(int __a, const vector bool int *__b)
+{
+  return (vector bool int)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector float __ATTRS_o_ai
+vec_ld(int __a, const vector float *__b)
+{
+  return (vector float)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector float __ATTRS_o_ai
+vec_ld(int __a, const float *__b)
+{
+  return (vector float)__builtin_altivec_lvx(__a, __b);
+}
+
+/* vec_lvx */
+
+static vector signed char __ATTRS_o_ai
+vec_lvx(int __a, const vector signed char *__b)
+{
+  return (vector signed char)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_lvx(int __a, const signed char *__b)
+{
+  return (vector signed char)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvx(int __a, const vector unsigned char *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvx(int __a, const unsigned char *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_lvx(int __a, const vector bool char *__b)
+{
+  return (vector bool char)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_lvx(int __a, const vector short *__b)
+{
+  return (vector short)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_lvx(int __a, const short *__b)
+{
+  return (vector short)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvx(int __a, const vector unsigned short *__b)
+{
+  return (vector unsigned short)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvx(int __a, const unsigned short *__b)
+{
+  return (vector unsigned short)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_lvx(int __a, const vector bool short *__b)
+{
+  return (vector bool short)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_lvx(int __a, const vector pixel *__b)
+{
+  return (vector pixel)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_lvx(int __a, const vector int *__b)
+{
+  return (vector int)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_lvx(int __a, const int *__b)
+{
+  return (vector int)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvx(int __a, const vector unsigned int *__b)
+{
+  return (vector unsigned int)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvx(int __a, const unsigned int *__b)
+{
+  return (vector unsigned int)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_lvx(int __a, const vector bool int *__b)
+{
+  return (vector bool int)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector float __ATTRS_o_ai
+vec_lvx(int __a, const vector float *__b)
+{
+  return (vector float)__builtin_altivec_lvx(__a, __b);
+}
+
+static vector float __ATTRS_o_ai
+vec_lvx(int __a, const float *__b)
+{
+  return (vector float)__builtin_altivec_lvx(__a, __b);
+}
+
+/* vec_lde */
+
+static vector signed char __ATTRS_o_ai
+vec_lde(int __a, const signed char *__b)
+{
+  return (vector signed char)__builtin_altivec_lvebx(__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lde(int __a, const unsigned char *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvebx(__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_lde(int __a, const short *__b)
+{
+  return (vector short)__builtin_altivec_lvehx(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lde(int __a, const unsigned short *__b)
+{
+  return (vector unsigned short)__builtin_altivec_lvehx(__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_lde(int __a, const int *__b)
+{
+  return (vector int)__builtin_altivec_lvewx(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lde(int __a, const unsigned int *__b)
+{
+  return (vector unsigned int)__builtin_altivec_lvewx(__a, __b);
+}
+
+static vector float __ATTRS_o_ai
+vec_lde(int __a, const float *__b)
+{
+  return (vector float)__builtin_altivec_lvewx(__a, __b);
+}
+
+/* vec_lvebx */
+
+static vector signed char __ATTRS_o_ai
+vec_lvebx(int __a, const signed char *__b)
+{
+  return (vector signed char)__builtin_altivec_lvebx(__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvebx(int __a, const unsigned char *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvebx(__a, __b);
+}
+
+/* vec_lvehx */
+
+static vector short __ATTRS_o_ai
+vec_lvehx(int __a, const short *__b)
+{
+  return (vector short)__builtin_altivec_lvehx(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvehx(int __a, const unsigned short *__b)
+{
+  return (vector unsigned short)__builtin_altivec_lvehx(__a, __b);
+}
+
+/* vec_lvewx */
+
+static vector int __ATTRS_o_ai
+vec_lvewx(int __a, const int *__b)
+{
+  return (vector int)__builtin_altivec_lvewx(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvewx(int __a, const unsigned int *__b)
+{
+  return (vector unsigned int)__builtin_altivec_lvewx(__a, __b);
+}
+
+static vector float __ATTRS_o_ai
+vec_lvewx(int __a, const float *__b)
+{
+  return (vector float)__builtin_altivec_lvewx(__a, __b);
+}
+
+/* vec_ldl */
+
+static vector signed char __ATTRS_o_ai
+vec_ldl(int __a, const vector signed char *__b)
+{
+  return (vector signed char)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_ldl(int __a, const signed char *__b)
+{
+  return (vector signed char)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_ldl(int __a, const vector unsigned char *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_ldl(int __a, const unsigned char *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_ldl(int __a, const vector bool char *__b)
+{
+  return (vector bool char)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_ldl(int __a, const vector short *__b)
+{
+  return (vector short)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_ldl(int __a, const short *__b)
+{
+  return (vector short)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_ldl(int __a, const vector unsigned short *__b)
+{
+  return (vector unsigned short)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_ldl(int __a, const unsigned short *__b)
+{
+  return (vector unsigned short)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_ldl(int __a, const vector bool short *__b)
+{
+  return (vector bool short)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_ldl(int __a, const vector pixel *__b)
+{
+  return (vector pixel short)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_ldl(int __a, const vector int *__b)
+{
+  return (vector int)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_ldl(int __a, const int *__b)
+{
+  return (vector int)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_ldl(int __a, const vector unsigned int *__b)
+{
+  return (vector unsigned int)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_ldl(int __a, const unsigned int *__b)
+{
+  return (vector unsigned int)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_ldl(int __a, const vector bool int *__b)
+{
+  return (vector bool int)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector float __ATTRS_o_ai
+vec_ldl(int __a, const vector float *__b)
+{
+  return (vector float)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector float __ATTRS_o_ai
+vec_ldl(int __a, const float *__b)
+{
+  return (vector float)__builtin_altivec_lvxl(__a, __b);
+}
+
+/* vec_lvxl */
+
+static vector signed char __ATTRS_o_ai
+vec_lvxl(int __a, const vector signed char *__b)
+{
+  return (vector signed char)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_lvxl(int __a, const signed char *__b)
+{
+  return (vector signed char)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvxl(int __a, const vector unsigned char *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvxl(int __a, const unsigned char *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_lvxl(int __a, const vector bool char *__b)
+{
+  return (vector bool char)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_lvxl(int __a, const vector short *__b)
+{
+  return (vector short)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_lvxl(int __a, const short *__b)
+{
+  return (vector short)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvxl(int __a, const vector unsigned short *__b)
+{
+  return (vector unsigned short)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvxl(int __a, const unsigned short *__b)
+{
+  return (vector unsigned short)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_lvxl(int __a, const vector bool short *__b)
+{
+  return (vector bool short)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_lvxl(int __a, const vector pixel *__b)
+{
+  return (vector pixel)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_lvxl(int __a, const vector int *__b)
+{
+  return (vector int)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_lvxl(int __a, const int *__b)
+{
+  return (vector int)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvxl(int __a, const vector unsigned int *__b)
+{
+  return (vector unsigned int)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvxl(int __a, const unsigned int *__b)
+{
+  return (vector unsigned int)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_lvxl(int __a, const vector bool int *__b)
+{
+  return (vector bool int)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector float __ATTRS_o_ai
+vec_lvxl(int __a, const vector float *__b)
+{
+  return (vector float)__builtin_altivec_lvxl(__a, __b);
+}
+
+static vector float __ATTRS_o_ai
+vec_lvxl(int __a, const float *__b)
+{
+  return (vector float)__builtin_altivec_lvxl(__a, __b);
+}
+
+/* vec_loge */
+
+static vector float __attribute__((__always_inline__))
+vec_loge(vector float __a)
+{
+  return __builtin_altivec_vlogefp(__a);
+}
+
+/* vec_vlogefp */
+
+static vector float __attribute__((__always_inline__))
+vec_vlogefp(vector float __a)
+{
+  return __builtin_altivec_vlogefp(__a);
+}
+
+/* vec_lvsl */
+
+#ifdef __LITTLE_ENDIAN__
+static vector unsigned char __ATTRS_o_ai
+__attribute__((deprecated("use assignment for unaligned little endian \
+loads/stores")))
+vec_lvsl(int __a, const signed char *__b)
+{
+  vector unsigned char mask = 
+    (vector unsigned char)__builtin_altivec_lvsl(__a, __b);
+  vector unsigned char reverse = {15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0};
+  return vec_perm(mask, mask, reverse);
+}
+#else
+static vector unsigned char __ATTRS_o_ai
+vec_lvsl(int __a, const signed char *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsl(__a, __b);
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+static vector unsigned char __ATTRS_o_ai
+__attribute__((deprecated("use assignment for unaligned little endian \
+loads/stores")))
+vec_lvsl(int __a, const unsigned char *__b)
+{
+  vector unsigned char mask =
+    (vector unsigned char)__builtin_altivec_lvsl(__a, __b);
+  vector unsigned char reverse = {15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0};
+  return vec_perm(mask, mask, reverse);
+}
+#else
+static vector unsigned char __ATTRS_o_ai
+vec_lvsl(int __a, const unsigned char *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsl(__a, __b);
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+static vector unsigned char __ATTRS_o_ai
+__attribute__((deprecated("use assignment for unaligned little endian \
+loads/stores")))
+vec_lvsl(int __a, const short *__b)
+{
+  vector unsigned char mask =
+    (vector unsigned char)__builtin_altivec_lvsl(__a, __b);
+  vector unsigned char reverse = {15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0};
+  return vec_perm(mask, mask, reverse);
+}
+#else
+static vector unsigned char __ATTRS_o_ai
+vec_lvsl(int __a, const short *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsl(__a, __b);
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+static vector unsigned char __ATTRS_o_ai
+__attribute__((deprecated("use assignment for unaligned little endian \
+loads/stores")))
+vec_lvsl(int __a, const unsigned short *__b)
+{
+  vector unsigned char mask =
+    (vector unsigned char)__builtin_altivec_lvsl(__a, __b);
+  vector unsigned char reverse = {15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0};
+  return vec_perm(mask, mask, reverse);
+}
+#else
+static vector unsigned char __ATTRS_o_ai
+vec_lvsl(int __a, const unsigned short *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsl(__a, __b);
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+static vector unsigned char __ATTRS_o_ai
+__attribute__((deprecated("use assignment for unaligned little endian \
+loads/stores")))
+vec_lvsl(int __a, const int *__b)
+{
+  vector unsigned char mask =
+    (vector unsigned char)__builtin_altivec_lvsl(__a, __b);
+  vector unsigned char reverse = {15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0};
+  return vec_perm(mask, mask, reverse);
+}
+#else
+static vector unsigned char __ATTRS_o_ai
+vec_lvsl(int __a, const int *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsl(__a, __b);
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+static vector unsigned char __ATTRS_o_ai
+__attribute__((deprecated("use assignment for unaligned little endian \
+loads/stores")))
+vec_lvsl(int __a, const unsigned int *__b)
+{
+  vector unsigned char mask =
+    (vector unsigned char)__builtin_altivec_lvsl(__a, __b);
+  vector unsigned char reverse = {15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0};
+  return vec_perm(mask, mask, reverse);
+}
+#else
+static vector unsigned char __ATTRS_o_ai
+vec_lvsl(int __a, const unsigned int *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsl(__a, __b);
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+static vector unsigned char __ATTRS_o_ai
+__attribute__((deprecated("use assignment for unaligned little endian \
+loads/stores")))
+vec_lvsl(int __a, const float *__b)
+{
+  vector unsigned char mask =
+    (vector unsigned char)__builtin_altivec_lvsl(__a, __b);
+  vector unsigned char reverse = {15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0};
+  return vec_perm(mask, mask, reverse);
+}
+#else
+static vector unsigned char __ATTRS_o_ai
+vec_lvsl(int __a, const float *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsl(__a, __b);
+}
+#endif
+
+/* vec_lvsr */
+
+#ifdef __LITTLE_ENDIAN__
+static vector unsigned char __ATTRS_o_ai
+__attribute__((deprecated("use assignment for unaligned little endian \
+loads/stores")))
+vec_lvsr(int __a, const signed char *__b)
+{
+  vector unsigned char mask =
+    (vector unsigned char)__builtin_altivec_lvsr(__a, __b);
+  vector unsigned char reverse = {15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0};
+  return vec_perm(mask, mask, reverse);
+}
+#else
+static vector unsigned char __ATTRS_o_ai
+vec_lvsr(int __a, const signed char *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsr(__a, __b);
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+static vector unsigned char __ATTRS_o_ai
+__attribute__((deprecated("use assignment for unaligned little endian \
+loads/stores")))
+vec_lvsr(int __a, const unsigned char *__b)
+{
+  vector unsigned char mask =
+    (vector unsigned char)__builtin_altivec_lvsr(__a, __b);
+  vector unsigned char reverse = {15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0};
+  return vec_perm(mask, mask, reverse);
+}
+#else
+static vector unsigned char __ATTRS_o_ai
+vec_lvsr(int __a, const unsigned char *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsr(__a, __b);
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+static vector unsigned char __ATTRS_o_ai
+__attribute__((deprecated("use assignment for unaligned little endian \
+loads/stores")))
+vec_lvsr(int __a, const short *__b)
+{
+  vector unsigned char mask =
+    (vector unsigned char)__builtin_altivec_lvsr(__a, __b);
+  vector unsigned char reverse = {15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0};
+  return vec_perm(mask, mask, reverse);
+}
+#else
+static vector unsigned char __ATTRS_o_ai
+vec_lvsr(int __a, const short *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsr(__a, __b);
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+static vector unsigned char __ATTRS_o_ai
+__attribute__((deprecated("use assignment for unaligned little endian \
+loads/stores")))
+vec_lvsr(int __a, const unsigned short *__b)
+{
+  vector unsigned char mask =
+    (vector unsigned char)__builtin_altivec_lvsr(__a, __b);
+  vector unsigned char reverse = {15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0};
+  return vec_perm(mask, mask, reverse);
+}
+#else
+static vector unsigned char __ATTRS_o_ai
+vec_lvsr(int __a, const unsigned short *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsr(__a, __b);
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+static vector unsigned char __ATTRS_o_ai
+__attribute__((deprecated("use assignment for unaligned little endian \
+loads/stores")))
+vec_lvsr(int __a, const int *__b)
+{
+  vector unsigned char mask =
+    (vector unsigned char)__builtin_altivec_lvsr(__a, __b);
+  vector unsigned char reverse = {15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0};
+  return vec_perm(mask, mask, reverse);
+}
+#else
+static vector unsigned char __ATTRS_o_ai
+vec_lvsr(int __a, const int *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsr(__a, __b);
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+static vector unsigned char __ATTRS_o_ai
+__attribute__((deprecated("use assignment for unaligned little endian \
+loads/stores")))
+vec_lvsr(int __a, const unsigned int *__b)
+{
+  vector unsigned char mask =
+    (vector unsigned char)__builtin_altivec_lvsr(__a, __b);
+  vector unsigned char reverse = {15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0};
+  return vec_perm(mask, mask, reverse);
+}
+#else
+static vector unsigned char __ATTRS_o_ai
+vec_lvsr(int __a, const unsigned int *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsr(__a, __b);
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+static vector unsigned char __ATTRS_o_ai
+__attribute__((deprecated("use assignment for unaligned little endian \
+loads/stores")))
+vec_lvsr(int __a, const float *__b)
+{
+  vector unsigned char mask =
+    (vector unsigned char)__builtin_altivec_lvsr(__a, __b);
+  vector unsigned char reverse = {15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0};
+  return vec_perm(mask, mask, reverse);
+}
+#else
+static vector unsigned char __ATTRS_o_ai
+vec_lvsr(int __a, const float *__b)
+{
+  return (vector unsigned char)__builtin_altivec_lvsr(__a, __b);
+}
+#endif
+
+/* vec_madd */
+
+static vector float __attribute__((__always_inline__))
+vec_madd(vector float __a, vector float __b, vector float __c)
+{
+  return __builtin_altivec_vmaddfp(__a, __b, __c);
+}
+
+/* vec_vmaddfp */
+
+static vector float __attribute__((__always_inline__))
+vec_vmaddfp(vector float __a, vector float __b, vector float __c)
+{
+  return __builtin_altivec_vmaddfp(__a, __b, __c);
+}
+
+/* vec_madds */
+
+static vector signed short __attribute__((__always_inline__))
+vec_madds(vector signed short __a, vector signed short __b, vector signed short __c)
+{
+  return __builtin_altivec_vmhaddshs(__a, __b, __c);
+}
+
+/* vec_vmhaddshs */
+static vector signed short __attribute__((__always_inline__))
+vec_vmhaddshs(vector signed short __a,
+              vector signed short __b,
+              vector signed short __c)
+{
+  return __builtin_altivec_vmhaddshs(__a, __b, __c);
+}
+
+/* vec_max */
+
+static vector signed char __ATTRS_o_ai
+vec_max(vector signed char __a, vector signed char __b)
+{
+  return __builtin_altivec_vmaxsb(__a, __b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_max(vector bool char __a, vector signed char __b)
+{
+  return __builtin_altivec_vmaxsb((vector signed char)__a, __b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_max(vector signed char __a, vector bool char __b)
+{
+  return __builtin_altivec_vmaxsb(__a, (vector signed char)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_max(vector unsigned char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vmaxub(__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_max(vector bool char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vmaxub((vector unsigned char)__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_max(vector unsigned char __a, vector bool char __b)
+{
+  return __builtin_altivec_vmaxub(__a, (vector unsigned char)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_max(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vmaxsh(__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_max(vector bool short __a, vector short __b)
+{
+  return __builtin_altivec_vmaxsh((vector short)__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_max(vector short __a, vector bool short __b)
+{
+  return __builtin_altivec_vmaxsh(__a, (vector short)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_max(vector unsigned short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vmaxuh(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_max(vector bool short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vmaxuh((vector unsigned short)__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_max(vector unsigned short __a, vector bool short __b)
+{
+  return __builtin_altivec_vmaxuh(__a, (vector unsigned short)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_max(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vmaxsw(__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_max(vector bool int __a, vector int __b)
+{
+  return __builtin_altivec_vmaxsw((vector int)__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_max(vector int __a, vector bool int __b)
+{
+  return __builtin_altivec_vmaxsw(__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_max(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vmaxuw(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_max(vector bool int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vmaxuw((vector unsigned int)__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_max(vector unsigned int __a, vector bool int __b)
+{
+  return __builtin_altivec_vmaxuw(__a, (vector unsigned int)__b);
+}
+
+static vector float __ATTRS_o_ai
+vec_max(vector float __a, vector float __b)
+{
+#ifdef __VSX__
+  return __builtin_vsx_xvmaxsp(__a, __b);
+#else
+  return __builtin_altivec_vmaxfp(__a, __b);
+#endif
+}
+
+#ifdef __VSX__
+static vector double __ATTRS_o_ai
+vec_max(vector double __a, vector double __b)
+{
+  return __builtin_vsx_xvmaxdp(__a, __b);
+}
+#endif
+
+/* vec_vmaxsb */
+
+static vector signed char __ATTRS_o_ai
+vec_vmaxsb(vector signed char __a, vector signed char __b)
+{
+  return __builtin_altivec_vmaxsb(__a, __b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vmaxsb(vector bool char __a, vector signed char __b)
+{
+  return __builtin_altivec_vmaxsb((vector signed char)__a, __b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vmaxsb(vector signed char __a, vector bool char __b)
+{
+  return __builtin_altivec_vmaxsb(__a, (vector signed char)__b);
+}
+
+/* vec_vmaxub */
+
+static vector unsigned char __ATTRS_o_ai
+vec_vmaxub(vector unsigned char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vmaxub(__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vmaxub(vector bool char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vmaxub((vector unsigned char)__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vmaxub(vector unsigned char __a, vector bool char __b)
+{
+  return __builtin_altivec_vmaxub(__a, (vector unsigned char)__b);
+}
+
+/* vec_vmaxsh */
+
+static vector short __ATTRS_o_ai
+vec_vmaxsh(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vmaxsh(__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vmaxsh(vector bool short __a, vector short __b)
+{
+  return __builtin_altivec_vmaxsh((vector short)__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vmaxsh(vector short __a, vector bool short __b)
+{
+  return __builtin_altivec_vmaxsh(__a, (vector short)__b);
+}
+
+/* vec_vmaxuh */
+
+static vector unsigned short __ATTRS_o_ai
+vec_vmaxuh(vector unsigned short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vmaxuh(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vmaxuh(vector bool short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vmaxuh((vector unsigned short)__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vmaxuh(vector unsigned short __a, vector bool short __b)
+{
+  return __builtin_altivec_vmaxuh(__a, (vector unsigned short)__b);
+}
+
+/* vec_vmaxsw */
+
+static vector int __ATTRS_o_ai
+vec_vmaxsw(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vmaxsw(__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vmaxsw(vector bool int __a, vector int __b)
+{
+  return __builtin_altivec_vmaxsw((vector int)__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vmaxsw(vector int __a, vector bool int __b)
+{
+  return __builtin_altivec_vmaxsw(__a, (vector int)__b);
+}
+
+/* vec_vmaxuw */
+
+static vector unsigned int __ATTRS_o_ai
+vec_vmaxuw(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vmaxuw(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vmaxuw(vector bool int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vmaxuw((vector unsigned int)__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vmaxuw(vector unsigned int __a, vector bool int __b)
+{
+  return __builtin_altivec_vmaxuw(__a, (vector unsigned int)__b);
+}
+
+/* vec_vmaxfp */
+
+static vector float __attribute__((__always_inline__))
+vec_vmaxfp(vector float __a, vector float __b)
+{
+#ifdef __VSX__
+  return __builtin_vsx_xvmaxsp(__a, __b);
+#else
+  return __builtin_altivec_vmaxfp(__a, __b);
+#endif
+}
+
+/* vec_mergeh */
+
+static vector signed char __ATTRS_o_ai
+vec_mergeh(vector signed char __a, vector signed char __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x10, 0x01, 0x11, 0x02, 0x12, 0x03, 0x13, 
+     0x04, 0x14, 0x05, 0x15, 0x06, 0x16, 0x07, 0x17));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_mergeh(vector unsigned char __a, vector unsigned char __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x10, 0x01, 0x11, 0x02, 0x12, 0x03, 0x13, 
+     0x04, 0x14, 0x05, 0x15, 0x06, 0x16, 0x07, 0x17));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_mergeh(vector bool char __a, vector bool char __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x10, 0x01, 0x11, 0x02, 0x12, 0x03, 0x13, 
+     0x04, 0x14, 0x05, 0x15, 0x06, 0x16, 0x07, 0x17));
+}
+
+static vector short __ATTRS_o_ai
+vec_mergeh(vector short __a, vector short __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x10, 0x11, 0x02, 0x03, 0x12, 0x13,
+     0x04, 0x05, 0x14, 0x15, 0x06, 0x07, 0x16, 0x17));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_mergeh(vector unsigned short __a, vector unsigned short __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x10, 0x11, 0x02, 0x03, 0x12, 0x13,
+     0x04, 0x05, 0x14, 0x15, 0x06, 0x07, 0x16, 0x17));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_mergeh(vector bool short __a, vector bool short __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x10, 0x11, 0x02, 0x03, 0x12, 0x13,
+     0x04, 0x05, 0x14, 0x15, 0x06, 0x07, 0x16, 0x17));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_mergeh(vector pixel __a, vector pixel __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x10, 0x11, 0x02, 0x03, 0x12, 0x13,
+     0x04, 0x05, 0x14, 0x15, 0x06, 0x07, 0x16, 0x17));
+}
+
+static vector int __ATTRS_o_ai
+vec_mergeh(vector int __a, vector int __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x02, 0x03, 0x10, 0x11, 0x12, 0x13,
+     0x04, 0x05, 0x06, 0x07, 0x14, 0x15, 0x16, 0x17));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_mergeh(vector unsigned int __a, vector unsigned int __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x02, 0x03, 0x10, 0x11, 0x12, 0x13,
+     0x04, 0x05, 0x06, 0x07, 0x14, 0x15, 0x16, 0x17));
+}
+
+static vector bool int __ATTRS_o_ai
+vec_mergeh(vector bool int __a, vector bool int __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x02, 0x03, 0x10, 0x11, 0x12, 0x13,
+     0x04, 0x05, 0x06, 0x07, 0x14, 0x15, 0x16, 0x17));
+}
+
+static vector float __ATTRS_o_ai
+vec_mergeh(vector float __a, vector float __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x02, 0x03, 0x10, 0x11, 0x12, 0x13,
+     0x04, 0x05, 0x06, 0x07, 0x14, 0x15, 0x16, 0x17));
+}
+
+/* vec_vmrghb */
+
+#define __builtin_altivec_vmrghb vec_vmrghb
+
+static vector signed char __ATTRS_o_ai
+vec_vmrghb(vector signed char __a, vector signed char __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x10, 0x01, 0x11, 0x02, 0x12, 0x03, 0x13, 
+     0x04, 0x14, 0x05, 0x15, 0x06, 0x16, 0x07, 0x17));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vmrghb(vector unsigned char __a, vector unsigned char __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x10, 0x01, 0x11, 0x02, 0x12, 0x03, 0x13, 
+     0x04, 0x14, 0x05, 0x15, 0x06, 0x16, 0x07, 0x17));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vmrghb(vector bool char __a, vector bool char __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x10, 0x01, 0x11, 0x02, 0x12, 0x03, 0x13, 
+     0x04, 0x14, 0x05, 0x15, 0x06, 0x16, 0x07, 0x17));
+}
+
+/* vec_vmrghh */
+
+#define __builtin_altivec_vmrghh vec_vmrghh
+
+static vector short __ATTRS_o_ai
+vec_vmrghh(vector short __a, vector short __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x10, 0x11, 0x02, 0x03, 0x12, 0x13,
+     0x04, 0x05, 0x14, 0x15, 0x06, 0x07, 0x16, 0x17));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vmrghh(vector unsigned short __a, vector unsigned short __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x10, 0x11, 0x02, 0x03, 0x12, 0x13,
+     0x04, 0x05, 0x14, 0x15, 0x06, 0x07, 0x16, 0x17));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vmrghh(vector bool short __a, vector bool short __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x10, 0x11, 0x02, 0x03, 0x12, 0x13,
+     0x04, 0x05, 0x14, 0x15, 0x06, 0x07, 0x16, 0x17));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vmrghh(vector pixel __a, vector pixel __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x10, 0x11, 0x02, 0x03, 0x12, 0x13,
+     0x04, 0x05, 0x14, 0x15, 0x06, 0x07, 0x16, 0x17));
+}
+
+/* vec_vmrghw */
+
+#define __builtin_altivec_vmrghw vec_vmrghw
+
+static vector int __ATTRS_o_ai
+vec_vmrghw(vector int __a, vector int __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x02, 0x03, 0x10, 0x11, 0x12, 0x13,
+     0x04, 0x05, 0x06, 0x07, 0x14, 0x15, 0x16, 0x17));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vmrghw(vector unsigned int __a, vector unsigned int __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x02, 0x03, 0x10, 0x11, 0x12, 0x13,
+     0x04, 0x05, 0x06, 0x07, 0x14, 0x15, 0x16, 0x17));
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vmrghw(vector bool int __a, vector bool int __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x02, 0x03, 0x10, 0x11, 0x12, 0x13,
+     0x04, 0x05, 0x06, 0x07, 0x14, 0x15, 0x16, 0x17));
+}
+
+static vector float __ATTRS_o_ai
+vec_vmrghw(vector float __a, vector float __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x02, 0x03, 0x10, 0x11, 0x12, 0x13,
+     0x04, 0x05, 0x06, 0x07, 0x14, 0x15, 0x16, 0x17));
+}
+
+/* vec_mergel */
+
+static vector signed char __ATTRS_o_ai
+vec_mergel(vector signed char __a, vector signed char __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x18, 0x09, 0x19, 0x0A, 0x1A, 0x0B, 0x1B, 
+     0x0C, 0x1C, 0x0D, 0x1D, 0x0E, 0x1E, 0x0F, 0x1F));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_mergel(vector unsigned char __a, vector unsigned char __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x18, 0x09, 0x19, 0x0A, 0x1A, 0x0B, 0x1B, 
+     0x0C, 0x1C, 0x0D, 0x1D, 0x0E, 0x1E, 0x0F, 0x1F));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_mergel(vector bool char __a, vector bool char __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x18, 0x09, 0x19, 0x0A, 0x1A, 0x0B, 0x1B, 
+     0x0C, 0x1C, 0x0D, 0x1D, 0x0E, 0x1E, 0x0F, 0x1F));
+}
+
+static vector short __ATTRS_o_ai
+vec_mergel(vector short __a, vector short __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x09, 0x18, 0x19, 0x0A, 0x0B, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x1C, 0x1D, 0x0E, 0x0F, 0x1E, 0x1F));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_mergel(vector unsigned short __a, vector unsigned short __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x09, 0x18, 0x19, 0x0A, 0x0B, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x1C, 0x1D, 0x0E, 0x0F, 0x1E, 0x1F));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_mergel(vector bool short __a, vector bool short __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x09, 0x18, 0x19, 0x0A, 0x0B, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x1C, 0x1D, 0x0E, 0x0F, 0x1E, 0x1F));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_mergel(vector pixel __a, vector pixel __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x09, 0x18, 0x19, 0x0A, 0x0B, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x1C, 0x1D, 0x0E, 0x0F, 0x1E, 0x1F));
+}
+
+static vector int __ATTRS_o_ai
+vec_mergel(vector int __a, vector int __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x09, 0x0A, 0x0B, 0x18, 0x19, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x0E, 0x0F, 0x1C, 0x1D, 0x1E, 0x1F));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_mergel(vector unsigned int __a, vector unsigned int __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x09, 0x0A, 0x0B, 0x18, 0x19, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x0E, 0x0F, 0x1C, 0x1D, 0x1E, 0x1F));
+}
+
+static vector bool int __ATTRS_o_ai
+vec_mergel(vector bool int __a, vector bool int __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x09, 0x0A, 0x0B, 0x18, 0x19, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x0E, 0x0F, 0x1C, 0x1D, 0x1E, 0x1F));
+}
+
+static vector float __ATTRS_o_ai
+vec_mergel(vector float __a, vector float __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x09, 0x0A, 0x0B, 0x18, 0x19, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x0E, 0x0F, 0x1C, 0x1D, 0x1E, 0x1F));
+}
+
+/* vec_vmrglb */
+
+#define __builtin_altivec_vmrglb vec_vmrglb
+
+static vector signed char __ATTRS_o_ai
+vec_vmrglb(vector signed char __a, vector signed char __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x18, 0x09, 0x19, 0x0A, 0x1A, 0x0B, 0x1B, 
+     0x0C, 0x1C, 0x0D, 0x1D, 0x0E, 0x1E, 0x0F, 0x1F));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vmrglb(vector unsigned char __a, vector unsigned char __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x18, 0x09, 0x19, 0x0A, 0x1A, 0x0B, 0x1B, 
+     0x0C, 0x1C, 0x0D, 0x1D, 0x0E, 0x1E, 0x0F, 0x1F));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vmrglb(vector bool char __a, vector bool char __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x18, 0x09, 0x19, 0x0A, 0x1A, 0x0B, 0x1B, 
+     0x0C, 0x1C, 0x0D, 0x1D, 0x0E, 0x1E, 0x0F, 0x1F));
+}
+
+/* vec_vmrglh */
+
+#define __builtin_altivec_vmrglh vec_vmrglh
+
+static vector short __ATTRS_o_ai
+vec_vmrglh(vector short __a, vector short __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x09, 0x18, 0x19, 0x0A, 0x0B, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x1C, 0x1D, 0x0E, 0x0F, 0x1E, 0x1F));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vmrglh(vector unsigned short __a, vector unsigned short __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x09, 0x18, 0x19, 0x0A, 0x0B, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x1C, 0x1D, 0x0E, 0x0F, 0x1E, 0x1F));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vmrglh(vector bool short __a, vector bool short __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x09, 0x18, 0x19, 0x0A, 0x0B, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x1C, 0x1D, 0x0E, 0x0F, 0x1E, 0x1F));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vmrglh(vector pixel __a, vector pixel __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x09, 0x18, 0x19, 0x0A, 0x0B, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x1C, 0x1D, 0x0E, 0x0F, 0x1E, 0x1F));
+}
+
+/* vec_vmrglw */
+
+#define __builtin_altivec_vmrglw vec_vmrglw
+
+static vector int __ATTRS_o_ai
+vec_vmrglw(vector int __a, vector int __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x09, 0x0A, 0x0B, 0x18, 0x19, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x0E, 0x0F, 0x1C, 0x1D, 0x1E, 0x1F));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vmrglw(vector unsigned int __a, vector unsigned int __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x09, 0x0A, 0x0B, 0x18, 0x19, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x0E, 0x0F, 0x1C, 0x1D, 0x1E, 0x1F));
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vmrglw(vector bool int __a, vector bool int __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x09, 0x0A, 0x0B, 0x18, 0x19, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x0E, 0x0F, 0x1C, 0x1D, 0x1E, 0x1F));
+}
+
+static vector float __ATTRS_o_ai
+vec_vmrglw(vector float __a, vector float __b)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (0x08, 0x09, 0x0A, 0x0B, 0x18, 0x19, 0x1A, 0x1B,
+     0x0C, 0x0D, 0x0E, 0x0F, 0x1C, 0x1D, 0x1E, 0x1F));
+}
+
+/* vec_mfvscr */
+
+static vector unsigned short __attribute__((__always_inline__))
+vec_mfvscr(void)
+{
+  return __builtin_altivec_mfvscr();
+}
+
+/* vec_min */
+
+static vector signed char __ATTRS_o_ai
+vec_min(vector signed char __a, vector signed char __b)
+{
+  return __builtin_altivec_vminsb(__a, __b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_min(vector bool char __a, vector signed char __b)
+{
+  return __builtin_altivec_vminsb((vector signed char)__a, __b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_min(vector signed char __a, vector bool char __b)
+{
+  return __builtin_altivec_vminsb(__a, (vector signed char)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_min(vector unsigned char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vminub(__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_min(vector bool char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vminub((vector unsigned char)__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_min(vector unsigned char __a, vector bool char __b)
+{
+  return __builtin_altivec_vminub(__a, (vector unsigned char)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_min(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vminsh(__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_min(vector bool short __a, vector short __b)
+{
+  return __builtin_altivec_vminsh((vector short)__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_min(vector short __a, vector bool short __b)
+{
+  return __builtin_altivec_vminsh(__a, (vector short)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_min(vector unsigned short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vminuh(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_min(vector bool short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vminuh((vector unsigned short)__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_min(vector unsigned short __a, vector bool short __b)
+{
+  return __builtin_altivec_vminuh(__a, (vector unsigned short)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_min(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vminsw(__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_min(vector bool int __a, vector int __b)
+{
+  return __builtin_altivec_vminsw((vector int)__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_min(vector int __a, vector bool int __b)
+{
+  return __builtin_altivec_vminsw(__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_min(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vminuw(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_min(vector bool int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vminuw((vector unsigned int)__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_min(vector unsigned int __a, vector bool int __b)
+{
+  return __builtin_altivec_vminuw(__a, (vector unsigned int)__b);
+}
+
+static vector float __ATTRS_o_ai
+vec_min(vector float __a, vector float __b)
+{
+#ifdef __VSX__
+  return __builtin_vsx_xvminsp(__a, __b);
+#else
+  return __builtin_altivec_vminfp(__a, __b);
+#endif
+}
+
+#ifdef __VSX__
+static vector double __ATTRS_o_ai
+vec_min(vector double __a, vector double __b)
+{
+  return __builtin_vsx_xvmindp(__a, __b);
+}
+#endif
+
+/* vec_vminsb */
+
+static vector signed char __ATTRS_o_ai
+vec_vminsb(vector signed char __a, vector signed char __b)
+{
+  return __builtin_altivec_vminsb(__a, __b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vminsb(vector bool char __a, vector signed char __b)
+{
+  return __builtin_altivec_vminsb((vector signed char)__a, __b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vminsb(vector signed char __a, vector bool char __b)
+{
+  return __builtin_altivec_vminsb(__a, (vector signed char)__b);
+}
+
+/* vec_vminub */
+
+static vector unsigned char __ATTRS_o_ai
+vec_vminub(vector unsigned char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vminub(__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vminub(vector bool char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vminub((vector unsigned char)__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vminub(vector unsigned char __a, vector bool char __b)
+{
+  return __builtin_altivec_vminub(__a, (vector unsigned char)__b);
+}
+
+/* vec_vminsh */
+
+static vector short __ATTRS_o_ai
+vec_vminsh(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vminsh(__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vminsh(vector bool short __a, vector short __b)
+{
+  return __builtin_altivec_vminsh((vector short)__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vminsh(vector short __a, vector bool short __b)
+{
+  return __builtin_altivec_vminsh(__a, (vector short)__b);
+}
+
+/* vec_vminuh */
+
+static vector unsigned short __ATTRS_o_ai
+vec_vminuh(vector unsigned short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vminuh(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vminuh(vector bool short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vminuh((vector unsigned short)__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vminuh(vector unsigned short __a, vector bool short __b)
+{
+  return __builtin_altivec_vminuh(__a, (vector unsigned short)__b);
+}
+
+/* vec_vminsw */
+
+static vector int __ATTRS_o_ai
+vec_vminsw(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vminsw(__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vminsw(vector bool int __a, vector int __b)
+{
+  return __builtin_altivec_vminsw((vector int)__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vminsw(vector int __a, vector bool int __b)
+{
+  return __builtin_altivec_vminsw(__a, (vector int)__b);
+}
+
+/* vec_vminuw */
+
+static vector unsigned int __ATTRS_o_ai
+vec_vminuw(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vminuw(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vminuw(vector bool int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vminuw((vector unsigned int)__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vminuw(vector unsigned int __a, vector bool int __b)
+{
+  return __builtin_altivec_vminuw(__a, (vector unsigned int)__b);
+}
+
+/* vec_vminfp */
+
+static vector float __attribute__((__always_inline__))
+vec_vminfp(vector float __a, vector float __b)
+{
+#ifdef __VSX__
+  return __builtin_vsx_xvminsp(__a, __b);
+#else
+  return __builtin_altivec_vminfp(__a, __b);
+#endif
+}
+
+/* vec_mladd */
+
+#define __builtin_altivec_vmladduhm vec_mladd
+
+static vector short __ATTRS_o_ai
+vec_mladd(vector short __a, vector short __b, vector short __c)
+{
+  return __a * __b + __c;
+}
+
+static vector short __ATTRS_o_ai
+vec_mladd(vector short __a, vector unsigned short __b, vector unsigned short __c)
+{
+  return __a * (vector short)__b + (vector short)__c;
+}
+
+static vector short __ATTRS_o_ai
+vec_mladd(vector unsigned short __a, vector short __b, vector short __c)
+{
+  return (vector short)__a * __b + __c;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_mladd(vector unsigned short __a,
+          vector unsigned short __b,
+          vector unsigned short __c)
+{
+  return __a * __b + __c;
+}
+
+/* vec_vmladduhm */
+
+static vector short __ATTRS_o_ai
+vec_vmladduhm(vector short __a, vector short __b, vector short __c)
+{
+  return __a * __b + __c;
+}
+
+static vector short __ATTRS_o_ai
+vec_vmladduhm(vector short __a, vector unsigned short __b, vector unsigned short __c)
+{
+  return __a * (vector short)__b + (vector short)__c;
+}
+
+static vector short __ATTRS_o_ai
+vec_vmladduhm(vector unsigned short __a, vector short __b, vector short __c)
+{
+  return (vector short)__a * __b + __c;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vmladduhm(vector unsigned short __a,
+              vector unsigned short __b,
+              vector unsigned short __c)
+{
+  return __a * __b + __c;
+}
+
+/* vec_mradds */
+
+static vector short __attribute__((__always_inline__))
+vec_mradds(vector short __a, vector short __b, vector short __c)
+{
+  return __builtin_altivec_vmhraddshs(__a, __b, __c);
+}
+
+/* vec_vmhraddshs */
+
+static vector short __attribute__((__always_inline__))
+vec_vmhraddshs(vector short __a, vector short __b, vector short __c)
+{
+  return __builtin_altivec_vmhraddshs(__a, __b, __c);
+}
+
+/* vec_msum */
+
+static vector int __ATTRS_o_ai
+vec_msum(vector signed char __a, vector unsigned char __b, vector int __c)
+{
+  return __builtin_altivec_vmsummbm(__a, __b, __c);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_msum(vector unsigned char __a, vector unsigned char __b, vector unsigned int __c)
+{
+  return __builtin_altivec_vmsumubm(__a, __b, __c);
+}
+
+static vector int __ATTRS_o_ai
+vec_msum(vector short __a, vector short __b, vector int __c)
+{
+  return __builtin_altivec_vmsumshm(__a, __b, __c);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_msum(vector unsigned short __a,
+         vector unsigned short __b,
+         vector unsigned int __c)
+{
+  return __builtin_altivec_vmsumuhm(__a, __b, __c);
+}
+
+/* vec_vmsummbm */
+
+static vector int __attribute__((__always_inline__))
+vec_vmsummbm(vector signed char __a, vector unsigned char __b, vector int __c)
+{
+  return __builtin_altivec_vmsummbm(__a, __b, __c);
+}
+
+/* vec_vmsumubm */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_vmsumubm(vector unsigned char __a,
+             vector unsigned char __b,
+             vector unsigned int __c)
+{
+  return __builtin_altivec_vmsumubm(__a, __b, __c);
+}
+
+/* vec_vmsumshm */
+
+static vector int __attribute__((__always_inline__))
+vec_vmsumshm(vector short __a, vector short __b, vector int __c)
+{
+  return __builtin_altivec_vmsumshm(__a, __b, __c);
+}
+
+/* vec_vmsumuhm */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_vmsumuhm(vector unsigned short __a,
+             vector unsigned short __b,
+             vector unsigned int __c)
+{
+  return __builtin_altivec_vmsumuhm(__a, __b, __c);
+}
+
+/* vec_msums */
+
+static vector int __ATTRS_o_ai
+vec_msums(vector short __a, vector short __b, vector int __c)
+{
+  return __builtin_altivec_vmsumshs(__a, __b, __c);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_msums(vector unsigned short __a,
+          vector unsigned short __b,
+          vector unsigned int __c)
+{
+  return __builtin_altivec_vmsumuhs(__a, __b, __c);
+}
+
+/* vec_vmsumshs */
+
+static vector int __attribute__((__always_inline__))
+vec_vmsumshs(vector short __a, vector short __b, vector int __c)
+{
+  return __builtin_altivec_vmsumshs(__a, __b, __c);
+}
+
+/* vec_vmsumuhs */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_vmsumuhs(vector unsigned short __a,
+             vector unsigned short __b,
+             vector unsigned int __c)
+{
+  return __builtin_altivec_vmsumuhs(__a, __b, __c);
+}
+
+/* vec_mtvscr */
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector signed char __a)
+{
+  __builtin_altivec_mtvscr((vector int)__a);
+}
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector unsigned char __a)
+{
+  __builtin_altivec_mtvscr((vector int)__a);
+}
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector bool char __a)
+{
+  __builtin_altivec_mtvscr((vector int)__a);
+}
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector short __a)
+{
+  __builtin_altivec_mtvscr((vector int)__a);
+}
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector unsigned short __a)
+{
+  __builtin_altivec_mtvscr((vector int)__a);
+}
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector bool short __a)
+{
+  __builtin_altivec_mtvscr((vector int)__a);
+}
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector pixel __a)
+{
+  __builtin_altivec_mtvscr((vector int)__a);
+}
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector int __a)
+{
+  __builtin_altivec_mtvscr((vector int)__a);
+}
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector unsigned int __a)
+{
+  __builtin_altivec_mtvscr((vector int)__a);
+}
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector bool int __a)
+{
+  __builtin_altivec_mtvscr((vector int)__a);
+}
+
+static void __ATTRS_o_ai
+vec_mtvscr(vector float __a)
+{
+  __builtin_altivec_mtvscr((vector int)__a);
+}
+
+/* The vmulos* and vmules* instructions have a big endian bias, so
+   we must reverse the meaning of "even" and "odd" for little endian.  */
+
+/* vec_mule */
+
+static vector short __ATTRS_o_ai
+vec_mule(vector signed char __a, vector signed char __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vmulosb(__a, __b);
+#else
+  return __builtin_altivec_vmulesb(__a, __b);
+#endif
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_mule(vector unsigned char __a, vector unsigned char __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vmuloub(__a, __b);
+#else
+  return __builtin_altivec_vmuleub(__a, __b);
+#endif
+}
+
+static vector int __ATTRS_o_ai
+vec_mule(vector short __a, vector short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vmulosh(__a, __b);
+#else
+  return __builtin_altivec_vmulesh(__a, __b);
+#endif
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_mule(vector unsigned short __a, vector unsigned short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vmulouh(__a, __b);
+#else
+  return __builtin_altivec_vmuleuh(__a, __b);
+#endif
+}
+
+/* vec_vmulesb */
+
+static vector short __attribute__((__always_inline__))
+vec_vmulesb(vector signed char __a, vector signed char __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vmulosb(__a, __b);
+#else
+  return __builtin_altivec_vmulesb(__a, __b);
+#endif
+}
+
+/* vec_vmuleub */
+
+static vector unsigned short __attribute__((__always_inline__))
+vec_vmuleub(vector unsigned char __a, vector unsigned char __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vmuloub(__a, __b);
+#else
+  return __builtin_altivec_vmuleub(__a, __b);
+#endif
+}
+
+/* vec_vmulesh */
+
+static vector int __attribute__((__always_inline__))
+vec_vmulesh(vector short __a, vector short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vmulosh(__a, __b);
+#else
+  return __builtin_altivec_vmulesh(__a, __b);
+#endif
+}
+
+/* vec_vmuleuh */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_vmuleuh(vector unsigned short __a, vector unsigned short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vmulouh(__a, __b);
+#else
+  return __builtin_altivec_vmuleuh(__a, __b);
+#endif
+}
+
+/* vec_mulo */
+
+static vector short __ATTRS_o_ai
+vec_mulo(vector signed char __a, vector signed char __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vmulesb(__a, __b);
+#else
+  return __builtin_altivec_vmulosb(__a, __b);
+#endif
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_mulo(vector unsigned char __a, vector unsigned char __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vmuleub(__a, __b);
+#else
+  return __builtin_altivec_vmuloub(__a, __b);
+#endif
+}
+
+static vector int __ATTRS_o_ai
+vec_mulo(vector short __a, vector short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vmulesh(__a, __b);
+#else
+  return __builtin_altivec_vmulosh(__a, __b);
+#endif
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_mulo(vector unsigned short __a, vector unsigned short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vmuleuh(__a, __b);
+#else
+  return __builtin_altivec_vmulouh(__a, __b);
+#endif
+}
+
+/* vec_vmulosb */
+
+static vector short __attribute__((__always_inline__))
+vec_vmulosb(vector signed char __a, vector signed char __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vmulesb(__a, __b);
+#else
+  return __builtin_altivec_vmulosb(__a, __b);
+#endif
+}
+
+/* vec_vmuloub */
+
+static vector unsigned short __attribute__((__always_inline__))
+vec_vmuloub(vector unsigned char __a, vector unsigned char __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vmuleub(__a, __b);
+#else
+  return __builtin_altivec_vmuloub(__a, __b);
+#endif
+}
+
+/* vec_vmulosh */
+
+static vector int __attribute__((__always_inline__))
+vec_vmulosh(vector short __a, vector short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vmulesh(__a, __b);
+#else
+  return __builtin_altivec_vmulosh(__a, __b);
+#endif
+}
+
+/* vec_vmulouh */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_vmulouh(vector unsigned short __a, vector unsigned short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vmuleuh(__a, __b);
+#else
+  return __builtin_altivec_vmulouh(__a, __b);
+#endif
+}
+
+/* vec_nmsub */
+
+static vector float __attribute__((__always_inline__))
+vec_nmsub(vector float __a, vector float __b, vector float __c)
+{
+  return __builtin_altivec_vnmsubfp(__a, __b, __c);
+}
+
+/* vec_vnmsubfp */
+
+static vector float __attribute__((__always_inline__))
+vec_vnmsubfp(vector float __a, vector float __b, vector float __c)
+{
+  return __builtin_altivec_vnmsubfp(__a, __b, __c);
+}
+
+/* vec_nor */
+
+#define __builtin_altivec_vnor vec_nor
+
+static vector signed char __ATTRS_o_ai
+vec_nor(vector signed char __a, vector signed char __b)
+{
+  return ~(__a | __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_nor(vector unsigned char __a, vector unsigned char __b)
+{
+  return ~(__a | __b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_nor(vector bool char __a, vector bool char __b)
+{
+  return ~(__a | __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_nor(vector short __a, vector short __b)
+{
+  return ~(__a | __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_nor(vector unsigned short __a, vector unsigned short __b)
+{
+  return ~(__a | __b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_nor(vector bool short __a, vector bool short __b)
+{
+  return ~(__a | __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_nor(vector int __a, vector int __b)
+{
+  return ~(__a | __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_nor(vector unsigned int __a, vector unsigned int __b)
+{
+  return ~(__a | __b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_nor(vector bool int __a, vector bool int __b)
+{
+  return ~(__a | __b);
+}
+
+static vector float __ATTRS_o_ai
+vec_nor(vector float __a, vector float __b)
+{
+  vector unsigned int __res = ~((vector unsigned int)__a | (vector unsigned int)__b);
+  return (vector float)__res;
+}
+
+/* vec_vnor */
+
+static vector signed char __ATTRS_o_ai
+vec_vnor(vector signed char __a, vector signed char __b)
+{
+  return ~(__a | __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vnor(vector unsigned char __a, vector unsigned char __b)
+{
+  return ~(__a | __b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vnor(vector bool char __a, vector bool char __b)
+{
+  return ~(__a | __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vnor(vector short __a, vector short __b)
+{
+  return ~(__a | __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vnor(vector unsigned short __a, vector unsigned short __b)
+{
+  return ~(__a | __b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vnor(vector bool short __a, vector bool short __b)
+{
+  return ~(__a | __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vnor(vector int __a, vector int __b)
+{
+  return ~(__a | __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vnor(vector unsigned int __a, vector unsigned int __b)
+{
+  return ~(__a | __b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vnor(vector bool int __a, vector bool int __b)
+{
+  return ~(__a | __b);
+}
+
+static vector float __ATTRS_o_ai
+vec_vnor(vector float __a, vector float __b)
+{
+  vector unsigned int __res = ~((vector unsigned int)__a | (vector unsigned int)__b);
+  return (vector float)__res;
+}
+
+/* vec_or */
+
+#define __builtin_altivec_vor vec_or
+
+static vector signed char __ATTRS_o_ai
+vec_or(vector signed char __a, vector signed char __b)
+{
+  return __a | __b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_or(vector bool char __a, vector signed char __b)
+{
+  return (vector signed char)__a | __b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_or(vector signed char __a, vector bool char __b)
+{
+  return __a | (vector signed char)__b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_or(vector unsigned char __a, vector unsigned char __b)
+{
+  return __a | __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_or(vector bool char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)__a | __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_or(vector unsigned char __a, vector bool char __b)
+{
+  return __a | (vector unsigned char)__b;
+}
+
+static vector bool char __ATTRS_o_ai
+vec_or(vector bool char __a, vector bool char __b)
+{
+  return __a | __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_or(vector short __a, vector short __b)
+{
+  return __a | __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_or(vector bool short __a, vector short __b)
+{
+  return (vector short)__a | __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_or(vector short __a, vector bool short __b)
+{
+  return __a | (vector short)__b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_or(vector unsigned short __a, vector unsigned short __b)
+{
+  return __a | __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_or(vector bool short __a, vector unsigned short __b)
+{
+  return (vector unsigned short)__a | __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_or(vector unsigned short __a, vector bool short __b)
+{
+  return __a | (vector unsigned short)__b;
+}
+
+static vector bool short __ATTRS_o_ai
+vec_or(vector bool short __a, vector bool short __b)
+{
+  return __a | __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_or(vector int __a, vector int __b)
+{
+  return __a | __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_or(vector bool int __a, vector int __b)
+{
+  return (vector int)__a | __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_or(vector int __a, vector bool int __b)
+{
+  return __a | (vector int)__b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_or(vector unsigned int __a, vector unsigned int __b)
+{
+  return __a | __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_or(vector bool int __a, vector unsigned int __b)
+{
+  return (vector unsigned int)__a | __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_or(vector unsigned int __a, vector bool int __b)
+{
+  return __a | (vector unsigned int)__b;
+}
+
+static vector bool int __ATTRS_o_ai
+vec_or(vector bool int __a, vector bool int __b)
+{
+  return __a | __b;
+}
+
+static vector float __ATTRS_o_ai
+vec_or(vector float __a, vector float __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a | (vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+static vector float __ATTRS_o_ai
+vec_or(vector bool int __a, vector float __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a | (vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+static vector float __ATTRS_o_ai
+vec_or(vector float __a, vector bool int __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a | (vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+/* vec_vor */
+
+static vector signed char __ATTRS_o_ai
+vec_vor(vector signed char __a, vector signed char __b)
+{
+  return __a | __b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vor(vector bool char __a, vector signed char __b)
+{
+  return (vector signed char)__a | __b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vor(vector signed char __a, vector bool char __b)
+{
+  return __a | (vector signed char)__b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vor(vector unsigned char __a, vector unsigned char __b)
+{
+  return __a | __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vor(vector bool char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)__a | __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vor(vector unsigned char __a, vector bool char __b)
+{
+  return __a | (vector unsigned char)__b;
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vor(vector bool char __a, vector bool char __b)
+{
+  return __a | __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vor(vector short __a, vector short __b)
+{
+  return __a | __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vor(vector bool short __a, vector short __b)
+{
+  return (vector short)__a | __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vor(vector short __a, vector bool short __b)
+{
+  return __a | (vector short)__b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vor(vector unsigned short __a, vector unsigned short __b)
+{
+  return __a | __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vor(vector bool short __a, vector unsigned short __b)
+{
+  return (vector unsigned short)__a | __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vor(vector unsigned short __a, vector bool short __b)
+{
+  return __a | (vector unsigned short)__b;
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vor(vector bool short __a, vector bool short __b)
+{
+  return __a | __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vor(vector int __a, vector int __b)
+{
+  return __a | __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vor(vector bool int __a, vector int __b)
+{
+  return (vector int)__a | __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vor(vector int __a, vector bool int __b)
+{
+  return __a | (vector int)__b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vor(vector unsigned int __a, vector unsigned int __b)
+{
+  return __a | __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vor(vector bool int __a, vector unsigned int __b)
+{
+  return (vector unsigned int)__a | __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vor(vector unsigned int __a, vector bool int __b)
+{
+  return __a | (vector unsigned int)__b;
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vor(vector bool int __a, vector bool int __b)
+{
+  return __a | __b;
+}
+
+static vector float __ATTRS_o_ai
+vec_vor(vector float __a, vector float __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a | (vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+static vector float __ATTRS_o_ai
+vec_vor(vector bool int __a, vector float __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a | (vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+static vector float __ATTRS_o_ai
+vec_vor(vector float __a, vector bool int __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a | (vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+/* vec_pack */
+
+/* The various vector pack instructions have a big-endian bias, so for
+   little endian we must handle reversed element numbering.  */
+
+static vector signed char __ATTRS_o_ai
+vec_pack(vector signed short __a, vector signed short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector signed char)vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x02, 0x04, 0x06, 0x08, 0x0A, 0x0C, 0x0E,
+     0x10, 0x12, 0x14, 0x16, 0x18, 0x1A, 0x1C, 0x1E));
+#else
+  return (vector signed char)vec_perm(__a, __b, (vector unsigned char)
+    (0x01, 0x03, 0x05, 0x07, 0x09, 0x0B, 0x0D, 0x0F,
+     0x11, 0x13, 0x15, 0x17, 0x19, 0x1B, 0x1D, 0x1F));
+#endif
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_pack(vector unsigned short __a, vector unsigned short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector unsigned char)vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x02, 0x04, 0x06, 0x08, 0x0A, 0x0C, 0x0E,
+     0x10, 0x12, 0x14, 0x16, 0x18, 0x1A, 0x1C, 0x1E));
+#else
+  return (vector unsigned char)vec_perm(__a, __b, (vector unsigned char)
+    (0x01, 0x03, 0x05, 0x07, 0x09, 0x0B, 0x0D, 0x0F,
+     0x11, 0x13, 0x15, 0x17, 0x19, 0x1B, 0x1D, 0x1F));
+#endif
+}
+
+static vector bool char __ATTRS_o_ai
+vec_pack(vector bool short __a, vector bool short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector bool char)vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x02, 0x04, 0x06, 0x08, 0x0A, 0x0C, 0x0E,
+     0x10, 0x12, 0x14, 0x16, 0x18, 0x1A, 0x1C, 0x1E));
+#else
+  return (vector bool char)vec_perm(__a, __b, (vector unsigned char)
+    (0x01, 0x03, 0x05, 0x07, 0x09, 0x0B, 0x0D, 0x0F,
+     0x11, 0x13, 0x15, 0x17, 0x19, 0x1B, 0x1D, 0x1F));
+#endif
+}
+
+static vector short __ATTRS_o_ai
+vec_pack(vector int __a, vector int __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector short)vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x04, 0x05, 0x08, 0x09, 0x0C, 0x0D,
+     0x10, 0x11, 0x14, 0x15, 0x18, 0x19, 0x1C, 0x1D));
+#else
+  return (vector short)vec_perm(__a, __b, (vector unsigned char)
+    (0x02, 0x03, 0x06, 0x07, 0x0A, 0x0B, 0x0E, 0x0F,
+     0x12, 0x13, 0x16, 0x17, 0x1A, 0x1B, 0x1E, 0x1F));
+#endif
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_pack(vector unsigned int __a, vector unsigned int __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector unsigned short)vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x04, 0x05, 0x08, 0x09, 0x0C, 0x0D,
+     0x10, 0x11, 0x14, 0x15, 0x18, 0x19, 0x1C, 0x1D));
+#else
+  return (vector unsigned short)vec_perm(__a, __b, (vector unsigned char)
+    (0x02, 0x03, 0x06, 0x07, 0x0A, 0x0B, 0x0E, 0x0F,
+     0x12, 0x13, 0x16, 0x17, 0x1A, 0x1B, 0x1E, 0x1F));
+#endif
+}
+
+static vector bool short __ATTRS_o_ai
+vec_pack(vector bool int __a, vector bool int __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector bool short)vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x04, 0x05, 0x08, 0x09, 0x0C, 0x0D,
+     0x10, 0x11, 0x14, 0x15, 0x18, 0x19, 0x1C, 0x1D));
+#else
+  return (vector bool short)vec_perm(__a, __b, (vector unsigned char)
+    (0x02, 0x03, 0x06, 0x07, 0x0A, 0x0B, 0x0E, 0x0F,
+     0x12, 0x13, 0x16, 0x17, 0x1A, 0x1B, 0x1E, 0x1F));
+#endif
+}
+
+/* vec_vpkuhum */
+
+#define __builtin_altivec_vpkuhum vec_vpkuhum
+
+static vector signed char __ATTRS_o_ai
+vec_vpkuhum(vector signed short __a, vector signed short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector signed char)vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x02, 0x04, 0x06, 0x08, 0x0A, 0x0C, 0x0E,
+     0x10, 0x12, 0x14, 0x16, 0x18, 0x1A, 0x1C, 0x1E));
+#else
+  return (vector signed char)vec_perm(__a, __b, (vector unsigned char)
+    (0x01, 0x03, 0x05, 0x07, 0x09, 0x0B, 0x0D, 0x0F,
+     0x11, 0x13, 0x15, 0x17, 0x19, 0x1B, 0x1D, 0x1F));
+#endif
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vpkuhum(vector unsigned short __a, vector unsigned short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector unsigned char)vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x02, 0x04, 0x06, 0x08, 0x0A, 0x0C, 0x0E,
+     0x10, 0x12, 0x14, 0x16, 0x18, 0x1A, 0x1C, 0x1E));
+#else
+  return (vector unsigned char)vec_perm(__a, __b, (vector unsigned char)
+    (0x01, 0x03, 0x05, 0x07, 0x09, 0x0B, 0x0D, 0x0F,
+     0x11, 0x13, 0x15, 0x17, 0x19, 0x1B, 0x1D, 0x1F));
+#endif
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vpkuhum(vector bool short __a, vector bool short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector bool char)vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x02, 0x04, 0x06, 0x08, 0x0A, 0x0C, 0x0E,
+     0x10, 0x12, 0x14, 0x16, 0x18, 0x1A, 0x1C, 0x1E));
+#else
+  return (vector bool char)vec_perm(__a, __b, (vector unsigned char)
+    (0x01, 0x03, 0x05, 0x07, 0x09, 0x0B, 0x0D, 0x0F,
+     0x11, 0x13, 0x15, 0x17, 0x19, 0x1B, 0x1D, 0x1F));
+#endif
+}
+
+/* vec_vpkuwum */
+
+#define __builtin_altivec_vpkuwum vec_vpkuwum
+
+static vector short __ATTRS_o_ai
+vec_vpkuwum(vector int __a, vector int __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector short)vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x04, 0x05, 0x08, 0x09, 0x0C, 0x0D,
+     0x10, 0x11, 0x14, 0x15, 0x18, 0x19, 0x1C, 0x1D));
+#else
+  return (vector short)vec_perm(__a, __b, (vector unsigned char)
+    (0x02, 0x03, 0x06, 0x07, 0x0A, 0x0B, 0x0E, 0x0F,
+     0x12, 0x13, 0x16, 0x17, 0x1A, 0x1B, 0x1E, 0x1F));
+#endif
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vpkuwum(vector unsigned int __a, vector unsigned int __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector unsigned short)vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x04, 0x05, 0x08, 0x09, 0x0C, 0x0D,
+     0x10, 0x11, 0x14, 0x15, 0x18, 0x19, 0x1C, 0x1D));
+#else
+  return (vector unsigned short)vec_perm(__a, __b, (vector unsigned char)
+    (0x02, 0x03, 0x06, 0x07, 0x0A, 0x0B, 0x0E, 0x0F,
+     0x12, 0x13, 0x16, 0x17, 0x1A, 0x1B, 0x1E, 0x1F));
+#endif
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vpkuwum(vector bool int __a, vector bool int __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector bool short)vec_perm(__a, __b, (vector unsigned char)
+    (0x00, 0x01, 0x04, 0x05, 0x08, 0x09, 0x0C, 0x0D,
+     0x10, 0x11, 0x14, 0x15, 0x18, 0x19, 0x1C, 0x1D));
+#else
+  return (vector bool short)vec_perm(__a, __b, (vector unsigned char)
+    (0x02, 0x03, 0x06, 0x07, 0x0A, 0x0B, 0x0E, 0x0F,
+     0x12, 0x13, 0x16, 0x17, 0x1A, 0x1B, 0x1E, 0x1F));
+#endif
+}
+
+/* vec_packpx */
+
+static vector pixel __attribute__((__always_inline__))
+vec_packpx(vector unsigned int __a, vector unsigned int __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector pixel)__builtin_altivec_vpkpx(__b, __a);
+#else
+  return (vector pixel)__builtin_altivec_vpkpx(__a, __b);
+#endif
+}
+
+/* vec_vpkpx */
+
+static vector pixel __attribute__((__always_inline__))
+vec_vpkpx(vector unsigned int __a, vector unsigned int __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector pixel)__builtin_altivec_vpkpx(__b, __a);
+#else
+  return (vector pixel)__builtin_altivec_vpkpx(__a, __b);
+#endif
+}
+
+/* vec_packs */
+
+static vector signed char __ATTRS_o_ai
+vec_packs(vector short __a, vector short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vpkshss(__b, __a);
+#else
+  return __builtin_altivec_vpkshss(__a, __b);
+#endif
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_packs(vector unsigned short __a, vector unsigned short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vpkuhus(__b, __a);
+#else
+  return __builtin_altivec_vpkuhus(__a, __b);
+#endif
+}
+
+static vector signed short __ATTRS_o_ai
+vec_packs(vector int __a, vector int __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vpkswss(__b, __a);
+#else
+  return __builtin_altivec_vpkswss(__a, __b);
+#endif
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_packs(vector unsigned int __a, vector unsigned int __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vpkuwus(__b, __a);
+#else
+  return __builtin_altivec_vpkuwus(__a, __b);
+#endif
+}
+
+/* vec_vpkshss */
+
+static vector signed char __attribute__((__always_inline__))
+vec_vpkshss(vector short __a, vector short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vpkshss(__b, __a);
+#else
+  return __builtin_altivec_vpkshss(__a, __b);
+#endif
+}
+
+/* vec_vpkuhus */
+
+static vector unsigned char __attribute__((__always_inline__))
+vec_vpkuhus(vector unsigned short __a, vector unsigned short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vpkuhus(__b, __a);
+#else
+  return __builtin_altivec_vpkuhus(__a, __b);
+#endif
+}
+
+/* vec_vpkswss */
+
+static vector signed short __attribute__((__always_inline__))
+vec_vpkswss(vector int __a, vector int __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vpkswss(__b, __a);
+#else
+  return __builtin_altivec_vpkswss(__a, __b);
+#endif
+}
+
+/* vec_vpkuwus */
+
+static vector unsigned short __attribute__((__always_inline__))
+vec_vpkuwus(vector unsigned int __a, vector unsigned int __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vpkuwus(__b, __a);
+#else
+  return __builtin_altivec_vpkuwus(__a, __b);
+#endif
+}
+
+/* vec_packsu */
+
+static vector unsigned char __ATTRS_o_ai
+vec_packsu(vector short __a, vector short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vpkshus(__b, __a);
+#else
+  return __builtin_altivec_vpkshus(__a, __b);
+#endif
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_packsu(vector unsigned short __a, vector unsigned short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vpkuhus(__b, __a);
+#else
+  return __builtin_altivec_vpkuhus(__a, __b);
+#endif
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_packsu(vector int __a, vector int __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vpkswus(__b, __a);
+#else
+  return __builtin_altivec_vpkswus(__a, __b);
+#endif
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_packsu(vector unsigned int __a, vector unsigned int __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vpkuwus(__b, __a);
+#else
+  return __builtin_altivec_vpkuwus(__a, __b);
+#endif
+}
+
+/* vec_vpkshus */
+
+static vector unsigned char __ATTRS_o_ai
+vec_vpkshus(vector short __a, vector short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vpkshus(__b, __a);
+#else
+  return __builtin_altivec_vpkshus(__a, __b);
+#endif
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vpkshus(vector unsigned short __a, vector unsigned short __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vpkuhus(__b, __a);
+#else
+  return __builtin_altivec_vpkuhus(__a, __b);
+#endif
+}
+
+/* vec_vpkswus */
+
+static vector unsigned short __ATTRS_o_ai
+vec_vpkswus(vector int __a, vector int __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vpkswus(__b, __a);
+#else
+  return __builtin_altivec_vpkswus(__a, __b);
+#endif
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vpkswus(vector unsigned int __a, vector unsigned int __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vpkuwus(__b, __a);
+#else
+  return __builtin_altivec_vpkuwus(__a, __b);
+#endif
+}
+
+/* vec_perm */
+
+// The vperm instruction is defined architecturally with a big-endian bias.
+// For little endian, we swap the input operands and invert the permute
+// control vector.  Only the rightmost 5 bits matter, so we could use
+// a vector of all 31s instead of all 255s to perform the inversion.
+// However, when the PCV is not a constant, using 255 has an advantage
+// in that the vec_xor can be recognized as a vec_nor (and for P8 and
+// later, possibly a vec_nand).
+
+vector signed char __ATTRS_o_ai
+vec_perm(vector signed char __a, vector signed char __b, vector unsigned char __c)
+{
+#ifdef __LITTLE_ENDIAN__
+  vector unsigned char __d = {255,255,255,255,255,255,255,255,
+                              255,255,255,255,255,255,255,255};
+  __d = vec_xor(__c, __d);
+  return (vector signed char)
+           __builtin_altivec_vperm_4si((vector int)__b, (vector int)__a, __d);
+#else
+  return (vector signed char)
+           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __c);
+#endif
+}
+
+vector unsigned char __ATTRS_o_ai
+vec_perm(vector unsigned char __a,
+         vector unsigned char __b,
+         vector unsigned char __c)
+{
+#ifdef __LITTLE_ENDIAN__
+  vector unsigned char __d = {255,255,255,255,255,255,255,255,
+                              255,255,255,255,255,255,255,255};
+  __d = vec_xor(__c, __d);
+  return (vector unsigned char)
+           __builtin_altivec_vperm_4si((vector int)__b, (vector int)__a, __d);
+#else
+  return (vector unsigned char)
+           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __c);
+#endif
+}
+
+vector bool char __ATTRS_o_ai
+vec_perm(vector bool char __a, vector bool char __b, vector unsigned char __c)
+{
+#ifdef __LITTLE_ENDIAN__
+  vector unsigned char __d = {255,255,255,255,255,255,255,255,
+                              255,255,255,255,255,255,255,255};
+  __d = vec_xor(__c, __d);
+  return (vector bool char)
+           __builtin_altivec_vperm_4si((vector int)__b, (vector int)__a, __d);
+#else
+  return (vector bool char)
+           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __c);
+#endif
+}
+
+vector short __ATTRS_o_ai
+vec_perm(vector short __a, vector short __b, vector unsigned char __c)
+{
+#ifdef __LITTLE_ENDIAN__
+  vector unsigned char __d = {255,255,255,255,255,255,255,255,
+                              255,255,255,255,255,255,255,255};
+  __d = vec_xor(__c, __d);
+  return (vector short)
+           __builtin_altivec_vperm_4si((vector int)__b, (vector int)__a, __d);
+#else
+  return (vector short)
+           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __c);
+#endif
+}
+
+vector unsigned short __ATTRS_o_ai
+vec_perm(vector unsigned short __a,
+         vector unsigned short __b,
+         vector unsigned char __c)
+{
+#ifdef __LITTLE_ENDIAN__
+  vector unsigned char __d = {255,255,255,255,255,255,255,255,
+                              255,255,255,255,255,255,255,255};
+  __d = vec_xor(__c, __d);
+  return (vector unsigned short)
+           __builtin_altivec_vperm_4si((vector int)__b, (vector int)__a, __d);
+#else
+  return (vector unsigned short)
+           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __c);
+#endif
+}
+
+vector bool short __ATTRS_o_ai
+vec_perm(vector bool short __a, vector bool short __b, vector unsigned char __c)
+{
+#ifdef __LITTLE_ENDIAN__
+  vector unsigned char __d = {255,255,255,255,255,255,255,255,
+                              255,255,255,255,255,255,255,255};
+  __d = vec_xor(__c, __d);
+  return (vector bool short)
+           __builtin_altivec_vperm_4si((vector int)__b, (vector int)__a, __d);
+#else
+  return (vector bool short)
+           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __c);
+#endif
+}
+
+vector pixel __ATTRS_o_ai
+vec_perm(vector pixel __a, vector pixel __b, vector unsigned char __c)
+{
+#ifdef __LITTLE_ENDIAN__
+  vector unsigned char __d = {255,255,255,255,255,255,255,255,
+                              255,255,255,255,255,255,255,255};
+  __d = vec_xor(__c, __d);
+  return (vector pixel)
+           __builtin_altivec_vperm_4si((vector int)__b, (vector int)__a, __d);
+#else
+  return (vector pixel)
+           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __c);
+#endif
+}
+
+vector int __ATTRS_o_ai
+vec_perm(vector int __a, vector int __b, vector unsigned char __c)
+{
+#ifdef __LITTLE_ENDIAN__
+  vector unsigned char __d = {255,255,255,255,255,255,255,255,
+                              255,255,255,255,255,255,255,255};
+  __d = vec_xor(__c, __d);
+  return (vector int)__builtin_altivec_vperm_4si(__b, __a, __d);
+#else
+  return (vector int)__builtin_altivec_vperm_4si(__a, __b, __c);
+#endif
+}
+
+vector unsigned int __ATTRS_o_ai
+vec_perm(vector unsigned int __a, vector unsigned int __b, vector unsigned char __c)
+{
+#ifdef __LITTLE_ENDIAN__
+  vector unsigned char __d = {255,255,255,255,255,255,255,255,
+                              255,255,255,255,255,255,255,255};
+  __d = vec_xor(__c, __d);
+  return (vector unsigned int)
+           __builtin_altivec_vperm_4si((vector int)__b, (vector int)__a, __d);
+#else
+  return (vector unsigned int)
+           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __c);
+#endif
+}
+
+vector bool int __ATTRS_o_ai
+vec_perm(vector bool int __a, vector bool int __b, vector unsigned char __c)
+{
+#ifdef __LITTLE_ENDIAN__
+  vector unsigned char __d = {255,255,255,255,255,255,255,255,
+                              255,255,255,255,255,255,255,255};
+  __d = vec_xor(__c, __d);
+  return (vector bool int)
+           __builtin_altivec_vperm_4si((vector int)__b, (vector int)__a, __d);
+#else
+  return (vector bool int)
+           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __c);
+#endif
+}
+
+vector float __ATTRS_o_ai
+vec_perm(vector float __a, vector float __b, vector unsigned char __c)
+{
+#ifdef __LITTLE_ENDIAN__
+  vector unsigned char __d = {255,255,255,255,255,255,255,255,
+                              255,255,255,255,255,255,255,255};
+  __d = vec_xor(__c, __d);
+  return (vector float)
+           __builtin_altivec_vperm_4si((vector int)__b, (vector int)__a, __d);
+#else
+  return (vector float)
+           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __c);
+#endif
+}
+
+#ifdef __VSX__
+vector long long __ATTRS_o_ai
+vec_perm(vector long long __a, vector long long __b, vector unsigned char __c)
+{
+#ifdef __LITTLE_ENDIAN__
+  vector unsigned char __d = {255,255,255,255,255,255,255,255,
+                              255,255,255,255,255,255,255,255};
+  __d = vec_xor(__c, __d);
+  return (vector long long)__builtin_altivec_vperm_4si(__b, __a, __d);
+#else
+  return (vector long long)__builtin_altivec_vperm_4si(__a, __b, __c);
+#endif
+}
+
+vector unsigned long long __ATTRS_o_ai
+vec_perm(vector unsigned long long __a, vector unsigned long long __b,
+         vector unsigned char __c)
+{
+#ifdef __LITTLE_ENDIAN__
+  vector unsigned char __d = {255,255,255,255,255,255,255,255,
+                              255,255,255,255,255,255,255,255};
+  __d = vec_xor(__c, __d);
+  return (vector unsigned long long)
+           __builtin_altivec_vperm_4si((vector int)__b, (vector int)__a, __d);
+#else
+  return (vector unsigned long long)
+           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __c);
+#endif
+}
+
+vector double __ATTRS_o_ai
+vec_perm(vector double __a, vector double __b, vector unsigned char __c)
+{
+#ifdef __LITTLE_ENDIAN__
+  vector unsigned char __d = {255,255,255,255,255,255,255,255,
+                              255,255,255,255,255,255,255,255};
+  __d = vec_xor(__c, __d);
+  return (vector double)
+           __builtin_altivec_vperm_4si((vector int)__b, (vector int)__a, __d);
+#else
+  return (vector double)
+           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __c);
+#endif
+}
+#endif
+
+/* vec_vperm */
+
+static vector signed char __ATTRS_o_ai
+vec_vperm(vector signed char __a, vector signed char __b, vector unsigned char __c)
+{
+  return vec_perm(__a, __b, __c);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vperm(vector unsigned char __a,
+          vector unsigned char __b,
+          vector unsigned char __c)
+{
+  return vec_perm(__a, __b, __c);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vperm(vector bool char __a, vector bool char __b, vector unsigned char __c)
+{
+  return vec_perm(__a, __b, __c);
+}
+
+static vector short __ATTRS_o_ai
+vec_vperm(vector short __a, vector short __b, vector unsigned char __c)
+{
+  return vec_perm(__a, __b, __c);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vperm(vector unsigned short __a,
+          vector unsigned short __b,
+          vector unsigned char __c)
+{
+  return vec_perm(__a, __b, __c);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vperm(vector bool short __a, vector bool short __b, vector unsigned char __c)
+{
+  return vec_perm(__a, __b, __c);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vperm(vector pixel __a, vector pixel __b, vector unsigned char __c)
+{
+  return vec_perm(__a, __b, __c);
+}
+
+static vector int __ATTRS_o_ai
+vec_vperm(vector int __a, vector int __b, vector unsigned char __c)
+{
+  return vec_perm(__a, __b, __c);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vperm(vector unsigned int __a, vector unsigned int __b, vector unsigned char __c)
+{
+  return vec_perm(__a, __b, __c);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vperm(vector bool int __a, vector bool int __b, vector unsigned char __c)
+{
+  return vec_perm(__a, __b, __c);
+}
+
+static vector float __ATTRS_o_ai
+vec_vperm(vector float __a, vector float __b, vector unsigned char __c)
+{
+  return vec_perm(__a, __b, __c);
+}
+
+#ifdef __VSX__
+static vector long long __ATTRS_o_ai
+vec_vperm(vector long long __a, vector long long __b, vector unsigned char __c)
+{
+  return vec_perm(__a, __b, __c);
+}
+
+static vector unsigned long long __ATTRS_o_ai
+vec_vperm(vector unsigned long long __a, vector unsigned long long __b,
+          vector unsigned char __c)
+{
+  return vec_perm(__a, __b, __c);
+}
+
+static vector double __ATTRS_o_ai
+vec_vperm(vector double __a, vector double __b, vector unsigned char __c)
+{
+  return vec_perm(__a, __b, __c);
+}
+#endif
+
+/* vec_re */
+
+static vector float __attribute__((__always_inline__))
+vec_re(vector float __a)
+{
+  return __builtin_altivec_vrefp(__a);
+}
+
+/* vec_vrefp */
+
+static vector float __attribute__((__always_inline__))
+vec_vrefp(vector float __a)
+{
+  return __builtin_altivec_vrefp(__a);
+}
+
+/* vec_rl */
+
+static vector signed char __ATTRS_o_ai
+vec_rl(vector signed char __a, vector unsigned char __b)
+{
+  return (vector signed char)__builtin_altivec_vrlb((vector char)__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_rl(vector unsigned char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)__builtin_altivec_vrlb((vector char)__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_rl(vector short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vrlh(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_rl(vector unsigned short __a, vector unsigned short __b)
+{
+  return (vector unsigned short)__builtin_altivec_vrlh((vector short)__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_rl(vector int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vrlw(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_rl(vector unsigned int __a, vector unsigned int __b)
+{
+  return (vector unsigned int)__builtin_altivec_vrlw((vector int)__a, __b);
+}
+
+/* vec_vrlb */
+
+static vector signed char __ATTRS_o_ai
+vec_vrlb(vector signed char __a, vector unsigned char __b)
+{
+  return (vector signed char)__builtin_altivec_vrlb((vector char)__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vrlb(vector unsigned char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)__builtin_altivec_vrlb((vector char)__a, __b);
+}
+
+/* vec_vrlh */
+
+static vector short __ATTRS_o_ai
+vec_vrlh(vector short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vrlh(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vrlh(vector unsigned short __a, vector unsigned short __b)
+{
+  return (vector unsigned short)__builtin_altivec_vrlh((vector short)__a, __b);
+}
+
+/* vec_vrlw */
+
+static vector int __ATTRS_o_ai
+vec_vrlw(vector int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vrlw(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vrlw(vector unsigned int __a, vector unsigned int __b)
+{
+  return (vector unsigned int)__builtin_altivec_vrlw((vector int)__a, __b);
+}
+
+/* vec_round */
+
+static vector float __attribute__((__always_inline__))
+vec_round(vector float __a)
+{
+  return __builtin_altivec_vrfin(__a);
+}
+
+/* vec_vrfin */
+
+static vector float __attribute__((__always_inline__))
+vec_vrfin(vector float __a)
+{
+  return __builtin_altivec_vrfin(__a);
+}
+
+/* vec_rsqrte */
+
+static __vector float __attribute__((__always_inline__))
+vec_rsqrte(vector float __a)
+{
+  return __builtin_altivec_vrsqrtefp(__a);
+}
+
+/* vec_vrsqrtefp */
+
+static __vector float __attribute__((__always_inline__))
+vec_vrsqrtefp(vector float __a)
+{
+  return __builtin_altivec_vrsqrtefp(__a);
+}
+
+/* vec_sel */
+
+#define __builtin_altivec_vsel_4si vec_sel
+
+static vector signed char __ATTRS_o_ai
+vec_sel(vector signed char __a, vector signed char __b, vector unsigned char __c)
+{
+  return (__a & ~(vector signed char)__c) | (__b & (vector signed char)__c);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_sel(vector signed char __a, vector signed char __b, vector bool char __c)
+{
+  return (__a & ~(vector signed char)__c) | (__b & (vector signed char)__c);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sel(vector unsigned char __a, vector unsigned char __b, vector unsigned char __c)
+{
+  return (__a & ~__c) | (__b & __c);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sel(vector unsigned char __a, vector unsigned char __b, vector bool char __c)
+{
+  return (__a & ~(vector unsigned char)__c) | (__b & (vector unsigned char)__c);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_sel(vector bool char __a, vector bool char __b, vector unsigned char __c)
+{
+  return (__a & ~(vector bool char)__c) | (__b & (vector bool char)__c);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_sel(vector bool char __a, vector bool char __b, vector bool char __c)
+{
+  return (__a & ~__c) | (__b & __c);
+}
+
+static vector short __ATTRS_o_ai
+vec_sel(vector short __a, vector short __b, vector unsigned short __c)
+{
+  return (__a & ~(vector short)__c) | (__b & (vector short)__c);
+}
+
+static vector short __ATTRS_o_ai
+vec_sel(vector short __a, vector short __b, vector bool short __c)
+{
+  return (__a & ~(vector short)__c) | (__b & (vector short)__c);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sel(vector unsigned short __a,
+        vector unsigned short __b,
+        vector unsigned short __c)
+{
+  return (__a & ~__c) | (__b & __c);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sel(vector unsigned short __a, vector unsigned short __b, vector bool short __c)
+{
+  return (__a & ~(vector unsigned short)__c) | (__b & (vector unsigned short)__c);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_sel(vector bool short __a, vector bool short __b, vector unsigned short __c)
+{
+  return (__a & ~(vector bool short)__c) | (__b & (vector bool short)__c);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_sel(vector bool short __a, vector bool short __b, vector bool short __c)
+{
+  return (__a & ~__c) | (__b & __c);
+}
+
+static vector int __ATTRS_o_ai
+vec_sel(vector int __a, vector int __b, vector unsigned int __c)
+{
+  return (__a & ~(vector int)__c) | (__b & (vector int)__c);
+}
+
+static vector int __ATTRS_o_ai
+vec_sel(vector int __a, vector int __b, vector bool int __c)
+{
+  return (__a & ~(vector int)__c) | (__b & (vector int)__c);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sel(vector unsigned int __a, vector unsigned int __b, vector unsigned int __c)
+{
+  return (__a & ~__c) | (__b & __c);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sel(vector unsigned int __a, vector unsigned int __b, vector bool int __c)
+{
+  return (__a & ~(vector unsigned int)__c) | (__b & (vector unsigned int)__c);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_sel(vector bool int __a, vector bool int __b, vector unsigned int __c)
+{
+  return (__a & ~(vector bool int)__c) | (__b & (vector bool int)__c);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_sel(vector bool int __a, vector bool int __b, vector bool int __c)
+{
+  return (__a & ~__c) | (__b & __c);
+}
+
+static vector float __ATTRS_o_ai
+vec_sel(vector float __a, vector float __b, vector unsigned int __c)
+{
+  vector int __res = ((vector int)__a & ~(vector int)__c)
+                   | ((vector int)__b & (vector int)__c);
+  return (vector float)__res;
+}
+
+static vector float __ATTRS_o_ai
+vec_sel(vector float __a, vector float __b, vector bool int __c)
+{
+  vector int __res = ((vector int)__a & ~(vector int)__c)
+                   | ((vector int)__b & (vector int)__c);
+  return (vector float)__res;
+}
+
+/* vec_vsel */
+
+static vector signed char __ATTRS_o_ai
+vec_vsel(vector signed char __a, vector signed char __b, vector unsigned char __c)
+{
+  return (__a & ~(vector signed char)__c) | (__b & (vector signed char)__c);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vsel(vector signed char __a, vector signed char __b, vector bool char __c)
+{
+  return (__a & ~(vector signed char)__c) | (__b & (vector signed char)__c);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsel(vector unsigned char __a, vector unsigned char __b, vector unsigned char __c)
+{
+  return (__a & ~__c) | (__b & __c);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsel(vector unsigned char __a, vector unsigned char __b, vector bool char __c)
+{
+  return (__a & ~(vector unsigned char)__c) | (__b & (vector unsigned char)__c);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vsel(vector bool char __a, vector bool char __b, vector unsigned char __c)
+{
+  return (__a & ~(vector bool char)__c) | (__b & (vector bool char)__c);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vsel(vector bool char __a, vector bool char __b, vector bool char __c)
+{
+  return (__a & ~__c) | (__b & __c);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsel(vector short __a, vector short __b, vector unsigned short __c)
+{
+  return (__a & ~(vector short)__c) | (__b & (vector short)__c);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsel(vector short __a, vector short __b, vector bool short __c)
+{
+  return (__a & ~(vector short)__c) | (__b & (vector short)__c);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsel(vector unsigned short __a,
+         vector unsigned short __b,
+         vector unsigned short __c)
+{
+  return (__a & ~__c) | (__b & __c);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsel(vector unsigned short __a, vector unsigned short __b, vector bool short __c)
+{
+  return (__a & ~(vector unsigned short)__c) | (__b & (vector unsigned short)__c);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vsel(vector bool short __a, vector bool short __b, vector unsigned short __c)
+{
+  return (__a & ~(vector bool short)__c) | (__b & (vector bool short)__c);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vsel(vector bool short __a, vector bool short __b, vector bool short __c)
+{
+  return (__a & ~__c) | (__b & __c);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsel(vector int __a, vector int __b, vector unsigned int __c)
+{
+  return (__a & ~(vector int)__c) | (__b & (vector int)__c);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsel(vector int __a, vector int __b, vector bool int __c)
+{
+  return (__a & ~(vector int)__c) | (__b & (vector int)__c);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsel(vector unsigned int __a, vector unsigned int __b, vector unsigned int __c)
+{
+  return (__a & ~__c) | (__b & __c);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsel(vector unsigned int __a, vector unsigned int __b, vector bool int __c)
+{
+  return (__a & ~(vector unsigned int)__c) | (__b & (vector unsigned int)__c);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vsel(vector bool int __a, vector bool int __b, vector unsigned int __c)
+{
+  return (__a & ~(vector bool int)__c) | (__b & (vector bool int)__c);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vsel(vector bool int __a, vector bool int __b, vector bool int __c)
+{
+  return (__a & ~__c) | (__b & __c);
+}
+
+static vector float __ATTRS_o_ai
+vec_vsel(vector float __a, vector float __b, vector unsigned int __c)
+{
+  vector int __res = ((vector int)__a & ~(vector int)__c)
+                   | ((vector int)__b & (vector int)__c);
+  return (vector float)__res;
+}
+
+static vector float __ATTRS_o_ai
+vec_vsel(vector float __a, vector float __b, vector bool int __c)
+{
+  vector int __res = ((vector int)__a & ~(vector int)__c)
+                   | ((vector int)__b & (vector int)__c);
+  return (vector float)__res;
+}
+
+/* vec_sl */
+
+static vector signed char __ATTRS_o_ai
+vec_sl(vector signed char __a, vector unsigned char __b)
+{
+  return __a << (vector signed char)__b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sl(vector unsigned char __a, vector unsigned char __b)
+{
+  return __a << __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_sl(vector short __a, vector unsigned short __b)
+{
+  return __a << (vector short)__b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sl(vector unsigned short __a, vector unsigned short __b)
+{
+  return __a << __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_sl(vector int __a, vector unsigned int __b)
+{
+  return __a << (vector int)__b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sl(vector unsigned int __a, vector unsigned int __b)
+{
+  return __a << __b;
+}
+
+/* vec_vslb */
+
+#define __builtin_altivec_vslb vec_vslb
+
+static vector signed char __ATTRS_o_ai
+vec_vslb(vector signed char __a, vector unsigned char __b)
+{
+  return vec_sl(__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vslb(vector unsigned char __a, vector unsigned char __b)
+{
+  return vec_sl(__a, __b);
+}
+
+/* vec_vslh */
+
+#define __builtin_altivec_vslh vec_vslh
+
+static vector short __ATTRS_o_ai
+vec_vslh(vector short __a, vector unsigned short __b)
+{
+  return vec_sl(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vslh(vector unsigned short __a, vector unsigned short __b)
+{
+  return vec_sl(__a, __b);
+}
+
+/* vec_vslw */
+
+#define __builtin_altivec_vslw vec_vslw
+
+static vector int __ATTRS_o_ai
+vec_vslw(vector int __a, vector unsigned int __b)
+{
+  return vec_sl(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vslw(vector unsigned int __a, vector unsigned int __b)
+{
+  return vec_sl(__a, __b);
+}
+
+/* vec_sld */
+
+#define __builtin_altivec_vsldoi_4si vec_sld
+
+static vector signed char __ATTRS_o_ai
+vec_sld(vector signed char __a, vector signed char __b, unsigned char __c)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (__c,   __c+1, __c+2,  __c+3,  __c+4,  __c+5,  __c+6,  __c+7,
+     __c+8, __c+9, __c+10, __c+11, __c+12, __c+13, __c+14, __c+15));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sld(vector unsigned char __a, vector unsigned char __b, unsigned char __c)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (__c,   __c+1, __c+2,  __c+3,  __c+4,  __c+5,  __c+6,  __c+7,
+     __c+8, __c+9, __c+10, __c+11, __c+12, __c+13, __c+14, __c+15));
+}
+
+static vector short __ATTRS_o_ai
+vec_sld(vector short __a, vector short __b, unsigned char __c)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (__c,   __c+1, __c+2,  __c+3,  __c+4,  __c+5,  __c+6,  __c+7,
+     __c+8, __c+9, __c+10, __c+11, __c+12, __c+13, __c+14, __c+15));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sld(vector unsigned short __a, vector unsigned short __b, unsigned char __c)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (__c,   __c+1, __c+2,  __c+3,  __c+4,  __c+5,  __c+6,  __c+7,
+     __c+8, __c+9, __c+10, __c+11, __c+12, __c+13, __c+14, __c+15));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_sld(vector pixel __a, vector pixel __b, unsigned char __c)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (__c,   __c+1, __c+2,  __c+3,  __c+4,  __c+5,  __c+6,  __c+7,
+     __c+8, __c+9, __c+10, __c+11, __c+12, __c+13, __c+14, __c+15));
+}
+
+static vector int __ATTRS_o_ai
+vec_sld(vector int __a, vector int __b, unsigned char __c)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (__c,   __c+1, __c+2,  __c+3,  __c+4,  __c+5,  __c+6,  __c+7,
+     __c+8, __c+9, __c+10, __c+11, __c+12, __c+13, __c+14, __c+15));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sld(vector unsigned int __a, vector unsigned int __b, unsigned char __c)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (__c,   __c+1, __c+2,  __c+3,  __c+4,  __c+5,  __c+6,  __c+7,
+     __c+8, __c+9, __c+10, __c+11, __c+12, __c+13, __c+14, __c+15));
+}
+
+static vector float __ATTRS_o_ai
+vec_sld(vector float __a, vector float __b, unsigned char __c)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (__c,   __c+1, __c+2,  __c+3,  __c+4,  __c+5,  __c+6,  __c+7,
+     __c+8, __c+9, __c+10, __c+11, __c+12, __c+13, __c+14, __c+15));
+}
+
+/* vec_vsldoi */
+
+static vector signed char __ATTRS_o_ai
+vec_vsldoi(vector signed char __a, vector signed char __b, unsigned char __c)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (__c,   __c+1, __c+2,  __c+3,  __c+4,  __c+5,  __c+6,  __c+7,
+     __c+8, __c+9, __c+10, __c+11, __c+12, __c+13, __c+14, __c+15));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsldoi(vector unsigned char __a, vector unsigned char __b, unsigned char __c)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (__c,   __c+1, __c+2,  __c+3,  __c+4,  __c+5,  __c+6,  __c+7,
+     __c+8, __c+9, __c+10, __c+11, __c+12, __c+13, __c+14, __c+15));
+}
+
+static vector short __ATTRS_o_ai
+vec_vsldoi(vector short __a, vector short __b, unsigned char __c)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (__c,   __c+1, __c+2,  __c+3,  __c+4,  __c+5,  __c+6,  __c+7,
+     __c+8, __c+9, __c+10, __c+11, __c+12, __c+13, __c+14, __c+15));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsldoi(vector unsigned short __a, vector unsigned short __b, unsigned char __c)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (__c,   __c+1, __c+2,  __c+3,  __c+4,  __c+5,  __c+6,  __c+7,
+     __c+8, __c+9, __c+10, __c+11, __c+12, __c+13, __c+14, __c+15));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vsldoi(vector pixel __a, vector pixel __b, unsigned char __c)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (__c,   __c+1, __c+2,  __c+3,  __c+4,  __c+5,  __c+6,  __c+7,
+     __c+8, __c+9, __c+10, __c+11, __c+12, __c+13, __c+14, __c+15));
+}
+
+static vector int __ATTRS_o_ai
+vec_vsldoi(vector int __a, vector int __b, unsigned char __c)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (__c,   __c+1, __c+2,  __c+3,  __c+4,  __c+5,  __c+6,  __c+7,
+     __c+8, __c+9, __c+10, __c+11, __c+12, __c+13, __c+14, __c+15));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsldoi(vector unsigned int __a, vector unsigned int __b, unsigned char __c)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (__c,   __c+1, __c+2,  __c+3,  __c+4,  __c+5,  __c+6,  __c+7,
+     __c+8, __c+9, __c+10, __c+11, __c+12, __c+13, __c+14, __c+15));
+}
+
+static vector float __ATTRS_o_ai
+vec_vsldoi(vector float __a, vector float __b, unsigned char __c)
+{
+  return vec_perm(__a, __b, (vector unsigned char)
+    (__c,   __c+1, __c+2,  __c+3,  __c+4,  __c+5,  __c+6,  __c+7,
+     __c+8, __c+9, __c+10, __c+11, __c+12, __c+13, __c+14, __c+15));
+}
+
+/* vec_sll */
+
+static vector signed char __ATTRS_o_ai
+vec_sll(vector signed char __a, vector unsigned char __b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_sll(vector signed char __a, vector unsigned short __b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_sll(vector signed char __a, vector unsigned int __b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sll(vector unsigned char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sll(vector unsigned char __a, vector unsigned short __b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sll(vector unsigned char __a, vector unsigned int __b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_sll(vector bool char __a, vector unsigned char __b)
+{
+  return (vector bool char)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_sll(vector bool char __a, vector unsigned short __b)
+{
+  return (vector bool char)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_sll(vector bool char __a, vector unsigned int __b)
+{
+  return (vector bool char)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_sll(vector short __a, vector unsigned char __b)
+{
+  return (vector short)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_sll(vector short __a, vector unsigned short __b)
+{
+  return (vector short)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_sll(vector short __a, vector unsigned int __b)
+{
+  return (vector short)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sll(vector unsigned short __a, vector unsigned char __b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sll(vector unsigned short __a, vector unsigned short __b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sll(vector unsigned short __a, vector unsigned int __b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_sll(vector bool short __a, vector unsigned char __b)
+{
+  return (vector bool short)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_sll(vector bool short __a, vector unsigned short __b)
+{
+  return (vector bool short)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_sll(vector bool short __a, vector unsigned int __b)
+{
+  return (vector bool short)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_sll(vector pixel __a, vector unsigned char __b)
+{
+  return (vector pixel)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_sll(vector pixel __a, vector unsigned short __b)
+{
+  return (vector pixel)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_sll(vector pixel __a, vector unsigned int __b)
+{
+  return (vector pixel)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_sll(vector int __a, vector unsigned char __b)
+{
+  return (vector int)__builtin_altivec_vsl(__a, (vector int)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_sll(vector int __a, vector unsigned short __b)
+{
+  return (vector int)__builtin_altivec_vsl(__a, (vector int)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_sll(vector int __a, vector unsigned int __b)
+{
+  return (vector int)__builtin_altivec_vsl(__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sll(vector unsigned int __a, vector unsigned char __b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sll(vector unsigned int __a, vector unsigned short __b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sll(vector unsigned int __a, vector unsigned int __b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_sll(vector bool int __a, vector unsigned char __b)
+{
+  return (vector bool int)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_sll(vector bool int __a, vector unsigned short __b)
+{
+  return (vector bool int)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_sll(vector bool int __a, vector unsigned int __b)
+{
+  return (vector bool int)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+/* vec_vsl */
+
+static vector signed char __ATTRS_o_ai
+vec_vsl(vector signed char __a, vector unsigned char __b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vsl(vector signed char __a, vector unsigned short __b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vsl(vector signed char __a, vector unsigned int __b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsl(vector unsigned char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsl(vector unsigned char __a, vector unsigned short __b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsl(vector unsigned char __a, vector unsigned int __b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vsl(vector bool char __a, vector unsigned char __b)
+{
+  return (vector bool char)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vsl(vector bool char __a, vector unsigned short __b)
+{
+  return (vector bool char)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vsl(vector bool char __a, vector unsigned int __b)
+{
+  return (vector bool char)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsl(vector short __a, vector unsigned char __b)
+{
+  return (vector short)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsl(vector short __a, vector unsigned short __b)
+{
+  return (vector short)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsl(vector short __a, vector unsigned int __b)
+{
+  return (vector short)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsl(vector unsigned short __a, vector unsigned char __b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsl(vector unsigned short __a, vector unsigned short __b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsl(vector unsigned short __a, vector unsigned int __b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vsl(vector bool short __a, vector unsigned char __b)
+{
+  return (vector bool short)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vsl(vector bool short __a, vector unsigned short __b)
+{
+  return (vector bool short)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vsl(vector bool short __a, vector unsigned int __b)
+{
+  return (vector bool short)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vsl(vector pixel __a, vector unsigned char __b)
+{
+  return (vector pixel)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vsl(vector pixel __a, vector unsigned short __b)
+{
+  return (vector pixel)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vsl(vector pixel __a, vector unsigned int __b)
+{
+  return (vector pixel)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsl(vector int __a, vector unsigned char __b)
+{
+  return (vector int)__builtin_altivec_vsl(__a, (vector int)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsl(vector int __a, vector unsigned short __b)
+{
+  return (vector int)__builtin_altivec_vsl(__a, (vector int)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsl(vector int __a, vector unsigned int __b)
+{
+  return (vector int)__builtin_altivec_vsl(__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsl(vector unsigned int __a, vector unsigned char __b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsl(vector unsigned int __a, vector unsigned short __b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsl(vector unsigned int __a, vector unsigned int __b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vsl(vector bool int __a, vector unsigned char __b)
+{
+  return (vector bool int)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vsl(vector bool int __a, vector unsigned short __b)
+{
+  return (vector bool int)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vsl(vector bool int __a, vector unsigned int __b)
+{
+  return (vector bool int)__builtin_altivec_vsl((vector int)__a, (vector int)__b);
+}
+
+/* vec_slo */
+
+static vector signed char __ATTRS_o_ai
+vec_slo(vector signed char __a, vector signed char __b)
+{
+  return (vector signed char)
+           __builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_slo(vector signed char __a, vector unsigned char __b)
+{
+  return (vector signed char)
+           __builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_slo(vector unsigned char __a, vector signed char __b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_slo(vector unsigned char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_slo(vector short __a, vector signed char __b)
+{
+  return (vector short)__builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_slo(vector short __a, vector unsigned char __b)
+{
+  return (vector short)__builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_slo(vector unsigned short __a, vector signed char __b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_slo(vector unsigned short __a, vector unsigned char __b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_slo(vector pixel __a, vector signed char __b)
+{
+  return (vector pixel)__builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_slo(vector pixel __a, vector unsigned char __b)
+{
+  return (vector pixel)__builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_slo(vector int __a, vector signed char __b)
+{
+  return (vector int)__builtin_altivec_vslo(__a, (vector int)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_slo(vector int __a, vector unsigned char __b)
+{
+  return (vector int)__builtin_altivec_vslo(__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_slo(vector unsigned int __a, vector signed char __b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_slo(vector unsigned int __a, vector unsigned char __b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector float __ATTRS_o_ai
+vec_slo(vector float __a, vector signed char __b)
+{
+  return (vector float)__builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector float __ATTRS_o_ai
+vec_slo(vector float __a, vector unsigned char __b)
+{
+  return (vector float)__builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+/* vec_vslo */
+
+static vector signed char __ATTRS_o_ai
+vec_vslo(vector signed char __a, vector signed char __b)
+{
+  return (vector signed char)
+           __builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vslo(vector signed char __a, vector unsigned char __b)
+{
+  return (vector signed char)
+           __builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vslo(vector unsigned char __a, vector signed char __b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vslo(vector unsigned char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vslo(vector short __a, vector signed char __b)
+{
+  return (vector short)__builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vslo(vector short __a, vector unsigned char __b)
+{
+  return (vector short)__builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vslo(vector unsigned short __a, vector signed char __b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vslo(vector unsigned short __a, vector unsigned char __b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vslo(vector pixel __a, vector signed char __b)
+{
+  return (vector pixel)__builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vslo(vector pixel __a, vector unsigned char __b)
+{
+  return (vector pixel)__builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vslo(vector int __a, vector signed char __b)
+{
+  return (vector int)__builtin_altivec_vslo(__a, (vector int)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vslo(vector int __a, vector unsigned char __b)
+{
+  return (vector int)__builtin_altivec_vslo(__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vslo(vector unsigned int __a, vector signed char __b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vslo(vector unsigned int __a, vector unsigned char __b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector float __ATTRS_o_ai
+vec_vslo(vector float __a, vector signed char __b)
+{
+  return (vector float)__builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+static vector float __ATTRS_o_ai
+vec_vslo(vector float __a, vector unsigned char __b)
+{
+  return (vector float)__builtin_altivec_vslo((vector int)__a, (vector int)__b);
+}
+
+/* vec_splat */
+
+static vector signed char __ATTRS_o_ai
+vec_splat(vector signed char __a, unsigned char __b)
+{
+  return vec_perm(__a, __a, (vector unsigned char)(__b));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_splat(vector unsigned char __a, unsigned char __b)
+{
+  return vec_perm(__a, __a, (vector unsigned char)(__b));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_splat(vector bool char __a, unsigned char __b)
+{
+  return vec_perm(__a, __a, (vector unsigned char)(__b));
+}
+
+static vector short __ATTRS_o_ai
+vec_splat(vector short __a, unsigned char __b)
+{ 
+  __b *= 2;
+  unsigned char b1=__b+1;
+  return vec_perm(__a, __a, (vector unsigned char)
+    (__b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_splat(vector unsigned short __a, unsigned char __b)
+{ 
+  __b *= 2;
+  unsigned char b1=__b+1;
+  return vec_perm(__a, __a, (vector unsigned char)
+    (__b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_splat(vector bool short __a, unsigned char __b)
+{ 
+  __b *= 2;
+  unsigned char b1=__b+1;
+  return vec_perm(__a, __a, (vector unsigned char)
+    (__b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_splat(vector pixel __a, unsigned char __b)
+{ 
+  __b *= 2;
+  unsigned char b1=__b+1;
+  return vec_perm(__a, __a, (vector unsigned char)
+    (__b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1));
+}
+
+static vector int __ATTRS_o_ai
+vec_splat(vector int __a, unsigned char __b)
+{ 
+  __b *= 4;
+  unsigned char b1=__b+1, b2=__b+2, b3=__b+3;
+  return vec_perm(__a, __a, (vector unsigned char)
+    (__b, b1, b2, b3, __b, b1, b2, b3, __b, b1, b2, b3, __b, b1, b2, b3));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_splat(vector unsigned int __a, unsigned char __b)
+{ 
+  __b *= 4;
+  unsigned char b1=__b+1, b2=__b+2, b3=__b+3;
+  return vec_perm(__a, __a, (vector unsigned char)
+    (__b, b1, b2, b3, __b, b1, b2, b3, __b, b1, b2, b3, __b, b1, b2, b3));
+}
+
+static vector bool int __ATTRS_o_ai
+vec_splat(vector bool int __a, unsigned char __b)
+{ 
+  __b *= 4;
+  unsigned char b1=__b+1, b2=__b+2, b3=__b+3;
+  return vec_perm(__a, __a, (vector unsigned char)
+    (__b, b1, b2, b3, __b, b1, b2, b3, __b, b1, b2, b3, __b, b1, b2, b3));
+}
+
+static vector float __ATTRS_o_ai
+vec_splat(vector float __a, unsigned char __b)
+{ 
+  __b *= 4;
+  unsigned char b1=__b+1, b2=__b+2, b3=__b+3;
+  return vec_perm(__a, __a, (vector unsigned char)
+    (__b, b1, b2, b3, __b, b1, b2, b3, __b, b1, b2, b3, __b, b1, b2, b3));
+}
+
+/* vec_vspltb */
+
+#define __builtin_altivec_vspltb vec_vspltb
+
+static vector signed char __ATTRS_o_ai
+vec_vspltb(vector signed char __a, unsigned char __b)
+{
+  return vec_perm(__a, __a, (vector unsigned char)(__b));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vspltb(vector unsigned char __a, unsigned char __b)
+{
+  return vec_perm(__a, __a, (vector unsigned char)(__b));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vspltb(vector bool char __a, unsigned char __b)
+{
+  return vec_perm(__a, __a, (vector unsigned char)(__b));
+}
+
+/* vec_vsplth */
+
+#define __builtin_altivec_vsplth vec_vsplth
+
+static vector short __ATTRS_o_ai
+vec_vsplth(vector short __a, unsigned char __b)
+{
+  __b *= 2;
+  unsigned char b1=__b+1;
+  return vec_perm(__a, __a, (vector unsigned char)
+    (__b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsplth(vector unsigned short __a, unsigned char __b)
+{
+  __b *= 2;
+  unsigned char b1=__b+1;
+  return vec_perm(__a, __a, (vector unsigned char)
+    (__b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vsplth(vector bool short __a, unsigned char __b)
+{
+  __b *= 2;
+  unsigned char b1=__b+1;
+  return vec_perm(__a, __a, (vector unsigned char)
+    (__b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vsplth(vector pixel __a, unsigned char __b)
+{
+  __b *= 2;
+  unsigned char b1=__b+1;
+  return vec_perm(__a, __a, (vector unsigned char)
+    (__b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1, __b, b1));
+}
+
+/* vec_vspltw */
+
+#define __builtin_altivec_vspltw vec_vspltw
+
+static vector int __ATTRS_o_ai
+vec_vspltw(vector int __a, unsigned char __b)
+{
+  __b *= 4;
+  unsigned char b1=__b+1, b2=__b+2, b3=__b+3;
+  return vec_perm(__a, __a, (vector unsigned char)
+    (__b, b1, b2, b3, __b, b1, b2, b3, __b, b1, b2, b3, __b, b1, b2, b3));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vspltw(vector unsigned int __a, unsigned char __b)
+{
+  __b *= 4;
+  unsigned char b1=__b+1, b2=__b+2, b3=__b+3;
+  return vec_perm(__a, __a, (vector unsigned char)
+    (__b, b1, b2, b3, __b, b1, b2, b3, __b, b1, b2, b3, __b, b1, b2, b3));
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vspltw(vector bool int __a, unsigned char __b)
+{
+  __b *= 4;
+  unsigned char b1=__b+1, b2=__b+2, b3=__b+3;
+  return vec_perm(__a, __a, (vector unsigned char)
+    (__b, b1, b2, b3, __b, b1, b2, b3, __b, b1, b2, b3, __b, b1, b2, b3));
+}
+
+static vector float __ATTRS_o_ai
+vec_vspltw(vector float __a, unsigned char __b)
+{
+  __b *= 4;
+  unsigned char b1=__b+1, b2=__b+2, b3=__b+3;
+  return vec_perm(__a, __a, (vector unsigned char)
+    (__b, b1, b2, b3, __b, b1, b2, b3, __b, b1, b2, b3, __b, b1, b2, b3));
+}
+
+/* vec_splat_s8 */
+
+#define __builtin_altivec_vspltisb vec_splat_s8
+
+// FIXME: parameter should be treated as 5-bit signed literal
+static vector signed char __ATTRS_o_ai
+vec_splat_s8(signed char __a)
+{
+  return (vector signed char)(__a);
+}
+
+/* vec_vspltisb */
+
+// FIXME: parameter should be treated as 5-bit signed literal
+static vector signed char __ATTRS_o_ai
+vec_vspltisb(signed char __a)
+{
+  return (vector signed char)(__a);
+}
+
+/* vec_splat_s16 */
+
+#define __builtin_altivec_vspltish vec_splat_s16
+
+// FIXME: parameter should be treated as 5-bit signed literal
+static vector short __ATTRS_o_ai
+vec_splat_s16(signed char __a)
+{
+  return (vector short)(__a);
+}
+
+/* vec_vspltish */
+
+// FIXME: parameter should be treated as 5-bit signed literal
+static vector short __ATTRS_o_ai
+vec_vspltish(signed char __a)
+{
+  return (vector short)(__a);
+}
+
+/* vec_splat_s32 */
+
+#define __builtin_altivec_vspltisw vec_splat_s32
+
+// FIXME: parameter should be treated as 5-bit signed literal
+static vector int __ATTRS_o_ai
+vec_splat_s32(signed char __a)
+{
+  return (vector int)(__a);
+}
+
+/* vec_vspltisw */
+
+// FIXME: parameter should be treated as 5-bit signed literal
+static vector int __ATTRS_o_ai
+vec_vspltisw(signed char __a)
+{
+  return (vector int)(__a);
+}
+
+/* vec_splat_u8 */
+
+// FIXME: parameter should be treated as 5-bit signed literal
+static vector unsigned char __ATTRS_o_ai
+vec_splat_u8(unsigned char __a)
+{
+  return (vector unsigned char)(__a);
+}
+
+/* vec_splat_u16 */
+
+// FIXME: parameter should be treated as 5-bit signed literal
+static vector unsigned short __ATTRS_o_ai
+vec_splat_u16(signed char __a)
+{
+  return (vector unsigned short)(__a);
+}
+
+/* vec_splat_u32 */
+
+// FIXME: parameter should be treated as 5-bit signed literal
+static vector unsigned int __ATTRS_o_ai
+vec_splat_u32(signed char __a)
+{
+  return (vector unsigned int)(__a);
+}
+
+/* vec_sr */
+
+static vector signed char __ATTRS_o_ai
+vec_sr(vector signed char __a, vector unsigned char __b)
+{
+  return __a >> (vector signed char)__b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sr(vector unsigned char __a, vector unsigned char __b)
+{
+  return __a >> __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_sr(vector short __a, vector unsigned short __b)
+{
+  return __a >> (vector short)__b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sr(vector unsigned short __a, vector unsigned short __b)
+{
+  return __a >> __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_sr(vector int __a, vector unsigned int __b)
+{
+  return __a >> (vector int)__b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sr(vector unsigned int __a, vector unsigned int __b)
+{
+  return __a >> __b;
+}
+
+/* vec_vsrb */
+
+#define __builtin_altivec_vsrb vec_vsrb
+
+static vector signed char __ATTRS_o_ai
+vec_vsrb(vector signed char __a, vector unsigned char __b)
+{
+  return __a >> (vector signed char)__b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsrb(vector unsigned char __a, vector unsigned char __b)
+{
+  return __a >> __b;
+}
+
+/* vec_vsrh */
+
+#define __builtin_altivec_vsrh vec_vsrh
+
+static vector short __ATTRS_o_ai
+vec_vsrh(vector short __a, vector unsigned short __b)
+{
+  return __a >> (vector short)__b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsrh(vector unsigned short __a, vector unsigned short __b)
+{
+  return __a >> __b;
+}
+
+/* vec_vsrw */
+
+#define __builtin_altivec_vsrw vec_vsrw
+
+static vector int __ATTRS_o_ai
+vec_vsrw(vector int __a, vector unsigned int __b)
+{
+  return __a >> (vector int)__b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsrw(vector unsigned int __a, vector unsigned int __b)
+{
+  return __a >> __b;
+}
+
+/* vec_sra */
+
+static vector signed char __ATTRS_o_ai
+vec_sra(vector signed char __a, vector unsigned char __b)
+{
+  return (vector signed char)__builtin_altivec_vsrab((vector char)__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sra(vector unsigned char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)__builtin_altivec_vsrab((vector char)__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_sra(vector short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vsrah(__a, (vector unsigned short)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sra(vector unsigned short __a, vector unsigned short __b)
+{
+  return (vector unsigned short)__builtin_altivec_vsrah((vector short)__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_sra(vector int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vsraw(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sra(vector unsigned int __a, vector unsigned int __b)
+{
+  return (vector unsigned int)__builtin_altivec_vsraw((vector int)__a, __b);
+}
+
+/* vec_vsrab */
+
+static vector signed char __ATTRS_o_ai
+vec_vsrab(vector signed char __a, vector unsigned char __b)
+{
+  return (vector signed char)__builtin_altivec_vsrab((vector char)__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsrab(vector unsigned char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)__builtin_altivec_vsrab((vector char)__a, __b);
+}
+
+/* vec_vsrah */
+
+static vector short __ATTRS_o_ai
+vec_vsrah(vector short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vsrah(__a, (vector unsigned short)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsrah(vector unsigned short __a, vector unsigned short __b)
+{
+  return (vector unsigned short)__builtin_altivec_vsrah((vector short)__a, __b);
+}
+
+/* vec_vsraw */
+
+static vector int __ATTRS_o_ai
+vec_vsraw(vector int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vsraw(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsraw(vector unsigned int __a, vector unsigned int __b)
+{
+  return (vector unsigned int)__builtin_altivec_vsraw((vector int)__a, __b);
+}
+
+/* vec_srl */
+
+static vector signed char __ATTRS_o_ai
+vec_srl(vector signed char __a, vector unsigned char __b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_srl(vector signed char __a, vector unsigned short __b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_srl(vector signed char __a, vector unsigned int __b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_srl(vector unsigned char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_srl(vector unsigned char __a, vector unsigned short __b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_srl(vector unsigned char __a, vector unsigned int __b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_srl(vector bool char __a, vector unsigned char __b)
+{
+  return (vector bool char)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_srl(vector bool char __a, vector unsigned short __b)
+{
+  return (vector bool char)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_srl(vector bool char __a, vector unsigned int __b)
+{
+  return (vector bool char)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_srl(vector short __a, vector unsigned char __b)
+{
+  return (vector short)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_srl(vector short __a, vector unsigned short __b)
+{
+  return (vector short)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_srl(vector short __a, vector unsigned int __b)
+{
+  return (vector short)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_srl(vector unsigned short __a, vector unsigned char __b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_srl(vector unsigned short __a, vector unsigned short __b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_srl(vector unsigned short __a, vector unsigned int __b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_srl(vector bool short __a, vector unsigned char __b)
+{
+  return (vector bool short)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_srl(vector bool short __a, vector unsigned short __b)
+{
+  return (vector bool short)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_srl(vector bool short __a, vector unsigned int __b)
+{
+  return (vector bool short)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_srl(vector pixel __a, vector unsigned char __b)
+{
+  return (vector pixel)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_srl(vector pixel __a, vector unsigned short __b)
+{
+  return (vector pixel)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_srl(vector pixel __a, vector unsigned int __b)
+{
+  return (vector pixel)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_srl(vector int __a, vector unsigned char __b)
+{
+  return (vector int)__builtin_altivec_vsr(__a, (vector int)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_srl(vector int __a, vector unsigned short __b)
+{
+  return (vector int)__builtin_altivec_vsr(__a, (vector int)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_srl(vector int __a, vector unsigned int __b)
+{
+  return (vector int)__builtin_altivec_vsr(__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_srl(vector unsigned int __a, vector unsigned char __b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_srl(vector unsigned int __a, vector unsigned short __b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_srl(vector unsigned int __a, vector unsigned int __b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_srl(vector bool int __a, vector unsigned char __b)
+{
+  return (vector bool int)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_srl(vector bool int __a, vector unsigned short __b)
+{
+  return (vector bool int)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_srl(vector bool int __a, vector unsigned int __b)
+{
+  return (vector bool int)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+/* vec_vsr */
+
+static vector signed char __ATTRS_o_ai
+vec_vsr(vector signed char __a, vector unsigned char __b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vsr(vector signed char __a, vector unsigned short __b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vsr(vector signed char __a, vector unsigned int __b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsr(vector unsigned char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsr(vector unsigned char __a, vector unsigned short __b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsr(vector unsigned char __a, vector unsigned int __b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vsr(vector bool char __a, vector unsigned char __b)
+{
+  return (vector bool char)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vsr(vector bool char __a, vector unsigned short __b)
+{
+  return (vector bool char)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vsr(vector bool char __a, vector unsigned int __b)
+{
+  return (vector bool char)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsr(vector short __a, vector unsigned char __b)
+{
+  return (vector short)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsr(vector short __a, vector unsigned short __b)
+{
+  return (vector short)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsr(vector short __a, vector unsigned int __b)
+{
+  return (vector short)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsr(vector unsigned short __a, vector unsigned char __b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsr(vector unsigned short __a, vector unsigned short __b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsr(vector unsigned short __a, vector unsigned int __b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vsr(vector bool short __a, vector unsigned char __b)
+{
+  return (vector bool short)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vsr(vector bool short __a, vector unsigned short __b)
+{
+  return (vector bool short)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vsr(vector bool short __a, vector unsigned int __b)
+{
+  return (vector bool short)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vsr(vector pixel __a, vector unsigned char __b)
+{
+  return (vector pixel)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vsr(vector pixel __a, vector unsigned short __b)
+{
+  return (vector pixel)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vsr(vector pixel __a, vector unsigned int __b)
+{
+  return (vector pixel)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsr(vector int __a, vector unsigned char __b)
+{
+  return (vector int)__builtin_altivec_vsr(__a, (vector int)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsr(vector int __a, vector unsigned short __b)
+{
+  return (vector int)__builtin_altivec_vsr(__a, (vector int)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsr(vector int __a, vector unsigned int __b)
+{
+  return (vector int)__builtin_altivec_vsr(__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsr(vector unsigned int __a, vector unsigned char __b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsr(vector unsigned int __a, vector unsigned short __b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsr(vector unsigned int __a, vector unsigned int __b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vsr(vector bool int __a, vector unsigned char __b)
+{
+  return (vector bool int)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vsr(vector bool int __a, vector unsigned short __b)
+{
+  return (vector bool int)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vsr(vector bool int __a, vector unsigned int __b)
+{
+  return (vector bool int)__builtin_altivec_vsr((vector int)__a, (vector int)__b);
+}
+
+/* vec_sro */
+
+static vector signed char __ATTRS_o_ai
+vec_sro(vector signed char __a, vector signed char __b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_sro(vector signed char __a, vector unsigned char __b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sro(vector unsigned char __a, vector signed char __b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sro(vector unsigned char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_sro(vector short __a, vector signed char __b)
+{
+  return (vector short)__builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_sro(vector short __a, vector unsigned char __b)
+{
+  return (vector short)__builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sro(vector unsigned short __a, vector signed char __b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sro(vector unsigned short __a, vector unsigned char __b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_sro(vector pixel __a, vector signed char __b)
+{
+  return (vector pixel)__builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_sro(vector pixel __a, vector unsigned char __b)
+{
+  return (vector pixel)__builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_sro(vector int __a, vector signed char __b)
+{
+  return (vector int)__builtin_altivec_vsro(__a, (vector int)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_sro(vector int __a, vector unsigned char __b)
+{
+  return (vector int)__builtin_altivec_vsro(__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sro(vector unsigned int __a, vector signed char __b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sro(vector unsigned int __a, vector unsigned char __b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector float __ATTRS_o_ai
+vec_sro(vector float __a, vector signed char __b)
+{
+  return (vector float)__builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector float __ATTRS_o_ai
+vec_sro(vector float __a, vector unsigned char __b)
+{
+  return (vector float)__builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+/* vec_vsro */
+
+static vector signed char __ATTRS_o_ai
+vec_vsro(vector signed char __a, vector signed char __b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vsro(vector signed char __a, vector unsigned char __b)
+{
+  return (vector signed char)
+           __builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsro(vector unsigned char __a, vector signed char __b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsro(vector unsigned char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)
+           __builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsro(vector short __a, vector signed char __b)
+{
+  return (vector short)__builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsro(vector short __a, vector unsigned char __b)
+{
+  return (vector short)__builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsro(vector unsigned short __a, vector signed char __b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsro(vector unsigned short __a, vector unsigned char __b)
+{
+  return (vector unsigned short)
+           __builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vsro(vector pixel __a, vector signed char __b)
+{
+  return (vector pixel)__builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector pixel __ATTRS_o_ai
+vec_vsro(vector pixel __a, vector unsigned char __b)
+{
+  return (vector pixel)__builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsro(vector int __a, vector signed char __b)
+{
+  return (vector int)__builtin_altivec_vsro(__a, (vector int)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsro(vector int __a, vector unsigned char __b)
+{
+  return (vector int)__builtin_altivec_vsro(__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsro(vector unsigned int __a, vector signed char __b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsro(vector unsigned int __a, vector unsigned char __b)
+{
+  return (vector unsigned int)
+           __builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector float __ATTRS_o_ai
+vec_vsro(vector float __a, vector signed char __b)
+{
+  return (vector float)__builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+static vector float __ATTRS_o_ai
+vec_vsro(vector float __a, vector unsigned char __b)
+{
+  return (vector float)__builtin_altivec_vsro((vector int)__a, (vector int)__b);
+}
+
+/* vec_st */
+
+static void __ATTRS_o_ai
+vec_st(vector signed char __a, int __b, vector signed char *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector signed char __a, int __b, signed char *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector unsigned char __a, int __b, vector unsigned char *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector unsigned char __a, int __b, unsigned char *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector bool char __a, int __b, signed char *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector bool char __a, int __b, unsigned char *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector bool char __a, int __b, vector bool char *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector short __a, int __b, vector short *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector short __a, int __b, short *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector unsigned short __a, int __b, vector unsigned short *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector unsigned short __a, int __b, unsigned short *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector bool short __a, int __b, short *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector bool short __a, int __b, unsigned short *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector bool short __a, int __b, vector bool short *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector pixel __a, int __b, short *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector pixel __a, int __b, unsigned short *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector pixel __a, int __b, vector pixel *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector int __a, int __b, vector int *__c)
+{
+  __builtin_altivec_stvx(__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector int __a, int __b, int *__c)
+{
+  __builtin_altivec_stvx(__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector unsigned int __a, int __b, vector unsigned int *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector unsigned int __a, int __b, unsigned int *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector bool int __a, int __b, int *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector bool int __a, int __b, unsigned int *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector bool int __a, int __b, vector bool int *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector float __a, int __b, vector float *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_st(vector float __a, int __b, float *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+/* vec_stvx */
+
+static void __ATTRS_o_ai
+vec_stvx(vector signed char __a, int __b, vector signed char *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector signed char __a, int __b, signed char *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector unsigned char __a, int __b, vector unsigned char *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector unsigned char __a, int __b, unsigned char *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector bool char __a, int __b, signed char *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector bool char __a, int __b, unsigned char *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector bool char __a, int __b, vector bool char *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector short __a, int __b, vector short *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector short __a, int __b, short *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector unsigned short __a, int __b, vector unsigned short *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector unsigned short __a, int __b, unsigned short *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector bool short __a, int __b, short *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector bool short __a, int __b, unsigned short *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector bool short __a, int __b, vector bool short *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector pixel __a, int __b, short *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector pixel __a, int __b, unsigned short *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector pixel __a, int __b, vector pixel *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector int __a, int __b, vector int *__c)
+{
+  __builtin_altivec_stvx(__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector int __a, int __b, int *__c)
+{
+  __builtin_altivec_stvx(__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector unsigned int __a, int __b, vector unsigned int *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector unsigned int __a, int __b, unsigned int *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector bool int __a, int __b, int *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector bool int __a, int __b, unsigned int *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector bool int __a, int __b, vector bool int *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector float __a, int __b, vector float *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvx(vector float __a, int __b, float *__c)
+{
+  __builtin_altivec_stvx((vector int)__a, __b, __c);
+}
+
+/* vec_ste */
+
+static void __ATTRS_o_ai
+vec_ste(vector signed char __a, int __b, signed char *__c)
+{
+  __builtin_altivec_stvebx((vector char)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector unsigned char __a, int __b, unsigned char *__c)
+{
+  __builtin_altivec_stvebx((vector char)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector bool char __a, int __b, signed char *__c)
+{
+  __builtin_altivec_stvebx((vector char)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector bool char __a, int __b, unsigned char *__c)
+{
+  __builtin_altivec_stvebx((vector char)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector short __a, int __b, short *__c)
+{
+  __builtin_altivec_stvehx(__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector unsigned short __a, int __b, unsigned short *__c)
+{
+  __builtin_altivec_stvehx((vector short)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector bool short __a, int __b, short *__c)
+{
+  __builtin_altivec_stvehx((vector short)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector bool short __a, int __b, unsigned short *__c)
+{
+  __builtin_altivec_stvehx((vector short)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector pixel __a, int __b, short *__c)
+{
+  __builtin_altivec_stvehx((vector short)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector pixel __a, int __b, unsigned short *__c)
+{
+  __builtin_altivec_stvehx((vector short)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector int __a, int __b, int *__c)
+{
+  __builtin_altivec_stvewx(__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector unsigned int __a, int __b, unsigned int *__c)
+{
+  __builtin_altivec_stvewx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector bool int __a, int __b, int *__c)
+{
+  __builtin_altivec_stvewx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector bool int __a, int __b, unsigned int *__c)
+{
+  __builtin_altivec_stvewx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_ste(vector float __a, int __b, float *__c)
+{
+  __builtin_altivec_stvewx((vector int)__a, __b, __c);
+}
+
+/* vec_stvebx */
+
+static void __ATTRS_o_ai
+vec_stvebx(vector signed char __a, int __b, signed char *__c)
+{
+  __builtin_altivec_stvebx((vector char)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvebx(vector unsigned char __a, int __b, unsigned char *__c)
+{
+  __builtin_altivec_stvebx((vector char)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvebx(vector bool char __a, int __b, signed char *__c)
+{
+  __builtin_altivec_stvebx((vector char)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvebx(vector bool char __a, int __b, unsigned char *__c)
+{
+  __builtin_altivec_stvebx((vector char)__a, __b, __c);
+}
+
+/* vec_stvehx */
+
+static void __ATTRS_o_ai
+vec_stvehx(vector short __a, int __b, short *__c)
+{
+  __builtin_altivec_stvehx(__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvehx(vector unsigned short __a, int __b, unsigned short *__c)
+{
+  __builtin_altivec_stvehx((vector short)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvehx(vector bool short __a, int __b, short *__c)
+{
+  __builtin_altivec_stvehx((vector short)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvehx(vector bool short __a, int __b, unsigned short *__c)
+{
+  __builtin_altivec_stvehx((vector short)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvehx(vector pixel __a, int __b, short *__c)
+{
+  __builtin_altivec_stvehx((vector short)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvehx(vector pixel __a, int __b, unsigned short *__c)
+{
+  __builtin_altivec_stvehx((vector short)__a, __b, __c);
+}
+
+/* vec_stvewx */
+
+static void __ATTRS_o_ai
+vec_stvewx(vector int __a, int __b, int *__c)
+{
+  __builtin_altivec_stvewx(__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvewx(vector unsigned int __a, int __b, unsigned int *__c)
+{
+  __builtin_altivec_stvewx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvewx(vector bool int __a, int __b, int *__c)
+{
+  __builtin_altivec_stvewx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvewx(vector bool int __a, int __b, unsigned int *__c)
+{
+  __builtin_altivec_stvewx((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvewx(vector float __a, int __b, float *__c)
+{
+  __builtin_altivec_stvewx((vector int)__a, __b, __c);
+}
+
+/* vec_stl */
+
+static void __ATTRS_o_ai
+vec_stl(vector signed char __a, int __b, vector signed char *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector signed char __a, int __b, signed char *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector unsigned char __a, int __b, vector unsigned char *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector unsigned char __a, int __b, unsigned char *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector bool char __a, int __b, signed char *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector bool char __a, int __b, unsigned char *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector bool char __a, int __b, vector bool char *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector short __a, int __b, vector short *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector short __a, int __b, short *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector unsigned short __a, int __b, vector unsigned short *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector unsigned short __a, int __b, unsigned short *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector bool short __a, int __b, short *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector bool short __a, int __b, unsigned short *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector bool short __a, int __b, vector bool short *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector pixel __a, int __b, short *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector pixel __a, int __b, unsigned short *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector pixel __a, int __b, vector pixel *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector int __a, int __b, vector int *__c)
+{
+  __builtin_altivec_stvxl(__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector int __a, int __b, int *__c)
+{
+  __builtin_altivec_stvxl(__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector unsigned int __a, int __b, vector unsigned int *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector unsigned int __a, int __b, unsigned int *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector bool int __a, int __b, int *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector bool int __a, int __b, unsigned int *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector bool int __a, int __b, vector bool int *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector float __a, int __b, vector float *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stl(vector float __a, int __b, float *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+/* vec_stvxl */
+
+static void __ATTRS_o_ai
+vec_stvxl(vector signed char __a, int __b, vector signed char *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector signed char __a, int __b, signed char *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector unsigned char __a, int __b, vector unsigned char *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector unsigned char __a, int __b, unsigned char *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector bool char __a, int __b, signed char *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector bool char __a, int __b, unsigned char *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector bool char __a, int __b, vector bool char *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector short __a, int __b, vector short *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector short __a, int __b, short *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector unsigned short __a, int __b, vector unsigned short *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector unsigned short __a, int __b, unsigned short *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector bool short __a, int __b, short *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector bool short __a, int __b, unsigned short *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector bool short __a, int __b, vector bool short *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector pixel __a, int __b, short *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector pixel __a, int __b, unsigned short *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector pixel __a, int __b, vector pixel *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector int __a, int __b, vector int *__c)
+{
+  __builtin_altivec_stvxl(__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector int __a, int __b, int *__c)
+{
+  __builtin_altivec_stvxl(__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector unsigned int __a, int __b, vector unsigned int *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector unsigned int __a, int __b, unsigned int *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector bool int __a, int __b, int *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector bool int __a, int __b, unsigned int *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector bool int __a, int __b, vector bool int *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector float __a, int __b, vector float *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvxl(vector float __a, int __b, float *__c)
+{
+  __builtin_altivec_stvxl((vector int)__a, __b, __c);
+}
+
+/* vec_sub */
+
+static vector signed char __ATTRS_o_ai
+vec_sub(vector signed char __a, vector signed char __b)
+{
+  return __a - __b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_sub(vector bool char __a, vector signed char __b)
+{
+  return (vector signed char)__a - __b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_sub(vector signed char __a, vector bool char __b)
+{
+  return __a - (vector signed char)__b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sub(vector unsigned char __a, vector unsigned char __b)
+{
+  return __a - __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sub(vector bool char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)__a - __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_sub(vector unsigned char __a, vector bool char __b)
+{
+  return __a - (vector unsigned char)__b;
+}
+
+static vector short __ATTRS_o_ai
+vec_sub(vector short __a, vector short __b)
+{
+  return __a - __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_sub(vector bool short __a, vector short __b)
+{
+  return (vector short)__a - __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_sub(vector short __a, vector bool short __b)
+{
+  return __a - (vector short)__b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sub(vector unsigned short __a, vector unsigned short __b)
+{
+  return __a - __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sub(vector bool short __a, vector unsigned short __b)
+{
+  return (vector unsigned short)__a - __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_sub(vector unsigned short __a, vector bool short __b)
+{
+  return __a - (vector unsigned short)__b;
+}
+
+static vector int __ATTRS_o_ai
+vec_sub(vector int __a, vector int __b)
+{
+  return __a - __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_sub(vector bool int __a, vector int __b)
+{
+  return (vector int)__a - __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_sub(vector int __a, vector bool int __b)
+{
+  return __a - (vector int)__b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sub(vector unsigned int __a, vector unsigned int __b)
+{
+  return __a - __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sub(vector bool int __a, vector unsigned int __b)
+{
+  return (vector unsigned int)__a - __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sub(vector unsigned int __a, vector bool int __b)
+{
+  return __a - (vector unsigned int)__b;
+}
+
+static vector float __ATTRS_o_ai
+vec_sub(vector float __a, vector float __b)
+{
+  return __a - __b;
+}
+
+/* vec_vsububm */
+
+#define __builtin_altivec_vsububm vec_vsububm
+
+static vector signed char __ATTRS_o_ai
+vec_vsububm(vector signed char __a, vector signed char __b)
+{
+  return __a - __b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vsububm(vector bool char __a, vector signed char __b)
+{
+  return (vector signed char)__a - __b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vsububm(vector signed char __a, vector bool char __b)
+{
+  return __a - (vector signed char)__b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsububm(vector unsigned char __a, vector unsigned char __b)
+{
+  return __a - __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsububm(vector bool char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)__a - __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsububm(vector unsigned char __a, vector bool char __b)
+{
+  return __a - (vector unsigned char)__b;
+}
+
+/* vec_vsubuhm */
+
+#define __builtin_altivec_vsubuhm vec_vsubuhm
+
+static vector short __ATTRS_o_ai
+vec_vsubuhm(vector short __a, vector short __b)
+{
+  return __a - __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vsubuhm(vector bool short __a, vector short __b)
+{
+  return (vector short)__a - __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vsubuhm(vector short __a, vector bool short __b)
+{
+  return __a - (vector short)__b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsubuhm(vector unsigned short __a, vector unsigned short __b)
+{
+  return __a - __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsubuhm(vector bool short __a, vector unsigned short __b)
+{
+  return (vector unsigned short)__a - __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsubuhm(vector unsigned short __a, vector bool short __b)
+{
+  return __a - (vector unsigned short)__b;
+}
+
+/* vec_vsubuwm */
+
+#define __builtin_altivec_vsubuwm vec_vsubuwm
+
+static vector int __ATTRS_o_ai
+vec_vsubuwm(vector int __a, vector int __b)
+{
+  return __a - __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vsubuwm(vector bool int __a, vector int __b)
+{
+  return (vector int)__a - __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vsubuwm(vector int __a, vector bool int __b)
+{
+  return __a - (vector int)__b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsubuwm(vector unsigned int __a, vector unsigned int __b)
+{
+  return __a - __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsubuwm(vector bool int __a, vector unsigned int __b)
+{
+  return (vector unsigned int)__a - __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsubuwm(vector unsigned int __a, vector bool int __b)
+{
+  return __a - (vector unsigned int)__b;
+}
+
+/* vec_vsubfp */
+
+#define __builtin_altivec_vsubfp vec_vsubfp
+
+static vector float __attribute__((__always_inline__))
+vec_vsubfp(vector float __a, vector float __b)
+{
+  return __a - __b;
+}
+
+/* vec_subc */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_subc(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vsubcuw(__a, __b);
+}
+
+/* vec_vsubcuw */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_vsubcuw(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vsubcuw(__a, __b);
+}
+
+/* vec_subs */
+
+static vector signed char __ATTRS_o_ai
+vec_subs(vector signed char __a, vector signed char __b)
+{
+  return __builtin_altivec_vsubsbs(__a, __b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_subs(vector bool char __a, vector signed char __b)
+{
+  return __builtin_altivec_vsubsbs((vector signed char)__a, __b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_subs(vector signed char __a, vector bool char __b)
+{
+  return __builtin_altivec_vsubsbs(__a, (vector signed char)__b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_subs(vector unsigned char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vsububs(__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_subs(vector bool char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vsububs((vector unsigned char)__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_subs(vector unsigned char __a, vector bool char __b)
+{
+  return __builtin_altivec_vsububs(__a, (vector unsigned char)__b);
+}
+
+static vector short __ATTRS_o_ai
+vec_subs(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vsubshs(__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_subs(vector bool short __a, vector short __b)
+{
+  return __builtin_altivec_vsubshs((vector short)__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_subs(vector short __a, vector bool short __b)
+{
+  return __builtin_altivec_vsubshs(__a, (vector short)__b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_subs(vector unsigned short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vsubuhs(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_subs(vector bool short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vsubuhs((vector unsigned short)__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_subs(vector unsigned short __a, vector bool short __b)
+{
+  return __builtin_altivec_vsubuhs(__a, (vector unsigned short)__b);
+}
+
+static vector int __ATTRS_o_ai
+vec_subs(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vsubsws(__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_subs(vector bool int __a, vector int __b)
+{
+  return __builtin_altivec_vsubsws((vector int)__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_subs(vector int __a, vector bool int __b)
+{
+  return __builtin_altivec_vsubsws(__a, (vector int)__b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_subs(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vsubuws(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_subs(vector bool int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vsubuws((vector unsigned int)__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_subs(vector unsigned int __a, vector bool int __b)
+{
+  return __builtin_altivec_vsubuws(__a, (vector unsigned int)__b);
+}
+
+/* vec_vsubsbs */
+
+static vector signed char __ATTRS_o_ai
+vec_vsubsbs(vector signed char __a, vector signed char __b)
+{
+  return __builtin_altivec_vsubsbs(__a, __b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vsubsbs(vector bool char __a, vector signed char __b)
+{
+  return __builtin_altivec_vsubsbs((vector signed char)__a, __b);
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vsubsbs(vector signed char __a, vector bool char __b)
+{
+  return __builtin_altivec_vsubsbs(__a, (vector signed char)__b);
+}
+
+/* vec_vsububs */
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsububs(vector unsigned char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vsububs(__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsububs(vector bool char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vsububs((vector unsigned char)__a, __b);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vsububs(vector unsigned char __a, vector bool char __b)
+{
+  return __builtin_altivec_vsububs(__a, (vector unsigned char)__b);
+}
+
+/* vec_vsubshs */
+
+static vector short __ATTRS_o_ai
+vec_vsubshs(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vsubshs(__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsubshs(vector bool short __a, vector short __b)
+{
+  return __builtin_altivec_vsubshs((vector short)__a, __b);
+}
+
+static vector short __ATTRS_o_ai
+vec_vsubshs(vector short __a, vector bool short __b)
+{
+  return __builtin_altivec_vsubshs(__a, (vector short)__b);
+}
+
+/* vec_vsubuhs */
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsubuhs(vector unsigned short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vsubuhs(__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsubuhs(vector bool short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vsubuhs((vector unsigned short)__a, __b);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vsubuhs(vector unsigned short __a, vector bool short __b)
+{
+  return __builtin_altivec_vsubuhs(__a, (vector unsigned short)__b);
+}
+
+/* vec_vsubsws */
+
+static vector int __ATTRS_o_ai
+vec_vsubsws(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vsubsws(__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsubsws(vector bool int __a, vector int __b)
+{
+  return __builtin_altivec_vsubsws((vector int)__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_vsubsws(vector int __a, vector bool int __b)
+{
+  return __builtin_altivec_vsubsws(__a, (vector int)__b);
+}
+
+/* vec_vsubuws */
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsubuws(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vsubuws(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsubuws(vector bool int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vsubuws((vector unsigned int)__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsubuws(vector unsigned int __a, vector bool int __b)
+{
+  return __builtin_altivec_vsubuws(__a, (vector unsigned int)__b);
+}
+
+/* vec_sum4s */
+
+static vector int __ATTRS_o_ai
+vec_sum4s(vector signed char __a, vector int __b)
+{
+  return __builtin_altivec_vsum4sbs(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_sum4s(vector unsigned char __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vsum4ubs(__a, __b);
+}
+
+static vector int __ATTRS_o_ai
+vec_sum4s(vector signed short __a, vector int __b)
+{
+  return __builtin_altivec_vsum4shs(__a, __b);
+}
+
+/* vec_vsum4sbs */
+
+static vector int __attribute__((__always_inline__))
+vec_vsum4sbs(vector signed char __a, vector int __b)
+{
+  return __builtin_altivec_vsum4sbs(__a, __b);
+}
+
+/* vec_vsum4ubs */
+
+static vector unsigned int __attribute__((__always_inline__))
+vec_vsum4ubs(vector unsigned char __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vsum4ubs(__a, __b);
+}
+
+/* vec_vsum4shs */
+
+static vector int __attribute__((__always_inline__))
+vec_vsum4shs(vector signed short __a, vector int __b)
+{
+  return __builtin_altivec_vsum4shs(__a, __b);
+}
+
+/* vec_sum2s */
+
+/* The vsum2sws instruction has a big-endian bias, so that the second
+   input vector and the result always reference big-endian elements
+   1 and 3 (little-endian element 0 and 2).  For ease of porting the
+   programmer wants elements 1 and 3 in both cases, so for little
+   endian we must perform some permutes.  */
+
+static vector signed int __attribute__((__always_inline__))
+vec_sum2s(vector int __a, vector int __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  vector int __c = (vector signed int)
+    vec_perm(__b, __b, (vector unsigned char)
+	     (4,5,6,7,0,1,2,3,12,13,14,15,8,9,10,11));
+  __c = __builtin_altivec_vsum2sws(__a, __c);
+  return (vector signed int)
+    vec_perm(__c, __c, (vector unsigned char)
+	     (4,5,6,7,0,1,2,3,12,13,14,15,8,9,10,11));
+#else
+  return __builtin_altivec_vsum2sws(__a, __b);
+#endif
+}
+
+/* vec_vsum2sws */
+
+static vector signed int __attribute__((__always_inline__))
+vec_vsum2sws(vector int __a, vector int __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  vector int __c = (vector signed int)
+    vec_perm(__b, __b, (vector unsigned char)
+	     (4,5,6,7,0,1,2,3,12,13,14,15,8,9,10,11));
+  __c = __builtin_altivec_vsum2sws(__a, __c);
+  return (vector signed int)
+    vec_perm(__c, __c, (vector unsigned char)
+	     (4,5,6,7,0,1,2,3,12,13,14,15,8,9,10,11));
+#else
+  return __builtin_altivec_vsum2sws(__a, __b);
+#endif
+}
+
+/* vec_sums */
+
+/* The vsumsws instruction has a big-endian bias, so that the second
+   input vector and the result always reference big-endian element 3
+   (little-endian element 0).  For ease of porting the programmer
+   wants element 3 in both cases, so for little endian we must perform
+   some permutes.  */
+
+static vector signed int __attribute__((__always_inline__))
+vec_sums(vector signed int __a, vector signed int __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  __b = (vector signed int)vec_splat(__b, 3);
+  __b = __builtin_altivec_vsumsws(__a, __b);
+  return (vector signed int)(0, 0, 0, __b[0]);
+#else
+  return __builtin_altivec_vsumsws(__a, __b);
+#endif
+}
+
+/* vec_vsumsws */
+
+static vector signed int __attribute__((__always_inline__))
+vec_vsumsws(vector signed int __a, vector signed int __b)
+{
+#ifdef __LITTLE_ENDIAN__
+  __b = (vector signed int)vec_splat(__b, 3);
+  __b = __builtin_altivec_vsumsws(__a, __b);
+  return (vector signed int)(0, 0, 0, __b[0]);
+#else
+  return __builtin_altivec_vsumsws(__a, __b);
+#endif
+}
+
+/* vec_trunc */
+
+static vector float __attribute__((__always_inline__))
+vec_trunc(vector float __a)
+{
+  return __builtin_altivec_vrfiz(__a);
+}
+
+/* vec_vrfiz */
+
+static vector float __attribute__((__always_inline__))
+vec_vrfiz(vector float __a)
+{
+  return __builtin_altivec_vrfiz(__a);
+}
+
+/* vec_unpackh */
+
+/* The vector unpack instructions all have a big-endian bias, so for
+   little endian we must reverse the meanings of "high" and "low."  */
+
+static vector short __ATTRS_o_ai
+vec_unpackh(vector signed char __a)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vupklsb((vector char)__a);
+#else
+  return __builtin_altivec_vupkhsb((vector char)__a);
+#endif
+}
+
+static vector bool short __ATTRS_o_ai
+vec_unpackh(vector bool char __a)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector bool short)__builtin_altivec_vupklsb((vector char)__a);
+#else
+  return (vector bool short)__builtin_altivec_vupkhsb((vector char)__a);
+#endif
+}
+
+static vector int __ATTRS_o_ai
+vec_unpackh(vector short __a)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vupklsh(__a);
+#else
+  return __builtin_altivec_vupkhsh(__a);
+#endif
+}
+
+static vector bool int __ATTRS_o_ai
+vec_unpackh(vector bool short __a)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector bool int)__builtin_altivec_vupklsh((vector short)__a);
+#else
+  return (vector bool int)__builtin_altivec_vupkhsh((vector short)__a);
+#endif
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_unpackh(vector pixel __a)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector unsigned int)__builtin_altivec_vupklpx((vector short)__a);
+#else
+  return (vector unsigned int)__builtin_altivec_vupkhpx((vector short)__a);
+#endif
+}
+
+/* vec_vupkhsb */
+
+static vector short __ATTRS_o_ai
+vec_vupkhsb(vector signed char __a)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vupklsb((vector char)__a);
+#else
+  return __builtin_altivec_vupkhsb((vector char)__a);
+#endif
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vupkhsb(vector bool char __a)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector bool short)__builtin_altivec_vupklsb((vector char)__a);
+#else
+  return (vector bool short)__builtin_altivec_vupkhsb((vector char)__a);
+#endif
+}
+
+/* vec_vupkhsh */
+
+static vector int __ATTRS_o_ai
+vec_vupkhsh(vector short __a)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vupklsh(__a);
+#else
+  return __builtin_altivec_vupkhsh(__a);
+#endif
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vupkhsh(vector bool short __a)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector bool int)__builtin_altivec_vupklsh((vector short)__a);
+#else
+  return (vector bool int)__builtin_altivec_vupkhsh((vector short)__a);
+#endif
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vupkhsh(vector pixel __a)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector unsigned int)__builtin_altivec_vupklpx((vector short)__a);
+#else
+  return (vector unsigned int)__builtin_altivec_vupkhpx((vector short)__a);
+#endif
+}
+
+/* vec_unpackl */
+
+static vector short __ATTRS_o_ai
+vec_unpackl(vector signed char __a)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vupkhsb((vector char)__a);
+#else
+  return __builtin_altivec_vupklsb((vector char)__a);
+#endif
+}
+
+static vector bool short __ATTRS_o_ai
+vec_unpackl(vector bool char __a)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector bool short)__builtin_altivec_vupkhsb((vector char)__a);
+#else
+  return (vector bool short)__builtin_altivec_vupklsb((vector char)__a);
+#endif
+}
+
+static vector int __ATTRS_o_ai
+vec_unpackl(vector short __a)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vupkhsh(__a);
+#else
+  return __builtin_altivec_vupklsh(__a);
+#endif
+}
+
+static vector bool int __ATTRS_o_ai
+vec_unpackl(vector bool short __a)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector bool int)__builtin_altivec_vupkhsh((vector short)__a);
+#else
+  return (vector bool int)__builtin_altivec_vupklsh((vector short)__a);
+#endif
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_unpackl(vector pixel __a)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector unsigned int)__builtin_altivec_vupkhpx((vector short)__a);
+#else
+  return (vector unsigned int)__builtin_altivec_vupklpx((vector short)__a);
+#endif
+}
+
+/* vec_vupklsb */
+
+static vector short __ATTRS_o_ai
+vec_vupklsb(vector signed char __a)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vupkhsb((vector char)__a);
+#else
+  return __builtin_altivec_vupklsb((vector char)__a);
+#endif
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vupklsb(vector bool char __a)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector bool short)__builtin_altivec_vupkhsb((vector char)__a);
+#else
+  return (vector bool short)__builtin_altivec_vupklsb((vector char)__a);
+#endif
+}
+
+/* vec_vupklsh */
+
+static vector int __ATTRS_o_ai
+vec_vupklsh(vector short __a)
+{
+#ifdef __LITTLE_ENDIAN__
+  return __builtin_altivec_vupkhsh(__a);
+#else
+  return __builtin_altivec_vupklsh(__a);
+#endif
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vupklsh(vector bool short __a)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector bool int)__builtin_altivec_vupkhsh((vector short)__a);
+#else
+  return (vector bool int)__builtin_altivec_vupklsh((vector short)__a);
+#endif
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vupklsh(vector pixel __a)
+{
+#ifdef __LITTLE_ENDIAN__
+  return (vector unsigned int)__builtin_altivec_vupkhpx((vector short)__a);
+#else
+  return (vector unsigned int)__builtin_altivec_vupklpx((vector short)__a);
+#endif
+}
+
+/* vec_vsx_ld */
+
+#ifdef __VSX__
+
+static vector signed int __ATTRS_o_ai
+vec_vsx_ld(int __a, const vector signed int *__b)
+{
+  return (vector signed int)__builtin_vsx_lxvw4x(__a, __b);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vsx_ld(int __a, const vector unsigned int *__b)
+{
+  return (vector unsigned int)__builtin_vsx_lxvw4x(__a, __b);
+}
+
+static vector float __ATTRS_o_ai
+vec_vsx_ld(int __a, const vector float *__b)
+{
+  return (vector float)__builtin_vsx_lxvw4x(__a, __b);
+}
+
+static vector signed long long __ATTRS_o_ai
+vec_vsx_ld(int __a, const vector signed long long *__b)
+{
+  return (vector signed long long)__builtin_vsx_lxvd2x(__a, __b);
+}
+
+static vector unsigned long long __ATTRS_o_ai
+vec_vsx_ld(int __a, const vector unsigned long long *__b)
+{
+  return (vector unsigned long long)__builtin_vsx_lxvd2x(__a, __b);
+}
+
+static vector double __ATTRS_o_ai
+vec_vsx_ld(int __a, const vector double *__b)
+{
+  return (vector double)__builtin_vsx_lxvd2x(__a, __b);
+}
+
+#endif
+
+/* vec_vsx_st */
+
+#ifdef __VSX__
+
+static void __ATTRS_o_ai
+vec_vsx_st(vector signed int __a, int __b, vector signed int *__c)
+{
+  __builtin_vsx_stxvw4x((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_vsx_st(vector unsigned int __a, int __b, vector unsigned int *__c)
+{
+  __builtin_vsx_stxvw4x((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_vsx_st(vector float __a, int __b, vector float *__c)
+{
+  __builtin_vsx_stxvw4x((vector int)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_vsx_st(vector signed long long __a, int __b, vector signed long long *__c)
+{
+  __builtin_vsx_stxvd2x((vector double)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_vsx_st(vector unsigned long long __a, int __b,
+           vector unsigned long long *__c)
+{
+  __builtin_vsx_stxvd2x((vector double)__a, __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_vsx_st(vector double __a, int __b, vector double *__c)
+{
+  __builtin_vsx_stxvd2x((vector double)__a, __b, __c);
+}
+
+#endif
+
+/* vec_xor */
+
+#define __builtin_altivec_vxor vec_xor
+
+static vector signed char __ATTRS_o_ai
+vec_xor(vector signed char __a, vector signed char __b)
+{
+  return __a ^ __b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_xor(vector bool char __a, vector signed char __b)
+{
+  return (vector signed char)__a ^ __b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_xor(vector signed char __a, vector bool char __b)
+{
+  return __a ^ (vector signed char)__b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_xor(vector unsigned char __a, vector unsigned char __b)
+{
+  return __a ^ __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_xor(vector bool char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)__a ^ __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_xor(vector unsigned char __a, vector bool char __b)
+{
+  return __a ^ (vector unsigned char)__b;
+}
+
+static vector bool char __ATTRS_o_ai
+vec_xor(vector bool char __a, vector bool char __b)
+{
+  return __a ^ __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_xor(vector short __a, vector short __b)
+{
+  return __a ^ __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_xor(vector bool short __a, vector short __b)
+{
+  return (vector short)__a ^ __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_xor(vector short __a, vector bool short __b)
+{
+  return __a ^ (vector short)__b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_xor(vector unsigned short __a, vector unsigned short __b)
+{
+  return __a ^ __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_xor(vector bool short __a, vector unsigned short __b)
+{
+  return (vector unsigned short)__a ^ __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_xor(vector unsigned short __a, vector bool short __b)
+{
+  return __a ^ (vector unsigned short)__b;
+}
+
+static vector bool short __ATTRS_o_ai
+vec_xor(vector bool short __a, vector bool short __b)
+{
+  return __a ^ __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_xor(vector int __a, vector int __b)
+{
+  return __a ^ __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_xor(vector bool int __a, vector int __b)
+{
+  return (vector int)__a ^ __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_xor(vector int __a, vector bool int __b)
+{
+  return __a ^ (vector int)__b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_xor(vector unsigned int __a, vector unsigned int __b)
+{
+  return __a ^ __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_xor(vector bool int __a, vector unsigned int __b)
+{
+  return (vector unsigned int)__a ^ __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_xor(vector unsigned int __a, vector bool int __b)
+{
+  return __a ^ (vector unsigned int)__b;
+}
+
+static vector bool int __ATTRS_o_ai
+vec_xor(vector bool int __a, vector bool int __b)
+{
+  return __a ^ __b;
+}
+
+static vector float __ATTRS_o_ai
+vec_xor(vector float __a, vector float __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a ^ (vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+static vector float __ATTRS_o_ai
+vec_xor(vector bool int __a, vector float __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a ^ (vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+static vector float __ATTRS_o_ai
+vec_xor(vector float __a, vector bool int __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a ^ (vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+/* vec_vxor */
+
+static vector signed char __ATTRS_o_ai
+vec_vxor(vector signed char __a, vector signed char __b)
+{
+  return __a ^ __b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vxor(vector bool char __a, vector signed char __b)
+{
+  return (vector signed char)__a ^ __b;
+}
+
+static vector signed char __ATTRS_o_ai
+vec_vxor(vector signed char __a, vector bool char __b)
+{
+  return __a ^ (vector signed char)__b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vxor(vector unsigned char __a, vector unsigned char __b)
+{
+  return __a ^ __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vxor(vector bool char __a, vector unsigned char __b)
+{
+  return (vector unsigned char)__a ^ __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_vxor(vector unsigned char __a, vector bool char __b)
+{
+  return __a ^ (vector unsigned char)__b;
+}
+
+static vector bool char __ATTRS_o_ai
+vec_vxor(vector bool char __a, vector bool char __b)
+{
+  return __a ^ __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vxor(vector short __a, vector short __b)
+{
+  return __a ^ __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vxor(vector bool short __a, vector short __b)
+{
+  return (vector short)__a ^ __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_vxor(vector short __a, vector bool short __b)
+{
+  return __a ^ (vector short)__b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vxor(vector unsigned short __a, vector unsigned short __b)
+{
+  return __a ^ __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vxor(vector bool short __a, vector unsigned short __b)
+{
+  return (vector unsigned short)__a ^ __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_vxor(vector unsigned short __a, vector bool short __b)
+{
+  return __a ^ (vector unsigned short)__b;
+}
+
+static vector bool short __ATTRS_o_ai
+vec_vxor(vector bool short __a, vector bool short __b)
+{
+  return __a ^ __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vxor(vector int __a, vector int __b)
+{
+  return __a ^ __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vxor(vector bool int __a, vector int __b)
+{
+  return (vector int)__a ^ __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_vxor(vector int __a, vector bool int __b)
+{
+  return __a ^ (vector int)__b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vxor(vector unsigned int __a, vector unsigned int __b)
+{
+  return __a ^ __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vxor(vector bool int __a, vector unsigned int __b)
+{
+  return (vector unsigned int)__a ^ __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_vxor(vector unsigned int __a, vector bool int __b)
+{
+  return __a ^ (vector unsigned int)__b;
+}
+
+static vector bool int __ATTRS_o_ai
+vec_vxor(vector bool int __a, vector bool int __b)
+{
+  return __a ^ __b;
+}
+
+static vector float __ATTRS_o_ai
+vec_vxor(vector float __a, vector float __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a ^ (vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+static vector float __ATTRS_o_ai
+vec_vxor(vector bool int __a, vector float __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a ^ (vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+static vector float __ATTRS_o_ai
+vec_vxor(vector float __a, vector bool int __b)
+{
+  vector unsigned int __res = (vector unsigned int)__a ^ (vector unsigned int)__b;
+  return (vector float)__res;
+}
+
+/* ------------------------ extensions for CBEA ----------------------------- */
+
+/* vec_extract */
+
+static signed char __ATTRS_o_ai
+vec_extract(vector signed char __a, int __b)
+{
+  return __a[__b];
+}
+
+static unsigned char __ATTRS_o_ai
+vec_extract(vector unsigned char __a, int __b)
+{
+  return __a[__b];
+}
+
+static short __ATTRS_o_ai
+vec_extract(vector short __a, int __b)
+{
+  return __a[__b];
+}
+
+static unsigned short __ATTRS_o_ai
+vec_extract(vector unsigned short __a, int __b)
+{
+  return __a[__b];
+}
+
+static int __ATTRS_o_ai
+vec_extract(vector int __a, int __b)
+{
+  return __a[__b];
+}
+
+static unsigned int __ATTRS_o_ai
+vec_extract(vector unsigned int __a, int __b)
+{
+  return __a[__b];
+}
+
+static float __ATTRS_o_ai
+vec_extract(vector float __a, int __b)
+{
+  return __a[__b];
+}
+
+/* vec_insert */
+
+static vector signed char __ATTRS_o_ai
+vec_insert(signed char __a, vector signed char __b, int __c)
+{
+  __b[__c] = __a;
+  return __b;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_insert(unsigned char __a, vector unsigned char __b, int __c)
+{
+  __b[__c] = __a;
+  return __b;
+}
+
+static vector short __ATTRS_o_ai
+vec_insert(short __a, vector short __b, int __c)
+{
+  __b[__c] = __a;
+  return __b;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_insert(unsigned short __a, vector unsigned short __b, int __c)
+{
+  __b[__c] = __a;
+  return __b;
+}
+
+static vector int __ATTRS_o_ai
+vec_insert(int __a, vector int __b, int __c)
+{
+  __b[__c] = __a;
+  return __b;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_insert(unsigned int __a, vector unsigned int __b, int __c)
+{
+  __b[__c] = __a;
+  return __b;
+}
+
+static vector float __ATTRS_o_ai
+vec_insert(float __a, vector float __b, int __c)
+{
+  __b[__c] = __a;
+  return __b;
+}
+
+/* vec_lvlx */
+
+static vector signed char __ATTRS_o_ai
+vec_lvlx(int __a, const signed char *__b)
+{
+  return vec_perm(vec_ld(__a, __b),
+                  (vector signed char)(0),
+                  vec_lvsl(__a, __b));
+}
+
+static vector signed char __ATTRS_o_ai
+vec_lvlx(int __a, const vector signed char *__b)
+{
+  return vec_perm(vec_ld(__a, __b),
+                  (vector signed char)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvlx(int __a, const unsigned char *__b)
+{
+  return vec_perm(vec_ld(__a, __b),
+                  (vector unsigned char)(0),
+                  vec_lvsl(__a, __b));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvlx(int __a, const vector unsigned char *__b)
+{
+  return vec_perm(vec_ld(__a, __b),
+                  (vector unsigned char)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_lvlx(int __a, const vector bool char *__b)
+{
+  return vec_perm(vec_ld(__a, __b),
+                  (vector bool char)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector short __ATTRS_o_ai
+vec_lvlx(int __a, const short *__b)
+{
+  return vec_perm(vec_ld(__a, __b),
+                  (vector short)(0),
+                  vec_lvsl(__a, __b));
+}
+
+static vector short __ATTRS_o_ai
+vec_lvlx(int __a, const vector short *__b)
+{
+  return vec_perm(vec_ld(__a, __b),
+                  (vector short)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvlx(int __a, const unsigned short *__b)
+{
+  return vec_perm(vec_ld(__a, __b),
+                  (vector unsigned short)(0),
+                  vec_lvsl(__a, __b));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvlx(int __a, const vector unsigned short *__b)
+{
+  return vec_perm(vec_ld(__a, __b),
+                  (vector unsigned short)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_lvlx(int __a, const vector bool short *__b)
+{
+  return vec_perm(vec_ld(__a, __b),
+                  (vector bool short)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_lvlx(int __a, const vector pixel *__b)
+{
+  return vec_perm(vec_ld(__a, __b),
+                  (vector pixel)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector int __ATTRS_o_ai
+vec_lvlx(int __a, const int *__b)
+{
+  return vec_perm(vec_ld(__a, __b),
+                  (vector int)(0),
+                  vec_lvsl(__a, __b));
+}
+
+static vector int __ATTRS_o_ai
+vec_lvlx(int __a, const vector int *__b)
+{
+  return vec_perm(vec_ld(__a, __b),
+                  (vector int)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvlx(int __a, const unsigned int *__b)
+{
+  return vec_perm(vec_ld(__a, __b),
+                  (vector unsigned int)(0),
+                  vec_lvsl(__a, __b));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvlx(int __a, const vector unsigned int *__b)
+{
+  return vec_perm(vec_ld(__a, __b),
+                  (vector unsigned int)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector bool int __ATTRS_o_ai
+vec_lvlx(int __a, const vector bool int *__b)
+{
+  return vec_perm(vec_ld(__a, __b),
+                  (vector bool int)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector float __ATTRS_o_ai
+vec_lvlx(int __a, const float *__b)
+{
+  return vec_perm(vec_ld(__a, __b),
+                  (vector float)(0),
+                  vec_lvsl(__a, __b));
+}
+
+static vector float __ATTRS_o_ai
+vec_lvlx(int __a, const vector float *__b)
+{
+  return vec_perm(vec_ld(__a, __b),
+                  (vector float)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+/* vec_lvlxl */
+
+static vector signed char __ATTRS_o_ai
+vec_lvlxl(int __a, const signed char *__b)
+{
+  return vec_perm(vec_ldl(__a, __b),
+                  (vector signed char)(0),
+                  vec_lvsl(__a, __b));
+}
+
+static vector signed char __ATTRS_o_ai
+vec_lvlxl(int __a, const vector signed char *__b)
+{
+  return vec_perm(vec_ldl(__a, __b),
+                  (vector signed char)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvlxl(int __a, const unsigned char *__b)
+{
+  return vec_perm(vec_ldl(__a, __b),
+                  (vector unsigned char)(0),
+                  vec_lvsl(__a, __b));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvlxl(int __a, const vector unsigned char *__b)
+{
+  return vec_perm(vec_ldl(__a, __b),
+                  (vector unsigned char)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_lvlxl(int __a, const vector bool char *__b)
+{
+  return vec_perm(vec_ldl(__a, __b),
+                  (vector bool char)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector short __ATTRS_o_ai
+vec_lvlxl(int __a, const short *__b)
+{
+  return vec_perm(vec_ldl(__a, __b),
+                  (vector short)(0),
+                  vec_lvsl(__a, __b));
+}
+
+static vector short __ATTRS_o_ai
+vec_lvlxl(int __a, const vector short *__b)
+{
+  return vec_perm(vec_ldl(__a, __b),
+                  (vector short)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvlxl(int __a, const unsigned short *__b)
+{
+  return vec_perm(vec_ldl(__a, __b),
+                  (vector unsigned short)(0),
+                  vec_lvsl(__a, __b));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvlxl(int __a, const vector unsigned short *__b)
+{
+  return vec_perm(vec_ldl(__a, __b),
+                  (vector unsigned short)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_lvlxl(int __a, const vector bool short *__b)
+{
+  return vec_perm(vec_ldl(__a, __b),
+                  (vector bool short)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_lvlxl(int __a, const vector pixel *__b)
+{
+  return vec_perm(vec_ldl(__a, __b),
+                  (vector pixel)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector int __ATTRS_o_ai
+vec_lvlxl(int __a, const int *__b)
+{
+  return vec_perm(vec_ldl(__a, __b),
+                  (vector int)(0),
+                  vec_lvsl(__a, __b));
+}
+
+static vector int __ATTRS_o_ai
+vec_lvlxl(int __a, const vector int *__b)
+{
+  return vec_perm(vec_ldl(__a, __b),
+                  (vector int)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvlxl(int __a, const unsigned int *__b)
+{
+  return vec_perm(vec_ldl(__a, __b),
+                  (vector unsigned int)(0),
+                  vec_lvsl(__a, __b));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvlxl(int __a, const vector unsigned int *__b)
+{
+  return vec_perm(vec_ldl(__a, __b),
+                  (vector unsigned int)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector bool int __ATTRS_o_ai
+vec_lvlxl(int __a, const vector bool int *__b)
+{
+  return vec_perm(vec_ldl(__a, __b),
+                  (vector bool int)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector float __ATTRS_o_ai
+vec_lvlxl(int __a, const float *__b)
+{
+  return vec_perm(vec_ldl(__a, __b),
+                  (vector float)(0),
+                  vec_lvsl(__a, __b));
+}
+
+static vector float __ATTRS_o_ai
+vec_lvlxl(int __a, vector float *__b)
+{
+  return vec_perm(vec_ldl(__a, __b),
+                  (vector float)(0),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+/* vec_lvrx */
+
+static vector signed char __ATTRS_o_ai
+vec_lvrx(int __a, const signed char *__b)
+{
+  return vec_perm((vector signed char)(0),
+                  vec_ld(__a, __b),
+                  vec_lvsl(__a, __b));
+}
+
+static vector signed char __ATTRS_o_ai
+vec_lvrx(int __a, const vector signed char *__b)
+{
+  return vec_perm((vector signed char)(0),
+                  vec_ld(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvrx(int __a, const unsigned char *__b)
+{
+  return vec_perm((vector unsigned char)(0),
+                  vec_ld(__a, __b),
+                  vec_lvsl(__a, __b));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvrx(int __a, const vector unsigned char *__b)
+{
+  return vec_perm((vector unsigned char)(0),
+                  vec_ld(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_lvrx(int __a, const vector bool char *__b)
+{
+  return vec_perm((vector bool char)(0),
+                  vec_ld(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector short __ATTRS_o_ai
+vec_lvrx(int __a, const short *__b)
+{
+  return vec_perm((vector short)(0),
+                  vec_ld(__a, __b),
+                  vec_lvsl(__a, __b));
+}
+
+static vector short __ATTRS_o_ai
+vec_lvrx(int __a, const vector short *__b)
+{
+  return vec_perm((vector short)(0),
+                  vec_ld(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvrx(int __a, const unsigned short *__b)
+{
+  return vec_perm((vector unsigned short)(0),
+                  vec_ld(__a, __b),
+                  vec_lvsl(__a, __b));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvrx(int __a, const vector unsigned short *__b)
+{
+  return vec_perm((vector unsigned short)(0),
+                  vec_ld(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_lvrx(int __a, const vector bool short *__b)
+{
+  return vec_perm((vector bool short)(0),
+                  vec_ld(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_lvrx(int __a, const vector pixel *__b)
+{
+  return vec_perm((vector pixel)(0),
+                  vec_ld(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector int __ATTRS_o_ai
+vec_lvrx(int __a, const int *__b)
+{
+  return vec_perm((vector int)(0),
+                  vec_ld(__a, __b),
+                  vec_lvsl(__a, __b));
+}
+
+static vector int __ATTRS_o_ai
+vec_lvrx(int __a, const vector int *__b)
+{
+  return vec_perm((vector int)(0),
+                  vec_ld(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvrx(int __a, const unsigned int *__b)
+{
+  return vec_perm((vector unsigned int)(0),
+                  vec_ld(__a, __b),
+                  vec_lvsl(__a, __b));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvrx(int __a, const vector unsigned int *__b)
+{
+  return vec_perm((vector unsigned int)(0),
+                  vec_ld(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector bool int __ATTRS_o_ai
+vec_lvrx(int __a, const vector bool int *__b)
+{
+  return vec_perm((vector bool int)(0),
+                  vec_ld(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector float __ATTRS_o_ai
+vec_lvrx(int __a, const float *__b)
+{
+  return vec_perm((vector float)(0),
+                  vec_ld(__a, __b),
+                  vec_lvsl(__a, __b));
+}
+
+static vector float __ATTRS_o_ai
+vec_lvrx(int __a, const vector float *__b)
+{
+  return vec_perm((vector float)(0),
+                  vec_ld(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+/* vec_lvrxl */
+
+static vector signed char __ATTRS_o_ai
+vec_lvrxl(int __a, const signed char *__b)
+{
+  return vec_perm((vector signed char)(0),
+                  vec_ldl(__a, __b),
+                  vec_lvsl(__a, __b));
+}
+
+static vector signed char __ATTRS_o_ai
+vec_lvrxl(int __a, const vector signed char *__b)
+{
+  return vec_perm((vector signed char)(0),
+                  vec_ldl(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvrxl(int __a, const unsigned char *__b)
+{
+  return vec_perm((vector unsigned char)(0),
+                  vec_ldl(__a, __b),
+                  vec_lvsl(__a, __b));
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_lvrxl(int __a, const vector unsigned char *__b)
+{
+  return vec_perm((vector unsigned char)(0),
+                  vec_ldl(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector bool char __ATTRS_o_ai
+vec_lvrxl(int __a, const vector bool char *__b)
+{
+  return vec_perm((vector bool char)(0),
+                  vec_ldl(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector short __ATTRS_o_ai
+vec_lvrxl(int __a, const short *__b)
+{
+  return vec_perm((vector short)(0),
+                  vec_ldl(__a, __b),
+                  vec_lvsl(__a, __b));
+}
+
+static vector short __ATTRS_o_ai
+vec_lvrxl(int __a, const vector short *__b)
+{
+  return vec_perm((vector short)(0),
+                  vec_ldl(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvrxl(int __a, const unsigned short *__b)
+{
+  return vec_perm((vector unsigned short)(0),
+                  vec_ldl(__a, __b),
+                  vec_lvsl(__a, __b));
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_lvrxl(int __a, const vector unsigned short *__b)
+{
+  return vec_perm((vector unsigned short)(0),
+                  vec_ldl(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector bool short __ATTRS_o_ai
+vec_lvrxl(int __a, const vector bool short *__b)
+{
+  return vec_perm((vector bool short)(0),
+                  vec_ldl(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector pixel __ATTRS_o_ai
+vec_lvrxl(int __a, const vector pixel *__b)
+{
+  return vec_perm((vector pixel)(0),
+                  vec_ldl(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector int __ATTRS_o_ai
+vec_lvrxl(int __a, const int *__b)
+{
+  return vec_perm((vector int)(0),
+                  vec_ldl(__a, __b),
+                  vec_lvsl(__a, __b));
+}
+
+static vector int __ATTRS_o_ai
+vec_lvrxl(int __a, const vector int *__b)
+{
+  return vec_perm((vector int)(0),
+                  vec_ldl(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvrxl(int __a, const unsigned int *__b)
+{
+  return vec_perm((vector unsigned int)(0),
+                  vec_ldl(__a, __b),
+                  vec_lvsl(__a, __b));
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_lvrxl(int __a, const vector unsigned int *__b)
+{
+  return vec_perm((vector unsigned int)(0),
+                  vec_ldl(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector bool int __ATTRS_o_ai
+vec_lvrxl(int __a, const vector bool int *__b)
+{
+  return vec_perm((vector bool int)(0),
+                  vec_ldl(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+static vector float __ATTRS_o_ai
+vec_lvrxl(int __a, const float *__b)
+{
+  return vec_perm((vector float)(0),
+                  vec_ldl(__a, __b),
+                  vec_lvsl(__a, __b));
+}
+
+static vector float __ATTRS_o_ai
+vec_lvrxl(int __a, const vector float *__b)
+{
+  return vec_perm((vector float)(0),
+                  vec_ldl(__a, __b),
+                  vec_lvsl(__a, (unsigned char *)__b));
+}
+
+/* vec_stvlx */
+
+static void __ATTRS_o_ai
+vec_stvlx(vector signed char __a, int __b, signed char *__c)
+{
+  return vec_st(vec_perm(vec_lvrx(__b, __c),
+                         __a,
+                         vec_lvsr(__b, __c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector signed char __a, int __b, vector signed char *__c)
+{
+  return vec_st(vec_perm(vec_lvrx(__b, __c),
+                         __a,
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector unsigned char __a, int __b, unsigned char *__c)
+{
+  return vec_st(vec_perm(vec_lvrx(__b, __c),
+                         __a,
+                         vec_lvsr(__b, __c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector unsigned char __a, int __b, vector unsigned char *__c)
+{
+  return vec_st(vec_perm(vec_lvrx(__b, __c),
+                         __a,
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector bool char __a, int __b, vector bool char *__c)
+{
+  return vec_st(vec_perm(vec_lvrx(__b, __c),
+                         __a,
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector short __a, int __b, short *__c)
+{
+  return vec_st(vec_perm(vec_lvrx(__b, __c),
+                         __a,
+                         vec_lvsr(__b, __c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector short __a, int __b, vector short *__c)
+{
+  return vec_st(vec_perm(vec_lvrx(__b, __c),
+                         __a,
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector unsigned short __a, int __b, unsigned short *__c)
+{
+  return vec_st(vec_perm(vec_lvrx(__b, __c),
+                         __a,
+                         vec_lvsr(__b, __c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector unsigned short __a, int __b, vector unsigned short *__c)
+{
+  return vec_st(vec_perm(vec_lvrx(__b, __c),
+                         __a,
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector bool short __a, int __b, vector bool short *__c)
+{
+  return vec_st(vec_perm(vec_lvrx(__b, __c),
+                         __a,
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector pixel __a, int __b, vector pixel *__c)
+{
+  return vec_st(vec_perm(vec_lvrx(__b, __c),
+                         __a,
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector int __a, int __b, int *__c)
+{
+  return vec_st(vec_perm(vec_lvrx(__b, __c),
+                         __a,
+                         vec_lvsr(__b, __c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector int __a, int __b, vector int *__c)
+{
+  return vec_st(vec_perm(vec_lvrx(__b, __c),
+                         __a,
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector unsigned int __a, int __b, unsigned int *__c)
+{
+  return vec_st(vec_perm(vec_lvrx(__b, __c),
+                         __a,
+                         vec_lvsr(__b, __c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector unsigned int __a, int __b, vector unsigned int *__c)
+{
+  return vec_st(vec_perm(vec_lvrx(__b, __c),
+                         __a,
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector bool int __a, int __b, vector bool int *__c)
+{
+  return vec_st(vec_perm(vec_lvrx(__b, __c),
+                         __a,
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlx(vector float __a, int __b, vector float *__c)
+{
+  return vec_st(vec_perm(vec_lvrx(__b, __c),
+                         __a,
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+/* vec_stvlxl */
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector signed char __a, int __b, signed char *__c)
+{
+  return vec_stl(vec_perm(vec_lvrx(__b, __c),
+                          __a,
+                          vec_lvsr(__b, __c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector signed char __a, int __b, vector signed char *__c)
+{
+  return vec_stl(vec_perm(vec_lvrx(__b, __c),
+                          __a,
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector unsigned char __a, int __b, unsigned char *__c)
+{
+  return vec_stl(vec_perm(vec_lvrx(__b, __c),
+                          __a,
+                          vec_lvsr(__b, __c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector unsigned char __a, int __b, vector unsigned char *__c)
+{
+  return vec_stl(vec_perm(vec_lvrx(__b, __c),
+                          __a,
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector bool char __a, int __b, vector bool char *__c)
+{
+  return vec_stl(vec_perm(vec_lvrx(__b, __c),
+                          __a,
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector short __a, int __b, short *__c)
+{
+  return vec_stl(vec_perm(vec_lvrx(__b, __c),
+                          __a,
+                          vec_lvsr(__b, __c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector short __a, int __b, vector short *__c)
+{
+  return vec_stl(vec_perm(vec_lvrx(__b, __c),
+                          __a,
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector unsigned short __a, int __b, unsigned short *__c)
+{
+  return vec_stl(vec_perm(vec_lvrx(__b, __c),
+                          __a,
+                          vec_lvsr(__b, __c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector unsigned short __a, int __b, vector unsigned short *__c)
+{
+  return vec_stl(vec_perm(vec_lvrx(__b, __c),
+                          __a,
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector bool short __a, int __b, vector bool short *__c)
+{
+  return vec_stl(vec_perm(vec_lvrx(__b, __c),
+                          __a,
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector pixel __a, int __b, vector pixel *__c)
+{
+  return vec_stl(vec_perm(vec_lvrx(__b, __c),
+                          __a,
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector int __a, int __b, int *__c)
+{
+  return vec_stl(vec_perm(vec_lvrx(__b, __c),
+                          __a,
+                          vec_lvsr(__b, __c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector int __a, int __b, vector int *__c)
+{
+  return vec_stl(vec_perm(vec_lvrx(__b, __c),
+                          __a,
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector unsigned int __a, int __b, unsigned int *__c)
+{
+  return vec_stl(vec_perm(vec_lvrx(__b, __c),
+                          __a,
+                          vec_lvsr(__b, __c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector unsigned int __a, int __b, vector unsigned int *__c)
+{
+  return vec_stl(vec_perm(vec_lvrx(__b, __c),
+                          __a,
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector bool int __a, int __b, vector bool int *__c)
+{
+  return vec_stl(vec_perm(vec_lvrx(__b, __c),
+                          __a,
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvlxl(vector float __a, int __b, vector float *__c)
+{
+  return vec_stl(vec_perm(vec_lvrx(__b, __c),
+                          __a,
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+/* vec_stvrx */
+
+static void __ATTRS_o_ai
+vec_stvrx(vector signed char __a, int __b, signed char *__c)
+{
+  return vec_st(vec_perm(__a,
+                         vec_lvlx(__b, __c),
+                         vec_lvsr(__b, __c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector signed char __a, int __b, vector signed char *__c)
+{
+  return vec_st(vec_perm(__a,
+                         vec_lvlx(__b, __c),
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector unsigned char __a, int __b, unsigned char *__c)
+{
+  return vec_st(vec_perm(__a,
+                         vec_lvlx(__b, __c),
+                         vec_lvsr(__b, __c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector unsigned char __a, int __b, vector unsigned char *__c)
+{
+  return vec_st(vec_perm(__a,
+                         vec_lvlx(__b, __c),
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector bool char __a, int __b, vector bool char *__c)
+{
+  return vec_st(vec_perm(__a,
+                         vec_lvlx(__b, __c),
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector short __a, int __b, short *__c)
+{
+  return vec_st(vec_perm(__a,
+                         vec_lvlx(__b, __c),
+                         vec_lvsr(__b, __c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector short __a, int __b, vector short *__c)
+{
+  return vec_st(vec_perm(__a,
+                         vec_lvlx(__b, __c),
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector unsigned short __a, int __b, unsigned short *__c)
+{
+  return vec_st(vec_perm(__a,
+                         vec_lvlx(__b, __c),
+                         vec_lvsr(__b, __c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector unsigned short __a, int __b, vector unsigned short *__c)
+{
+  return vec_st(vec_perm(__a,
+                         vec_lvlx(__b, __c),
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector bool short __a, int __b, vector bool short *__c)
+{
+  return vec_st(vec_perm(__a,
+                         vec_lvlx(__b, __c),
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector pixel __a, int __b, vector pixel *__c)
+{
+  return vec_st(vec_perm(__a,
+                         vec_lvlx(__b, __c),
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector int __a, int __b, int *__c)
+{
+  return vec_st(vec_perm(__a,
+                         vec_lvlx(__b, __c),
+                         vec_lvsr(__b, __c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector int __a, int __b, vector int *__c)
+{
+  return vec_st(vec_perm(__a,
+                         vec_lvlx(__b, __c),
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector unsigned int __a, int __b, unsigned int *__c)
+{
+  return vec_st(vec_perm(__a,
+                         vec_lvlx(__b, __c),
+                         vec_lvsr(__b, __c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector unsigned int __a, int __b, vector unsigned int *__c)
+{
+  return vec_st(vec_perm(__a,
+                         vec_lvlx(__b, __c),
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector bool int __a, int __b, vector bool int *__c)
+{
+  return vec_st(vec_perm(__a,
+                         vec_lvlx(__b, __c),
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrx(vector float __a, int __b, vector float *__c)
+{
+  return vec_st(vec_perm(__a,
+                         vec_lvlx(__b, __c),
+                         vec_lvsr(__b, (unsigned char *)__c)),
+                __b, __c);
+}
+
+/* vec_stvrxl */
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector signed char __a, int __b, signed char *__c)
+{
+  return vec_stl(vec_perm(__a,
+                          vec_lvlx(__b, __c),
+                          vec_lvsr(__b, __c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector signed char __a, int __b, vector signed char *__c)
+{
+  return vec_stl(vec_perm(__a,
+                          vec_lvlx(__b, __c),
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector unsigned char __a, int __b, unsigned char *__c)
+{
+  return vec_stl(vec_perm(__a,
+                          vec_lvlx(__b, __c),
+                          vec_lvsr(__b, __c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector unsigned char __a, int __b, vector unsigned char *__c)
+{
+  return vec_stl(vec_perm(__a,
+                          vec_lvlx(__b, __c),
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector bool char __a, int __b, vector bool char *__c)
+{
+  return vec_stl(vec_perm(__a,
+                          vec_lvlx(__b, __c),
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector short __a, int __b, short *__c)
+{
+  return vec_stl(vec_perm(__a,
+                          vec_lvlx(__b, __c),
+                          vec_lvsr(__b, __c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector short __a, int __b, vector short *__c)
+{
+  return vec_stl(vec_perm(__a,
+                          vec_lvlx(__b, __c),
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector unsigned short __a, int __b, unsigned short *__c)
+{
+  return vec_stl(vec_perm(__a,
+                          vec_lvlx(__b, __c),
+                          vec_lvsr(__b, __c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector unsigned short __a, int __b, vector unsigned short *__c)
+{
+  return vec_stl(vec_perm(__a,
+                          vec_lvlx(__b, __c),
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector bool short __a, int __b, vector bool short *__c)
+{
+  return vec_stl(vec_perm(__a,
+                          vec_lvlx(__b, __c),
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector pixel __a, int __b, vector pixel *__c)
+{
+  return vec_stl(vec_perm(__a,
+                          vec_lvlx(__b, __c),
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector int __a, int __b, int *__c)
+{
+  return vec_stl(vec_perm(__a,
+                          vec_lvlx(__b, __c),
+                          vec_lvsr(__b, __c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector int __a, int __b, vector int *__c)
+{
+  return vec_stl(vec_perm(__a,
+                          vec_lvlx(__b, __c),
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector unsigned int __a, int __b, unsigned int *__c)
+{
+  return vec_stl(vec_perm(__a,
+                          vec_lvlx(__b, __c),
+                          vec_lvsr(__b, __c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector unsigned int __a, int __b, vector unsigned int *__c)
+{
+  return vec_stl(vec_perm(__a,
+                          vec_lvlx(__b, __c),
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector bool int __a, int __b, vector bool int *__c)
+{
+  return vec_stl(vec_perm(__a,
+                          vec_lvlx(__b, __c),
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+static void __ATTRS_o_ai
+vec_stvrxl(vector float __a, int __b, vector float *__c)
+{
+  return vec_stl(vec_perm(__a,
+                          vec_lvlx(__b, __c),
+                          vec_lvsr(__b, (unsigned char *)__c)),
+                 __b, __c);
+}
+
+/* vec_promote */
+
+static vector signed char __ATTRS_o_ai
+vec_promote(signed char __a, int __b)
+{
+  vector signed char __res = (vector signed char)(0);
+  __res[__b] = __a;
+  return __res;
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_promote(unsigned char __a, int __b)
+{
+  vector unsigned char __res = (vector unsigned char)(0);
+  __res[__b] = __a;
+  return __res;
+}
+
+static vector short __ATTRS_o_ai
+vec_promote(short __a, int __b)
+{
+  vector short __res = (vector short)(0);
+  __res[__b] = __a;
+  return __res;
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_promote(unsigned short __a, int __b)
+{
+  vector unsigned short __res = (vector unsigned short)(0);
+  __res[__b] = __a;
+  return __res;
+}
+
+static vector int __ATTRS_o_ai
+vec_promote(int __a, int __b)
+{
+  vector int __res = (vector int)(0);
+  __res[__b] = __a;
+  return __res;
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_promote(unsigned int __a, int __b)
+{
+  vector unsigned int __res = (vector unsigned int)(0);
+  __res[__b] = __a;
+  return __res;
+}
+
+static vector float __ATTRS_o_ai
+vec_promote(float __a, int __b)
+{
+  vector float __res = (vector float)(0);
+  __res[__b] = __a;
+  return __res;
+}
+
+/* vec_splats */
+
+static vector signed char __ATTRS_o_ai
+vec_splats(signed char __a)
+{
+  return (vector signed char)(__a);
+}
+
+static vector unsigned char __ATTRS_o_ai
+vec_splats(unsigned char __a)
+{
+  return (vector unsigned char)(__a);
+}
+
+static vector short __ATTRS_o_ai
+vec_splats(short __a)
+{
+  return (vector short)(__a);
+}
+
+static vector unsigned short __ATTRS_o_ai
+vec_splats(unsigned short __a)
+{
+  return (vector unsigned short)(__a);
+}
+
+static vector int __ATTRS_o_ai
+vec_splats(int __a)
+{
+  return (vector int)(__a);
+}
+
+static vector unsigned int __ATTRS_o_ai
+vec_splats(unsigned int __a)
+{
+  return (vector unsigned int)(__a);
+}
+
+static vector float __ATTRS_o_ai
+vec_splats(float __a)
+{
+  return (vector float)(__a);
+}
+
+/* ----------------------------- predicates --------------------------------- */
+
+/* vec_all_eq */
+
+static int __ATTRS_o_ai
+vec_all_eq(vector signed char __a, vector signed char __b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_LT, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector signed char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_LT, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector unsigned char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_LT, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector unsigned char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_LT, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector bool char __a, vector signed char __b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_LT, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector bool char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_LT, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector bool char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_LT, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_LT, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_LT, __a, (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector unsigned short __a, vector unsigned short __b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_LT, (vector short)__a, (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector unsigned short __a, vector bool short __b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_LT, (vector short)__a, (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector bool short __a, vector short __b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_LT, (vector short)__a, (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector bool short __a, vector unsigned short __b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_LT, (vector short)__a, (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector bool short __a, vector bool short __b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_LT, (vector short)__a, (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector pixel __a, vector pixel __b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_LT, (vector short)__a, (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_LT, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_LT, __a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_LT, (vector int)__a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector unsigned int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_LT, (vector int)__a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector bool int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_LT, (vector int)__a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector bool int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_LT, (vector int)__a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector bool int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_LT, (vector int)__a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_eq(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpeqfp_p(__CR6_LT, __a, __b);
+}
+
+/* vec_all_ge */
+
+static int __ATTRS_o_ai
+vec_all_ge(vector signed char __a, vector signed char __b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_EQ, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector signed char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_EQ, (vector signed char)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector unsigned char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector unsigned char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ, (vector unsigned char)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector bool char __a, vector signed char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ,
+                                      (vector unsigned char)__b,
+                                      (vector unsigned char)__a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector bool char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ, __b, (vector unsigned char)__a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector bool char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ,
+                                      (vector unsigned char)__b,
+                                      (vector unsigned char)__a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_EQ, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_EQ, (vector short)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector unsigned short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector unsigned short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ, (vector unsigned short)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector bool short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ,
+                                      (vector unsigned short)__b,
+                                      (vector unsigned short)__a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector bool short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ, __b, (vector unsigned short)__a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector bool short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ,
+                                      (vector unsigned short)__b,
+                                      (vector unsigned short)__a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_EQ, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_EQ, (vector int)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector unsigned int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ, (vector unsigned int)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector bool int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ,
+                                      (vector unsigned int)__b,
+                                      (vector unsigned int)__a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector bool int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ, __b, (vector unsigned int)__a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector bool int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ,
+                                      (vector unsigned int)__b,
+                                      (vector unsigned int)__a);
+}
+
+static int __ATTRS_o_ai
+vec_all_ge(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpgefp_p(__CR6_LT, __a, __b);
+}
+
+/* vec_all_gt */
+
+static int __ATTRS_o_ai
+vec_all_gt(vector signed char __a, vector signed char __b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_LT, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector signed char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_LT, __a, (vector signed char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector unsigned char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector unsigned char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT, __a, (vector unsigned char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector bool char __a, vector signed char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT,
+                                      (vector unsigned char)__a,
+                                      (vector unsigned char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector bool char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT, (vector unsigned char)__a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector bool char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT,
+                                      (vector unsigned char)__a,
+                                      (vector unsigned char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_LT, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_LT, __a, (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector unsigned short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector unsigned short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT, __a, (vector unsigned short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector bool short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT,
+                                      (vector unsigned short)__a,
+                                      (vector unsigned short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector bool short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT, (vector unsigned short)__a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector bool short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT,
+                                      (vector unsigned short)__a,
+                                      (vector unsigned short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_LT, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_LT, __a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector unsigned int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT, __a, (vector unsigned int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector bool int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT,
+                                      (vector unsigned int)__a,
+                                      (vector unsigned int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector bool int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT, (vector unsigned int)__a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector bool int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT,
+                                      (vector unsigned int)__a,
+                                      (vector unsigned int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_gt(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpgtfp_p(__CR6_LT, __a, __b);
+}
+
+/* vec_all_in */
+
+static int __attribute__((__always_inline__))
+vec_all_in(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpbfp_p(__CR6_EQ, __a, __b);
+}
+
+/* vec_all_le */
+
+static int __ATTRS_o_ai
+vec_all_le(vector signed char __a, vector signed char __b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_EQ, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector signed char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_EQ, __a, (vector signed char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector unsigned char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector unsigned char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ, __a, (vector unsigned char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector bool char __a, vector signed char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ,
+                                      (vector unsigned char)__a,
+                                      (vector unsigned char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector bool char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ, (vector unsigned char)__a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector bool char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ,
+                                      (vector unsigned char)__a,
+                                      (vector unsigned char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_EQ, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_EQ, __a, (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector unsigned short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector unsigned short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ, __a, (vector unsigned short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector bool short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ,
+                                      (vector unsigned short)__a,
+                                      (vector unsigned short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector bool short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ, (vector unsigned short)__a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector bool short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ,
+                                      (vector unsigned short)__a,
+                                      (vector unsigned short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_EQ, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_EQ, __a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector unsigned int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ, __a, (vector unsigned int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector bool int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ,
+                                      (vector unsigned int)__a,
+                                      (vector unsigned int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector bool int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ, (vector unsigned int)__a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector bool int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ,
+                                      (vector unsigned int)__a,
+                                      (vector unsigned int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_le(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpgefp_p(__CR6_LT, __b, __a);
+}
+
+/* vec_all_lt */
+
+static int __ATTRS_o_ai
+vec_all_lt(vector signed char __a, vector signed char __b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_LT, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector signed char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_LT, (vector signed char)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector unsigned char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector unsigned char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT, (vector unsigned char)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector bool char __a, vector signed char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT,
+                                      (vector unsigned char)__b,
+                                      (vector unsigned char)__a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector bool char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT, __b, (vector unsigned char)__a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector bool char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT,
+                                      (vector unsigned char)__b,
+                                      (vector unsigned char)__a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_LT, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_LT, (vector short)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector unsigned short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector unsigned short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT, (vector unsigned short)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector bool short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT,
+                                      (vector unsigned short)__b,
+                                      (vector unsigned short)__a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector bool short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT, __b, (vector unsigned short)__a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector bool short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT,
+                                      (vector unsigned short)__b,
+                                      (vector unsigned short)__a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_LT, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_LT, (vector int)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector unsigned int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT, (vector unsigned int)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector bool int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT,
+                                      (vector unsigned int)__b,
+                                      (vector unsigned int)__a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector bool int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT, __b, (vector unsigned int)__a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector bool int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT,
+                                      (vector unsigned int)__b,
+                                      (vector unsigned int)__a);
+}
+
+static int __ATTRS_o_ai
+vec_all_lt(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpgtfp_p(__CR6_LT, __b, __a);
+}
+
+/* vec_all_nan */
+
+static int __attribute__((__always_inline__))
+vec_all_nan(vector float __a)
+{
+  return __builtin_altivec_vcmpeqfp_p(__CR6_EQ, __a, __a);
+}
+
+/* vec_all_ne */
+
+static int __ATTRS_o_ai
+vec_all_ne(vector signed char __a, vector signed char __b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_EQ, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector signed char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_EQ, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector unsigned char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_EQ, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector unsigned char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_EQ, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector bool char __a, vector signed char __b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_EQ, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector bool char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_EQ, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector bool char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpequb_p(__CR6_EQ, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_EQ, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_EQ, __a, (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector unsigned short __a, vector unsigned short __b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_EQ, (vector short)__a, (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector unsigned short __a, vector bool short __b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_EQ, (vector short)__a, (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector bool short __a, vector short __b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_EQ, (vector short)__a, (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector bool short __a, vector unsigned short __b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_EQ, (vector short)__a, (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector bool short __a, vector bool short __b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_EQ, (vector short)__a, (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector pixel __a, vector pixel __b)
+{
+  return
+    __builtin_altivec_vcmpequh_p(__CR6_EQ, (vector short)__a, (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_EQ, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_EQ, __a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_EQ, (vector int)__a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector unsigned int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_EQ, (vector int)__a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector bool int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_EQ, (vector int)__a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector bool int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_EQ, (vector int)__a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector bool int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_EQ, (vector int)__a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_all_ne(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpeqfp_p(__CR6_EQ, __a, __b);
+}
+
+/* vec_all_nge */
+
+static int __attribute__((__always_inline__))
+vec_all_nge(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpgefp_p(__CR6_EQ, __a, __b);
+}
+
+/* vec_all_ngt */
+
+static int __attribute__((__always_inline__))
+vec_all_ngt(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpgtfp_p(__CR6_EQ, __a, __b);
+}
+
+/* vec_all_nle */
+
+static int __attribute__((__always_inline__))
+vec_all_nle(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpgefp_p(__CR6_EQ, __b, __a);
+}
+
+/* vec_all_nlt */
+
+static int __attribute__((__always_inline__))
+vec_all_nlt(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpgtfp_p(__CR6_EQ, __b, __a);
+}
+
+/* vec_all_numeric */
+
+static int __attribute__((__always_inline__))
+vec_all_numeric(vector float __a)
+{
+  return __builtin_altivec_vcmpeqfp_p(__CR6_LT, __a, __a);
+}
+
+/* vec_any_eq */
+
+static int __ATTRS_o_ai
+vec_any_eq(vector signed char __a, vector signed char __b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_EQ_REV, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector signed char __a, vector bool char __b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_EQ_REV, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector unsigned char __a, vector unsigned char __b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_EQ_REV, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector unsigned char __a, vector bool char __b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_EQ_REV, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector bool char __a, vector signed char __b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_EQ_REV, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector bool char __a, vector unsigned char __b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_EQ_REV, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector bool char __a, vector bool char __b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_EQ_REV, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_EQ_REV, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_EQ_REV, __a, (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector unsigned short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_EQ_REV, 
+                                      (vector short)__a,
+                                      (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector unsigned short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_EQ_REV, 
+                                      (vector short)__a,
+                                      (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector bool short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_EQ_REV,
+                                      (vector short)__a,
+                                      (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector bool short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_EQ_REV,
+                                      (vector short)__a,
+                                      (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector bool short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_EQ_REV,
+                                      (vector short)__a,
+                                      (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector pixel __a, vector pixel __b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_EQ_REV, 
+                                      (vector short)__a,
+                                      (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_EQ_REV, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_EQ_REV, __a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector unsigned int __a, vector unsigned int __b)
+{
+  return
+    __builtin_altivec_vcmpequw_p(__CR6_EQ_REV, (vector int)__a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector unsigned int __a, vector bool int __b)
+{
+  return
+    __builtin_altivec_vcmpequw_p(__CR6_EQ_REV, (vector int)__a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector bool int __a, vector int __b)
+{
+  return
+    __builtin_altivec_vcmpequw_p(__CR6_EQ_REV, (vector int)__a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector bool int __a, vector unsigned int __b)
+{
+  return
+    __builtin_altivec_vcmpequw_p(__CR6_EQ_REV, (vector int)__a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector bool int __a, vector bool int __b)
+{
+  return
+    __builtin_altivec_vcmpequw_p(__CR6_EQ_REV, (vector int)__a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_eq(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpeqfp_p(__CR6_EQ_REV, __a, __b);
+}
+
+/* vec_any_ge */
+
+static int __ATTRS_o_ai
+vec_any_ge(vector signed char __a, vector signed char __b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_LT_REV, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector signed char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_LT_REV, (vector signed char)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector unsigned char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT_REV, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector unsigned char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT_REV, (vector unsigned char)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector bool char __a, vector signed char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT_REV,
+                                      (vector unsigned char)__b,
+                                      (vector unsigned char)__a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector bool char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT_REV, __b, (vector unsigned char)__a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector bool char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT_REV,
+                                      (vector unsigned char)__b,
+                                      (vector unsigned char)__a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_LT_REV, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_LT_REV, (vector short)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector unsigned short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT_REV, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector unsigned short __a, vector bool short __b)
+{
+  return
+    __builtin_altivec_vcmpgtuh_p(__CR6_LT_REV, (vector unsigned short)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector bool short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT_REV,
+                                      (vector unsigned short)__b,
+                                      (vector unsigned short)__a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector bool short __a, vector unsigned short __b)
+{
+  return 
+    __builtin_altivec_vcmpgtuh_p(__CR6_LT_REV, __b, (vector unsigned short)__a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector bool short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT_REV,
+                                      (vector unsigned short)__b,
+                                      (vector unsigned short)__a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_LT_REV, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_LT_REV, (vector int)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT_REV, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector unsigned int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT_REV, (vector unsigned int)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector bool int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT_REV,
+                                      (vector unsigned int)__b,
+                                      (vector unsigned int)__a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector bool int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT_REV, __b, (vector unsigned int)__a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector bool int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT_REV,
+                                      (vector unsigned int)__b,
+                                      (vector unsigned int)__a);
+}
+
+static int __ATTRS_o_ai
+vec_any_ge(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpgefp_p(__CR6_EQ_REV, __a, __b);
+}
+
+/* vec_any_gt */
+
+static int __ATTRS_o_ai
+vec_any_gt(vector signed char __a, vector signed char __b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_EQ_REV, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector signed char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_EQ_REV, __a, (vector signed char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector unsigned char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ_REV, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector unsigned char __a, vector bool char __b)
+{
+  return 
+    __builtin_altivec_vcmpgtub_p(__CR6_EQ_REV, __a, (vector unsigned char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector bool char __a, vector signed char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ_REV,
+                                      (vector unsigned char)__a,
+                                      (vector unsigned char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector bool char __a, vector unsigned char __b)
+{
+  return 
+    __builtin_altivec_vcmpgtub_p(__CR6_EQ_REV, (vector unsigned char)__a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector bool char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ_REV,
+                                      (vector unsigned char)__a,
+                                      (vector unsigned char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_EQ_REV, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_EQ_REV, __a, (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector unsigned short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ_REV, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector unsigned short __a, vector bool short __b)
+{
+  return 
+    __builtin_altivec_vcmpgtuh_p(__CR6_EQ_REV, __a, (vector unsigned short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector bool short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ_REV,
+                                      (vector unsigned short)__a,
+                                      (vector unsigned short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector bool short __a, vector unsigned short __b)
+{
+  return
+    __builtin_altivec_vcmpgtuh_p(__CR6_EQ_REV, (vector unsigned short)__a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector bool short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ_REV,
+                                      (vector unsigned short)__a,
+                                      (vector unsigned short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_EQ_REV, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_EQ_REV, __a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ_REV, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector unsigned int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ_REV, __a, (vector unsigned int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector bool int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ_REV,
+                                      (vector unsigned int)__a,
+                                      (vector unsigned int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector bool int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ_REV, (vector unsigned int)__a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector bool int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ_REV,
+                                      (vector unsigned int)__a,
+                                      (vector unsigned int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_gt(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpgtfp_p(__CR6_EQ_REV, __a, __b);
+}
+
+/* vec_any_le */
+
+static int __ATTRS_o_ai
+vec_any_le(vector signed char __a, vector signed char __b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_LT_REV, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector signed char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_LT_REV, __a, (vector signed char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector unsigned char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT_REV, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector unsigned char __a, vector bool char __b)
+{
+  return 
+    __builtin_altivec_vcmpgtub_p(__CR6_LT_REV, __a, (vector unsigned char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector bool char __a, vector signed char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT_REV,
+                                      (vector unsigned char)__a,
+                                      (vector unsigned char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector bool char __a, vector unsigned char __b)
+{
+  return 
+    __builtin_altivec_vcmpgtub_p(__CR6_LT_REV, (vector unsigned char)__a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector bool char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_LT_REV,
+                                      (vector unsigned char)__a,
+                                      (vector unsigned char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_LT_REV, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_LT_REV, __a, (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector unsigned short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT_REV, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector unsigned short __a, vector bool short __b)
+{
+  return 
+    __builtin_altivec_vcmpgtuh_p(__CR6_LT_REV, __a, (vector unsigned short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector bool short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT_REV,
+                                      (vector unsigned short)__a,
+                                      (vector unsigned short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector bool short __a, vector unsigned short __b)
+{
+  return 
+    __builtin_altivec_vcmpgtuh_p(__CR6_LT_REV, (vector unsigned short)__a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector bool short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_LT_REV,
+                                      (vector unsigned short)__a,
+                                      (vector unsigned short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_LT_REV, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_LT_REV, __a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT_REV, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector unsigned int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT_REV, __a, (vector unsigned int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector bool int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT_REV,
+                                      (vector unsigned int)__a,
+                                      (vector unsigned int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector bool int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT_REV, (vector unsigned int)__a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector bool int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_LT_REV,
+                                      (vector unsigned int)__a,
+                                      (vector unsigned int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_le(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpgefp_p(__CR6_EQ_REV, __b, __a);
+}
+
+/* vec_any_lt */
+
+static int __ATTRS_o_ai
+vec_any_lt(vector signed char __a, vector signed char __b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_EQ_REV, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector signed char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtsb_p(__CR6_EQ_REV, (vector signed char)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector unsigned char __a, vector unsigned char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ_REV, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector unsigned char __a, vector bool char __b)
+{
+  return 
+    __builtin_altivec_vcmpgtub_p(__CR6_EQ_REV, (vector unsigned char)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector bool char __a, vector signed char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ_REV,
+                                      (vector unsigned char)__b,
+                                      (vector unsigned char)__a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector bool char __a, vector unsigned char __b)
+{
+  return 
+    __builtin_altivec_vcmpgtub_p(__CR6_EQ_REV, __b, (vector unsigned char)__a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector bool char __a, vector bool char __b)
+{
+  return __builtin_altivec_vcmpgtub_p(__CR6_EQ_REV,
+                                      (vector unsigned char)__b,
+                                      (vector unsigned char)__a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_EQ_REV, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpgtsh_p(__CR6_EQ_REV, (vector short)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector unsigned short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ_REV, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector unsigned short __a, vector bool short __b)
+{
+  return 
+    __builtin_altivec_vcmpgtuh_p(__CR6_EQ_REV, (vector unsigned short)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector bool short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ_REV,
+                                      (vector unsigned short)__b,
+                                      (vector unsigned short)__a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector bool short __a, vector unsigned short __b)
+{
+  return 
+    __builtin_altivec_vcmpgtuh_p(__CR6_EQ_REV, __b, (vector unsigned short)__a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector bool short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpgtuh_p(__CR6_EQ_REV,
+                                      (vector unsigned short)__b,
+                                      (vector unsigned short)__a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_EQ_REV, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtsw_p(__CR6_EQ_REV, (vector int)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector unsigned int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ_REV, __b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector unsigned int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ_REV, (vector unsigned int)__b, __a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector bool int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ_REV,
+                                      (vector unsigned int)__b,
+                                      (vector unsigned int)__a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector bool int __a, vector unsigned int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ_REV, __b, (vector unsigned int)__a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector bool int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpgtuw_p(__CR6_EQ_REV,
+                                      (vector unsigned int)__b,
+                                      (vector unsigned int)__a);
+}
+
+static int __ATTRS_o_ai
+vec_any_lt(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpgtfp_p(__CR6_EQ_REV, __b, __a);
+}
+
+/* vec_any_nan */
+
+static int __attribute__((__always_inline__))
+vec_any_nan(vector float __a)
+{
+  return __builtin_altivec_vcmpeqfp_p(__CR6_LT_REV, __a, __a);
+}
+
+/* vec_any_ne */
+
+static int __ATTRS_o_ai
+vec_any_ne(vector signed char __a, vector signed char __b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_LT_REV, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector signed char __a, vector bool char __b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_LT_REV, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector unsigned char __a, vector unsigned char __b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_LT_REV, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector unsigned char __a, vector bool char __b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_LT_REV, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector bool char __a, vector signed char __b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_LT_REV, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector bool char __a, vector unsigned char __b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_LT_REV, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector bool char __a, vector bool char __b)
+{
+  return
+    __builtin_altivec_vcmpequb_p(__CR6_LT_REV, (vector char)__a, (vector char)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_LT_REV, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_LT_REV, __a, (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector unsigned short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_LT_REV, 
+                                      (vector short)__a,
+                                      (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector unsigned short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_LT_REV,
+                                      (vector short)__a,
+                                      (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector bool short __a, vector short __b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_LT_REV,
+                                      (vector short)__a,
+                                      (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector bool short __a, vector unsigned short __b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_LT_REV,
+                                      (vector short)__a,
+                                      (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector bool short __a, vector bool short __b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_LT_REV,
+                                      (vector short)__a,
+                                      (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector pixel __a, vector pixel __b)
+{
+  return __builtin_altivec_vcmpequh_p(__CR6_LT_REV,
+                                      (vector short)__a,
+                                      (vector short)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector int __a, vector int __b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_LT_REV, __a, __b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector int __a, vector bool int __b)
+{
+  return __builtin_altivec_vcmpequw_p(__CR6_LT_REV, __a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector unsigned int __a, vector unsigned int __b)
+{
+  return
+    __builtin_altivec_vcmpequw_p(__CR6_LT_REV, (vector int)__a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector unsigned int __a, vector bool int __b)
+{
+  return
+    __builtin_altivec_vcmpequw_p(__CR6_LT_REV, (vector int)__a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector bool int __a, vector int __b)
+{
+  return
+    __builtin_altivec_vcmpequw_p(__CR6_LT_REV, (vector int)__a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector bool int __a, vector unsigned int __b)
+{
+  return
+    __builtin_altivec_vcmpequw_p(__CR6_LT_REV, (vector int)__a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector bool int __a, vector bool int __b)
+{
+  return
+    __builtin_altivec_vcmpequw_p(__CR6_LT_REV, (vector int)__a, (vector int)__b);
+}
+
+static int __ATTRS_o_ai
+vec_any_ne(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpeqfp_p(__CR6_LT_REV, __a, __b);
+}
+
+/* vec_any_nge */
+
+static int __attribute__((__always_inline__))
+vec_any_nge(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpgefp_p(__CR6_LT_REV, __a, __b);
+}
+
+/* vec_any_ngt */
+
+static int __attribute__((__always_inline__))
+vec_any_ngt(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpgtfp_p(__CR6_LT_REV, __a, __b);
+}
+
+/* vec_any_nle */
+
+static int __attribute__((__always_inline__))
+vec_any_nle(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpgefp_p(__CR6_LT_REV, __b, __a);
+}
+
+/* vec_any_nlt */
+
+static int __attribute__((__always_inline__))
+vec_any_nlt(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpgtfp_p(__CR6_LT_REV, __b, __a);
+}
+
+/* vec_any_numeric */
+
+static int __attribute__((__always_inline__))
+vec_any_numeric(vector float __a)
+{
+  return __builtin_altivec_vcmpeqfp_p(__CR6_EQ_REV, __a, __a);
+}
+
+/* vec_any_out */
+
+static int __attribute__((__always_inline__))
+vec_any_out(vector float __a, vector float __b)
+{
+  return __builtin_altivec_vcmpbfp_p(__CR6_EQ_REV, __a, __b);
+}
+
+#undef __ATTRS_o_ai
+
+#endif /* __ALTIVEC_H */
diff --git a/lib/clang/3.6/include/ammintrin.h b/lib/clang/3.6/include/ammintrin.h
new file mode 100644
index 0000000..d87b9cd
--- /dev/null
+++ b/lib/clang/3.6/include/ammintrin.h
@@ -0,0 +1,68 @@
+/*===---- ammintrin.h - SSE4a intrinsics -----------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __AMMINTRIN_H
+#define __AMMINTRIN_H
+
+#ifndef __SSE4A__
+#error "SSE4A instruction set not enabled"
+#else
+
+#include <pmmintrin.h>
+
+#define _mm_extracti_si64(x, len, idx) \
+  ((__m128i)__builtin_ia32_extrqi((__v2di)(__m128i)(x), \
+                                  (char)(len), (char)(idx)))
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_extract_si64(__m128i __x, __m128i __y)
+{
+  return (__m128i)__builtin_ia32_extrq((__v2di)__x, (__v16qi)__y);
+}
+
+#define _mm_inserti_si64(x, y, len, idx) \
+  ((__m128i)__builtin_ia32_insertqi((__v2di)(__m128i)(x), \
+                                    (__v2di)(__m128i)(y), \
+                                    (char)(len), (char)(idx)))
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_insert_si64(__m128i __x, __m128i __y)
+{
+  return (__m128i)__builtin_ia32_insertq((__v2di)__x, (__v2di)__y);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_stream_sd(double *__p, __m128d __a)
+{
+  __builtin_ia32_movntsd(__p, (__v2df)__a);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_stream_ss(float *__p, __m128 __a)
+{
+  __builtin_ia32_movntss(__p, (__v4sf)__a);
+}
+
+#endif /* __SSE4A__ */
+
+#endif /* __AMMINTRIN_H */
diff --git a/lib/clang/3.6/include/arm_acle.h b/lib/clang/3.6/include/arm_acle.h
new file mode 100644
index 0000000..814df2c
--- /dev/null
+++ b/lib/clang/3.6/include/arm_acle.h
@@ -0,0 +1,296 @@
+/*===---- arm_acle.h - ARM Non-Neon intrinsics -----------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __ARM_ACLE_H
+#define __ARM_ACLE_H
+
+#ifndef __ARM_ACLE
+#error "ACLE intrinsics support not enabled."
+#endif
+
+#include <stdint.h>
+
+#if defined(__cplusplus)
+extern "C" {
+#endif
+
+/* 8 SYNCHRONIZATION, BARRIER AND HINT INTRINSICS */
+/* 8.3 Memory barriers */
+#if !defined(_MSC_VER)
+#define __dmb(i) __builtin_arm_dmb(i)
+#define __dsb(i) __builtin_arm_dsb(i)
+#define __isb(i) __builtin_arm_isb(i)
+#endif
+
+/* 8.4 Hints */
+
+#if !defined(_MSC_VER)
+static __inline__ void __attribute__((always_inline, nodebug)) __wfi(void) {
+  __builtin_arm_wfi();
+}
+
+static __inline__ void __attribute__((always_inline, nodebug)) __wfe(void) {
+  __builtin_arm_wfe();
+}
+
+static __inline__ void __attribute__((always_inline, nodebug)) __sev(void) {
+  __builtin_arm_sev();
+}
+
+static __inline__ void __attribute__((always_inline, nodebug)) __sevl(void) {
+  __builtin_arm_sevl();
+}
+
+static __inline__ void __attribute__((always_inline, nodebug)) __yield(void) {
+  __builtin_arm_yield();
+}
+#endif
+
+#if __ARM_32BIT_STATE
+#define __dbg(t) __builtin_arm_dbg(t)
+#endif
+
+/* 8.5 Swap */
+static __inline__ uint32_t __attribute__((always_inline, nodebug))
+  __swp(uint32_t x, volatile uint32_t *p) {
+  uint32_t v;
+  do v = __builtin_arm_ldrex(p); while (__builtin_arm_strex(x, p));
+  return v;
+}
+
+/* 8.6 Memory prefetch intrinsics */
+/* 8.6.1 Data prefetch */
+#define __pld(addr) __pldx(0, 0, 0, addr)
+
+#if __ARM_32BIT_STATE
+#define __pldx(access_kind, cache_level, retention_policy, addr) \
+  __builtin_arm_prefetch(addr, access_kind, 1)
+#else
+#define __pldx(access_kind, cache_level, retention_policy, addr) \
+  __builtin_arm_prefetch(addr, access_kind, cache_level, retention_policy, 1)
+#endif
+
+/* 8.6.2 Instruction prefetch */
+#define __pli(addr) __plix(0, 0, addr)
+
+#if __ARM_32BIT_STATE
+#define __plix(cache_level, retention_policy, addr) \
+  __builtin_arm_prefetch(addr, 0, 0)
+#else
+#define __plix(cache_level, retention_policy, addr) \
+  __builtin_arm_prefetch(addr, 0, cache_level, retention_policy, 0)
+#endif
+
+/* 8.7 NOP */
+static __inline__ void __attribute__((always_inline, nodebug)) __nop(void) {
+  __builtin_arm_nop();
+}
+
+/* 9 DATA-PROCESSING INTRINSICS */
+/* 9.2 Miscellaneous data-processing intrinsics */
+/* ROR */
+static __inline__ uint32_t __attribute__((always_inline, nodebug))
+  __ror(uint32_t x, uint32_t y) {
+  y %= 32;
+  if (y == 0)  return x;
+  return (x >> y) | (x << (32 - y));
+}
+
+static __inline__ uint64_t __attribute__((always_inline, nodebug))
+  __rorll(uint64_t x, uint32_t y) {
+  y %= 64;
+  if (y == 0)  return x;
+  return (x >> y) | (x << (64 - y));
+}
+
+static __inline__ unsigned long __attribute__((always_inline, nodebug))
+  __rorl(unsigned long x, uint32_t y) {
+#if __SIZEOF_LONG__ == 4
+  return __ror(x, y);
+#else
+  return __rorll(x, y);
+#endif
+}
+
+
+/* CLZ */
+static __inline__ uint32_t __attribute__((always_inline, nodebug))
+  __clz(uint32_t t) {
+  return __builtin_clz(t);
+}
+
+static __inline__ unsigned long __attribute__((always_inline, nodebug))
+  __clzl(unsigned long t) {
+  return __builtin_clzl(t);
+}
+
+static __inline__ uint64_t __attribute__((always_inline, nodebug))
+  __clzll(uint64_t t) {
+  return __builtin_clzll(t);
+}
+
+/* REV */
+static __inline__ uint32_t __attribute__((always_inline, nodebug))
+  __rev(uint32_t t) {
+  return __builtin_bswap32(t);
+}
+
+static __inline__ unsigned long __attribute__((always_inline, nodebug))
+  __revl(unsigned long t) {
+#if __SIZEOF_LONG__ == 4
+  return __builtin_bswap32(t);
+#else
+  return __builtin_bswap64(t);
+#endif
+}
+
+static __inline__ uint64_t __attribute__((always_inline, nodebug))
+  __revll(uint64_t t) {
+  return __builtin_bswap64(t);
+}
+
+/* REV16 */
+static __inline__ uint32_t __attribute__((always_inline, nodebug))
+  __rev16(uint32_t t) {
+  return __ror(__rev(t), 16);
+}
+
+static __inline__ unsigned long __attribute__((always_inline, nodebug))
+  __rev16l(unsigned long t) {
+    return __rorl(__revl(t), sizeof(long) / 2);
+}
+
+static __inline__ uint64_t __attribute__((always_inline, nodebug))
+  __rev16ll(uint64_t t) {
+  return __rorll(__revll(t), 32);
+}
+
+/* REVSH */
+static __inline__ int16_t __attribute__((always_inline, nodebug))
+  __revsh(int16_t t) {
+  return __builtin_bswap16(t);
+}
+
+/* RBIT */
+static __inline__ uint32_t __attribute__((always_inline, nodebug))
+  __rbit(uint32_t t) {
+  return __builtin_arm_rbit(t);
+}
+
+static __inline__ uint64_t __attribute__((always_inline, nodebug))
+  __rbitll(uint64_t t) {
+#if __ARM_32BIT_STATE
+  return (((uint64_t) __builtin_arm_rbit(t)) << 32) |
+    __builtin_arm_rbit(t >> 32);
+#else
+  return __builtin_arm_rbit64(t);
+#endif
+}
+
+static __inline__ unsigned long __attribute__((always_inline, nodebug))
+  __rbitl(unsigned long t) {
+#if __SIZEOF_LONG__ == 4
+  return __rbit(t);
+#else
+  return __rbitll(t);
+#endif
+}
+
+/*
+ * 9.4 Saturating intrinsics
+ *
+ * FIXME: Change guard to their corrosponding __ARM_FEATURE flag when Q flag
+ * intrinsics are implemented and the flag is enabled.
+ */
+/* 9.4.1 Width-specified saturation intrinsics */
+#if __ARM_32BIT_STATE
+#define __ssat(x, y) __builtin_arm_ssat(x, y)
+#define __usat(x, y) __builtin_arm_usat(x, y)
+#endif
+
+/* 9.4.2 Saturating addition and subtraction intrinsics */
+#if __ARM_32BIT_STATE
+static __inline__ int32_t __attribute__((always_inline, nodebug))
+  __qadd(int32_t t, int32_t v) {
+  return __builtin_arm_qadd(t, v);
+}
+
+static __inline__ int32_t __attribute__((always_inline, nodebug))
+  __qsub(int32_t t, int32_t v) {
+  return __builtin_arm_qsub(t, v);
+}
+
+static __inline__ int32_t __attribute__((always_inline, nodebug))
+__qdbl(int32_t t) {
+  return __builtin_arm_qadd(t, t);
+}
+#endif
+
+/* 9.7 CRC32 intrinsics */
+#if __ARM_FEATURE_CRC32
+static __inline__ uint32_t __attribute__((always_inline, nodebug))
+  __crc32b(uint32_t a, uint8_t b) {
+  return __builtin_arm_crc32b(a, b);
+}
+
+static __inline__ uint32_t __attribute__((always_inline, nodebug))
+  __crc32h(uint32_t a, uint16_t b) {
+  return __builtin_arm_crc32h(a, b);
+}
+
+static __inline__ uint32_t __attribute__((always_inline, nodebug))
+  __crc32w(uint32_t a, uint32_t b) {
+  return __builtin_arm_crc32w(a, b);
+}
+
+static __inline__ uint32_t __attribute__((always_inline, nodebug))
+  __crc32d(uint32_t a, uint64_t b) {
+  return __builtin_arm_crc32d(a, b);
+}
+
+static __inline__ uint32_t __attribute__((always_inline, nodebug))
+  __crc32cb(uint32_t a, uint8_t b) {
+  return __builtin_arm_crc32cb(a, b);
+}
+
+static __inline__ uint32_t __attribute__((always_inline, nodebug))
+  __crc32ch(uint32_t a, uint16_t b) {
+  return __builtin_arm_crc32ch(a, b);
+}
+
+static __inline__ uint32_t __attribute__((always_inline, nodebug))
+  __crc32cw(uint32_t a, uint32_t b) {
+  return __builtin_arm_crc32cw(a, b);
+}
+
+static __inline__ uint32_t __attribute__((always_inline, nodebug))
+  __crc32cd(uint32_t a, uint64_t b) {
+  return __builtin_arm_crc32cd(a, b);
+}
+#endif
+
+#if defined(__cplusplus)
+}
+#endif
+
+#endif /* __ARM_ACLE_H */
diff --git a/lib/clang/3.6/include/arm_neon.h b/lib/clang/3.6/include/arm_neon.h
new file mode 100644
index 0000000..cba1c19
--- /dev/null
+++ b/lib/clang/3.6/include/arm_neon.h
@@ -0,0 +1,68419 @@
+/*===---- arm_neon.h - ARM Neon intrinsics ---------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __ARM_NEON_H
+#define __ARM_NEON_H
+
+#if !defined(__ARM_NEON)
+#error "NEON support not enabled"
+#endif
+
+#include <stdint.h>
+
+typedef float float32_t;
+typedef __fp16 float16_t;
+#ifdef __aarch64__
+typedef double float64_t;
+#endif
+
+#ifdef __aarch64__
+typedef uint8_t poly8_t;
+typedef uint16_t poly16_t;
+typedef uint64_t poly64_t;
+typedef __uint128_t poly128_t;
+#else
+typedef int8_t poly8_t;
+typedef int16_t poly16_t;
+#endif
+typedef __attribute__((neon_vector_type(8))) int8_t int8x8_t;
+typedef __attribute__((neon_vector_type(16))) int8_t int8x16_t;
+typedef __attribute__((neon_vector_type(4))) int16_t int16x4_t;
+typedef __attribute__((neon_vector_type(8))) int16_t int16x8_t;
+typedef __attribute__((neon_vector_type(2))) int32_t int32x2_t;
+typedef __attribute__((neon_vector_type(4))) int32_t int32x4_t;
+typedef __attribute__((neon_vector_type(1))) int64_t int64x1_t;
+typedef __attribute__((neon_vector_type(2))) int64_t int64x2_t;
+typedef __attribute__((neon_vector_type(8))) uint8_t uint8x8_t;
+typedef __attribute__((neon_vector_type(16))) uint8_t uint8x16_t;
+typedef __attribute__((neon_vector_type(4))) uint16_t uint16x4_t;
+typedef __attribute__((neon_vector_type(8))) uint16_t uint16x8_t;
+typedef __attribute__((neon_vector_type(2))) uint32_t uint32x2_t;
+typedef __attribute__((neon_vector_type(4))) uint32_t uint32x4_t;
+typedef __attribute__((neon_vector_type(1))) uint64_t uint64x1_t;
+typedef __attribute__((neon_vector_type(2))) uint64_t uint64x2_t;
+typedef __attribute__((neon_vector_type(4))) float16_t float16x4_t;
+typedef __attribute__((neon_vector_type(8))) float16_t float16x8_t;
+typedef __attribute__((neon_vector_type(2))) float32_t float32x2_t;
+typedef __attribute__((neon_vector_type(4))) float32_t float32x4_t;
+#ifdef __aarch64__
+typedef __attribute__((neon_vector_type(1))) float64_t float64x1_t;
+typedef __attribute__((neon_vector_type(2))) float64_t float64x2_t;
+#endif
+typedef __attribute__((neon_polyvector_type(8))) poly8_t poly8x8_t;
+typedef __attribute__((neon_polyvector_type(16))) poly8_t poly8x16_t;
+typedef __attribute__((neon_polyvector_type(4))) poly16_t poly16x4_t;
+typedef __attribute__((neon_polyvector_type(8))) poly16_t poly16x8_t;
+#ifdef __aarch64__
+typedef __attribute__((neon_polyvector_type(1))) poly64_t poly64x1_t;
+typedef __attribute__((neon_polyvector_type(2))) poly64_t poly64x2_t;
+#endif
+
+typedef struct int8x8x2_t {
+  int8x8_t val[2];
+} int8x8x2_t;
+
+typedef struct int8x16x2_t {
+  int8x16_t val[2];
+} int8x16x2_t;
+
+typedef struct int16x4x2_t {
+  int16x4_t val[2];
+} int16x4x2_t;
+
+typedef struct int16x8x2_t {
+  int16x8_t val[2];
+} int16x8x2_t;
+
+typedef struct int32x2x2_t {
+  int32x2_t val[2];
+} int32x2x2_t;
+
+typedef struct int32x4x2_t {
+  int32x4_t val[2];
+} int32x4x2_t;
+
+typedef struct int64x1x2_t {
+  int64x1_t val[2];
+} int64x1x2_t;
+
+typedef struct int64x2x2_t {
+  int64x2_t val[2];
+} int64x2x2_t;
+
+typedef struct uint8x8x2_t {
+  uint8x8_t val[2];
+} uint8x8x2_t;
+
+typedef struct uint8x16x2_t {
+  uint8x16_t val[2];
+} uint8x16x2_t;
+
+typedef struct uint16x4x2_t {
+  uint16x4_t val[2];
+} uint16x4x2_t;
+
+typedef struct uint16x8x2_t {
+  uint16x8_t val[2];
+} uint16x8x2_t;
+
+typedef struct uint32x2x2_t {
+  uint32x2_t val[2];
+} uint32x2x2_t;
+
+typedef struct uint32x4x2_t {
+  uint32x4_t val[2];
+} uint32x4x2_t;
+
+typedef struct uint64x1x2_t {
+  uint64x1_t val[2];
+} uint64x1x2_t;
+
+typedef struct uint64x2x2_t {
+  uint64x2_t val[2];
+} uint64x2x2_t;
+
+typedef struct float16x4x2_t {
+  float16x4_t val[2];
+} float16x4x2_t;
+
+typedef struct float16x8x2_t {
+  float16x8_t val[2];
+} float16x8x2_t;
+
+typedef struct float32x2x2_t {
+  float32x2_t val[2];
+} float32x2x2_t;
+
+typedef struct float32x4x2_t {
+  float32x4_t val[2];
+} float32x4x2_t;
+
+#ifdef __aarch64__
+typedef struct float64x1x2_t {
+  float64x1_t val[2];
+} float64x1x2_t;
+
+typedef struct float64x2x2_t {
+  float64x2_t val[2];
+} float64x2x2_t;
+
+#endif
+typedef struct poly8x8x2_t {
+  poly8x8_t val[2];
+} poly8x8x2_t;
+
+typedef struct poly8x16x2_t {
+  poly8x16_t val[2];
+} poly8x16x2_t;
+
+typedef struct poly16x4x2_t {
+  poly16x4_t val[2];
+} poly16x4x2_t;
+
+typedef struct poly16x8x2_t {
+  poly16x8_t val[2];
+} poly16x8x2_t;
+
+#ifdef __aarch64__
+typedef struct poly64x1x2_t {
+  poly64x1_t val[2];
+} poly64x1x2_t;
+
+typedef struct poly64x2x2_t {
+  poly64x2_t val[2];
+} poly64x2x2_t;
+
+#endif
+typedef struct int8x8x3_t {
+  int8x8_t val[3];
+} int8x8x3_t;
+
+typedef struct int8x16x3_t {
+  int8x16_t val[3];
+} int8x16x3_t;
+
+typedef struct int16x4x3_t {
+  int16x4_t val[3];
+} int16x4x3_t;
+
+typedef struct int16x8x3_t {
+  int16x8_t val[3];
+} int16x8x3_t;
+
+typedef struct int32x2x3_t {
+  int32x2_t val[3];
+} int32x2x3_t;
+
+typedef struct int32x4x3_t {
+  int32x4_t val[3];
+} int32x4x3_t;
+
+typedef struct int64x1x3_t {
+  int64x1_t val[3];
+} int64x1x3_t;
+
+typedef struct int64x2x3_t {
+  int64x2_t val[3];
+} int64x2x3_t;
+
+typedef struct uint8x8x3_t {
+  uint8x8_t val[3];
+} uint8x8x3_t;
+
+typedef struct uint8x16x3_t {
+  uint8x16_t val[3];
+} uint8x16x3_t;
+
+typedef struct uint16x4x3_t {
+  uint16x4_t val[3];
+} uint16x4x3_t;
+
+typedef struct uint16x8x3_t {
+  uint16x8_t val[3];
+} uint16x8x3_t;
+
+typedef struct uint32x2x3_t {
+  uint32x2_t val[3];
+} uint32x2x3_t;
+
+typedef struct uint32x4x3_t {
+  uint32x4_t val[3];
+} uint32x4x3_t;
+
+typedef struct uint64x1x3_t {
+  uint64x1_t val[3];
+} uint64x1x3_t;
+
+typedef struct uint64x2x3_t {
+  uint64x2_t val[3];
+} uint64x2x3_t;
+
+typedef struct float16x4x3_t {
+  float16x4_t val[3];
+} float16x4x3_t;
+
+typedef struct float16x8x3_t {
+  float16x8_t val[3];
+} float16x8x3_t;
+
+typedef struct float32x2x3_t {
+  float32x2_t val[3];
+} float32x2x3_t;
+
+typedef struct float32x4x3_t {
+  float32x4_t val[3];
+} float32x4x3_t;
+
+#ifdef __aarch64__
+typedef struct float64x1x3_t {
+  float64x1_t val[3];
+} float64x1x3_t;
+
+typedef struct float64x2x3_t {
+  float64x2_t val[3];
+} float64x2x3_t;
+
+#endif
+typedef struct poly8x8x3_t {
+  poly8x8_t val[3];
+} poly8x8x3_t;
+
+typedef struct poly8x16x3_t {
+  poly8x16_t val[3];
+} poly8x16x3_t;
+
+typedef struct poly16x4x3_t {
+  poly16x4_t val[3];
+} poly16x4x3_t;
+
+typedef struct poly16x8x3_t {
+  poly16x8_t val[3];
+} poly16x8x3_t;
+
+#ifdef __aarch64__
+typedef struct poly64x1x3_t {
+  poly64x1_t val[3];
+} poly64x1x3_t;
+
+typedef struct poly64x2x3_t {
+  poly64x2_t val[3];
+} poly64x2x3_t;
+
+#endif
+typedef struct int8x8x4_t {
+  int8x8_t val[4];
+} int8x8x4_t;
+
+typedef struct int8x16x4_t {
+  int8x16_t val[4];
+} int8x16x4_t;
+
+typedef struct int16x4x4_t {
+  int16x4_t val[4];
+} int16x4x4_t;
+
+typedef struct int16x8x4_t {
+  int16x8_t val[4];
+} int16x8x4_t;
+
+typedef struct int32x2x4_t {
+  int32x2_t val[4];
+} int32x2x4_t;
+
+typedef struct int32x4x4_t {
+  int32x4_t val[4];
+} int32x4x4_t;
+
+typedef struct int64x1x4_t {
+  int64x1_t val[4];
+} int64x1x4_t;
+
+typedef struct int64x2x4_t {
+  int64x2_t val[4];
+} int64x2x4_t;
+
+typedef struct uint8x8x4_t {
+  uint8x8_t val[4];
+} uint8x8x4_t;
+
+typedef struct uint8x16x4_t {
+  uint8x16_t val[4];
+} uint8x16x4_t;
+
+typedef struct uint16x4x4_t {
+  uint16x4_t val[4];
+} uint16x4x4_t;
+
+typedef struct uint16x8x4_t {
+  uint16x8_t val[4];
+} uint16x8x4_t;
+
+typedef struct uint32x2x4_t {
+  uint32x2_t val[4];
+} uint32x2x4_t;
+
+typedef struct uint32x4x4_t {
+  uint32x4_t val[4];
+} uint32x4x4_t;
+
+typedef struct uint64x1x4_t {
+  uint64x1_t val[4];
+} uint64x1x4_t;
+
+typedef struct uint64x2x4_t {
+  uint64x2_t val[4];
+} uint64x2x4_t;
+
+typedef struct float16x4x4_t {
+  float16x4_t val[4];
+} float16x4x4_t;
+
+typedef struct float16x8x4_t {
+  float16x8_t val[4];
+} float16x8x4_t;
+
+typedef struct float32x2x4_t {
+  float32x2_t val[4];
+} float32x2x4_t;
+
+typedef struct float32x4x4_t {
+  float32x4_t val[4];
+} float32x4x4_t;
+
+#ifdef __aarch64__
+typedef struct float64x1x4_t {
+  float64x1_t val[4];
+} float64x1x4_t;
+
+typedef struct float64x2x4_t {
+  float64x2_t val[4];
+} float64x2x4_t;
+
+#endif
+typedef struct poly8x8x4_t {
+  poly8x8_t val[4];
+} poly8x8x4_t;
+
+typedef struct poly8x16x4_t {
+  poly8x16_t val[4];
+} poly8x16x4_t;
+
+typedef struct poly16x4x4_t {
+  poly16x4_t val[4];
+} poly16x4x4_t;
+
+typedef struct poly16x8x4_t {
+  poly16x8_t val[4];
+} poly16x8x4_t;
+
+#ifdef __aarch64__
+typedef struct poly64x1x4_t {
+  poly64x1_t val[4];
+} poly64x1x4_t;
+
+typedef struct poly64x2x4_t {
+  poly64x2_t val[4];
+} poly64x2x4_t;
+
+#endif
+
+#define __ai static inline __attribute__((__always_inline__, __nodebug__))
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vabdq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vabdq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint8x16_t __noswap_vabdq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vabdq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vabdq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint32x4_t __noswap_vabdq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vabdq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vabdq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint16x8_t __noswap_vabdq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vabdq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vabdq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int8x16_t __noswap_vabdq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vabdq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vabdq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vabdq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vabdq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int32x4_t __noswap_vabdq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vabdq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vabdq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int16x8_t __noswap_vabdq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vabd_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vabd_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint8x8_t __noswap_vabd_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vabd_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vabd_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai uint32x2_t __noswap_vabd_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vabd_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vabd_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint16x4_t __noswap_vabd_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vabd_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vabd_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int8x8_t __noswap_vabd_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vabd_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vabd_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vabd_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vabd_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int32x2_t __noswap_vabd_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vabd_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vabd_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int16x4_t __noswap_vabd_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vabsq_s8(int8x16_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vabsq_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vabsq_f32(float32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vabsq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vabsq_s32(int32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vabsq_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vabsq_s16(int16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vabsq_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vabs_s8(int8x8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vabs_v((int8x8_t)__p0, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vabs_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vabs_v((int8x8_t)__rev0, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vabs_f32(float32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vabs_v((int8x8_t)__p0, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vabs_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vabs_v((int8x8_t)__rev0, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vabs_s32(int32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vabs_v((int8x8_t)__p0, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vabs_s32(int32x2_t __p0) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vabs_v((int8x8_t)__rev0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vabs_s16(int16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vabs_v((int8x8_t)__p0, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vabs_s16(int16x4_t __p0) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vabs_v((int8x8_t)__rev0, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#else
+__ai uint8x16_t vaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __rev0 + __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#else
+__ai uint32x4_t vaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 + __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#else
+__ai uint64x2_t vaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = __rev0 + __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#else
+__ai uint16x8_t vaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 + __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vaddq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#else
+__ai int8x16_t vaddq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __rev0 + __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vaddq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#else
+__ai float32x4_t vaddq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = __rev0 + __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vaddq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#else
+__ai int32x4_t vaddq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 + __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vaddq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#else
+__ai int64x2_t vaddq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = __rev0 + __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vaddq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#else
+__ai int16x8_t vaddq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 + __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#else
+__ai uint8x8_t vadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __rev0 + __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#else
+__ai uint32x2_t vadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = __rev0 + __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#else
+__ai uint64x1_t vadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#else
+__ai uint16x4_t vadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __rev0 + __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vadd_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#else
+__ai int8x8_t vadd_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __rev0 + __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vadd_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#else
+__ai float32x2_t vadd_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = __rev0 + __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vadd_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#else
+__ai int32x2_t vadd_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = __rev0 + __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vadd_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#else
+__ai int64x1_t vadd_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vadd_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#else
+__ai int16x4_t vadd_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __rev0 + __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vaddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vaddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint16x4_t __noswap_vaddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vaddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vaddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai uint32x2_t __noswap_vaddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vaddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vaddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint8x8_t __noswap_vaddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vaddhn_s32(int32x4_t __p0, int32x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vaddhn_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int16x4_t __noswap_vaddhn_s32(int32x4_t __p0, int32x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vaddhn_s64(int64x2_t __p0, int64x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vaddhn_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int32x2_t __noswap_vaddhn_s64(int64x2_t __p0, int64x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vaddhn_s16(int16x8_t __p0, int16x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vaddhn_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int8x8_t __noswap_vaddhn_s16(int16x8_t __p0, int16x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vandq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = __p0 & __p1;
+  return __ret;
+}
+#else
+__ai uint8x16_t vandq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __rev0 & __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vandq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = __p0 & __p1;
+  return __ret;
+}
+#else
+__ai uint32x4_t vandq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 & __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vandq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = __p0 & __p1;
+  return __ret;
+}
+#else
+__ai uint64x2_t vandq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = __rev0 & __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vandq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = __p0 & __p1;
+  return __ret;
+}
+#else
+__ai uint16x8_t vandq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 & __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vandq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = __p0 & __p1;
+  return __ret;
+}
+#else
+__ai int8x16_t vandq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __rev0 & __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vandq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = __p0 & __p1;
+  return __ret;
+}
+#else
+__ai int32x4_t vandq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 & __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vandq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = __p0 & __p1;
+  return __ret;
+}
+#else
+__ai int64x2_t vandq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = __rev0 & __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vandq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = __p0 & __p1;
+  return __ret;
+}
+#else
+__ai int16x8_t vandq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 & __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vand_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = __p0 & __p1;
+  return __ret;
+}
+#else
+__ai uint8x8_t vand_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __rev0 & __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vand_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = __p0 & __p1;
+  return __ret;
+}
+#else
+__ai uint32x2_t vand_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = __rev0 & __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vand_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = __p0 & __p1;
+  return __ret;
+}
+#else
+__ai uint64x1_t vand_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = __p0 & __p1;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vand_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = __p0 & __p1;
+  return __ret;
+}
+#else
+__ai uint16x4_t vand_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __rev0 & __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vand_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = __p0 & __p1;
+  return __ret;
+}
+#else
+__ai int8x8_t vand_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __rev0 & __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vand_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = __p0 & __p1;
+  return __ret;
+}
+#else
+__ai int32x2_t vand_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = __rev0 & __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vand_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = __p0 & __p1;
+  return __ret;
+}
+#else
+__ai int64x1_t vand_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = __p0 & __p1;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vand_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = __p0 & __p1;
+  return __ret;
+}
+#else
+__ai int16x4_t vand_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __rev0 & __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vbicq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = __p0 & ~__p1;
+  return __ret;
+}
+#else
+__ai uint8x16_t vbicq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __rev0 & ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vbicq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = __p0 & ~__p1;
+  return __ret;
+}
+#else
+__ai uint32x4_t vbicq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 & ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vbicq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = __p0 & ~__p1;
+  return __ret;
+}
+#else
+__ai uint64x2_t vbicq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = __rev0 & ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vbicq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = __p0 & ~__p1;
+  return __ret;
+}
+#else
+__ai uint16x8_t vbicq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 & ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vbicq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = __p0 & ~__p1;
+  return __ret;
+}
+#else
+__ai int8x16_t vbicq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __rev0 & ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vbicq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = __p0 & ~__p1;
+  return __ret;
+}
+#else
+__ai int32x4_t vbicq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 & ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vbicq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = __p0 & ~__p1;
+  return __ret;
+}
+#else
+__ai int64x2_t vbicq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = __rev0 & ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vbicq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = __p0 & ~__p1;
+  return __ret;
+}
+#else
+__ai int16x8_t vbicq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 & ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vbic_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = __p0 & ~__p1;
+  return __ret;
+}
+#else
+__ai uint8x8_t vbic_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __rev0 & ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vbic_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = __p0 & ~__p1;
+  return __ret;
+}
+#else
+__ai uint32x2_t vbic_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = __rev0 & ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vbic_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = __p0 & ~__p1;
+  return __ret;
+}
+#else
+__ai uint64x1_t vbic_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = __p0 & ~__p1;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vbic_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = __p0 & ~__p1;
+  return __ret;
+}
+#else
+__ai uint16x4_t vbic_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __rev0 & ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vbic_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = __p0 & ~__p1;
+  return __ret;
+}
+#else
+__ai int8x8_t vbic_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __rev0 & ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vbic_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = __p0 & ~__p1;
+  return __ret;
+}
+#else
+__ai int32x2_t vbic_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = __rev0 & ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vbic_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = __p0 & ~__p1;
+  return __ret;
+}
+#else
+__ai int64x1_t vbic_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = __p0 & ~__p1;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vbic_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = __p0 & ~__p1;
+  return __ret;
+}
+#else
+__ai int16x4_t vbic_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __rev0 & ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vbsl_p8(uint8x8_t __p0, poly8x8_t __p1, poly8x8_t __p2) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vbsl_p8(uint8x8_t __p0, poly8x8_t __p1, poly8x8_t __p2) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vbsl_p16(uint16x4_t __p0, poly16x4_t __p1, poly16x4_t __p2) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 5);
+  return __ret;
+}
+#else
+__ai poly16x4_t vbsl_p16(uint16x4_t __p0, poly16x4_t __p1, poly16x4_t __p2) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  poly16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  poly16x4_t __ret;
+  __ret = (poly16x4_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 5);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vbslq_p8(uint8x16_t __p0, poly8x16_t __p1, poly8x16_t __p2) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 36);
+  return __ret;
+}
+#else
+__ai poly8x16_t vbslq_p8(uint8x16_t __p0, poly8x16_t __p1, poly8x16_t __p2) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 36);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vbslq_p16(uint16x8_t __p0, poly16x8_t __p1, poly16x8_t __p2) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 37);
+  return __ret;
+}
+#else
+__ai poly16x8_t vbslq_p16(uint16x8_t __p0, poly16x8_t __p1, poly16x8_t __p2) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __ret;
+  __ret = (poly16x8_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 37);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vbslq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vbslq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vbslq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vbslq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vbslq_u64(uint64x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vbslq_u64(uint64x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vbslq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vbslq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vbslq_s8(uint8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vbslq_s8(uint8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vbslq_f32(uint32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vbslq_f32(uint32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vbslq_s32(uint32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vbslq_s32(uint32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vbslq_s64(uint64x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vbslq_s64(uint64x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vbslq_s16(uint16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vbslq_s16(uint16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vbsl_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vbsl_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vbsl_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vbsl_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vbsl_u64(uint64x1_t __p0, uint64x1_t __p1, uint64x1_t __p2) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vbsl_u64(uint64x1_t __p0, uint64x1_t __p1, uint64x1_t __p2) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vbsl_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vbsl_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vbsl_s8(uint8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vbsl_s8(uint8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vbsl_f32(uint32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vbsl_f32(uint32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vbsl_s32(uint32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vbsl_s32(uint32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vbsl_s64(uint64x1_t __p0, int64x1_t __p1, int64x1_t __p2) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 3);
+  return __ret;
+}
+#else
+__ai int64x1_t vbsl_s64(uint64x1_t __p0, int64x1_t __p1, int64x1_t __p2) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vbsl_s16(uint16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vbsl_s16(uint16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcageq_f32(float32x4_t __p0, float32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcageq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcageq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcageq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcage_f32(float32x2_t __p0, float32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcage_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcage_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcage_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcagtq_f32(float32x4_t __p0, float32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcagtq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcagtq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcagtq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcagt_f32(float32x2_t __p0, float32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcagt_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcagt_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcagt_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcaleq_f32(float32x4_t __p0, float32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcaleq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcaleq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcaleq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcale_f32(float32x2_t __p0, float32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcale_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcale_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcale_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcaltq_f32(float32x4_t __p0, float32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcaltq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcaltq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcaltq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcalt_f32(float32x2_t __p0, float32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcalt_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcalt_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcalt_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vceq_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint8x8_t vceq_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__rev0 == __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vceqq_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint8x16_t vceqq_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__rev0 == __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vceqq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint8x16_t vceqq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__rev0 == __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vceqq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vceqq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__rev0 == __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vceqq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint16x8_t vceqq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__rev0 == __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vceqq_s8(int8x16_t __p0, int8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint8x16_t vceqq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__rev0 == __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vceqq_f32(float32x4_t __p0, float32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vceqq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__rev0 == __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vceqq_s32(int32x4_t __p0, int32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vceqq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__rev0 == __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vceqq_s16(int16x8_t __p0, int16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint16x8_t vceqq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__rev0 == __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vceq_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint8x8_t vceq_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__rev0 == __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vceq_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint32x2_t vceq_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__rev0 == __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vceq_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint16x4_t vceq_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__rev0 == __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vceq_s8(int8x8_t __p0, int8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint8x8_t vceq_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__rev0 == __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vceq_f32(float32x2_t __p0, float32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint32x2_t vceq_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__rev0 == __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vceq_s32(int32x2_t __p0, int32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint32x2_t vceq_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__rev0 == __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vceq_s16(int16x4_t __p0, int16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint16x4_t vceq_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__rev0 == __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vcgeq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0 >= __p1);
+  return __ret;
+}
+#else
+__ai uint8x16_t vcgeq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__rev0 >= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcgeq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0 >= __p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcgeq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__rev0 >= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vcgeq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0 >= __p1);
+  return __ret;
+}
+#else
+__ai uint16x8_t vcgeq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__rev0 >= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vcgeq_s8(int8x16_t __p0, int8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0 >= __p1);
+  return __ret;
+}
+#else
+__ai uint8x16_t vcgeq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__rev0 >= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcgeq_f32(float32x4_t __p0, float32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0 >= __p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcgeq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__rev0 >= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcgeq_s32(int32x4_t __p0, int32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0 >= __p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcgeq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__rev0 >= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vcgeq_s16(int16x8_t __p0, int16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0 >= __p1);
+  return __ret;
+}
+#else
+__ai uint16x8_t vcgeq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__rev0 >= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vcge_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0 >= __p1);
+  return __ret;
+}
+#else
+__ai uint8x8_t vcge_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__rev0 >= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcge_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0 >= __p1);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcge_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__rev0 >= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vcge_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0 >= __p1);
+  return __ret;
+}
+#else
+__ai uint16x4_t vcge_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__rev0 >= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vcge_s8(int8x8_t __p0, int8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0 >= __p1);
+  return __ret;
+}
+#else
+__ai uint8x8_t vcge_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__rev0 >= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcge_f32(float32x2_t __p0, float32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0 >= __p1);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcge_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__rev0 >= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcge_s32(int32x2_t __p0, int32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0 >= __p1);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcge_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__rev0 >= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vcge_s16(int16x4_t __p0, int16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0 >= __p1);
+  return __ret;
+}
+#else
+__ai uint16x4_t vcge_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__rev0 >= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vcgtq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0 > __p1);
+  return __ret;
+}
+#else
+__ai uint8x16_t vcgtq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__rev0 > __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcgtq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0 > __p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcgtq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__rev0 > __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vcgtq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0 > __p1);
+  return __ret;
+}
+#else
+__ai uint16x8_t vcgtq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__rev0 > __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vcgtq_s8(int8x16_t __p0, int8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0 > __p1);
+  return __ret;
+}
+#else
+__ai uint8x16_t vcgtq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__rev0 > __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcgtq_f32(float32x4_t __p0, float32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0 > __p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcgtq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__rev0 > __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcgtq_s32(int32x4_t __p0, int32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0 > __p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcgtq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__rev0 > __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vcgtq_s16(int16x8_t __p0, int16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0 > __p1);
+  return __ret;
+}
+#else
+__ai uint16x8_t vcgtq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__rev0 > __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vcgt_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0 > __p1);
+  return __ret;
+}
+#else
+__ai uint8x8_t vcgt_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__rev0 > __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcgt_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0 > __p1);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcgt_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__rev0 > __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vcgt_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0 > __p1);
+  return __ret;
+}
+#else
+__ai uint16x4_t vcgt_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__rev0 > __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vcgt_s8(int8x8_t __p0, int8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0 > __p1);
+  return __ret;
+}
+#else
+__ai uint8x8_t vcgt_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__rev0 > __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcgt_f32(float32x2_t __p0, float32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0 > __p1);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcgt_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__rev0 > __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcgt_s32(int32x2_t __p0, int32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0 > __p1);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcgt_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__rev0 > __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vcgt_s16(int16x4_t __p0, int16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0 > __p1);
+  return __ret;
+}
+#else
+__ai uint16x4_t vcgt_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__rev0 > __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vcleq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0 <= __p1);
+  return __ret;
+}
+#else
+__ai uint8x16_t vcleq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__rev0 <= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcleq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0 <= __p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcleq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__rev0 <= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vcleq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0 <= __p1);
+  return __ret;
+}
+#else
+__ai uint16x8_t vcleq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__rev0 <= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vcleq_s8(int8x16_t __p0, int8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0 <= __p1);
+  return __ret;
+}
+#else
+__ai uint8x16_t vcleq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__rev0 <= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcleq_f32(float32x4_t __p0, float32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0 <= __p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcleq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__rev0 <= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcleq_s32(int32x4_t __p0, int32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0 <= __p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcleq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__rev0 <= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vcleq_s16(int16x8_t __p0, int16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0 <= __p1);
+  return __ret;
+}
+#else
+__ai uint16x8_t vcleq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__rev0 <= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vcle_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0 <= __p1);
+  return __ret;
+}
+#else
+__ai uint8x8_t vcle_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__rev0 <= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcle_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0 <= __p1);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcle_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__rev0 <= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vcle_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0 <= __p1);
+  return __ret;
+}
+#else
+__ai uint16x4_t vcle_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__rev0 <= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vcle_s8(int8x8_t __p0, int8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0 <= __p1);
+  return __ret;
+}
+#else
+__ai uint8x8_t vcle_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__rev0 <= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcle_f32(float32x2_t __p0, float32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0 <= __p1);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcle_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__rev0 <= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcle_s32(int32x2_t __p0, int32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0 <= __p1);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcle_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__rev0 <= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vcle_s16(int16x4_t __p0, int16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0 <= __p1);
+  return __ret;
+}
+#else
+__ai uint16x4_t vcle_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__rev0 <= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vclsq_s8(int8x16_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vclsq_v((int8x16_t)__p0, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vclsq_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vclsq_v((int8x16_t)__rev0, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vclsq_s32(int32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vclsq_v((int8x16_t)__p0, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vclsq_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vclsq_v((int8x16_t)__rev0, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vclsq_s16(int16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vclsq_v((int8x16_t)__p0, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vclsq_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vclsq_v((int8x16_t)__rev0, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vcls_s8(int8x8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vcls_v((int8x8_t)__p0, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vcls_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vcls_v((int8x8_t)__rev0, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vcls_s32(int32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vcls_v((int8x8_t)__p0, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vcls_s32(int32x2_t __p0) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vcls_v((int8x8_t)__rev0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vcls_s16(int16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vcls_v((int8x8_t)__p0, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vcls_s16(int16x4_t __p0) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vcls_v((int8x8_t)__rev0, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vcltq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0 < __p1);
+  return __ret;
+}
+#else
+__ai uint8x16_t vcltq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__rev0 < __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcltq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0 < __p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcltq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__rev0 < __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vcltq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0 < __p1);
+  return __ret;
+}
+#else
+__ai uint16x8_t vcltq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__rev0 < __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vcltq_s8(int8x16_t __p0, int8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0 < __p1);
+  return __ret;
+}
+#else
+__ai uint8x16_t vcltq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__rev0 < __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcltq_f32(float32x4_t __p0, float32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0 < __p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcltq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__rev0 < __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcltq_s32(int32x4_t __p0, int32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0 < __p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcltq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__rev0 < __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vcltq_s16(int16x8_t __p0, int16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0 < __p1);
+  return __ret;
+}
+#else
+__ai uint16x8_t vcltq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__rev0 < __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vclt_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0 < __p1);
+  return __ret;
+}
+#else
+__ai uint8x8_t vclt_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__rev0 < __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vclt_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0 < __p1);
+  return __ret;
+}
+#else
+__ai uint32x2_t vclt_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__rev0 < __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vclt_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0 < __p1);
+  return __ret;
+}
+#else
+__ai uint16x4_t vclt_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__rev0 < __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vclt_s8(int8x8_t __p0, int8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0 < __p1);
+  return __ret;
+}
+#else
+__ai uint8x8_t vclt_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__rev0 < __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vclt_f32(float32x2_t __p0, float32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0 < __p1);
+  return __ret;
+}
+#else
+__ai uint32x2_t vclt_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__rev0 < __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vclt_s32(int32x2_t __p0, int32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0 < __p1);
+  return __ret;
+}
+#else
+__ai uint32x2_t vclt_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__rev0 < __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vclt_s16(int16x4_t __p0, int16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0 < __p1);
+  return __ret;
+}
+#else
+__ai uint16x4_t vclt_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__rev0 < __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vclzq_u8(uint8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vclzq_u8(uint8x16_t __p0) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vclzq_u32(uint32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vclzq_u32(uint32x4_t __p0) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vclzq_u16(uint16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vclzq_u16(uint16x8_t __p0) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vclzq_s8(int8x16_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vclzq_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vclzq_s32(int32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vclzq_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vclzq_s16(int16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vclzq_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vclz_u8(uint8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vclz_v((int8x8_t)__p0, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vclz_u8(uint8x8_t __p0) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vclz_u32(uint32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vclz_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vclz_u32(uint32x2_t __p0) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vclz_u16(uint16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vclz_v((int8x8_t)__p0, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vclz_u16(uint16x4_t __p0) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vclz_s8(int8x8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vclz_v((int8x8_t)__p0, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vclz_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vclz_s32(int32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vclz_v((int8x8_t)__p0, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vclz_s32(int32x2_t __p0) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vclz_s16(int16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vclz_v((int8x8_t)__p0, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vclz_s16(int16x4_t __p0) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vcnt_p8(poly8x8_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vcnt_v((int8x8_t)__p0, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vcnt_p8(poly8x8_t __p0) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vcnt_v((int8x8_t)__rev0, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vcntq_p8(poly8x16_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vcntq_v((int8x16_t)__p0, 36);
+  return __ret;
+}
+#else
+__ai poly8x16_t vcntq_p8(poly8x16_t __p0) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vcntq_v((int8x16_t)__rev0, 36);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vcntq_u8(uint8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vcntq_v((int8x16_t)__p0, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vcntq_u8(uint8x16_t __p0) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vcntq_v((int8x16_t)__rev0, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vcntq_s8(int8x16_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vcntq_v((int8x16_t)__p0, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vcntq_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vcntq_v((int8x16_t)__rev0, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vcnt_u8(uint8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vcnt_v((int8x8_t)__p0, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vcnt_u8(uint8x8_t __p0) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vcnt_v((int8x8_t)__rev0, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vcnt_s8(int8x8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vcnt_v((int8x8_t)__p0, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vcnt_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vcnt_v((int8x8_t)__rev0, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vcombine_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
+  return __ret;
+}
+#else
+__ai poly8x16_t vcombine_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vcombine_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
+  return __ret;
+}
+#else
+__ai poly16x8_t vcombine_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  poly16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vcombine_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
+  return __ret;
+}
+#else
+__ai uint8x16_t vcombine_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint8x16_t __noswap_vcombine_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcombine_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcombine_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint32x4_t __noswap_vcombine_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcombine_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcombine_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vcombine_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
+  return __ret;
+}
+#else
+__ai uint16x8_t vcombine_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint16x8_t __noswap_vcombine_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vcombine_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
+  return __ret;
+}
+#else
+__ai int8x16_t vcombine_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int8x16_t __noswap_vcombine_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vcombine_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
+  return __ret;
+}
+#else
+__ai float32x4_t vcombine_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai float32x4_t __noswap_vcombine_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vcombine_f16(float16x4_t __p0, float16x4_t __p1) {
+  float16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
+  return __ret;
+}
+#else
+__ai float16x8_t vcombine_f16(float16x4_t __p0, float16x4_t __p1) {
+  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai float16x8_t __noswap_vcombine_f16(float16x4_t __p0, float16x4_t __p1) {
+  float16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vcombine_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
+  return __ret;
+}
+#else
+__ai int32x4_t vcombine_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int32x4_t __noswap_vcombine_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vcombine_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
+  return __ret;
+}
+#else
+__ai int64x2_t vcombine_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vcombine_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
+  return __ret;
+}
+#else
+__ai int16x8_t vcombine_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int16x8_t __noswap_vcombine_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vcreate_p8(uint64_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vcreate_p8(uint64_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vcreate_p16(uint64_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vcreate_p16(uint64_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vcreate_u8(uint64_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vcreate_u8(uint64_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcreate_u32(uint64_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcreate_u32(uint64_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcreate_u64(uint64_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcreate_u64(uint64_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vcreate_u16(uint64_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vcreate_u16(uint64_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vcreate_s8(uint64_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vcreate_s8(uint64_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vcreate_f32(uint64_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vcreate_f32(uint64_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vcreate_f16(uint64_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vcreate_f16(uint64_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vcreate_s32(uint64_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vcreate_s32(uint64_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vcreate_s64(uint64_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vcreate_s64(uint64_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vcreate_s16(uint64_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vcreate_s16(uint64_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vcvt_f16_f32(float32x4_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t) __builtin_neon_vcvt_f16_v((int8x16_t)__p0, 8);
+  return __ret;
+}
+#else
+__ai float16x4_t vcvt_f16_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float16x4_t __ret;
+  __ret = (float16x4_t) __builtin_neon_vcvt_f16_v((int8x16_t)__rev0, 8);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai float16x4_t __noswap_vcvt_f16_f32(float32x4_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t) __builtin_neon_vcvt_f16_v((int8x16_t)__p0, 8);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vcvtq_f32_u32(uint32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vcvtq_f32_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai float32x4_t vcvtq_f32_u32(uint32x4_t __p0) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vcvtq_f32_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vcvtq_f32_s32(int32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vcvtq_f32_v((int8x16_t)__p0, 34);
+  return __ret;
+}
+#else
+__ai float32x4_t vcvtq_f32_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vcvtq_f32_v((int8x16_t)__rev0, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vcvt_f32_u32(uint32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vcvt_f32_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai float32x2_t vcvt_f32_u32(uint32x2_t __p0) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vcvt_f32_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vcvt_f32_s32(int32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vcvt_f32_v((int8x8_t)__p0, 2);
+  return __ret;
+}
+#else
+__ai float32x2_t vcvt_f32_s32(int32x2_t __p0) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vcvt_f32_v((int8x8_t)__rev0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vcvt_f32_f16(float16x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vcvt_f32_f16((int8x8_t)__p0, 8);
+  return __ret;
+}
+#else
+__ai float32x4_t vcvt_f32_f16(float16x4_t __p0) {
+  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vcvt_f32_f16((int8x8_t)__rev0, 8);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai float32x4_t __noswap_vcvt_f32_f16(float16x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vcvt_f32_f16((int8x8_t)__p0, 8);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvtq_n_f32_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vcvtq_n_f32_v((int8x16_t)__s0, __p1, 50); \
+  __ret; \
+})
+#else
+#define vcvtq_n_f32_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vcvtq_n_f32_v((int8x16_t)__rev0, __p1, 50); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvtq_n_f32_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vcvtq_n_f32_v((int8x16_t)__s0, __p1, 34); \
+  __ret; \
+})
+#else
+#define vcvtq_n_f32_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vcvtq_n_f32_v((int8x16_t)__rev0, __p1, 34); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvt_n_f32_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vcvt_n_f32_v((int8x8_t)__s0, __p1, 18); \
+  __ret; \
+})
+#else
+#define vcvt_n_f32_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vcvt_n_f32_v((int8x8_t)__rev0, __p1, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvt_n_f32_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vcvt_n_f32_v((int8x8_t)__s0, __p1, 2); \
+  __ret; \
+})
+#else
+#define vcvt_n_f32_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vcvt_n_f32_v((int8x8_t)__rev0, __p1, 2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvtq_n_s32_f32(__p0, __p1) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vcvtq_n_s32_v((int8x16_t)__s0, __p1, 34); \
+  __ret; \
+})
+#else
+#define vcvtq_n_s32_f32(__p0, __p1) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vcvtq_n_s32_v((int8x16_t)__rev0, __p1, 34); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvt_n_s32_f32(__p0, __p1) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vcvt_n_s32_v((int8x8_t)__s0, __p1, 2); \
+  __ret; \
+})
+#else
+#define vcvt_n_s32_f32(__p0, __p1) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vcvt_n_s32_v((int8x8_t)__rev0, __p1, 2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvtq_n_u32_f32(__p0, __p1) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vcvtq_n_u32_v((int8x16_t)__s0, __p1, 50); \
+  __ret; \
+})
+#else
+#define vcvtq_n_u32_f32(__p0, __p1) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vcvtq_n_u32_v((int8x16_t)__rev0, __p1, 50); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvt_n_u32_f32(__p0, __p1) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vcvt_n_u32_v((int8x8_t)__s0, __p1, 18); \
+  __ret; \
+})
+#else
+#define vcvt_n_u32_f32(__p0, __p1) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vcvt_n_u32_v((int8x8_t)__rev0, __p1, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vcvtq_s32_f32(float32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vcvtq_s32_v((int8x16_t)__p0, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vcvtq_s32_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vcvtq_s32_v((int8x16_t)__rev0, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vcvt_s32_f32(float32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vcvt_s32_v((int8x8_t)__p0, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vcvt_s32_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vcvt_s32_v((int8x8_t)__rev0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcvtq_u32_f32(float32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcvtq_u32_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcvtq_u32_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcvtq_u32_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcvt_u32_f32(float32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcvt_u32_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcvt_u32_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcvt_u32_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_lane_p8(__p0, __p1) __extension__ ({ \
+  poly8x8_t __s0 = __p0; \
+  poly8x8_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdup_lane_p8(__p0, __p1) __extension__ ({ \
+  poly8x8_t __s0 = __p0; \
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x8_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_lane_p16(__p0, __p1) __extension__ ({ \
+  poly16x4_t __s0 = __p0; \
+  poly16x4_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdup_lane_p16(__p0, __p1) __extension__ ({ \
+  poly16x4_t __s0 = __p0; \
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  poly16x4_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_lane_p8(__p0, __p1) __extension__ ({ \
+  poly8x8_t __s0 = __p0; \
+  poly8x16_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_lane_p8(__p0, __p1) __extension__ ({ \
+  poly8x8_t __s0 = __p0; \
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_lane_p16(__p0, __p1) __extension__ ({ \
+  poly16x4_t __s0 = __p0; \
+  poly16x8_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_lane_p16(__p0, __p1) __extension__ ({ \
+  poly16x4_t __s0 = __p0; \
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  poly16x8_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_lane_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x16_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_lane_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_lane_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x4_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_lane_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_lane_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x2_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_lane_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x2_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_lane_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x8_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_lane_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_lane_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x16_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_lane_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_lane_f32(__p0, __p1) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x4_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_lane_f32(__p0, __p1) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float32x4_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_lane_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x4_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_lane_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_lane_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x2_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_lane_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x2_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_lane_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x8_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_lane_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_lane_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdup_lane_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_lane_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdup_lane_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_lane_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x1_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
+  __ret; \
+})
+#else
+#define vdup_lane_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x1_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_lane_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdup_lane_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_lane_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdup_lane_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_lane_f32(__p0, __p1) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdup_lane_f32(__p0, __p1) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float32x2_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_lane_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdup_lane_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_lane_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x1_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
+  __ret; \
+})
+#else
+#define vdup_lane_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x1_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_lane_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdup_lane_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vdup_n_p8(poly8_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai poly8x8_t vdup_n_p8(poly8_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vdup_n_p16(poly16_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t) {__p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai poly16x4_t vdup_n_p16(poly16_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t) {__p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vdupq_n_p8(poly8_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai poly8x16_t vdupq_n_p8(poly8_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vdupq_n_p16(poly16_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai poly16x8_t vdupq_n_p16(poly16_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vdupq_n_u8(uint8_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai uint8x16_t vdupq_n_u8(uint8_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vdupq_n_u32(uint32_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) {__p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai uint32x4_t vdupq_n_u32(uint32_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) {__p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vdupq_n_u64(uint64_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) {__p0, __p0};
+  return __ret;
+}
+#else
+__ai uint64x2_t vdupq_n_u64(uint64_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) {__p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vdupq_n_u16(uint16_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai uint16x8_t vdupq_n_u16(uint16_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vdupq_n_s8(int8_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai int8x16_t vdupq_n_s8(int8_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vdupq_n_f32(float32_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) {__p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai float32x4_t vdupq_n_f32(float32_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) {__p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_n_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16x8_t __ret; \
+  __ret = (float16x8_t) {__s0, __s0, __s0, __s0, __s0, __s0, __s0, __s0}; \
+  __ret; \
+})
+#else
+#define vdupq_n_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16x8_t __ret; \
+  __ret = (float16x8_t) {__s0, __s0, __s0, __s0, __s0, __s0, __s0, __s0}; \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vdupq_n_s32(int32_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) {__p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai int32x4_t vdupq_n_s32(int32_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) {__p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vdupq_n_s64(int64_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) {__p0, __p0};
+  return __ret;
+}
+#else
+__ai int64x2_t vdupq_n_s64(int64_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) {__p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vdupq_n_s16(int16_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai int16x8_t vdupq_n_s16(int16_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vdup_n_u8(uint8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai uint8x8_t vdup_n_u8(uint8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vdup_n_u32(uint32_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) {__p0, __p0};
+  return __ret;
+}
+#else
+__ai uint32x2_t vdup_n_u32(uint32_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) {__p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vdup_n_u64(uint64_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) {__p0};
+  return __ret;
+}
+#else
+__ai uint64x1_t vdup_n_u64(uint64_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) {__p0};
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vdup_n_u16(uint16_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) {__p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai uint16x4_t vdup_n_u16(uint16_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) {__p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vdup_n_s8(int8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai int8x8_t vdup_n_s8(int8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vdup_n_f32(float32_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) {__p0, __p0};
+  return __ret;
+}
+#else
+__ai float32x2_t vdup_n_f32(float32_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) {__p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_n_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16x4_t __ret; \
+  __ret = (float16x4_t) {__s0, __s0, __s0, __s0}; \
+  __ret; \
+})
+#else
+#define vdup_n_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16x4_t __ret; \
+  __ret = (float16x4_t) {__s0, __s0, __s0, __s0}; \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vdup_n_s32(int32_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) {__p0, __p0};
+  return __ret;
+}
+#else
+__ai int32x2_t vdup_n_s32(int32_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) {__p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vdup_n_s64(int64_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) {__p0};
+  return __ret;
+}
+#else
+__ai int64x1_t vdup_n_s64(int64_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) {__p0};
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vdup_n_s16(int16_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) {__p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai int16x4_t vdup_n_s16(int16_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) {__p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t veorq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = __p0 ^ __p1;
+  return __ret;
+}
+#else
+__ai uint8x16_t veorq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __rev0 ^ __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t veorq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = __p0 ^ __p1;
+  return __ret;
+}
+#else
+__ai uint32x4_t veorq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 ^ __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t veorq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = __p0 ^ __p1;
+  return __ret;
+}
+#else
+__ai uint64x2_t veorq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = __rev0 ^ __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t veorq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = __p0 ^ __p1;
+  return __ret;
+}
+#else
+__ai uint16x8_t veorq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 ^ __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t veorq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = __p0 ^ __p1;
+  return __ret;
+}
+#else
+__ai int8x16_t veorq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __rev0 ^ __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t veorq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = __p0 ^ __p1;
+  return __ret;
+}
+#else
+__ai int32x4_t veorq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 ^ __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t veorq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = __p0 ^ __p1;
+  return __ret;
+}
+#else
+__ai int64x2_t veorq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = __rev0 ^ __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t veorq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = __p0 ^ __p1;
+  return __ret;
+}
+#else
+__ai int16x8_t veorq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 ^ __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t veor_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = __p0 ^ __p1;
+  return __ret;
+}
+#else
+__ai uint8x8_t veor_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __rev0 ^ __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t veor_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = __p0 ^ __p1;
+  return __ret;
+}
+#else
+__ai uint32x2_t veor_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = __rev0 ^ __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t veor_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = __p0 ^ __p1;
+  return __ret;
+}
+#else
+__ai uint64x1_t veor_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = __p0 ^ __p1;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t veor_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = __p0 ^ __p1;
+  return __ret;
+}
+#else
+__ai uint16x4_t veor_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __rev0 ^ __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t veor_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = __p0 ^ __p1;
+  return __ret;
+}
+#else
+__ai int8x8_t veor_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __rev0 ^ __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t veor_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = __p0 ^ __p1;
+  return __ret;
+}
+#else
+__ai int32x2_t veor_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = __rev0 ^ __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t veor_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = __p0 ^ __p1;
+  return __ret;
+}
+#else
+__ai int64x1_t veor_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = __p0 ^ __p1;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t veor_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = __p0 ^ __p1;
+  return __ret;
+}
+#else
+__ai int16x4_t veor_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __rev0 ^ __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vext_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8_t __s0 = __p0; \
+  poly8x8_t __s1 = __p1; \
+  poly8x8_t __ret; \
+  __ret = (poly8x8_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 4); \
+  __ret; \
+})
+#else
+#define vext_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8_t __s0 = __p0; \
+  poly8x8_t __s1 = __p1; \
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x8_t __ret; \
+  __ret = (poly8x8_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 4); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vext_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4_t __s0 = __p0; \
+  poly16x4_t __s1 = __p1; \
+  poly16x4_t __ret; \
+  __ret = (poly16x4_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 5); \
+  __ret; \
+})
+#else
+#define vext_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4_t __s0 = __p0; \
+  poly16x4_t __s1 = __p1; \
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  poly16x4_t __ret; \
+  __ret = (poly16x4_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 5); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vextq_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16_t __s0 = __p0; \
+  poly8x16_t __s1 = __p1; \
+  poly8x16_t __ret; \
+  __ret = (poly8x16_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 36); \
+  __ret; \
+})
+#else
+#define vextq_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16_t __s0 = __p0; \
+  poly8x16_t __s1 = __p1; \
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16_t __ret; \
+  __ret = (poly8x16_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 36); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vextq_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8_t __s0 = __p0; \
+  poly16x8_t __s1 = __p1; \
+  poly16x8_t __ret; \
+  __ret = (poly16x8_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 37); \
+  __ret; \
+})
+#else
+#define vextq_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8_t __s0 = __p0; \
+  poly16x8_t __s1 = __p1; \
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x8_t __ret; \
+  __ret = (poly16x8_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 37); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vextq_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __s1 = __p1; \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 48); \
+  __ret; \
+})
+#else
+#define vextq_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __s1 = __p1; \
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 48); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vextq_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 50); \
+  __ret; \
+})
+#else
+#define vextq_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 50); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vextq_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __s1 = __p1; \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 51); \
+  __ret; \
+})
+#else
+#define vextq_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __s1 = __p1; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 51); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vextq_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 49); \
+  __ret; \
+})
+#else
+#define vextq_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 49); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vextq_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __s1 = __p1; \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 32); \
+  __ret; \
+})
+#else
+#define vextq_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __s1 = __p1; \
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 32); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vextq_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 41); \
+  __ret; \
+})
+#else
+#define vextq_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 41); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vextq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 34); \
+  __ret; \
+})
+#else
+#define vextq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 34); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vextq_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __s1 = __p1; \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 35); \
+  __ret; \
+})
+#else
+#define vextq_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __s1 = __p1; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 35); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vextq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 33); \
+  __ret; \
+})
+#else
+#define vextq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 33); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vext_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __s1 = __p1; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 16); \
+  __ret; \
+})
+#else
+#define vext_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __s1 = __p1; \
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 16); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vext_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 18); \
+  __ret; \
+})
+#else
+#define vext_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vext_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x1_t __s1 = __p1; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
+  __ret; \
+})
+#else
+#define vext_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x1_t __s1 = __p1; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vext_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 17); \
+  __ret; \
+})
+#else
+#define vext_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 17); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vext_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __s1 = __p1; \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 0); \
+  __ret; \
+})
+#else
+#define vext_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __s1 = __p1; \
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 0); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vext_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 9); \
+  __ret; \
+})
+#else
+#define vext_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 9); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vext_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 2); \
+  __ret; \
+})
+#else
+#define vext_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vext_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x1_t __s1 = __p1; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
+  __ret; \
+})
+#else
+#define vext_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x1_t __s1 = __p1; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vext_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 1); \
+  __ret; \
+})
+#else
+#define vext_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vfmaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vfmaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vfmaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vfmaq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai float32x4_t __noswap_vfmaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vfmaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vfma_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vfma_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vfma_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai float32x2_t __noswap_vfma_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vget_high_p8(poly8x16_t __p0) {
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
+  return __ret;
+}
+#else
+__ai poly8x8_t vget_high_p8(poly8x16_t __p0) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 8, 9, 10, 11, 12, 13, 14, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai poly8x8_t __noswap_vget_high_p8(poly8x16_t __p0) {
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vget_high_p16(poly16x8_t __p0) {
+  poly16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
+  return __ret;
+}
+#else
+__ai poly16x4_t vget_high_p16(poly16x8_t __p0) {
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 4, 5, 6, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vget_high_u8(uint8x16_t __p0) {
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
+  return __ret;
+}
+#else
+__ai uint8x8_t vget_high_u8(uint8x16_t __p0) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 8, 9, 10, 11, 12, 13, 14, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint8x8_t __noswap_vget_high_u8(uint8x16_t __p0) {
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vget_high_u32(uint32x4_t __p0) {
+  uint32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
+  return __ret;
+}
+#else
+__ai uint32x2_t vget_high_u32(uint32x4_t __p0) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 2, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai uint32x2_t __noswap_vget_high_u32(uint32x4_t __p0) {
+  uint32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vget_high_u64(uint64x2_t __p0) {
+  uint64x1_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1);
+  return __ret;
+}
+#else
+__ai uint64x1_t vget_high_u64(uint64x2_t __p0) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x1_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vget_high_u16(uint16x8_t __p0) {
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
+  return __ret;
+}
+#else
+__ai uint16x4_t vget_high_u16(uint16x8_t __p0) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 4, 5, 6, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint16x4_t __noswap_vget_high_u16(uint16x8_t __p0) {
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vget_high_s8(int8x16_t __p0) {
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
+  return __ret;
+}
+#else
+__ai int8x8_t vget_high_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 8, 9, 10, 11, 12, 13, 14, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int8x8_t __noswap_vget_high_s8(int8x16_t __p0) {
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vget_high_f32(float32x4_t __p0) {
+  float32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
+  return __ret;
+}
+#else
+__ai float32x2_t vget_high_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 2, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai float32x2_t __noswap_vget_high_f32(float32x4_t __p0) {
+  float32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vget_high_f16(float16x8_t __p0) {
+  float16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
+  return __ret;
+}
+#else
+__ai float16x4_t vget_high_f16(float16x8_t __p0) {
+  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  float16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 4, 5, 6, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai float16x4_t __noswap_vget_high_f16(float16x8_t __p0) {
+  float16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vget_high_s32(int32x4_t __p0) {
+  int32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
+  return __ret;
+}
+#else
+__ai int32x2_t vget_high_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 2, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int32x2_t __noswap_vget_high_s32(int32x4_t __p0) {
+  int32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vget_high_s64(int64x2_t __p0) {
+  int64x1_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1);
+  return __ret;
+}
+#else
+__ai int64x1_t vget_high_s64(int64x2_t __p0) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x1_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vget_high_s16(int16x8_t __p0) {
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
+  return __ret;
+}
+#else
+__ai int16x4_t vget_high_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 4, 5, 6, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int16x4_t __noswap_vget_high_s16(int16x8_t __p0) {
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vget_lane_p8(__p0, __p1) __extension__ ({ \
+  poly8x8_t __s0 = __p0; \
+  poly8_t __ret; \
+  __ret = (poly8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vget_lane_p8(__p0, __p1) __extension__ ({ \
+  poly8x8_t __s0 = __p0; \
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8_t __ret; \
+  __ret = (poly8_t) __builtin_neon_vget_lane_i8((int8x8_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vget_lane_p8(__p0, __p1) __extension__ ({ \
+  poly8x8_t __s0 = __p0; \
+  poly8_t __ret; \
+  __ret = (poly8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vget_lane_p16(__p0, __p1) __extension__ ({ \
+  poly16x4_t __s0 = __p0; \
+  poly16_t __ret; \
+  __ret = (poly16_t) __builtin_neon_vget_lane_i16((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vget_lane_p16(__p0, __p1) __extension__ ({ \
+  poly16x4_t __s0 = __p0; \
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  poly16_t __ret; \
+  __ret = (poly16_t) __builtin_neon_vget_lane_i16((int8x8_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vget_lane_p16(__p0, __p1) __extension__ ({ \
+  poly16x4_t __s0 = __p0; \
+  poly16_t __ret; \
+  __ret = (poly16_t) __builtin_neon_vget_lane_i16((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vgetq_lane_p8(__p0, __p1) __extension__ ({ \
+  poly8x16_t __s0 = __p0; \
+  poly8_t __ret; \
+  __ret = (poly8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vgetq_lane_p8(__p0, __p1) __extension__ ({ \
+  poly8x16_t __s0 = __p0; \
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8_t __ret; \
+  __ret = (poly8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vgetq_lane_p8(__p0, __p1) __extension__ ({ \
+  poly8x16_t __s0 = __p0; \
+  poly8_t __ret; \
+  __ret = (poly8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vgetq_lane_p16(__p0, __p1) __extension__ ({ \
+  poly16x8_t __s0 = __p0; \
+  poly16_t __ret; \
+  __ret = (poly16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vgetq_lane_p16(__p0, __p1) __extension__ ({ \
+  poly16x8_t __s0 = __p0; \
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16_t __ret; \
+  __ret = (poly16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vgetq_lane_p16(__p0, __p1) __extension__ ({ \
+  poly16x8_t __s0 = __p0; \
+  poly16_t __ret; \
+  __ret = (poly16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vgetq_lane_u8(__p0, __p1) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vgetq_lane_u8(__p0, __p1) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vgetq_lane_u8(__p0, __p1) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vgetq_lane_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vgetq_lane_i32((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vgetq_lane_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vgetq_lane_i32((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vgetq_lane_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vgetq_lane_i32((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vgetq_lane_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vgetq_lane_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vgetq_lane_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vgetq_lane_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vgetq_lane_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vgetq_lane_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vgetq_lane_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vgetq_lane_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vgetq_lane_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vgetq_lane_f32(__p0, __p1) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32_t __ret; \
+  __ret = (float32_t) __builtin_neon_vgetq_lane_f32((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vgetq_lane_f32(__p0, __p1) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  float32_t __ret; \
+  __ret = (float32_t) __builtin_neon_vgetq_lane_f32((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vgetq_lane_f32(__p0, __p1) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32_t __ret; \
+  __ret = (float32_t) __builtin_neon_vgetq_lane_f32((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vgetq_lane_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vgetq_lane_i32((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vgetq_lane_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vgetq_lane_i32((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vgetq_lane_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vgetq_lane_i32((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vgetq_lane_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vgetq_lane_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vgetq_lane_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vgetq_lane_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vgetq_lane_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vgetq_lane_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vget_lane_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vget_lane_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vget_lane_i8((int8x8_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vget_lane_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vget_lane_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vget_lane_i32((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vget_lane_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vget_lane_i32((int8x8_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vget_lane_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vget_lane_i32((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vget_lane_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vget_lane_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#define __noswap_vget_lane_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vget_lane_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vget_lane_i16((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vget_lane_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vget_lane_i16((int8x8_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vget_lane_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vget_lane_i16((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vget_lane_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vget_lane_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vget_lane_i8((int8x8_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vget_lane_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vget_lane_f32(__p0, __p1) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32_t __ret; \
+  __ret = (float32_t) __builtin_neon_vget_lane_f32((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vget_lane_f32(__p0, __p1) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float32_t __ret; \
+  __ret = (float32_t) __builtin_neon_vget_lane_f32((int8x8_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vget_lane_f32(__p0, __p1) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32_t __ret; \
+  __ret = (float32_t) __builtin_neon_vget_lane_f32((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vget_lane_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vget_lane_i32((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vget_lane_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vget_lane_i32((int8x8_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vget_lane_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vget_lane_i32((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vget_lane_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vget_lane_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#define __noswap_vget_lane_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vget_lane_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vget_lane_i16((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vget_lane_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vget_lane_i16((int8x8_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vget_lane_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vget_lane_i16((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vget_low_p8(poly8x16_t __p0) {
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3, 4, 5, 6, 7);
+  return __ret;
+}
+#else
+__ai poly8x8_t vget_low_p8(poly8x16_t __p0) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3, 4, 5, 6, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vget_low_p16(poly16x8_t __p0) {
+  poly16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3);
+  return __ret;
+}
+#else
+__ai poly16x4_t vget_low_p16(poly16x8_t __p0) {
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vget_low_u8(uint8x16_t __p0) {
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3, 4, 5, 6, 7);
+  return __ret;
+}
+#else
+__ai uint8x8_t vget_low_u8(uint8x16_t __p0) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3, 4, 5, 6, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vget_low_u32(uint32x4_t __p0) {
+  uint32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 0, 1);
+  return __ret;
+}
+#else
+__ai uint32x2_t vget_low_u32(uint32x4_t __p0) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vget_low_u64(uint64x2_t __p0) {
+  uint64x1_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vget_low_u64(uint64x2_t __p0) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x1_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vget_low_u16(uint16x8_t __p0) {
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3);
+  return __ret;
+}
+#else
+__ai uint16x4_t vget_low_u16(uint16x8_t __p0) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vget_low_s8(int8x16_t __p0) {
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3, 4, 5, 6, 7);
+  return __ret;
+}
+#else
+__ai int8x8_t vget_low_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3, 4, 5, 6, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vget_low_f32(float32x4_t __p0) {
+  float32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 0, 1);
+  return __ret;
+}
+#else
+__ai float32x2_t vget_low_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vget_low_f16(float16x8_t __p0) {
+  float16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3);
+  return __ret;
+}
+#else
+__ai float16x4_t vget_low_f16(float16x8_t __p0) {
+  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  float16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vget_low_s32(int32x4_t __p0) {
+  int32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 0, 1);
+  return __ret;
+}
+#else
+__ai int32x2_t vget_low_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vget_low_s64(int64x2_t __p0) {
+  int64x1_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 0);
+  return __ret;
+}
+#else
+__ai int64x1_t vget_low_s64(int64x2_t __p0) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x1_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vget_low_s16(int16x8_t __p0) {
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3);
+  return __ret;
+}
+#else
+__ai int16x4_t vget_low_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vhaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vhaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vhaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vhaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vhaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vhaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vhaddq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vhaddq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vhaddq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vhaddq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vhaddq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vhaddq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vhadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vhadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vhadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vhadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vhadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vhadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vhadd_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vhadd_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vhadd_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vhadd_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vhadd_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vhadd_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vhsubq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vhsubq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vhsubq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vhsubq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vhsubq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vhsubq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vhsubq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vhsubq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vhsubq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vhsubq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vhsubq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vhsubq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vhsub_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vhsub_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vhsub_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vhsub_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vhsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vhsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vhsub_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vhsub_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vhsub_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vhsub_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vhsub_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vhsub_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_p8(__p0) __extension__ ({ \
+  poly8x8_t __ret; \
+  __ret = (poly8x8_t) __builtin_neon_vld1_v(__p0, 4); \
+  __ret; \
+})
+#else
+#define vld1_p8(__p0) __extension__ ({ \
+  poly8x8_t __ret; \
+  __ret = (poly8x8_t) __builtin_neon_vld1_v(__p0, 4); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_p16(__p0) __extension__ ({ \
+  poly16x4_t __ret; \
+  __ret = (poly16x4_t) __builtin_neon_vld1_v(__p0, 5); \
+  __ret; \
+})
+#else
+#define vld1_p16(__p0) __extension__ ({ \
+  poly16x4_t __ret; \
+  __ret = (poly16x4_t) __builtin_neon_vld1_v(__p0, 5); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_p8(__p0) __extension__ ({ \
+  poly8x16_t __ret; \
+  __ret = (poly8x16_t) __builtin_neon_vld1q_v(__p0, 36); \
+  __ret; \
+})
+#else
+#define vld1q_p8(__p0) __extension__ ({ \
+  poly8x16_t __ret; \
+  __ret = (poly8x16_t) __builtin_neon_vld1q_v(__p0, 36); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_p16(__p0) __extension__ ({ \
+  poly16x8_t __ret; \
+  __ret = (poly16x8_t) __builtin_neon_vld1q_v(__p0, 37); \
+  __ret; \
+})
+#else
+#define vld1q_p16(__p0) __extension__ ({ \
+  poly16x8_t __ret; \
+  __ret = (poly16x8_t) __builtin_neon_vld1q_v(__p0, 37); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_u8(__p0) __extension__ ({ \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vld1q_v(__p0, 48); \
+  __ret; \
+})
+#else
+#define vld1q_u8(__p0) __extension__ ({ \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vld1q_v(__p0, 48); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_u32(__p0) __extension__ ({ \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vld1q_v(__p0, 50); \
+  __ret; \
+})
+#else
+#define vld1q_u32(__p0) __extension__ ({ \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vld1q_v(__p0, 50); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_u64(__p0) __extension__ ({ \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vld1q_v(__p0, 51); \
+  __ret; \
+})
+#else
+#define vld1q_u64(__p0) __extension__ ({ \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vld1q_v(__p0, 51); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_u16(__p0) __extension__ ({ \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vld1q_v(__p0, 49); \
+  __ret; \
+})
+#else
+#define vld1q_u16(__p0) __extension__ ({ \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vld1q_v(__p0, 49); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_s8(__p0) __extension__ ({ \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vld1q_v(__p0, 32); \
+  __ret; \
+})
+#else
+#define vld1q_s8(__p0) __extension__ ({ \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vld1q_v(__p0, 32); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_f32(__p0) __extension__ ({ \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vld1q_v(__p0, 41); \
+  __ret; \
+})
+#else
+#define vld1q_f32(__p0) __extension__ ({ \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vld1q_v(__p0, 41); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_f16(__p0) __extension__ ({ \
+  float16x8_t __ret; \
+  __ret = (float16x8_t) __builtin_neon_vld1q_v(__p0, 40); \
+  __ret; \
+})
+#else
+#define vld1q_f16(__p0) __extension__ ({ \
+  float16x8_t __ret; \
+  __ret = (float16x8_t) __builtin_neon_vld1q_v(__p0, 40); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_s32(__p0) __extension__ ({ \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vld1q_v(__p0, 34); \
+  __ret; \
+})
+#else
+#define vld1q_s32(__p0) __extension__ ({ \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vld1q_v(__p0, 34); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_s64(__p0) __extension__ ({ \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vld1q_v(__p0, 35); \
+  __ret; \
+})
+#else
+#define vld1q_s64(__p0) __extension__ ({ \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vld1q_v(__p0, 35); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_s16(__p0) __extension__ ({ \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vld1q_v(__p0, 33); \
+  __ret; \
+})
+#else
+#define vld1q_s16(__p0) __extension__ ({ \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vld1q_v(__p0, 33); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_u8(__p0) __extension__ ({ \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vld1_v(__p0, 16); \
+  __ret; \
+})
+#else
+#define vld1_u8(__p0) __extension__ ({ \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vld1_v(__p0, 16); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_u32(__p0) __extension__ ({ \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vld1_v(__p0, 18); \
+  __ret; \
+})
+#else
+#define vld1_u32(__p0) __extension__ ({ \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vld1_v(__p0, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_u64(__p0) __extension__ ({ \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vld1_v(__p0, 19); \
+  __ret; \
+})
+#else
+#define vld1_u64(__p0) __extension__ ({ \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vld1_v(__p0, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_u16(__p0) __extension__ ({ \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vld1_v(__p0, 17); \
+  __ret; \
+})
+#else
+#define vld1_u16(__p0) __extension__ ({ \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vld1_v(__p0, 17); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_s8(__p0) __extension__ ({ \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vld1_v(__p0, 0); \
+  __ret; \
+})
+#else
+#define vld1_s8(__p0) __extension__ ({ \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vld1_v(__p0, 0); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_f32(__p0) __extension__ ({ \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vld1_v(__p0, 9); \
+  __ret; \
+})
+#else
+#define vld1_f32(__p0) __extension__ ({ \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vld1_v(__p0, 9); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_f16(__p0) __extension__ ({ \
+  float16x4_t __ret; \
+  __ret = (float16x4_t) __builtin_neon_vld1_v(__p0, 8); \
+  __ret; \
+})
+#else
+#define vld1_f16(__p0) __extension__ ({ \
+  float16x4_t __ret; \
+  __ret = (float16x4_t) __builtin_neon_vld1_v(__p0, 8); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_s32(__p0) __extension__ ({ \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vld1_v(__p0, 2); \
+  __ret; \
+})
+#else
+#define vld1_s32(__p0) __extension__ ({ \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vld1_v(__p0, 2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_s64(__p0) __extension__ ({ \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vld1_v(__p0, 3); \
+  __ret; \
+})
+#else
+#define vld1_s64(__p0) __extension__ ({ \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vld1_v(__p0, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_s16(__p0) __extension__ ({ \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vld1_v(__p0, 1); \
+  __ret; \
+})
+#else
+#define vld1_s16(__p0) __extension__ ({ \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vld1_v(__p0, 1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_dup_p8(__p0) __extension__ ({ \
+  poly8x8_t __ret; \
+  __ret = (poly8x8_t) __builtin_neon_vld1_dup_v(__p0, 4); \
+  __ret; \
+})
+#else
+#define vld1_dup_p8(__p0) __extension__ ({ \
+  poly8x8_t __ret; \
+  __ret = (poly8x8_t) __builtin_neon_vld1_dup_v(__p0, 4); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_dup_p16(__p0) __extension__ ({ \
+  poly16x4_t __ret; \
+  __ret = (poly16x4_t) __builtin_neon_vld1_dup_v(__p0, 5); \
+  __ret; \
+})
+#else
+#define vld1_dup_p16(__p0) __extension__ ({ \
+  poly16x4_t __ret; \
+  __ret = (poly16x4_t) __builtin_neon_vld1_dup_v(__p0, 5); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_dup_p8(__p0) __extension__ ({ \
+  poly8x16_t __ret; \
+  __ret = (poly8x16_t) __builtin_neon_vld1q_dup_v(__p0, 36); \
+  __ret; \
+})
+#else
+#define vld1q_dup_p8(__p0) __extension__ ({ \
+  poly8x16_t __ret; \
+  __ret = (poly8x16_t) __builtin_neon_vld1q_dup_v(__p0, 36); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_dup_p16(__p0) __extension__ ({ \
+  poly16x8_t __ret; \
+  __ret = (poly16x8_t) __builtin_neon_vld1q_dup_v(__p0, 37); \
+  __ret; \
+})
+#else
+#define vld1q_dup_p16(__p0) __extension__ ({ \
+  poly16x8_t __ret; \
+  __ret = (poly16x8_t) __builtin_neon_vld1q_dup_v(__p0, 37); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_dup_u8(__p0) __extension__ ({ \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vld1q_dup_v(__p0, 48); \
+  __ret; \
+})
+#else
+#define vld1q_dup_u8(__p0) __extension__ ({ \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vld1q_dup_v(__p0, 48); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_dup_u32(__p0) __extension__ ({ \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vld1q_dup_v(__p0, 50); \
+  __ret; \
+})
+#else
+#define vld1q_dup_u32(__p0) __extension__ ({ \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vld1q_dup_v(__p0, 50); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_dup_u64(__p0) __extension__ ({ \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vld1q_dup_v(__p0, 51); \
+  __ret; \
+})
+#else
+#define vld1q_dup_u64(__p0) __extension__ ({ \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vld1q_dup_v(__p0, 51); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_dup_u16(__p0) __extension__ ({ \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vld1q_dup_v(__p0, 49); \
+  __ret; \
+})
+#else
+#define vld1q_dup_u16(__p0) __extension__ ({ \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vld1q_dup_v(__p0, 49); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_dup_s8(__p0) __extension__ ({ \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vld1q_dup_v(__p0, 32); \
+  __ret; \
+})
+#else
+#define vld1q_dup_s8(__p0) __extension__ ({ \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vld1q_dup_v(__p0, 32); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_dup_f32(__p0) __extension__ ({ \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vld1q_dup_v(__p0, 41); \
+  __ret; \
+})
+#else
+#define vld1q_dup_f32(__p0) __extension__ ({ \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vld1q_dup_v(__p0, 41); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_dup_f16(__p0) __extension__ ({ \
+  float16x8_t __ret; \
+  __ret = (float16x8_t) __builtin_neon_vld1q_dup_v(__p0, 40); \
+  __ret; \
+})
+#else
+#define vld1q_dup_f16(__p0) __extension__ ({ \
+  float16x8_t __ret; \
+  __ret = (float16x8_t) __builtin_neon_vld1q_dup_v(__p0, 40); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_dup_s32(__p0) __extension__ ({ \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vld1q_dup_v(__p0, 34); \
+  __ret; \
+})
+#else
+#define vld1q_dup_s32(__p0) __extension__ ({ \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vld1q_dup_v(__p0, 34); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_dup_s64(__p0) __extension__ ({ \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vld1q_dup_v(__p0, 35); \
+  __ret; \
+})
+#else
+#define vld1q_dup_s64(__p0) __extension__ ({ \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vld1q_dup_v(__p0, 35); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_dup_s16(__p0) __extension__ ({ \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vld1q_dup_v(__p0, 33); \
+  __ret; \
+})
+#else
+#define vld1q_dup_s16(__p0) __extension__ ({ \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vld1q_dup_v(__p0, 33); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_dup_u8(__p0) __extension__ ({ \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vld1_dup_v(__p0, 16); \
+  __ret; \
+})
+#else
+#define vld1_dup_u8(__p0) __extension__ ({ \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vld1_dup_v(__p0, 16); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_dup_u32(__p0) __extension__ ({ \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vld1_dup_v(__p0, 18); \
+  __ret; \
+})
+#else
+#define vld1_dup_u32(__p0) __extension__ ({ \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vld1_dup_v(__p0, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_dup_u64(__p0) __extension__ ({ \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vld1_dup_v(__p0, 19); \
+  __ret; \
+})
+#else
+#define vld1_dup_u64(__p0) __extension__ ({ \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vld1_dup_v(__p0, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_dup_u16(__p0) __extension__ ({ \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vld1_dup_v(__p0, 17); \
+  __ret; \
+})
+#else
+#define vld1_dup_u16(__p0) __extension__ ({ \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vld1_dup_v(__p0, 17); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_dup_s8(__p0) __extension__ ({ \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vld1_dup_v(__p0, 0); \
+  __ret; \
+})
+#else
+#define vld1_dup_s8(__p0) __extension__ ({ \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vld1_dup_v(__p0, 0); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_dup_f32(__p0) __extension__ ({ \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vld1_dup_v(__p0, 9); \
+  __ret; \
+})
+#else
+#define vld1_dup_f32(__p0) __extension__ ({ \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vld1_dup_v(__p0, 9); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_dup_f16(__p0) __extension__ ({ \
+  float16x4_t __ret; \
+  __ret = (float16x4_t) __builtin_neon_vld1_dup_v(__p0, 8); \
+  __ret; \
+})
+#else
+#define vld1_dup_f16(__p0) __extension__ ({ \
+  float16x4_t __ret; \
+  __ret = (float16x4_t) __builtin_neon_vld1_dup_v(__p0, 8); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_dup_s32(__p0) __extension__ ({ \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vld1_dup_v(__p0, 2); \
+  __ret; \
+})
+#else
+#define vld1_dup_s32(__p0) __extension__ ({ \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vld1_dup_v(__p0, 2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_dup_s64(__p0) __extension__ ({ \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vld1_dup_v(__p0, 3); \
+  __ret; \
+})
+#else
+#define vld1_dup_s64(__p0) __extension__ ({ \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vld1_dup_v(__p0, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_dup_s16(__p0) __extension__ ({ \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vld1_dup_v(__p0, 1); \
+  __ret; \
+})
+#else
+#define vld1_dup_s16(__p0) __extension__ ({ \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vld1_dup_v(__p0, 1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8_t __s1 = __p1; \
+  poly8x8_t __ret; \
+  __ret = (poly8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 4); \
+  __ret; \
+})
+#else
+#define vld1_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8_t __s1 = __p1; \
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x8_t __ret; \
+  __ret = (poly8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 4); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4_t __s1 = __p1; \
+  poly16x4_t __ret; \
+  __ret = (poly16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 5); \
+  __ret; \
+})
+#else
+#define vld1_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4_t __s1 = __p1; \
+  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  poly16x4_t __ret; \
+  __ret = (poly16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 5); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16_t __s1 = __p1; \
+  poly8x16_t __ret; \
+  __ret = (poly8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 36); \
+  __ret; \
+})
+#else
+#define vld1q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16_t __s1 = __p1; \
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16_t __ret; \
+  __ret = (poly8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 36); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8_t __s1 = __p1; \
+  poly16x8_t __ret; \
+  __ret = (poly16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 37); \
+  __ret; \
+})
+#else
+#define vld1q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8_t __s1 = __p1; \
+  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x8_t __ret; \
+  __ret = (poly16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 37); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16_t __s1 = __p1; \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 48); \
+  __ret; \
+})
+#else
+#define vld1q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16_t __s1 = __p1; \
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 48); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 50); \
+  __ret; \
+})
+#else
+#define vld1q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 50); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2_t __s1 = __p1; \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 51); \
+  __ret; \
+})
+#else
+#define vld1q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2_t __s1 = __p1; \
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 51); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 49); \
+  __ret; \
+})
+#else
+#define vld1q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 49); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16_t __s1 = __p1; \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 32); \
+  __ret; \
+})
+#else
+#define vld1q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16_t __s1 = __p1; \
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 32); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4_t __s1 = __p1; \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 41); \
+  __ret; \
+})
+#else
+#define vld1q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4_t __s1 = __p1; \
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 41); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x8_t __s1 = __p1; \
+  float16x8_t __ret; \
+  __ret = (float16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 40); \
+  __ret; \
+})
+#else
+#define vld1q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x8_t __s1 = __p1; \
+  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x8_t __ret; \
+  __ret = (float16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 40); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 34); \
+  __ret; \
+})
+#else
+#define vld1q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 34); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2_t __s1 = __p1; \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 35); \
+  __ret; \
+})
+#else
+#define vld1q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2_t __s1 = __p1; \
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 35); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 33); \
+  __ret; \
+})
+#else
+#define vld1q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 33); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8_t __s1 = __p1; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 16); \
+  __ret; \
+})
+#else
+#define vld1_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8_t __s1 = __p1; \
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 16); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 18); \
+  __ret; \
+})
+#else
+#define vld1_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1_t __s1 = __p1; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 19); \
+  __ret; \
+})
+#else
+#define vld1_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1_t __s1 = __p1; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 17); \
+  __ret; \
+})
+#else
+#define vld1_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 17); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8_t __s1 = __p1; \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 0); \
+  __ret; \
+})
+#else
+#define vld1_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8_t __s1 = __p1; \
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 0); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2_t __s1 = __p1; \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 9); \
+  __ret; \
+})
+#else
+#define vld1_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2_t __s1 = __p1; \
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 9); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x4_t __s1 = __p1; \
+  float16x4_t __ret; \
+  __ret = (float16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 8); \
+  __ret; \
+})
+#else
+#define vld1_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x4_t __s1 = __p1; \
+  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  float16x4_t __ret; \
+  __ret = (float16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 8); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 2); \
+  __ret; \
+})
+#else
+#define vld1_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1_t __s1 = __p1; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 3); \
+  __ret; \
+})
+#else
+#define vld1_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1_t __s1 = __p1; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 1); \
+  __ret; \
+})
+#else
+#define vld1_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_p8(__p0) __extension__ ({ \
+  poly8x8x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 4); \
+  __ret; \
+})
+#else
+#define vld2_p8(__p0) __extension__ ({ \
+  poly8x8x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 4); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_p16(__p0) __extension__ ({ \
+  poly16x4x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 5); \
+  __ret; \
+})
+#else
+#define vld2_p16(__p0) __extension__ ({ \
+  poly16x4x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 5); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_p8(__p0) __extension__ ({ \
+  poly8x16x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 36); \
+  __ret; \
+})
+#else
+#define vld2q_p8(__p0) __extension__ ({ \
+  poly8x16x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 36); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_p16(__p0) __extension__ ({ \
+  poly16x8x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 37); \
+  __ret; \
+})
+#else
+#define vld2q_p16(__p0) __extension__ ({ \
+  poly16x8x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 37); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_u8(__p0) __extension__ ({ \
+  uint8x16x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 48); \
+  __ret; \
+})
+#else
+#define vld2q_u8(__p0) __extension__ ({ \
+  uint8x16x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 48); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_u32(__p0) __extension__ ({ \
+  uint32x4x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 50); \
+  __ret; \
+})
+#else
+#define vld2q_u32(__p0) __extension__ ({ \
+  uint32x4x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 50); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_u16(__p0) __extension__ ({ \
+  uint16x8x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 49); \
+  __ret; \
+})
+#else
+#define vld2q_u16(__p0) __extension__ ({ \
+  uint16x8x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 49); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_s8(__p0) __extension__ ({ \
+  int8x16x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 32); \
+  __ret; \
+})
+#else
+#define vld2q_s8(__p0) __extension__ ({ \
+  int8x16x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 32); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_f32(__p0) __extension__ ({ \
+  float32x4x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 41); \
+  __ret; \
+})
+#else
+#define vld2q_f32(__p0) __extension__ ({ \
+  float32x4x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 41); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_f16(__p0) __extension__ ({ \
+  float16x8x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 40); \
+  __ret; \
+})
+#else
+#define vld2q_f16(__p0) __extension__ ({ \
+  float16x8x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 40); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_s32(__p0) __extension__ ({ \
+  int32x4x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 34); \
+  __ret; \
+})
+#else
+#define vld2q_s32(__p0) __extension__ ({ \
+  int32x4x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 34); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_s16(__p0) __extension__ ({ \
+  int16x8x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 33); \
+  __ret; \
+})
+#else
+#define vld2q_s16(__p0) __extension__ ({ \
+  int16x8x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 33); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_u8(__p0) __extension__ ({ \
+  uint8x8x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 16); \
+  __ret; \
+})
+#else
+#define vld2_u8(__p0) __extension__ ({ \
+  uint8x8x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 16); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_u32(__p0) __extension__ ({ \
+  uint32x2x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 18); \
+  __ret; \
+})
+#else
+#define vld2_u32(__p0) __extension__ ({ \
+  uint32x2x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 18); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_u64(__p0) __extension__ ({ \
+  uint64x1x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 19); \
+  __ret; \
+})
+#else
+#define vld2_u64(__p0) __extension__ ({ \
+  uint64x1x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_u16(__p0) __extension__ ({ \
+  uint16x4x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 17); \
+  __ret; \
+})
+#else
+#define vld2_u16(__p0) __extension__ ({ \
+  uint16x4x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 17); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_s8(__p0) __extension__ ({ \
+  int8x8x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 0); \
+  __ret; \
+})
+#else
+#define vld2_s8(__p0) __extension__ ({ \
+  int8x8x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 0); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_f32(__p0) __extension__ ({ \
+  float32x2x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 9); \
+  __ret; \
+})
+#else
+#define vld2_f32(__p0) __extension__ ({ \
+  float32x2x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 9); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_f16(__p0) __extension__ ({ \
+  float16x4x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 8); \
+  __ret; \
+})
+#else
+#define vld2_f16(__p0) __extension__ ({ \
+  float16x4x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 8); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_s32(__p0) __extension__ ({ \
+  int32x2x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 2); \
+  __ret; \
+})
+#else
+#define vld2_s32(__p0) __extension__ ({ \
+  int32x2x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 2); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_s64(__p0) __extension__ ({ \
+  int64x1x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 3); \
+  __ret; \
+})
+#else
+#define vld2_s64(__p0) __extension__ ({ \
+  int64x1x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_s16(__p0) __extension__ ({ \
+  int16x4x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 1); \
+  __ret; \
+})
+#else
+#define vld2_s16(__p0) __extension__ ({ \
+  int16x4x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 1); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_dup_p8(__p0) __extension__ ({ \
+  poly8x8x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 4); \
+  __ret; \
+})
+#else
+#define vld2_dup_p8(__p0) __extension__ ({ \
+  poly8x8x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 4); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_dup_p16(__p0) __extension__ ({ \
+  poly16x4x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 5); \
+  __ret; \
+})
+#else
+#define vld2_dup_p16(__p0) __extension__ ({ \
+  poly16x4x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 5); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_dup_u8(__p0) __extension__ ({ \
+  uint8x8x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 16); \
+  __ret; \
+})
+#else
+#define vld2_dup_u8(__p0) __extension__ ({ \
+  uint8x8x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 16); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_dup_u32(__p0) __extension__ ({ \
+  uint32x2x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 18); \
+  __ret; \
+})
+#else
+#define vld2_dup_u32(__p0) __extension__ ({ \
+  uint32x2x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 18); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_dup_u64(__p0) __extension__ ({ \
+  uint64x1x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 19); \
+  __ret; \
+})
+#else
+#define vld2_dup_u64(__p0) __extension__ ({ \
+  uint64x1x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_dup_u16(__p0) __extension__ ({ \
+  uint16x4x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 17); \
+  __ret; \
+})
+#else
+#define vld2_dup_u16(__p0) __extension__ ({ \
+  uint16x4x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 17); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_dup_s8(__p0) __extension__ ({ \
+  int8x8x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 0); \
+  __ret; \
+})
+#else
+#define vld2_dup_s8(__p0) __extension__ ({ \
+  int8x8x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 0); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_dup_f32(__p0) __extension__ ({ \
+  float32x2x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 9); \
+  __ret; \
+})
+#else
+#define vld2_dup_f32(__p0) __extension__ ({ \
+  float32x2x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 9); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_dup_f16(__p0) __extension__ ({ \
+  float16x4x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 8); \
+  __ret; \
+})
+#else
+#define vld2_dup_f16(__p0) __extension__ ({ \
+  float16x4x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 8); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_dup_s32(__p0) __extension__ ({ \
+  int32x2x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 2); \
+  __ret; \
+})
+#else
+#define vld2_dup_s32(__p0) __extension__ ({ \
+  int32x2x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 2); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_dup_s64(__p0) __extension__ ({ \
+  int64x1x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 3); \
+  __ret; \
+})
+#else
+#define vld2_dup_s64(__p0) __extension__ ({ \
+  int64x1x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_dup_s16(__p0) __extension__ ({ \
+  int16x4x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 1); \
+  __ret; \
+})
+#else
+#define vld2_dup_s16(__p0) __extension__ ({ \
+  int16x4x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 1); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8x2_t __s1 = __p1; \
+  poly8x8x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 4); \
+  __ret; \
+})
+#else
+#define vld2_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8x2_t __s1 = __p1; \
+  poly8x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x8x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 4); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4x2_t __s1 = __p1; \
+  poly16x4x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 5); \
+  __ret; \
+})
+#else
+#define vld2_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4x2_t __s1 = __p1; \
+  poly16x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  poly16x4x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 5); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8x2_t __s1 = __p1; \
+  poly16x8x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 37); \
+  __ret; \
+})
+#else
+#define vld2q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8x2_t __s1 = __p1; \
+  poly16x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x8x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 37); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4x2_t __s1 = __p1; \
+  uint32x4x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 50); \
+  __ret; \
+})
+#else
+#define vld2q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4x2_t __s1 = __p1; \
+  uint32x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  uint32x4x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 50); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8x2_t __s1 = __p1; \
+  uint16x8x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 49); \
+  __ret; \
+})
+#else
+#define vld2q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8x2_t __s1 = __p1; \
+  uint16x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 49); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4x2_t __s1 = __p1; \
+  float32x4x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 41); \
+  __ret; \
+})
+#else
+#define vld2q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4x2_t __s1 = __p1; \
+  float32x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  float32x4x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 41); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x8x2_t __s1 = __p1; \
+  float16x8x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 40); \
+  __ret; \
+})
+#else
+#define vld2q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x8x2_t __s1 = __p1; \
+  float16x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x8x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 40); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4x2_t __s1 = __p1; \
+  int32x4x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 34); \
+  __ret; \
+})
+#else
+#define vld2q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4x2_t __s1 = __p1; \
+  int32x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  int32x4x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 34); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8x2_t __s1 = __p1; \
+  int16x8x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 33); \
+  __ret; \
+})
+#else
+#define vld2q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8x2_t __s1 = __p1; \
+  int16x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 33); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8x2_t __s1 = __p1; \
+  uint8x8x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 16); \
+  __ret; \
+})
+#else
+#define vld2_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8x2_t __s1 = __p1; \
+  uint8x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 16); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2x2_t __s1 = __p1; \
+  uint32x2x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 18); \
+  __ret; \
+})
+#else
+#define vld2_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2x2_t __s1 = __p1; \
+  uint32x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  uint32x2x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 18); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4x2_t __s1 = __p1; \
+  uint16x4x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 17); \
+  __ret; \
+})
+#else
+#define vld2_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4x2_t __s1 = __p1; \
+  uint16x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  uint16x4x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 17); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8x2_t __s1 = __p1; \
+  int8x8x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 0); \
+  __ret; \
+})
+#else
+#define vld2_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8x2_t __s1 = __p1; \
+  int8x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 0); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2x2_t __s1 = __p1; \
+  float32x2x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 9); \
+  __ret; \
+})
+#else
+#define vld2_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2x2_t __s1 = __p1; \
+  float32x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  float32x2x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 9); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x4x2_t __s1 = __p1; \
+  float16x4x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 8); \
+  __ret; \
+})
+#else
+#define vld2_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x4x2_t __s1 = __p1; \
+  float16x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  float16x4x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 8); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2x2_t __s1 = __p1; \
+  int32x2x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 2); \
+  __ret; \
+})
+#else
+#define vld2_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2x2_t __s1 = __p1; \
+  int32x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  int32x2x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 2); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4x2_t __s1 = __p1; \
+  int16x4x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 1); \
+  __ret; \
+})
+#else
+#define vld2_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4x2_t __s1 = __p1; \
+  int16x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  int16x4x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 1); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_p8(__p0) __extension__ ({ \
+  poly8x8x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 4); \
+  __ret; \
+})
+#else
+#define vld3_p8(__p0) __extension__ ({ \
+  poly8x8x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 4); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_p16(__p0) __extension__ ({ \
+  poly16x4x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 5); \
+  __ret; \
+})
+#else
+#define vld3_p16(__p0) __extension__ ({ \
+  poly16x4x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 5); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_p8(__p0) __extension__ ({ \
+  poly8x16x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 36); \
+  __ret; \
+})
+#else
+#define vld3q_p8(__p0) __extension__ ({ \
+  poly8x16x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 36); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_p16(__p0) __extension__ ({ \
+  poly16x8x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 37); \
+  __ret; \
+})
+#else
+#define vld3q_p16(__p0) __extension__ ({ \
+  poly16x8x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 37); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_u8(__p0) __extension__ ({ \
+  uint8x16x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 48); \
+  __ret; \
+})
+#else
+#define vld3q_u8(__p0) __extension__ ({ \
+  uint8x16x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 48); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_u32(__p0) __extension__ ({ \
+  uint32x4x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 50); \
+  __ret; \
+})
+#else
+#define vld3q_u32(__p0) __extension__ ({ \
+  uint32x4x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 50); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_u16(__p0) __extension__ ({ \
+  uint16x8x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 49); \
+  __ret; \
+})
+#else
+#define vld3q_u16(__p0) __extension__ ({ \
+  uint16x8x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 49); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_s8(__p0) __extension__ ({ \
+  int8x16x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 32); \
+  __ret; \
+})
+#else
+#define vld3q_s8(__p0) __extension__ ({ \
+  int8x16x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 32); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_f32(__p0) __extension__ ({ \
+  float32x4x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 41); \
+  __ret; \
+})
+#else
+#define vld3q_f32(__p0) __extension__ ({ \
+  float32x4x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 41); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_f16(__p0) __extension__ ({ \
+  float16x8x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 40); \
+  __ret; \
+})
+#else
+#define vld3q_f16(__p0) __extension__ ({ \
+  float16x8x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 40); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_s32(__p0) __extension__ ({ \
+  int32x4x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 34); \
+  __ret; \
+})
+#else
+#define vld3q_s32(__p0) __extension__ ({ \
+  int32x4x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 34); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_s16(__p0) __extension__ ({ \
+  int16x8x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 33); \
+  __ret; \
+})
+#else
+#define vld3q_s16(__p0) __extension__ ({ \
+  int16x8x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 33); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_u8(__p0) __extension__ ({ \
+  uint8x8x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 16); \
+  __ret; \
+})
+#else
+#define vld3_u8(__p0) __extension__ ({ \
+  uint8x8x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 16); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_u32(__p0) __extension__ ({ \
+  uint32x2x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 18); \
+  __ret; \
+})
+#else
+#define vld3_u32(__p0) __extension__ ({ \
+  uint32x2x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 18); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_u64(__p0) __extension__ ({ \
+  uint64x1x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 19); \
+  __ret; \
+})
+#else
+#define vld3_u64(__p0) __extension__ ({ \
+  uint64x1x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_u16(__p0) __extension__ ({ \
+  uint16x4x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 17); \
+  __ret; \
+})
+#else
+#define vld3_u16(__p0) __extension__ ({ \
+  uint16x4x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 17); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_s8(__p0) __extension__ ({ \
+  int8x8x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 0); \
+  __ret; \
+})
+#else
+#define vld3_s8(__p0) __extension__ ({ \
+  int8x8x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 0); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_f32(__p0) __extension__ ({ \
+  float32x2x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 9); \
+  __ret; \
+})
+#else
+#define vld3_f32(__p0) __extension__ ({ \
+  float32x2x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 9); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_f16(__p0) __extension__ ({ \
+  float16x4x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 8); \
+  __ret; \
+})
+#else
+#define vld3_f16(__p0) __extension__ ({ \
+  float16x4x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 8); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_s32(__p0) __extension__ ({ \
+  int32x2x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 2); \
+  __ret; \
+})
+#else
+#define vld3_s32(__p0) __extension__ ({ \
+  int32x2x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 2); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_s64(__p0) __extension__ ({ \
+  int64x1x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 3); \
+  __ret; \
+})
+#else
+#define vld3_s64(__p0) __extension__ ({ \
+  int64x1x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_s16(__p0) __extension__ ({ \
+  int16x4x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 1); \
+  __ret; \
+})
+#else
+#define vld3_s16(__p0) __extension__ ({ \
+  int16x4x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 1); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_dup_p8(__p0) __extension__ ({ \
+  poly8x8x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 4); \
+  __ret; \
+})
+#else
+#define vld3_dup_p8(__p0) __extension__ ({ \
+  poly8x8x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 4); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_dup_p16(__p0) __extension__ ({ \
+  poly16x4x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 5); \
+  __ret; \
+})
+#else
+#define vld3_dup_p16(__p0) __extension__ ({ \
+  poly16x4x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 5); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_dup_u8(__p0) __extension__ ({ \
+  uint8x8x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 16); \
+  __ret; \
+})
+#else
+#define vld3_dup_u8(__p0) __extension__ ({ \
+  uint8x8x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 16); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_dup_u32(__p0) __extension__ ({ \
+  uint32x2x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 18); \
+  __ret; \
+})
+#else
+#define vld3_dup_u32(__p0) __extension__ ({ \
+  uint32x2x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 18); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_dup_u64(__p0) __extension__ ({ \
+  uint64x1x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 19); \
+  __ret; \
+})
+#else
+#define vld3_dup_u64(__p0) __extension__ ({ \
+  uint64x1x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_dup_u16(__p0) __extension__ ({ \
+  uint16x4x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 17); \
+  __ret; \
+})
+#else
+#define vld3_dup_u16(__p0) __extension__ ({ \
+  uint16x4x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 17); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_dup_s8(__p0) __extension__ ({ \
+  int8x8x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 0); \
+  __ret; \
+})
+#else
+#define vld3_dup_s8(__p0) __extension__ ({ \
+  int8x8x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 0); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_dup_f32(__p0) __extension__ ({ \
+  float32x2x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 9); \
+  __ret; \
+})
+#else
+#define vld3_dup_f32(__p0) __extension__ ({ \
+  float32x2x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 9); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_dup_f16(__p0) __extension__ ({ \
+  float16x4x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 8); \
+  __ret; \
+})
+#else
+#define vld3_dup_f16(__p0) __extension__ ({ \
+  float16x4x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 8); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_dup_s32(__p0) __extension__ ({ \
+  int32x2x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 2); \
+  __ret; \
+})
+#else
+#define vld3_dup_s32(__p0) __extension__ ({ \
+  int32x2x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 2); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_dup_s64(__p0) __extension__ ({ \
+  int64x1x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 3); \
+  __ret; \
+})
+#else
+#define vld3_dup_s64(__p0) __extension__ ({ \
+  int64x1x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_dup_s16(__p0) __extension__ ({ \
+  int16x4x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 1); \
+  __ret; \
+})
+#else
+#define vld3_dup_s16(__p0) __extension__ ({ \
+  int16x4x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 1); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8x3_t __s1 = __p1; \
+  poly8x8x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 4); \
+  __ret; \
+})
+#else
+#define vld3_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8x3_t __s1 = __p1; \
+  poly8x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x8x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 4); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4x3_t __s1 = __p1; \
+  poly16x4x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 5); \
+  __ret; \
+})
+#else
+#define vld3_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4x3_t __s1 = __p1; \
+  poly16x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  poly16x4x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 5); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8x3_t __s1 = __p1; \
+  poly16x8x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 37); \
+  __ret; \
+})
+#else
+#define vld3q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8x3_t __s1 = __p1; \
+  poly16x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x8x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 37); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4x3_t __s1 = __p1; \
+  uint32x4x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 50); \
+  __ret; \
+})
+#else
+#define vld3q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4x3_t __s1 = __p1; \
+  uint32x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  uint32x4x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 50); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8x3_t __s1 = __p1; \
+  uint16x8x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 49); \
+  __ret; \
+})
+#else
+#define vld3q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8x3_t __s1 = __p1; \
+  uint16x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 49); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4x3_t __s1 = __p1; \
+  float32x4x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 41); \
+  __ret; \
+})
+#else
+#define vld3q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4x3_t __s1 = __p1; \
+  float32x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  float32x4x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 41); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x8x3_t __s1 = __p1; \
+  float16x8x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 40); \
+  __ret; \
+})
+#else
+#define vld3q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x8x3_t __s1 = __p1; \
+  float16x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x8x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 40); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4x3_t __s1 = __p1; \
+  int32x4x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 34); \
+  __ret; \
+})
+#else
+#define vld3q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4x3_t __s1 = __p1; \
+  int32x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  int32x4x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 34); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8x3_t __s1 = __p1; \
+  int16x8x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 33); \
+  __ret; \
+})
+#else
+#define vld3q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8x3_t __s1 = __p1; \
+  int16x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 33); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8x3_t __s1 = __p1; \
+  uint8x8x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 16); \
+  __ret; \
+})
+#else
+#define vld3_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8x3_t __s1 = __p1; \
+  uint8x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 16); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2x3_t __s1 = __p1; \
+  uint32x2x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 18); \
+  __ret; \
+})
+#else
+#define vld3_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2x3_t __s1 = __p1; \
+  uint32x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  uint32x2x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 18); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4x3_t __s1 = __p1; \
+  uint16x4x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 17); \
+  __ret; \
+})
+#else
+#define vld3_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4x3_t __s1 = __p1; \
+  uint16x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  uint16x4x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 17); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8x3_t __s1 = __p1; \
+  int8x8x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 0); \
+  __ret; \
+})
+#else
+#define vld3_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8x3_t __s1 = __p1; \
+  int8x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 0); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2x3_t __s1 = __p1; \
+  float32x2x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 9); \
+  __ret; \
+})
+#else
+#define vld3_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2x3_t __s1 = __p1; \
+  float32x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  float32x2x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 9); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x4x3_t __s1 = __p1; \
+  float16x4x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 8); \
+  __ret; \
+})
+#else
+#define vld3_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x4x3_t __s1 = __p1; \
+  float16x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  float16x4x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 8); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2x3_t __s1 = __p1; \
+  int32x2x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 2); \
+  __ret; \
+})
+#else
+#define vld3_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2x3_t __s1 = __p1; \
+  int32x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  int32x2x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 2); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4x3_t __s1 = __p1; \
+  int16x4x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 1); \
+  __ret; \
+})
+#else
+#define vld3_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4x3_t __s1 = __p1; \
+  int16x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  int16x4x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 1); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_p8(__p0) __extension__ ({ \
+  poly8x8x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 4); \
+  __ret; \
+})
+#else
+#define vld4_p8(__p0) __extension__ ({ \
+  poly8x8x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 4); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_p16(__p0) __extension__ ({ \
+  poly16x4x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 5); \
+  __ret; \
+})
+#else
+#define vld4_p16(__p0) __extension__ ({ \
+  poly16x4x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 5); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_p8(__p0) __extension__ ({ \
+  poly8x16x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 36); \
+  __ret; \
+})
+#else
+#define vld4q_p8(__p0) __extension__ ({ \
+  poly8x16x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 36); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_p16(__p0) __extension__ ({ \
+  poly16x8x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 37); \
+  __ret; \
+})
+#else
+#define vld4q_p16(__p0) __extension__ ({ \
+  poly16x8x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 37); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_u8(__p0) __extension__ ({ \
+  uint8x16x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 48); \
+  __ret; \
+})
+#else
+#define vld4q_u8(__p0) __extension__ ({ \
+  uint8x16x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 48); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_u32(__p0) __extension__ ({ \
+  uint32x4x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 50); \
+  __ret; \
+})
+#else
+#define vld4q_u32(__p0) __extension__ ({ \
+  uint32x4x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 50); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_u16(__p0) __extension__ ({ \
+  uint16x8x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 49); \
+  __ret; \
+})
+#else
+#define vld4q_u16(__p0) __extension__ ({ \
+  uint16x8x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 49); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_s8(__p0) __extension__ ({ \
+  int8x16x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 32); \
+  __ret; \
+})
+#else
+#define vld4q_s8(__p0) __extension__ ({ \
+  int8x16x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 32); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_f32(__p0) __extension__ ({ \
+  float32x4x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 41); \
+  __ret; \
+})
+#else
+#define vld4q_f32(__p0) __extension__ ({ \
+  float32x4x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 41); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_f16(__p0) __extension__ ({ \
+  float16x8x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 40); \
+  __ret; \
+})
+#else
+#define vld4q_f16(__p0) __extension__ ({ \
+  float16x8x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 40); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_s32(__p0) __extension__ ({ \
+  int32x4x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 34); \
+  __ret; \
+})
+#else
+#define vld4q_s32(__p0) __extension__ ({ \
+  int32x4x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 34); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_s16(__p0) __extension__ ({ \
+  int16x8x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 33); \
+  __ret; \
+})
+#else
+#define vld4q_s16(__p0) __extension__ ({ \
+  int16x8x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 33); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_u8(__p0) __extension__ ({ \
+  uint8x8x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 16); \
+  __ret; \
+})
+#else
+#define vld4_u8(__p0) __extension__ ({ \
+  uint8x8x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 16); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_u32(__p0) __extension__ ({ \
+  uint32x2x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 18); \
+  __ret; \
+})
+#else
+#define vld4_u32(__p0) __extension__ ({ \
+  uint32x2x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 18); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_u64(__p0) __extension__ ({ \
+  uint64x1x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 19); \
+  __ret; \
+})
+#else
+#define vld4_u64(__p0) __extension__ ({ \
+  uint64x1x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_u16(__p0) __extension__ ({ \
+  uint16x4x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 17); \
+  __ret; \
+})
+#else
+#define vld4_u16(__p0) __extension__ ({ \
+  uint16x4x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 17); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_s8(__p0) __extension__ ({ \
+  int8x8x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 0); \
+  __ret; \
+})
+#else
+#define vld4_s8(__p0) __extension__ ({ \
+  int8x8x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 0); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_f32(__p0) __extension__ ({ \
+  float32x2x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 9); \
+  __ret; \
+})
+#else
+#define vld4_f32(__p0) __extension__ ({ \
+  float32x2x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 9); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_f16(__p0) __extension__ ({ \
+  float16x4x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 8); \
+  __ret; \
+})
+#else
+#define vld4_f16(__p0) __extension__ ({ \
+  float16x4x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 8); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_s32(__p0) __extension__ ({ \
+  int32x2x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 2); \
+  __ret; \
+})
+#else
+#define vld4_s32(__p0) __extension__ ({ \
+  int32x2x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 2); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_s64(__p0) __extension__ ({ \
+  int64x1x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 3); \
+  __ret; \
+})
+#else
+#define vld4_s64(__p0) __extension__ ({ \
+  int64x1x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_s16(__p0) __extension__ ({ \
+  int16x4x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 1); \
+  __ret; \
+})
+#else
+#define vld4_s16(__p0) __extension__ ({ \
+  int16x4x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 1); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_dup_p8(__p0) __extension__ ({ \
+  poly8x8x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 4); \
+  __ret; \
+})
+#else
+#define vld4_dup_p8(__p0) __extension__ ({ \
+  poly8x8x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 4); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_dup_p16(__p0) __extension__ ({ \
+  poly16x4x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 5); \
+  __ret; \
+})
+#else
+#define vld4_dup_p16(__p0) __extension__ ({ \
+  poly16x4x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 5); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_dup_u8(__p0) __extension__ ({ \
+  uint8x8x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 16); \
+  __ret; \
+})
+#else
+#define vld4_dup_u8(__p0) __extension__ ({ \
+  uint8x8x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 16); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_dup_u32(__p0) __extension__ ({ \
+  uint32x2x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 18); \
+  __ret; \
+})
+#else
+#define vld4_dup_u32(__p0) __extension__ ({ \
+  uint32x2x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 18); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_dup_u64(__p0) __extension__ ({ \
+  uint64x1x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 19); \
+  __ret; \
+})
+#else
+#define vld4_dup_u64(__p0) __extension__ ({ \
+  uint64x1x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_dup_u16(__p0) __extension__ ({ \
+  uint16x4x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 17); \
+  __ret; \
+})
+#else
+#define vld4_dup_u16(__p0) __extension__ ({ \
+  uint16x4x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 17); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_dup_s8(__p0) __extension__ ({ \
+  int8x8x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 0); \
+  __ret; \
+})
+#else
+#define vld4_dup_s8(__p0) __extension__ ({ \
+  int8x8x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 0); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_dup_f32(__p0) __extension__ ({ \
+  float32x2x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 9); \
+  __ret; \
+})
+#else
+#define vld4_dup_f32(__p0) __extension__ ({ \
+  float32x2x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 9); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_dup_f16(__p0) __extension__ ({ \
+  float16x4x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 8); \
+  __ret; \
+})
+#else
+#define vld4_dup_f16(__p0) __extension__ ({ \
+  float16x4x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 8); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_dup_s32(__p0) __extension__ ({ \
+  int32x2x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 2); \
+  __ret; \
+})
+#else
+#define vld4_dup_s32(__p0) __extension__ ({ \
+  int32x2x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 2); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_dup_s64(__p0) __extension__ ({ \
+  int64x1x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 3); \
+  __ret; \
+})
+#else
+#define vld4_dup_s64(__p0) __extension__ ({ \
+  int64x1x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_dup_s16(__p0) __extension__ ({ \
+  int16x4x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 1); \
+  __ret; \
+})
+#else
+#define vld4_dup_s16(__p0) __extension__ ({ \
+  int16x4x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 1); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8x4_t __s1 = __p1; \
+  poly8x8x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 4); \
+  __ret; \
+})
+#else
+#define vld4_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8x4_t __s1 = __p1; \
+  poly8x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x8x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 4); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4x4_t __s1 = __p1; \
+  poly16x4x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 5); \
+  __ret; \
+})
+#else
+#define vld4_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4x4_t __s1 = __p1; \
+  poly16x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  poly16x4x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 5); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8x4_t __s1 = __p1; \
+  poly16x8x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 37); \
+  __ret; \
+})
+#else
+#define vld4q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8x4_t __s1 = __p1; \
+  poly16x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x8x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 37); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4x4_t __s1 = __p1; \
+  uint32x4x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 50); \
+  __ret; \
+})
+#else
+#define vld4q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4x4_t __s1 = __p1; \
+  uint32x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  uint32x4x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 50); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8x4_t __s1 = __p1; \
+  uint16x8x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 49); \
+  __ret; \
+})
+#else
+#define vld4q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8x4_t __s1 = __p1; \
+  uint16x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 49); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4x4_t __s1 = __p1; \
+  float32x4x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 41); \
+  __ret; \
+})
+#else
+#define vld4q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4x4_t __s1 = __p1; \
+  float32x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  float32x4x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 41); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x8x4_t __s1 = __p1; \
+  float16x8x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 40); \
+  __ret; \
+})
+#else
+#define vld4q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x8x4_t __s1 = __p1; \
+  float16x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x8x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 40); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4x4_t __s1 = __p1; \
+  int32x4x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 34); \
+  __ret; \
+})
+#else
+#define vld4q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4x4_t __s1 = __p1; \
+  int32x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  int32x4x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 34); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8x4_t __s1 = __p1; \
+  int16x8x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 33); \
+  __ret; \
+})
+#else
+#define vld4q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8x4_t __s1 = __p1; \
+  int16x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 33); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8x4_t __s1 = __p1; \
+  uint8x8x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 16); \
+  __ret; \
+})
+#else
+#define vld4_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8x4_t __s1 = __p1; \
+  uint8x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 16); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2x4_t __s1 = __p1; \
+  uint32x2x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 18); \
+  __ret; \
+})
+#else
+#define vld4_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2x4_t __s1 = __p1; \
+  uint32x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  uint32x2x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 18); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4x4_t __s1 = __p1; \
+  uint16x4x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 17); \
+  __ret; \
+})
+#else
+#define vld4_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4x4_t __s1 = __p1; \
+  uint16x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  uint16x4x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 17); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8x4_t __s1 = __p1; \
+  int8x8x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 0); \
+  __ret; \
+})
+#else
+#define vld4_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8x4_t __s1 = __p1; \
+  int8x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 0); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2x4_t __s1 = __p1; \
+  float32x2x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 9); \
+  __ret; \
+})
+#else
+#define vld4_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2x4_t __s1 = __p1; \
+  float32x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  float32x2x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 9); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x4x4_t __s1 = __p1; \
+  float16x4x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 8); \
+  __ret; \
+})
+#else
+#define vld4_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x4x4_t __s1 = __p1; \
+  float16x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  float16x4x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 8); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2x4_t __s1 = __p1; \
+  int32x2x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 2); \
+  __ret; \
+})
+#else
+#define vld4_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2x4_t __s1 = __p1; \
+  int32x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  int32x2x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 2); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4x4_t __s1 = __p1; \
+  int16x4x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 1); \
+  __ret; \
+})
+#else
+#define vld4_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4x4_t __s1 = __p1; \
+  int16x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  int16x4x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 1); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vmaxq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vmaxq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmaxq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vmaxq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vmaxq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vmaxq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vmaxq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vmaxq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vmaxq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vmaxq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmaxq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vmaxq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vmaxq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vmaxq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vmax_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vmax_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vmax_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vmax_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vmax_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vmax_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vmax_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vmax_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vmax_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vmax_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vmax_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vmax_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vmax_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vmax_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vminq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vminq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vminq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vminq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vminq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vminq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vminq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vminq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vminq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vminq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vminq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vminq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vminq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vminq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vmin_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vmin_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vmin_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vmin_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vmin_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vmin_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vmin_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vmin_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vmin_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vmin_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vmin_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vmin_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vmin_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vmin_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vmlaq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
+  uint8x16_t __ret;
+  __ret = __p0 + __p1 * __p2;
+  return __ret;
+}
+#else
+__ai uint8x16_t vmlaq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __rev0 + __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmlaq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint32x4_t __ret;
+  __ret = __p0 + __p1 * __p2;
+  return __ret;
+}
+#else
+__ai uint32x4_t vmlaq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 + __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vmlaq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint16x8_t __ret;
+  __ret = __p0 + __p1 * __p2;
+  return __ret;
+}
+#else
+__ai uint16x8_t vmlaq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 + __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vmlaq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
+  int8x16_t __ret;
+  __ret = __p0 + __p1 * __p2;
+  return __ret;
+}
+#else
+__ai int8x16_t vmlaq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __rev0 + __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vmlaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
+  float32x4_t __ret;
+  __ret = __p0 + __p1 * __p2;
+  return __ret;
+}
+#else
+__ai float32x4_t vmlaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = __rev0 + __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmlaq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int32x4_t __ret;
+  __ret = __p0 + __p1 * __p2;
+  return __ret;
+}
+#else
+__ai int32x4_t vmlaq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 + __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vmlaq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int16x8_t __ret;
+  __ret = __p0 + __p1 * __p2;
+  return __ret;
+}
+#else
+__ai int16x8_t vmlaq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 + __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vmla_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
+  uint8x8_t __ret;
+  __ret = __p0 + __p1 * __p2;
+  return __ret;
+}
+#else
+__ai uint8x8_t vmla_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __rev0 + __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vmla_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
+  uint32x2_t __ret;
+  __ret = __p0 + __p1 * __p2;
+  return __ret;
+}
+#else
+__ai uint32x2_t vmla_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  uint32x2_t __ret;
+  __ret = __rev0 + __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vmla_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
+  uint16x4_t __ret;
+  __ret = __p0 + __p1 * __p2;
+  return __ret;
+}
+#else
+__ai uint16x4_t vmla_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __rev0 + __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vmla_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
+  int8x8_t __ret;
+  __ret = __p0 + __p1 * __p2;
+  return __ret;
+}
+#else
+__ai int8x8_t vmla_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __rev0 + __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vmla_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
+  float32x2_t __ret;
+  __ret = __p0 + __p1 * __p2;
+  return __ret;
+}
+#else
+__ai float32x2_t vmla_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  float32x2_t __ret;
+  __ret = __rev0 + __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vmla_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int32x2_t __ret;
+  __ret = __p0 + __p1 * __p2;
+  return __ret;
+}
+#else
+__ai int32x2_t vmla_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  int32x2_t __ret;
+  __ret = __rev0 + __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vmla_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int16x4_t __ret;
+  __ret = __p0 + __p1 * __p2;
+  return __ret;
+}
+#else
+__ai int16x4_t vmla_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __rev0 + __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlaq_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x2_t __s2 = __p2; \
+  uint32x4_t __ret; \
+  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmlaq_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x2_t __s2 = __p2; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlaq_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x4_t __s2 = __p2; \
+  uint16x8_t __ret; \
+  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmlaq_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x4_t __s2 = __p2; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlaq_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x2_t __s2 = __p2; \
+  float32x4_t __ret; \
+  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmlaq_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x2_t __s2 = __p2; \
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  float32x4_t __ret; \
+  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlaq_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int32x4_t __ret; \
+  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmlaq_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlaq_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int16x8_t __ret; \
+  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmlaq_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmla_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __s2 = __p2; \
+  uint32x2_t __ret; \
+  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmla_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __s2 = __p2; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmla_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __s2 = __p2; \
+  uint16x4_t __ret; \
+  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmla_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __s2 = __p2; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmla_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x2_t __s2 = __p2; \
+  float32x2_t __ret; \
+  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmla_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x2_t __s2 = __p2; \
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  float32x2_t __ret; \
+  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmla_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int32x2_t __ret; \
+  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmla_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  int32x2_t __ret; \
+  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmla_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int16x4_t __ret; \
+  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmla_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmlaq_n_u32(uint32x4_t __p0, uint32x4_t __p1, uint32_t __p2) {
+  uint32x4_t __ret;
+  __ret = __p0 + __p1 * (uint32x4_t) {__p2, __p2, __p2, __p2};
+  return __ret;
+}
+#else
+__ai uint32x4_t vmlaq_n_u32(uint32x4_t __p0, uint32x4_t __p1, uint32_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 + __rev1 * (uint32x4_t) {__p2, __p2, __p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vmlaq_n_u16(uint16x8_t __p0, uint16x8_t __p1, uint16_t __p2) {
+  uint16x8_t __ret;
+  __ret = __p0 + __p1 * (uint16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
+  return __ret;
+}
+#else
+__ai uint16x8_t vmlaq_n_u16(uint16x8_t __p0, uint16x8_t __p1, uint16_t __p2) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 + __rev1 * (uint16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vmlaq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
+  float32x4_t __ret;
+  __ret = __p0 + __p1 * (float32x4_t) {__p2, __p2, __p2, __p2};
+  return __ret;
+}
+#else
+__ai float32x4_t vmlaq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = __rev0 + __rev1 * (float32x4_t) {__p2, __p2, __p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmlaq_n_s32(int32x4_t __p0, int32x4_t __p1, int32_t __p2) {
+  int32x4_t __ret;
+  __ret = __p0 + __p1 * (int32x4_t) {__p2, __p2, __p2, __p2};
+  return __ret;
+}
+#else
+__ai int32x4_t vmlaq_n_s32(int32x4_t __p0, int32x4_t __p1, int32_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 + __rev1 * (int32x4_t) {__p2, __p2, __p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vmlaq_n_s16(int16x8_t __p0, int16x8_t __p1, int16_t __p2) {
+  int16x8_t __ret;
+  __ret = __p0 + __p1 * (int16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
+  return __ret;
+}
+#else
+__ai int16x8_t vmlaq_n_s16(int16x8_t __p0, int16x8_t __p1, int16_t __p2) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 + __rev1 * (int16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vmla_n_u32(uint32x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
+  uint32x2_t __ret;
+  __ret = __p0 + __p1 * (uint32x2_t) {__p2, __p2};
+  return __ret;
+}
+#else
+__ai uint32x2_t vmla_n_u32(uint32x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = __rev0 + __rev1 * (uint32x2_t) {__p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vmla_n_u16(uint16x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
+  uint16x4_t __ret;
+  __ret = __p0 + __p1 * (uint16x4_t) {__p2, __p2, __p2, __p2};
+  return __ret;
+}
+#else
+__ai uint16x4_t vmla_n_u16(uint16x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __rev0 + __rev1 * (uint16x4_t) {__p2, __p2, __p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vmla_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
+  float32x2_t __ret;
+  __ret = __p0 + __p1 * (float32x2_t) {__p2, __p2};
+  return __ret;
+}
+#else
+__ai float32x2_t vmla_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = __rev0 + __rev1 * (float32x2_t) {__p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vmla_n_s32(int32x2_t __p0, int32x2_t __p1, int32_t __p2) {
+  int32x2_t __ret;
+  __ret = __p0 + __p1 * (int32x2_t) {__p2, __p2};
+  return __ret;
+}
+#else
+__ai int32x2_t vmla_n_s32(int32x2_t __p0, int32x2_t __p1, int32_t __p2) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = __rev0 + __rev1 * (int32x2_t) {__p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vmla_n_s16(int16x4_t __p0, int16x4_t __p1, int16_t __p2) {
+  int16x4_t __ret;
+  __ret = __p0 + __p1 * (int16x4_t) {__p2, __p2, __p2, __p2};
+  return __ret;
+}
+#else
+__ai int16x4_t vmla_n_s16(int16x4_t __p0, int16x4_t __p1, int16_t __p2) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __rev0 + __rev1 * (int16x4_t) {__p2, __p2, __p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vmlsq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
+  uint8x16_t __ret;
+  __ret = __p0 - __p1 * __p2;
+  return __ret;
+}
+#else
+__ai uint8x16_t vmlsq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __rev0 - __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmlsq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint32x4_t __ret;
+  __ret = __p0 - __p1 * __p2;
+  return __ret;
+}
+#else
+__ai uint32x4_t vmlsq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 - __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vmlsq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint16x8_t __ret;
+  __ret = __p0 - __p1 * __p2;
+  return __ret;
+}
+#else
+__ai uint16x8_t vmlsq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 - __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vmlsq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
+  int8x16_t __ret;
+  __ret = __p0 - __p1 * __p2;
+  return __ret;
+}
+#else
+__ai int8x16_t vmlsq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __rev0 - __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vmlsq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
+  float32x4_t __ret;
+  __ret = __p0 - __p1 * __p2;
+  return __ret;
+}
+#else
+__ai float32x4_t vmlsq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = __rev0 - __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmlsq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int32x4_t __ret;
+  __ret = __p0 - __p1 * __p2;
+  return __ret;
+}
+#else
+__ai int32x4_t vmlsq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 - __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vmlsq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int16x8_t __ret;
+  __ret = __p0 - __p1 * __p2;
+  return __ret;
+}
+#else
+__ai int16x8_t vmlsq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 - __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vmls_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
+  uint8x8_t __ret;
+  __ret = __p0 - __p1 * __p2;
+  return __ret;
+}
+#else
+__ai uint8x8_t vmls_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __rev0 - __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vmls_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
+  uint32x2_t __ret;
+  __ret = __p0 - __p1 * __p2;
+  return __ret;
+}
+#else
+__ai uint32x2_t vmls_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  uint32x2_t __ret;
+  __ret = __rev0 - __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vmls_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
+  uint16x4_t __ret;
+  __ret = __p0 - __p1 * __p2;
+  return __ret;
+}
+#else
+__ai uint16x4_t vmls_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __rev0 - __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vmls_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
+  int8x8_t __ret;
+  __ret = __p0 - __p1 * __p2;
+  return __ret;
+}
+#else
+__ai int8x8_t vmls_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __rev0 - __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vmls_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
+  float32x2_t __ret;
+  __ret = __p0 - __p1 * __p2;
+  return __ret;
+}
+#else
+__ai float32x2_t vmls_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  float32x2_t __ret;
+  __ret = __rev0 - __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vmls_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int32x2_t __ret;
+  __ret = __p0 - __p1 * __p2;
+  return __ret;
+}
+#else
+__ai int32x2_t vmls_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  int32x2_t __ret;
+  __ret = __rev0 - __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vmls_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int16x4_t __ret;
+  __ret = __p0 - __p1 * __p2;
+  return __ret;
+}
+#else
+__ai int16x4_t vmls_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __rev0 - __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsq_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x2_t __s2 = __p2; \
+  uint32x4_t __ret; \
+  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmlsq_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x2_t __s2 = __p2; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsq_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x4_t __s2 = __p2; \
+  uint16x8_t __ret; \
+  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmlsq_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x4_t __s2 = __p2; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsq_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x2_t __s2 = __p2; \
+  float32x4_t __ret; \
+  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmlsq_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x2_t __s2 = __p2; \
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  float32x4_t __ret; \
+  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsq_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int32x4_t __ret; \
+  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmlsq_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsq_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int16x8_t __ret; \
+  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmlsq_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmls_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __s2 = __p2; \
+  uint32x2_t __ret; \
+  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmls_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __s2 = __p2; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmls_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __s2 = __p2; \
+  uint16x4_t __ret; \
+  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmls_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __s2 = __p2; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmls_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x2_t __s2 = __p2; \
+  float32x2_t __ret; \
+  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmls_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x2_t __s2 = __p2; \
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  float32x2_t __ret; \
+  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmls_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int32x2_t __ret; \
+  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmls_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  int32x2_t __ret; \
+  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmls_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int16x4_t __ret; \
+  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmls_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmlsq_n_u32(uint32x4_t __p0, uint32x4_t __p1, uint32_t __p2) {
+  uint32x4_t __ret;
+  __ret = __p0 - __p1 * (uint32x4_t) {__p2, __p2, __p2, __p2};
+  return __ret;
+}
+#else
+__ai uint32x4_t vmlsq_n_u32(uint32x4_t __p0, uint32x4_t __p1, uint32_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 - __rev1 * (uint32x4_t) {__p2, __p2, __p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vmlsq_n_u16(uint16x8_t __p0, uint16x8_t __p1, uint16_t __p2) {
+  uint16x8_t __ret;
+  __ret = __p0 - __p1 * (uint16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
+  return __ret;
+}
+#else
+__ai uint16x8_t vmlsq_n_u16(uint16x8_t __p0, uint16x8_t __p1, uint16_t __p2) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 - __rev1 * (uint16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vmlsq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
+  float32x4_t __ret;
+  __ret = __p0 - __p1 * (float32x4_t) {__p2, __p2, __p2, __p2};
+  return __ret;
+}
+#else
+__ai float32x4_t vmlsq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = __rev0 - __rev1 * (float32x4_t) {__p2, __p2, __p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmlsq_n_s32(int32x4_t __p0, int32x4_t __p1, int32_t __p2) {
+  int32x4_t __ret;
+  __ret = __p0 - __p1 * (int32x4_t) {__p2, __p2, __p2, __p2};
+  return __ret;
+}
+#else
+__ai int32x4_t vmlsq_n_s32(int32x4_t __p0, int32x4_t __p1, int32_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 - __rev1 * (int32x4_t) {__p2, __p2, __p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vmlsq_n_s16(int16x8_t __p0, int16x8_t __p1, int16_t __p2) {
+  int16x8_t __ret;
+  __ret = __p0 - __p1 * (int16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
+  return __ret;
+}
+#else
+__ai int16x8_t vmlsq_n_s16(int16x8_t __p0, int16x8_t __p1, int16_t __p2) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 - __rev1 * (int16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vmls_n_u32(uint32x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
+  uint32x2_t __ret;
+  __ret = __p0 - __p1 * (uint32x2_t) {__p2, __p2};
+  return __ret;
+}
+#else
+__ai uint32x2_t vmls_n_u32(uint32x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = __rev0 - __rev1 * (uint32x2_t) {__p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vmls_n_u16(uint16x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
+  uint16x4_t __ret;
+  __ret = __p0 - __p1 * (uint16x4_t) {__p2, __p2, __p2, __p2};
+  return __ret;
+}
+#else
+__ai uint16x4_t vmls_n_u16(uint16x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __rev0 - __rev1 * (uint16x4_t) {__p2, __p2, __p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vmls_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
+  float32x2_t __ret;
+  __ret = __p0 - __p1 * (float32x2_t) {__p2, __p2};
+  return __ret;
+}
+#else
+__ai float32x2_t vmls_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = __rev0 - __rev1 * (float32x2_t) {__p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vmls_n_s32(int32x2_t __p0, int32x2_t __p1, int32_t __p2) {
+  int32x2_t __ret;
+  __ret = __p0 - __p1 * (int32x2_t) {__p2, __p2};
+  return __ret;
+}
+#else
+__ai int32x2_t vmls_n_s32(int32x2_t __p0, int32x2_t __p1, int32_t __p2) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = __rev0 - __rev1 * (int32x2_t) {__p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vmls_n_s16(int16x4_t __p0, int16x4_t __p1, int16_t __p2) {
+  int16x4_t __ret;
+  __ret = __p0 - __p1 * (int16x4_t) {__p2, __p2, __p2, __p2};
+  return __ret;
+}
+#else
+__ai int16x4_t vmls_n_s16(int16x4_t __p0, int16x4_t __p1, int16_t __p2) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __rev0 - __rev1 * (int16x4_t) {__p2, __p2, __p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vmov_n_p8(poly8_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai poly8x8_t vmov_n_p8(poly8_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vmov_n_p16(poly16_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t) {__p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai poly16x4_t vmov_n_p16(poly16_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t) {__p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vmovq_n_p8(poly8_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai poly8x16_t vmovq_n_p8(poly8_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vmovq_n_p16(poly16_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai poly16x8_t vmovq_n_p16(poly16_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vmovq_n_u8(uint8_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai uint8x16_t vmovq_n_u8(uint8_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmovq_n_u32(uint32_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) {__p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai uint32x4_t vmovq_n_u32(uint32_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) {__p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vmovq_n_u64(uint64_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) {__p0, __p0};
+  return __ret;
+}
+#else
+__ai uint64x2_t vmovq_n_u64(uint64_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) {__p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vmovq_n_u16(uint16_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai uint16x8_t vmovq_n_u16(uint16_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vmovq_n_s8(int8_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai int8x16_t vmovq_n_s8(int8_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vmovq_n_f32(float32_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) {__p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai float32x4_t vmovq_n_f32(float32_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) {__p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmovq_n_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16x8_t __ret; \
+  __ret = (float16x8_t) {__s0, __s0, __s0, __s0, __s0, __s0, __s0, __s0}; \
+  __ret; \
+})
+#else
+#define vmovq_n_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16x8_t __ret; \
+  __ret = (float16x8_t) {__s0, __s0, __s0, __s0, __s0, __s0, __s0, __s0}; \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmovq_n_s32(int32_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) {__p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai int32x4_t vmovq_n_s32(int32_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) {__p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vmovq_n_s64(int64_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) {__p0, __p0};
+  return __ret;
+}
+#else
+__ai int64x2_t vmovq_n_s64(int64_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) {__p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vmovq_n_s16(int16_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai int16x8_t vmovq_n_s16(int16_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vmov_n_u8(uint8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai uint8x8_t vmov_n_u8(uint8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vmov_n_u32(uint32_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) {__p0, __p0};
+  return __ret;
+}
+#else
+__ai uint32x2_t vmov_n_u32(uint32_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) {__p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vmov_n_u64(uint64_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) {__p0};
+  return __ret;
+}
+#else
+__ai uint64x1_t vmov_n_u64(uint64_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) {__p0};
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vmov_n_u16(uint16_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) {__p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai uint16x4_t vmov_n_u16(uint16_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) {__p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vmov_n_s8(int8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai int8x8_t vmov_n_s8(int8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vmov_n_f32(float32_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) {__p0, __p0};
+  return __ret;
+}
+#else
+__ai float32x2_t vmov_n_f32(float32_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) {__p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmov_n_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16x4_t __ret; \
+  __ret = (float16x4_t) {__s0, __s0, __s0, __s0}; \
+  __ret; \
+})
+#else
+#define vmov_n_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16x4_t __ret; \
+  __ret = (float16x4_t) {__s0, __s0, __s0, __s0}; \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vmov_n_s32(int32_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) {__p0, __p0};
+  return __ret;
+}
+#else
+__ai int32x2_t vmov_n_s32(int32_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) {__p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vmov_n_s64(int64_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) {__p0};
+  return __ret;
+}
+#else
+__ai int64x1_t vmov_n_s64(int64_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) {__p0};
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vmov_n_s16(int16_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) {__p0, __p0, __p0, __p0};
+  return __ret;
+}
+#else
+__ai int16x4_t vmov_n_s16(int16_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) {__p0, __p0, __p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vmovl_u8(uint8x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vmovl_u8(uint8x8_t __p0) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vmovl_v((int8x8_t)__rev0, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint16x8_t __noswap_vmovl_u8(uint8x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 49);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vmovl_u32(uint32x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vmovl_u32(uint32x2_t __p0) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vmovl_v((int8x8_t)__rev0, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai uint64x2_t __noswap_vmovl_u32(uint32x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 51);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmovl_u16(uint16x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vmovl_u16(uint16x4_t __p0) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vmovl_v((int8x8_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint32x4_t __noswap_vmovl_u16(uint16x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 50);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vmovl_s8(int8x8_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vmovl_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vmovl_v((int8x8_t)__rev0, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int16x8_t __noswap_vmovl_s8(int8x8_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 33);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vmovl_s32(int32x2_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vmovl_s32(int32x2_t __p0) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vmovl_v((int8x8_t)__rev0, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int64x2_t __noswap_vmovl_s32(int32x2_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 35);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmovl_s16(int16x4_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vmovl_s16(int16x4_t __p0) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vmovl_v((int8x8_t)__rev0, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int32x4_t __noswap_vmovl_s16(int16x4_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 34);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vmovn_u32(uint32x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vmovn_u32(uint32x4_t __p0) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vmovn_v((int8x16_t)__rev0, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint16x4_t __noswap_vmovn_u32(uint32x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 17);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vmovn_u64(uint64x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vmovn_u64(uint64x2_t __p0) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vmovn_v((int8x16_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai uint32x2_t __noswap_vmovn_u64(uint64x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 18);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vmovn_u16(uint16x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vmovn_u16(uint16x8_t __p0) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vmovn_v((int8x16_t)__rev0, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint8x8_t __noswap_vmovn_u16(uint16x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 16);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vmovn_s32(int32x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vmovn_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vmovn_v((int8x16_t)__rev0, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int16x4_t __noswap_vmovn_s32(int32x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vmovn_s64(int64x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vmovn_s64(int64x2_t __p0) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vmovn_v((int8x16_t)__rev0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int32x2_t __noswap_vmovn_s64(int64x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vmovn_s16(int16x8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vmovn_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vmovn_v((int8x16_t)__rev0, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int8x8_t __noswap_vmovn_s16(int16x8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vmulq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = __p0 * __p1;
+  return __ret;
+}
+#else
+__ai uint8x16_t vmulq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __rev0 * __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmulq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = __p0 * __p1;
+  return __ret;
+}
+#else
+__ai uint32x4_t vmulq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 * __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vmulq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = __p0 * __p1;
+  return __ret;
+}
+#else
+__ai uint16x8_t vmulq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 * __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vmulq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = __p0 * __p1;
+  return __ret;
+}
+#else
+__ai int8x16_t vmulq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __rev0 * __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vmulq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = __p0 * __p1;
+  return __ret;
+}
+#else
+__ai float32x4_t vmulq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = __rev0 * __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmulq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = __p0 * __p1;
+  return __ret;
+}
+#else
+__ai int32x4_t vmulq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 * __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vmulq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = __p0 * __p1;
+  return __ret;
+}
+#else
+__ai int16x8_t vmulq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 * __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vmul_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = __p0 * __p1;
+  return __ret;
+}
+#else
+__ai uint8x8_t vmul_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __rev0 * __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vmul_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = __p0 * __p1;
+  return __ret;
+}
+#else
+__ai uint32x2_t vmul_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = __rev0 * __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vmul_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = __p0 * __p1;
+  return __ret;
+}
+#else
+__ai uint16x4_t vmul_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __rev0 * __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vmul_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = __p0 * __p1;
+  return __ret;
+}
+#else
+__ai int8x8_t vmul_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __rev0 * __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vmul_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = __p0 * __p1;
+  return __ret;
+}
+#else
+__ai float32x2_t vmul_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = __rev0 * __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vmul_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = __p0 * __p1;
+  return __ret;
+}
+#else
+__ai int32x2_t vmul_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = __rev0 * __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vmul_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = __p0 * __p1;
+  return __ret;
+}
+#else
+__ai int16x4_t vmul_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __rev0 * __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vmul_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vmul_v((int8x8_t)__p0, (int8x8_t)__p1, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vmul_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vmul_v((int8x8_t)__rev0, (int8x8_t)__rev1, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vmulq_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vmulq_v((int8x16_t)__p0, (int8x16_t)__p1, 36);
+  return __ret;
+}
+#else
+__ai poly8x16_t vmulq_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vmulq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 36);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulq_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x4_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmulq_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulq_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x8_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmulq_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulq_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x4_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmulq_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float32x4_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmulq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x8_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmulq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmul_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmul_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmul_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmul_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmul_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x2_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmul_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float32x2_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmul_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmul_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x2_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmul_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmul_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmulq_n_u32(uint32x4_t __p0, uint32_t __p1) {
+  uint32x4_t __ret;
+  __ret = __p0 * (uint32x4_t) {__p1, __p1, __p1, __p1};
+  return __ret;
+}
+#else
+__ai uint32x4_t vmulq_n_u32(uint32x4_t __p0, uint32_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 * (uint32x4_t) {__p1, __p1, __p1, __p1};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vmulq_n_u16(uint16x8_t __p0, uint16_t __p1) {
+  uint16x8_t __ret;
+  __ret = __p0 * (uint16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1};
+  return __ret;
+}
+#else
+__ai uint16x8_t vmulq_n_u16(uint16x8_t __p0, uint16_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 * (uint16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1};
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vmulq_n_f32(float32x4_t __p0, float32_t __p1) {
+  float32x4_t __ret;
+  __ret = __p0 * (float32x4_t) {__p1, __p1, __p1, __p1};
+  return __ret;
+}
+#else
+__ai float32x4_t vmulq_n_f32(float32x4_t __p0, float32_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = __rev0 * (float32x4_t) {__p1, __p1, __p1, __p1};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmulq_n_s32(int32x4_t __p0, int32_t __p1) {
+  int32x4_t __ret;
+  __ret = __p0 * (int32x4_t) {__p1, __p1, __p1, __p1};
+  return __ret;
+}
+#else
+__ai int32x4_t vmulq_n_s32(int32x4_t __p0, int32_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 * (int32x4_t) {__p1, __p1, __p1, __p1};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vmulq_n_s16(int16x8_t __p0, int16_t __p1) {
+  int16x8_t __ret;
+  __ret = __p0 * (int16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1};
+  return __ret;
+}
+#else
+__ai int16x8_t vmulq_n_s16(int16x8_t __p0, int16_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 * (int16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1};
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vmul_n_u32(uint32x2_t __p0, uint32_t __p1) {
+  uint32x2_t __ret;
+  __ret = __p0 * (uint32x2_t) {__p1, __p1};
+  return __ret;
+}
+#else
+__ai uint32x2_t vmul_n_u32(uint32x2_t __p0, uint32_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = __rev0 * (uint32x2_t) {__p1, __p1};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vmul_n_u16(uint16x4_t __p0, uint16_t __p1) {
+  uint16x4_t __ret;
+  __ret = __p0 * (uint16x4_t) {__p1, __p1, __p1, __p1};
+  return __ret;
+}
+#else
+__ai uint16x4_t vmul_n_u16(uint16x4_t __p0, uint16_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __rev0 * (uint16x4_t) {__p1, __p1, __p1, __p1};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vmul_n_f32(float32x2_t __p0, float32_t __p1) {
+  float32x2_t __ret;
+  __ret = __p0 * (float32x2_t) {__p1, __p1};
+  return __ret;
+}
+#else
+__ai float32x2_t vmul_n_f32(float32x2_t __p0, float32_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __ret;
+  __ret = __rev0 * (float32x2_t) {__p1, __p1};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vmul_n_s32(int32x2_t __p0, int32_t __p1) {
+  int32x2_t __ret;
+  __ret = __p0 * (int32x2_t) {__p1, __p1};
+  return __ret;
+}
+#else
+__ai int32x2_t vmul_n_s32(int32x2_t __p0, int32_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __ret;
+  __ret = __rev0 * (int32x2_t) {__p1, __p1};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vmul_n_s16(int16x4_t __p0, int16_t __p1) {
+  int16x4_t __ret;
+  __ret = __p0 * (int16x4_t) {__p1, __p1, __p1, __p1};
+  return __ret;
+}
+#else
+__ai int16x4_t vmul_n_s16(int16x4_t __p0, int16_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __rev0 * (int16x4_t) {__p1, __p1, __p1, __p1};
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vmull_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 37);
+  return __ret;
+}
+#else
+__ai poly16x8_t vmull_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __ret;
+  __ret = (poly16x8_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 37);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai poly16x8_t __noswap_vmull_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 37);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vmull_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vmull_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint16x8_t __noswap_vmull_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 49);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vmull_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vmull_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai uint64x2_t __noswap_vmull_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 51);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmull_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vmull_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint32x4_t __noswap_vmull_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 50);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vmull_s8(int8x8_t __p0, int8x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vmull_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int16x8_t __noswap_vmull_s8(int8x8_t __p0, int8x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 33);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vmull_s32(int32x2_t __p0, int32x2_t __p1) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vmull_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int64x2_t __noswap_vmull_s32(int32x2_t __p0, int32x2_t __p1) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 35);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmull_s16(int16x4_t __p0, int16x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vmull_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int32x4_t __noswap_vmull_s16(int16x4_t __p0, int16x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 34);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmull_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint64x2_t __ret; \
+  __ret = vmull_u32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmull_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = __noswap_vmull_u32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmull_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint32x4_t __ret; \
+  __ret = vmull_u16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmull_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = __noswap_vmull_u16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmull_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int64x2_t __ret; \
+  __ret = vmull_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmull_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __noswap_vmull_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmull_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = vmull_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmull_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __noswap_vmull_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vmull_n_u32(uint32x2_t __p0, uint32_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)(uint32x2_t) {__p1, __p1}, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vmull_n_u32(uint32x2_t __p0, uint32_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)(uint32x2_t) {__p1, __p1}, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai uint64x2_t __noswap_vmull_n_u32(uint32x2_t __p0, uint32_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)(uint32x2_t) {__p1, __p1}, 51);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmull_n_u16(uint16x4_t __p0, uint16_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)(uint16x4_t) {__p1, __p1, __p1, __p1}, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vmull_n_u16(uint16x4_t __p0, uint16_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)(uint16x4_t) {__p1, __p1, __p1, __p1}, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint32x4_t __noswap_vmull_n_u16(uint16x4_t __p0, uint16_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)(uint16x4_t) {__p1, __p1, __p1, __p1}, 50);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vmull_n_s32(int32x2_t __p0, int32_t __p1) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)(int32x2_t) {__p1, __p1}, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vmull_n_s32(int32x2_t __p0, int32_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)(int32x2_t) {__p1, __p1}, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int64x2_t __noswap_vmull_n_s32(int32x2_t __p0, int32_t __p1) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)(int32x2_t) {__p1, __p1}, 35);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmull_n_s16(int16x4_t __p0, int16_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)(int16x4_t) {__p1, __p1, __p1, __p1}, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vmull_n_s16(int16x4_t __p0, int16_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)(int16x4_t) {__p1, __p1, __p1, __p1}, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int32x4_t __noswap_vmull_n_s16(int16x4_t __p0, int16_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)(int16x4_t) {__p1, __p1, __p1, __p1}, 34);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vmvn_p8(poly8x8_t __p0) {
+  poly8x8_t __ret;
+  __ret = ~__p0;
+  return __ret;
+}
+#else
+__ai poly8x8_t vmvn_p8(poly8x8_t __p0) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = ~__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vmvnq_p8(poly8x16_t __p0) {
+  poly8x16_t __ret;
+  __ret = ~__p0;
+  return __ret;
+}
+#else
+__ai poly8x16_t vmvnq_p8(poly8x16_t __p0) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = ~__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vmvnq_u8(uint8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = ~__p0;
+  return __ret;
+}
+#else
+__ai uint8x16_t vmvnq_u8(uint8x16_t __p0) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = ~__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmvnq_u32(uint32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = ~__p0;
+  return __ret;
+}
+#else
+__ai uint32x4_t vmvnq_u32(uint32x4_t __p0) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = ~__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vmvnq_u16(uint16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = ~__p0;
+  return __ret;
+}
+#else
+__ai uint16x8_t vmvnq_u16(uint16x8_t __p0) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = ~__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vmvnq_s8(int8x16_t __p0) {
+  int8x16_t __ret;
+  __ret = ~__p0;
+  return __ret;
+}
+#else
+__ai int8x16_t vmvnq_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = ~__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmvnq_s32(int32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = ~__p0;
+  return __ret;
+}
+#else
+__ai int32x4_t vmvnq_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = ~__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vmvnq_s16(int16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = ~__p0;
+  return __ret;
+}
+#else
+__ai int16x8_t vmvnq_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = ~__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vmvn_u8(uint8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = ~__p0;
+  return __ret;
+}
+#else
+__ai uint8x8_t vmvn_u8(uint8x8_t __p0) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = ~__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vmvn_u32(uint32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = ~__p0;
+  return __ret;
+}
+#else
+__ai uint32x2_t vmvn_u32(uint32x2_t __p0) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = ~__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vmvn_u16(uint16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = ~__p0;
+  return __ret;
+}
+#else
+__ai uint16x4_t vmvn_u16(uint16x4_t __p0) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = ~__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vmvn_s8(int8x8_t __p0) {
+  int8x8_t __ret;
+  __ret = ~__p0;
+  return __ret;
+}
+#else
+__ai int8x8_t vmvn_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = ~__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vmvn_s32(int32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = ~__p0;
+  return __ret;
+}
+#else
+__ai int32x2_t vmvn_s32(int32x2_t __p0) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __ret;
+  __ret = ~__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vmvn_s16(int16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = ~__p0;
+  return __ret;
+}
+#else
+__ai int16x4_t vmvn_s16(int16x4_t __p0) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = ~__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vnegq_s8(int8x16_t __p0) {
+  int8x16_t __ret;
+  __ret = -__p0;
+  return __ret;
+}
+#else
+__ai int8x16_t vnegq_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = -__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vnegq_f32(float32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = -__p0;
+  return __ret;
+}
+#else
+__ai float32x4_t vnegq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = -__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vnegq_s32(int32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = -__p0;
+  return __ret;
+}
+#else
+__ai int32x4_t vnegq_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = -__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vnegq_s16(int16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = -__p0;
+  return __ret;
+}
+#else
+__ai int16x8_t vnegq_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = -__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vneg_s8(int8x8_t __p0) {
+  int8x8_t __ret;
+  __ret = -__p0;
+  return __ret;
+}
+#else
+__ai int8x8_t vneg_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = -__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vneg_f32(float32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = -__p0;
+  return __ret;
+}
+#else
+__ai float32x2_t vneg_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __ret;
+  __ret = -__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vneg_s32(int32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = -__p0;
+  return __ret;
+}
+#else
+__ai int32x2_t vneg_s32(int32x2_t __p0) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __ret;
+  __ret = -__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vneg_s16(int16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = -__p0;
+  return __ret;
+}
+#else
+__ai int16x4_t vneg_s16(int16x4_t __p0) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = -__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vornq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = __p0 | ~__p1;
+  return __ret;
+}
+#else
+__ai uint8x16_t vornq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __rev0 | ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vornq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = __p0 | ~__p1;
+  return __ret;
+}
+#else
+__ai uint32x4_t vornq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 | ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vornq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = __p0 | ~__p1;
+  return __ret;
+}
+#else
+__ai uint64x2_t vornq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = __rev0 | ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vornq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = __p0 | ~__p1;
+  return __ret;
+}
+#else
+__ai uint16x8_t vornq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 | ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vornq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = __p0 | ~__p1;
+  return __ret;
+}
+#else
+__ai int8x16_t vornq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __rev0 | ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vornq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = __p0 | ~__p1;
+  return __ret;
+}
+#else
+__ai int32x4_t vornq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 | ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vornq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = __p0 | ~__p1;
+  return __ret;
+}
+#else
+__ai int64x2_t vornq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = __rev0 | ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vornq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = __p0 | ~__p1;
+  return __ret;
+}
+#else
+__ai int16x8_t vornq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 | ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vorn_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = __p0 | ~__p1;
+  return __ret;
+}
+#else
+__ai uint8x8_t vorn_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __rev0 | ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vorn_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = __p0 | ~__p1;
+  return __ret;
+}
+#else
+__ai uint32x2_t vorn_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = __rev0 | ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vorn_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = __p0 | ~__p1;
+  return __ret;
+}
+#else
+__ai uint64x1_t vorn_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = __p0 | ~__p1;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vorn_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = __p0 | ~__p1;
+  return __ret;
+}
+#else
+__ai uint16x4_t vorn_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __rev0 | ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vorn_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = __p0 | ~__p1;
+  return __ret;
+}
+#else
+__ai int8x8_t vorn_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __rev0 | ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vorn_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = __p0 | ~__p1;
+  return __ret;
+}
+#else
+__ai int32x2_t vorn_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = __rev0 | ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vorn_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = __p0 | ~__p1;
+  return __ret;
+}
+#else
+__ai int64x1_t vorn_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = __p0 | ~__p1;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vorn_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = __p0 | ~__p1;
+  return __ret;
+}
+#else
+__ai int16x4_t vorn_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __rev0 | ~__rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vorrq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = __p0 | __p1;
+  return __ret;
+}
+#else
+__ai uint8x16_t vorrq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __rev0 | __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vorrq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = __p0 | __p1;
+  return __ret;
+}
+#else
+__ai uint32x4_t vorrq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 | __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vorrq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = __p0 | __p1;
+  return __ret;
+}
+#else
+__ai uint64x2_t vorrq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = __rev0 | __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vorrq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = __p0 | __p1;
+  return __ret;
+}
+#else
+__ai uint16x8_t vorrq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 | __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vorrq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = __p0 | __p1;
+  return __ret;
+}
+#else
+__ai int8x16_t vorrq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __rev0 | __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vorrq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = __p0 | __p1;
+  return __ret;
+}
+#else
+__ai int32x4_t vorrq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 | __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vorrq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = __p0 | __p1;
+  return __ret;
+}
+#else
+__ai int64x2_t vorrq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = __rev0 | __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vorrq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = __p0 | __p1;
+  return __ret;
+}
+#else
+__ai int16x8_t vorrq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 | __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vorr_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = __p0 | __p1;
+  return __ret;
+}
+#else
+__ai uint8x8_t vorr_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __rev0 | __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vorr_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = __p0 | __p1;
+  return __ret;
+}
+#else
+__ai uint32x2_t vorr_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = __rev0 | __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vorr_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = __p0 | __p1;
+  return __ret;
+}
+#else
+__ai uint64x1_t vorr_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = __p0 | __p1;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vorr_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = __p0 | __p1;
+  return __ret;
+}
+#else
+__ai uint16x4_t vorr_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __rev0 | __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vorr_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = __p0 | __p1;
+  return __ret;
+}
+#else
+__ai int8x8_t vorr_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __rev0 | __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vorr_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = __p0 | __p1;
+  return __ret;
+}
+#else
+__ai int32x2_t vorr_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = __rev0 | __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vorr_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = __p0 | __p1;
+  return __ret;
+}
+#else
+__ai int64x1_t vorr_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = __p0 | __p1;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vorr_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = __p0 | __p1;
+  return __ret;
+}
+#else
+__ai int16x4_t vorr_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __rev0 | __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vpadalq_u8(uint16x8_t __p0, uint8x16_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vpadalq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vpadalq_u8(uint16x8_t __p0, uint8x16_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vpadalq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vpadalq_u32(uint64x2_t __p0, uint32x4_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vpadalq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vpadalq_u32(uint64x2_t __p0, uint32x4_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vpadalq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vpadalq_u16(uint32x4_t __p0, uint16x8_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vpadalq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vpadalq_u16(uint32x4_t __p0, uint16x8_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vpadalq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vpadalq_s8(int16x8_t __p0, int8x16_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vpadalq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vpadalq_s8(int16x8_t __p0, int8x16_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vpadalq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vpadalq_s32(int64x2_t __p0, int32x4_t __p1) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vpadalq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vpadalq_s32(int64x2_t __p0, int32x4_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vpadalq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vpadalq_s16(int32x4_t __p0, int16x8_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vpadalq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vpadalq_s16(int32x4_t __p0, int16x8_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vpadalq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vpadal_u8(uint16x4_t __p0, uint8x8_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vpadal_u8(uint16x4_t __p0, uint8x8_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vpadal_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vpadal_u32(uint64x1_t __p0, uint32x2_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vpadal_u32(uint64x1_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__rev1, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vpadal_u16(uint32x2_t __p0, uint16x4_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vpadal_u16(uint32x2_t __p0, uint16x4_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vpadal_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vpadal_s8(int16x4_t __p0, int8x8_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vpadal_s8(int16x4_t __p0, int8x8_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vpadal_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vpadal_s32(int64x1_t __p0, int32x2_t __p1) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
+  return __ret;
+}
+#else
+__ai int64x1_t vpadal_s32(int64x1_t __p0, int32x2_t __p1) {
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__rev1, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vpadal_s16(int32x2_t __p0, int16x4_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vpadal_s16(int32x2_t __p0, int16x4_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vpadal_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vpadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vpadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vpadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vpadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vpadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vpadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vpadd_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vpadd_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vpadd_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vpadd_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vpadd_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vpadd_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vpadd_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vpadd_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vpaddlq_u8(uint8x16_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vpaddlq_v((int8x16_t)__p0, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vpaddlq_u8(uint8x16_t __p0) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vpaddlq_v((int8x16_t)__rev0, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vpaddlq_u32(uint32x4_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vpaddlq_v((int8x16_t)__p0, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vpaddlq_u32(uint32x4_t __p0) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vpaddlq_v((int8x16_t)__rev0, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vpaddlq_u16(uint16x8_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vpaddlq_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vpaddlq_u16(uint16x8_t __p0) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vpaddlq_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vpaddlq_s8(int8x16_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vpaddlq_v((int8x16_t)__p0, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vpaddlq_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vpaddlq_v((int8x16_t)__rev0, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vpaddlq_s32(int32x4_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vpaddlq_v((int8x16_t)__p0, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vpaddlq_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vpaddlq_v((int8x16_t)__rev0, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vpaddlq_s16(int16x8_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vpaddlq_v((int8x16_t)__p0, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vpaddlq_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vpaddlq_v((int8x16_t)__rev0, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vpaddl_u8(uint8x8_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vpaddl_v((int8x8_t)__p0, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vpaddl_u8(uint8x8_t __p0) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vpaddl_v((int8x8_t)__rev0, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vpaddl_u32(uint32x2_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vpaddl_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vpaddl_u32(uint32x2_t __p0) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vpaddl_v((int8x8_t)__rev0, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vpaddl_u16(uint16x4_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vpaddl_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vpaddl_u16(uint16x4_t __p0) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vpaddl_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vpaddl_s8(int8x8_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vpaddl_v((int8x8_t)__p0, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vpaddl_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vpaddl_v((int8x8_t)__rev0, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vpaddl_s32(int32x2_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vpaddl_v((int8x8_t)__p0, 3);
+  return __ret;
+}
+#else
+__ai int64x1_t vpaddl_s32(int32x2_t __p0) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vpaddl_v((int8x8_t)__rev0, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vpaddl_s16(int16x4_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vpaddl_v((int8x8_t)__p0, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vpaddl_s16(int16x4_t __p0) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vpaddl_v((int8x8_t)__rev0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vpmax_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vpmax_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vpmax_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vpmax_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vpmax_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vpmax_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vpmax_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vpmax_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vpmax_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vpmax_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vpmax_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vpmax_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vpmax_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vpmax_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vpmin_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vpmin_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vpmin_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vpmin_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vpmin_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vpmin_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vpmin_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vpmin_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vpmin_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vpmin_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vpmin_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vpmin_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vpmin_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vpmin_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vqabsq_s8(int8x16_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqabsq_v((int8x16_t)__p0, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vqabsq_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqabsq_v((int8x16_t)__rev0, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqabsq_s32(int32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqabsq_v((int8x16_t)__p0, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vqabsq_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqabsq_v((int8x16_t)__rev0, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vqabsq_s16(int16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqabsq_v((int8x16_t)__p0, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vqabsq_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqabsq_v((int8x16_t)__rev0, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vqabs_s8(int8x8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqabs_v((int8x8_t)__p0, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vqabs_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqabs_v((int8x8_t)__rev0, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vqabs_s32(int32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqabs_v((int8x8_t)__p0, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vqabs_s32(int32x2_t __p0) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqabs_v((int8x8_t)__rev0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vqabs_s16(int16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqabs_v((int8x8_t)__p0, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vqabs_s16(int16x4_t __p0) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqabs_v((int8x8_t)__rev0, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vqaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vqaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vqaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vqaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vqaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vqaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vqaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vqaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vqaddq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vqaddq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqaddq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vqaddq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vqaddq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vqaddq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vqaddq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vqaddq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vqadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vqadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vqadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vqadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vqadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vqadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vqadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vqadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vqadd_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vqadd_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vqadd_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vqadd_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vqadd_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
+  return __ret;
+}
+#else
+__ai int64x1_t vqadd_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vqadd_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vqadd_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vqdmlal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqdmlal_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vqdmlal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqdmlal_v((int8x16_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int64x2_t __noswap_vqdmlal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqdmlal_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 35);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqdmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmlal_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vqdmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmlal_v((int8x16_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int32x4_t __noswap_vqdmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmlal_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 34);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlal_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int64x2_t __ret; \
+  __ret = vqdmlal_s32(__s0, __s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vqdmlal_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __noswap_vqdmlal_s32(__rev0, __rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlal_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int32x4_t __ret; \
+  __ret = vqdmlal_s16(__s0, __s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vqdmlal_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __noswap_vqdmlal_s16(__rev0, __rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vqdmlal_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqdmlal_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)(int32x2_t) {__p2, __p2}, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vqdmlal_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqdmlal_v((int8x16_t)__rev0, (int8x8_t)__rev1, (int8x8_t)(int32x2_t) {__p2, __p2}, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int64x2_t __noswap_vqdmlal_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqdmlal_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)(int32x2_t) {__p2, __p2}, 35);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqdmlal_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmlal_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)(int16x4_t) {__p2, __p2, __p2, __p2}, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vqdmlal_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmlal_v((int8x16_t)__rev0, (int8x8_t)__rev1, (int8x8_t)(int16x4_t) {__p2, __p2, __p2, __p2}, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int32x4_t __noswap_vqdmlal_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmlal_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)(int16x4_t) {__p2, __p2, __p2, __p2}, 34);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vqdmlsl_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqdmlsl_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vqdmlsl_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqdmlsl_v((int8x16_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int64x2_t __noswap_vqdmlsl_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqdmlsl_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 35);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqdmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmlsl_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vqdmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmlsl_v((int8x16_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int32x4_t __noswap_vqdmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmlsl_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 34);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlsl_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int64x2_t __ret; \
+  __ret = vqdmlsl_s32(__s0, __s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vqdmlsl_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __noswap_vqdmlsl_s32(__rev0, __rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlsl_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int32x4_t __ret; \
+  __ret = vqdmlsl_s16(__s0, __s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vqdmlsl_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __noswap_vqdmlsl_s16(__rev0, __rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vqdmlsl_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqdmlsl_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)(int32x2_t) {__p2, __p2}, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vqdmlsl_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqdmlsl_v((int8x16_t)__rev0, (int8x8_t)__rev1, (int8x8_t)(int32x2_t) {__p2, __p2}, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int64x2_t __noswap_vqdmlsl_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqdmlsl_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)(int32x2_t) {__p2, __p2}, 35);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqdmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmlsl_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)(int16x4_t) {__p2, __p2, __p2, __p2}, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vqdmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmlsl_v((int8x16_t)__rev0, (int8x8_t)__rev1, (int8x8_t)(int16x4_t) {__p2, __p2, __p2, __p2}, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int32x4_t __noswap_vqdmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmlsl_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)(int16x4_t) {__p2, __p2, __p2, __p2}, 34);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqdmulhq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vqdmulhq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmulhq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int32x4_t __noswap_vqdmulhq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vqdmulhq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vqdmulhq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqdmulhq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int16x8_t __noswap_vqdmulhq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vqdmulh_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vqdmulh_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqdmulh_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int32x2_t __noswap_vqdmulh_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vqdmulh_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vqdmulh_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqdmulh_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int16x4_t __noswap_vqdmulh_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulhq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = vqdmulhq_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqdmulhq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __noswap_vqdmulhq_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulhq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x8_t __ret; \
+  __ret = vqdmulhq_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqdmulhq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = __noswap_vqdmulhq_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulh_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __ret; \
+  __ret = vqdmulh_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqdmulh_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x2_t __ret; \
+  __ret = __noswap_vqdmulh_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulh_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __ret; \
+  __ret = vqdmulh_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqdmulh_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = __noswap_vqdmulh_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqdmulhq_n_s32(int32x4_t __p0, int32_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmulhq_v((int8x16_t)__p0, (int8x16_t)(int32x4_t) {__p1, __p1, __p1, __p1}, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vqdmulhq_n_s32(int32x4_t __p0, int32_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmulhq_v((int8x16_t)__rev0, (int8x16_t)(int32x4_t) {__p1, __p1, __p1, __p1}, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vqdmulhq_n_s16(int16x8_t __p0, int16_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqdmulhq_v((int8x16_t)__p0, (int8x16_t)(int16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1}, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vqdmulhq_n_s16(int16x8_t __p0, int16_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqdmulhq_v((int8x16_t)__rev0, (int8x16_t)(int16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1}, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vqdmulh_n_s32(int32x2_t __p0, int32_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqdmulh_v((int8x8_t)__p0, (int8x8_t)(int32x2_t) {__p1, __p1}, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vqdmulh_n_s32(int32x2_t __p0, int32_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqdmulh_v((int8x8_t)__rev0, (int8x8_t)(int32x2_t) {__p1, __p1}, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vqdmulh_n_s16(int16x4_t __p0, int16_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqdmulh_v((int8x8_t)__p0, (int8x8_t)(int16x4_t) {__p1, __p1, __p1, __p1}, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vqdmulh_n_s16(int16x4_t __p0, int16_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqdmulh_v((int8x8_t)__rev0, (int8x8_t)(int16x4_t) {__p1, __p1, __p1, __p1}, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vqdmull_s32(int32x2_t __p0, int32x2_t __p1) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqdmull_v((int8x8_t)__p0, (int8x8_t)__p1, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vqdmull_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqdmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int64x2_t __noswap_vqdmull_s32(int32x2_t __p0, int32x2_t __p1) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqdmull_v((int8x8_t)__p0, (int8x8_t)__p1, 35);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqdmull_s16(int16x4_t __p0, int16x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmull_v((int8x8_t)__p0, (int8x8_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vqdmull_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int32x4_t __noswap_vqdmull_s16(int16x4_t __p0, int16x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmull_v((int8x8_t)__p0, (int8x8_t)__p1, 34);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmull_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int64x2_t __ret; \
+  __ret = vqdmull_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqdmull_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __noswap_vqdmull_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmull_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = vqdmull_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqdmull_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __noswap_vqdmull_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vqdmull_n_s32(int32x2_t __p0, int32_t __p1) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqdmull_v((int8x8_t)__p0, (int8x8_t)(int32x2_t) {__p1, __p1}, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vqdmull_n_s32(int32x2_t __p0, int32_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqdmull_v((int8x8_t)__rev0, (int8x8_t)(int32x2_t) {__p1, __p1}, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int64x2_t __noswap_vqdmull_n_s32(int32x2_t __p0, int32_t __p1) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqdmull_v((int8x8_t)__p0, (int8x8_t)(int32x2_t) {__p1, __p1}, 35);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqdmull_n_s16(int16x4_t __p0, int16_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmull_v((int8x8_t)__p0, (int8x8_t)(int16x4_t) {__p1, __p1, __p1, __p1}, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vqdmull_n_s16(int16x4_t __p0, int16_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmull_v((int8x8_t)__rev0, (int8x8_t)(int16x4_t) {__p1, __p1, __p1, __p1}, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int32x4_t __noswap_vqdmull_n_s16(int16x4_t __p0, int16_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqdmull_v((int8x8_t)__p0, (int8x8_t)(int16x4_t) {__p1, __p1, __p1, __p1}, 34);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vqmovn_u32(uint32x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vqmovn_u32(uint32x4_t __p0) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vqmovn_v((int8x16_t)__rev0, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint16x4_t __noswap_vqmovn_u32(uint32x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 17);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vqmovn_u64(uint64x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vqmovn_u64(uint64x2_t __p0) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vqmovn_v((int8x16_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai uint32x2_t __noswap_vqmovn_u64(uint64x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 18);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vqmovn_u16(uint16x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vqmovn_u16(uint16x8_t __p0) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqmovn_v((int8x16_t)__rev0, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint8x8_t __noswap_vqmovn_u16(uint16x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 16);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vqmovn_s32(int32x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vqmovn_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqmovn_v((int8x16_t)__rev0, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int16x4_t __noswap_vqmovn_s32(int32x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vqmovn_s64(int64x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vqmovn_s64(int64x2_t __p0) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqmovn_v((int8x16_t)__rev0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int32x2_t __noswap_vqmovn_s64(int64x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vqmovn_s16(int16x8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vqmovn_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqmovn_v((int8x16_t)__rev0, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int8x8_t __noswap_vqmovn_s16(int16x8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vqmovun_s32(int32x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vqmovun_v((int8x16_t)__p0, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vqmovun_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vqmovun_v((int8x16_t)__rev0, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint16x4_t __noswap_vqmovun_s32(int32x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vqmovun_v((int8x16_t)__p0, 17);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vqmovun_s64(int64x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vqmovun_v((int8x16_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vqmovun_s64(int64x2_t __p0) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vqmovun_v((int8x16_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai uint32x2_t __noswap_vqmovun_s64(int64x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vqmovun_v((int8x16_t)__p0, 18);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vqmovun_s16(int16x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqmovun_v((int8x16_t)__p0, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vqmovun_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqmovun_v((int8x16_t)__rev0, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint8x8_t __noswap_vqmovun_s16(int16x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqmovun_v((int8x16_t)__p0, 16);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vqnegq_s8(int8x16_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqnegq_v((int8x16_t)__p0, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vqnegq_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqnegq_v((int8x16_t)__rev0, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqnegq_s32(int32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqnegq_v((int8x16_t)__p0, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vqnegq_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqnegq_v((int8x16_t)__rev0, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vqnegq_s16(int16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqnegq_v((int8x16_t)__p0, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vqnegq_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqnegq_v((int8x16_t)__rev0, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vqneg_s8(int8x8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqneg_v((int8x8_t)__p0, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vqneg_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqneg_v((int8x8_t)__rev0, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vqneg_s32(int32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqneg_v((int8x8_t)__p0, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vqneg_s32(int32x2_t __p0) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqneg_v((int8x8_t)__rev0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vqneg_s16(int16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqneg_v((int8x8_t)__p0, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vqneg_s16(int16x4_t __p0) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqneg_v((int8x8_t)__rev0, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqrdmulhq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vqrdmulhq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int32x4_t __noswap_vqrdmulhq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vqrdmulhq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vqrdmulhq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int16x8_t __noswap_vqrdmulhq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vqrdmulh_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqrdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vqrdmulh_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqrdmulh_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int32x2_t __noswap_vqrdmulh_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqrdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vqrdmulh_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqrdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vqrdmulh_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqrdmulh_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int16x4_t __noswap_vqrdmulh_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqrdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmulhq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = vqrdmulhq_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqrdmulhq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __noswap_vqrdmulhq_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmulhq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x8_t __ret; \
+  __ret = vqrdmulhq_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqrdmulhq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = __noswap_vqrdmulhq_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmulh_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __ret; \
+  __ret = vqrdmulh_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqrdmulh_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x2_t __ret; \
+  __ret = __noswap_vqrdmulh_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmulh_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __ret; \
+  __ret = vqrdmulh_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqrdmulh_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = __noswap_vqrdmulh_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqrdmulhq_n_s32(int32x4_t __p0, int32_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__p0, (int8x16_t)(int32x4_t) {__p1, __p1, __p1, __p1}, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vqrdmulhq_n_s32(int32x4_t __p0, int32_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__rev0, (int8x16_t)(int32x4_t) {__p1, __p1, __p1, __p1}, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vqrdmulhq_n_s16(int16x8_t __p0, int16_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__p0, (int8x16_t)(int16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1}, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vqrdmulhq_n_s16(int16x8_t __p0, int16_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__rev0, (int8x16_t)(int16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1}, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vqrdmulh_n_s32(int32x2_t __p0, int32_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqrdmulh_v((int8x8_t)__p0, (int8x8_t)(int32x2_t) {__p1, __p1}, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vqrdmulh_n_s32(int32x2_t __p0, int32_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqrdmulh_v((int8x8_t)__rev0, (int8x8_t)(int32x2_t) {__p1, __p1}, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vqrdmulh_n_s16(int16x4_t __p0, int16_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqrdmulh_v((int8x8_t)__p0, (int8x8_t)(int16x4_t) {__p1, __p1, __p1, __p1}, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vqrdmulh_n_s16(int16x4_t __p0, int16_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqrdmulh_v((int8x8_t)__rev0, (int8x8_t)(int16x4_t) {__p1, __p1, __p1, __p1}, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vqrshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vqrshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vqrshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vqrshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vqrshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vqrshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vqrshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vqrshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vqrshlq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vqrshlq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqrshlq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vqrshlq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vqrshlq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vqrshlq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vqrshlq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vqrshlq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vqrshl_u8(uint8x8_t __p0, int8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vqrshl_u8(uint8x8_t __p0, int8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vqrshl_u32(uint32x2_t __p0, int32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vqrshl_u32(uint32x2_t __p0, int32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vqrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vqrshl_u64(uint64x1_t __p0, int64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vqrshl_u64(uint64x1_t __p0, int64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vqrshl_u16(uint16x4_t __p0, int16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vqrshl_u16(uint16x4_t __p0, int16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vqrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vqrshl_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vqrshl_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vqrshl_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vqrshl_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vqrshl_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
+  return __ret;
+}
+#else
+__ai int64x1_t vqrshl_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vqrshl_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vqrshl_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrn_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 17); \
+  __ret; \
+})
+#else
+#define vqrshrn_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__rev0, __p1, 17); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vqrshrn_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 17); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrn_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 18); \
+  __ret; \
+})
+#else
+#define vqrshrn_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__rev0, __p1, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vqrshrn_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 18); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrn_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 16); \
+  __ret; \
+})
+#else
+#define vqrshrn_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__rev0, __p1, 16); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vqrshrn_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 16); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrn_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 1); \
+  __ret; \
+})
+#else
+#define vqrshrn_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__rev0, __p1, 1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vqrshrn_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrn_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 2); \
+  __ret; \
+})
+#else
+#define vqrshrn_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__rev0, __p1, 2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vqrshrn_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrn_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 0); \
+  __ret; \
+})
+#else
+#define vqrshrn_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__rev0, __p1, 0); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vqrshrn_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrun_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__s0, __p1, 17); \
+  __ret; \
+})
+#else
+#define vqrshrun_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__rev0, __p1, 17); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vqrshrun_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__s0, __p1, 17); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrun_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__s0, __p1, 18); \
+  __ret; \
+})
+#else
+#define vqrshrun_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__rev0, __p1, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vqrshrun_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__s0, __p1, 18); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrun_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__s0, __p1, 16); \
+  __ret; \
+})
+#else
+#define vqrshrun_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__rev0, __p1, 16); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vqrshrun_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__s0, __p1, 16); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vqshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vqshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vqshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vqshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vqshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vqshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vqshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vqshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vqshlq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vqshlq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqshlq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vqshlq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vqshlq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vqshlq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vqshlq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vqshlq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vqshl_u8(uint8x8_t __p0, int8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vqshl_u8(uint8x8_t __p0, int8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vqshl_u32(uint32x2_t __p0, int32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vqshl_u32(uint32x2_t __p0, int32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vqshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vqshl_u64(uint64x1_t __p0, int64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vqshl_u64(uint64x1_t __p0, int64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vqshl_u16(uint16x4_t __p0, int16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vqshl_u16(uint16x4_t __p0, int16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vqshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vqshl_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vqshl_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vqshl_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vqshl_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vqshl_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
+  return __ret;
+}
+#else
+__ai int64x1_t vqshl_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vqshl_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vqshl_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshlq_n_u8(__p0, __p1) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 48); \
+  __ret; \
+})
+#else
+#define vqshlq_n_u8(__p0, __p1) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 48); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshlq_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 50); \
+  __ret; \
+})
+#else
+#define vqshlq_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 50); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshlq_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 51); \
+  __ret; \
+})
+#else
+#define vqshlq_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 51); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshlq_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 49); \
+  __ret; \
+})
+#else
+#define vqshlq_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 49); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshlq_n_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 32); \
+  __ret; \
+})
+#else
+#define vqshlq_n_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 32); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshlq_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 34); \
+  __ret; \
+})
+#else
+#define vqshlq_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 34); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshlq_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 35); \
+  __ret; \
+})
+#else
+#define vqshlq_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 35); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshlq_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 33); \
+  __ret; \
+})
+#else
+#define vqshlq_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 33); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshl_n_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 16); \
+  __ret; \
+})
+#else
+#define vqshl_n_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vqshl_n_v((int8x8_t)__rev0, __p1, 16); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshl_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 18); \
+  __ret; \
+})
+#else
+#define vqshl_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vqshl_n_v((int8x8_t)__rev0, __p1, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshl_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 19); \
+  __ret; \
+})
+#else
+#define vqshl_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshl_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 17); \
+  __ret; \
+})
+#else
+#define vqshl_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vqshl_n_v((int8x8_t)__rev0, __p1, 17); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshl_n_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 0); \
+  __ret; \
+})
+#else
+#define vqshl_n_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vqshl_n_v((int8x8_t)__rev0, __p1, 0); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshl_n_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 2); \
+  __ret; \
+})
+#else
+#define vqshl_n_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vqshl_n_v((int8x8_t)__rev0, __p1, 2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshl_n_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 3); \
+  __ret; \
+})
+#else
+#define vqshl_n_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshl_n_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 1); \
+  __ret; \
+})
+#else
+#define vqshl_n_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vqshl_n_v((int8x8_t)__rev0, __p1, 1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshluq_n_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vqshluq_n_v((int8x16_t)__s0, __p1, 48); \
+  __ret; \
+})
+#else
+#define vqshluq_n_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vqshluq_n_v((int8x16_t)__rev0, __p1, 48); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshluq_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vqshluq_n_v((int8x16_t)__s0, __p1, 50); \
+  __ret; \
+})
+#else
+#define vqshluq_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vqshluq_n_v((int8x16_t)__rev0, __p1, 50); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshluq_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vqshluq_n_v((int8x16_t)__s0, __p1, 51); \
+  __ret; \
+})
+#else
+#define vqshluq_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vqshluq_n_v((int8x16_t)__rev0, __p1, 51); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshluq_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vqshluq_n_v((int8x16_t)__s0, __p1, 49); \
+  __ret; \
+})
+#else
+#define vqshluq_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vqshluq_n_v((int8x16_t)__rev0, __p1, 49); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshlu_n_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vqshlu_n_v((int8x8_t)__s0, __p1, 16); \
+  __ret; \
+})
+#else
+#define vqshlu_n_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vqshlu_n_v((int8x8_t)__rev0, __p1, 16); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshlu_n_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vqshlu_n_v((int8x8_t)__s0, __p1, 18); \
+  __ret; \
+})
+#else
+#define vqshlu_n_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vqshlu_n_v((int8x8_t)__rev0, __p1, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshlu_n_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vqshlu_n_v((int8x8_t)__s0, __p1, 19); \
+  __ret; \
+})
+#else
+#define vqshlu_n_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vqshlu_n_v((int8x8_t)__s0, __p1, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshlu_n_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vqshlu_n_v((int8x8_t)__s0, __p1, 17); \
+  __ret; \
+})
+#else
+#define vqshlu_n_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vqshlu_n_v((int8x8_t)__rev0, __p1, 17); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrn_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 17); \
+  __ret; \
+})
+#else
+#define vqshrn_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vqshrn_n_v((int8x16_t)__rev0, __p1, 17); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vqshrn_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 17); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrn_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 18); \
+  __ret; \
+})
+#else
+#define vqshrn_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vqshrn_n_v((int8x16_t)__rev0, __p1, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vqshrn_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 18); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrn_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 16); \
+  __ret; \
+})
+#else
+#define vqshrn_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vqshrn_n_v((int8x16_t)__rev0, __p1, 16); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vqshrn_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 16); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrn_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 1); \
+  __ret; \
+})
+#else
+#define vqshrn_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vqshrn_n_v((int8x16_t)__rev0, __p1, 1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vqshrn_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrn_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 2); \
+  __ret; \
+})
+#else
+#define vqshrn_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vqshrn_n_v((int8x16_t)__rev0, __p1, 2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vqshrn_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrn_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 0); \
+  __ret; \
+})
+#else
+#define vqshrn_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vqshrn_n_v((int8x16_t)__rev0, __p1, 0); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vqshrn_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrun_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vqshrun_n_v((int8x16_t)__s0, __p1, 17); \
+  __ret; \
+})
+#else
+#define vqshrun_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vqshrun_n_v((int8x16_t)__rev0, __p1, 17); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vqshrun_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vqshrun_n_v((int8x16_t)__s0, __p1, 17); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrun_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vqshrun_n_v((int8x16_t)__s0, __p1, 18); \
+  __ret; \
+})
+#else
+#define vqshrun_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vqshrun_n_v((int8x16_t)__rev0, __p1, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vqshrun_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vqshrun_n_v((int8x16_t)__s0, __p1, 18); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrun_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vqshrun_n_v((int8x16_t)__s0, __p1, 16); \
+  __ret; \
+})
+#else
+#define vqshrun_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vqshrun_n_v((int8x16_t)__rev0, __p1, 16); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vqshrun_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vqshrun_n_v((int8x16_t)__s0, __p1, 16); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vqsubq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vqsubq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vqsubq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vqsubq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vqsubq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vqsubq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vqsubq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vqsubq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vqsubq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vqsubq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqsubq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vqsubq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vqsubq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vqsubq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vqsubq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vqsubq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vqsub_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vqsub_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vqsub_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vqsub_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vqsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vqsub_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vqsub_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vqsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vqsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vqsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vqsub_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vqsub_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vqsub_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vqsub_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vqsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vqsub_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
+  return __ret;
+}
+#else
+__ai int64x1_t vqsub_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vqsub_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vqsub_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vqsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vraddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vraddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vraddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint16x4_t __noswap_vraddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vraddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vraddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vraddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai uint32x2_t __noswap_vraddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vraddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vraddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vraddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint8x8_t __noswap_vraddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vraddhn_s32(int32x4_t __p0, int32x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vraddhn_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vraddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int16x4_t __noswap_vraddhn_s32(int32x4_t __p0, int32x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vraddhn_s64(int64x2_t __p0, int64x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vraddhn_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vraddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int32x2_t __noswap_vraddhn_s64(int64x2_t __p0, int64x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vraddhn_s16(int16x8_t __p0, int16x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vraddhn_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vraddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int8x8_t __noswap_vraddhn_s16(int16x8_t __p0, int16x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vrecpeq_u32(uint32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vrecpeq_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vrecpeq_u32(uint32x4_t __p0) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vrecpeq_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vrecpeq_f32(float32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrecpeq_v((int8x16_t)__p0, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vrecpeq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrecpeq_v((int8x16_t)__rev0, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vrecpe_u32(uint32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vrecpe_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vrecpe_u32(uint32x2_t __p0) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vrecpe_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vrecpe_f32(float32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrecpe_v((int8x8_t)__p0, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vrecpe_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrecpe_v((int8x8_t)__rev0, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vrecpsq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrecpsq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vrecpsq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrecpsq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vrecps_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrecps_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vrecps_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrecps_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vrev16_p8(poly8x8_t __p0) {
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6);
+  return __ret;
+}
+#else
+__ai poly8x8_t vrev16_p8(poly8x8_t __p0) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vrev16q_p8(poly8x16_t __p0) {
+  poly8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14);
+  return __ret;
+}
+#else
+__ai poly8x16_t vrev16q_p8(poly8x16_t __p0) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vrev16q_u8(uint8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14);
+  return __ret;
+}
+#else
+__ai uint8x16_t vrev16q_u8(uint8x16_t __p0) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vrev16q_s8(int8x16_t __p0) {
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14);
+  return __ret;
+}
+#else
+__ai int8x16_t vrev16q_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vrev16_u8(uint8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6);
+  return __ret;
+}
+#else
+__ai uint8x8_t vrev16_u8(uint8x8_t __p0) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vrev16_s8(int8x8_t __p0) {
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6);
+  return __ret;
+}
+#else
+__ai int8x8_t vrev16_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vrev32_p8(poly8x8_t __p0) {
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vrev32_p8(poly8x8_t __p0) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vrev32_p16(poly16x4_t __p0) {
+  poly16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2);
+  return __ret;
+}
+#else
+__ai poly16x4_t vrev32_p16(poly16x4_t __p0) {
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  poly16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vrev32q_p8(poly8x16_t __p0) {
+  poly8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12);
+  return __ret;
+}
+#else
+__ai poly8x16_t vrev32q_p8(poly8x16_t __p0) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vrev32q_p16(poly16x8_t __p0) {
+  poly16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6);
+  return __ret;
+}
+#else
+__ai poly16x8_t vrev32q_p16(poly16x8_t __p0) {
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vrev32q_u8(uint8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12);
+  return __ret;
+}
+#else
+__ai uint8x16_t vrev32q_u8(uint8x16_t __p0) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vrev32q_u16(uint16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6);
+  return __ret;
+}
+#else
+__ai uint16x8_t vrev32q_u16(uint16x8_t __p0) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vrev32q_s8(int8x16_t __p0) {
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12);
+  return __ret;
+}
+#else
+__ai int8x16_t vrev32q_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vrev32q_s16(int16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6);
+  return __ret;
+}
+#else
+__ai int16x8_t vrev32q_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vrev32_u8(uint8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4);
+  return __ret;
+}
+#else
+__ai uint8x8_t vrev32_u8(uint8x8_t __p0) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vrev32_u16(uint16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2);
+  return __ret;
+}
+#else
+__ai uint16x4_t vrev32_u16(uint16x4_t __p0) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vrev32_s8(int8x8_t __p0) {
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4);
+  return __ret;
+}
+#else
+__ai int8x8_t vrev32_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vrev32_s16(int16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2);
+  return __ret;
+}
+#else
+__ai int16x4_t vrev32_s16(int16x4_t __p0) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vrev64_p8(poly8x8_t __p0) {
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vrev64_p8(poly8x8_t __p0) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vrev64_p16(poly16x4_t __p0) {
+  poly16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vrev64_p16(poly16x4_t __p0) {
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  poly16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vrev64q_p8(poly8x16_t __p0) {
+  poly8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8);
+  return __ret;
+}
+#else
+__ai poly8x16_t vrev64q_p8(poly8x16_t __p0) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vrev64q_p16(poly16x8_t __p0) {
+  poly16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4);
+  return __ret;
+}
+#else
+__ai poly16x8_t vrev64q_p16(poly16x8_t __p0) {
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vrev64q_u8(uint8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8);
+  return __ret;
+}
+#else
+__ai uint8x16_t vrev64q_u8(uint8x16_t __p0) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vrev64q_u32(uint32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2);
+  return __ret;
+}
+#else
+__ai uint32x4_t vrev64q_u32(uint32x4_t __p0) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vrev64q_u16(uint16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4);
+  return __ret;
+}
+#else
+__ai uint16x8_t vrev64q_u16(uint16x8_t __p0) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vrev64q_s8(int8x16_t __p0) {
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8);
+  return __ret;
+}
+#else
+__ai int8x16_t vrev64q_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vrev64q_f32(float32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2);
+  return __ret;
+}
+#else
+__ai float32x4_t vrev64q_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vrev64q_s32(int32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2);
+  return __ret;
+}
+#else
+__ai int32x4_t vrev64q_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vrev64q_s16(int16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4);
+  return __ret;
+}
+#else
+__ai int16x8_t vrev64q_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vrev64_u8(uint8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vrev64_u8(uint8x8_t __p0) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vrev64_u32(uint32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1, 0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vrev64_u32(uint32x2_t __p0) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vrev64_u16(uint16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vrev64_u16(uint16x4_t __p0) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vrev64_s8(int8x8_t __p0) {
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vrev64_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vrev64_f32(float32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1, 0);
+  return __ret;
+}
+#else
+__ai float32x2_t vrev64_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vrev64_s32(int32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1, 0);
+  return __ret;
+}
+#else
+__ai int32x2_t vrev64_s32(int32x2_t __p0) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vrev64_s16(int16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  return __ret;
+}
+#else
+__ai int16x4_t vrev64_s16(int16x4_t __p0) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vrhaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vrhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vrhaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vrhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vrhaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vrhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vrhaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vrhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vrhaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vrhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vrhaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vrhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vrhaddq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vrhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vrhaddq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vrhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vrhaddq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vrhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vrhaddq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vrhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vrhaddq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vrhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vrhaddq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vrhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vrhadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vrhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vrhadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vrhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vrhadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vrhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vrhadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vrhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vrhadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vrhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vrhadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vrhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vrhadd_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vrhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vrhadd_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vrhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vrhadd_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vrhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vrhadd_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vrhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vrhadd_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vrhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vrhadd_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vrhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vrshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vrshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vrshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vrshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vrshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vrshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vrshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vrshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vrshlq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vrshlq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vrshlq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vrshlq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vrshlq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vrshlq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vrshlq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vrshlq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vrshl_u8(uint8x8_t __p0, int8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vrshl_u8(uint8x8_t __p0, int8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vrshl_u32(uint32x2_t __p0, int32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vrshl_u32(uint32x2_t __p0, int32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vrshl_u64(uint64x1_t __p0, int64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vrshl_u64(uint64x1_t __p0, int64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vrshl_u16(uint16x4_t __p0, int16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vrshl_u16(uint16x4_t __p0, int16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vrshl_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vrshl_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vrshl_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vrshl_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vrshl_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
+  return __ret;
+}
+#else
+__ai int64x1_t vrshl_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vrshl_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vrshl_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrq_n_u8(__p0, __p1) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 48); \
+  __ret; \
+})
+#else
+#define vrshrq_n_u8(__p0, __p1) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 48); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrq_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 50); \
+  __ret; \
+})
+#else
+#define vrshrq_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 50); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrq_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 51); \
+  __ret; \
+})
+#else
+#define vrshrq_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 51); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrq_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 49); \
+  __ret; \
+})
+#else
+#define vrshrq_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 49); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrq_n_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 32); \
+  __ret; \
+})
+#else
+#define vrshrq_n_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 32); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrq_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 34); \
+  __ret; \
+})
+#else
+#define vrshrq_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 34); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrq_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 35); \
+  __ret; \
+})
+#else
+#define vrshrq_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 35); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrq_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 33); \
+  __ret; \
+})
+#else
+#define vrshrq_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 33); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshr_n_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 16); \
+  __ret; \
+})
+#else
+#define vrshr_n_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vrshr_n_v((int8x8_t)__rev0, __p1, 16); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshr_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 18); \
+  __ret; \
+})
+#else
+#define vrshr_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vrshr_n_v((int8x8_t)__rev0, __p1, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshr_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 19); \
+  __ret; \
+})
+#else
+#define vrshr_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshr_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 17); \
+  __ret; \
+})
+#else
+#define vrshr_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vrshr_n_v((int8x8_t)__rev0, __p1, 17); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshr_n_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 0); \
+  __ret; \
+})
+#else
+#define vrshr_n_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vrshr_n_v((int8x8_t)__rev0, __p1, 0); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshr_n_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 2); \
+  __ret; \
+})
+#else
+#define vrshr_n_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vrshr_n_v((int8x8_t)__rev0, __p1, 2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshr_n_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 3); \
+  __ret; \
+})
+#else
+#define vrshr_n_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshr_n_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 1); \
+  __ret; \
+})
+#else
+#define vrshr_n_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vrshr_n_v((int8x8_t)__rev0, __p1, 1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrn_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 17); \
+  __ret; \
+})
+#else
+#define vrshrn_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vrshrn_n_v((int8x16_t)__rev0, __p1, 17); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vrshrn_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 17); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrn_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 18); \
+  __ret; \
+})
+#else
+#define vrshrn_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vrshrn_n_v((int8x16_t)__rev0, __p1, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vrshrn_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 18); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrn_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 16); \
+  __ret; \
+})
+#else
+#define vrshrn_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vrshrn_n_v((int8x16_t)__rev0, __p1, 16); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vrshrn_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 16); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrn_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 1); \
+  __ret; \
+})
+#else
+#define vrshrn_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vrshrn_n_v((int8x16_t)__rev0, __p1, 1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vrshrn_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrn_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 2); \
+  __ret; \
+})
+#else
+#define vrshrn_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vrshrn_n_v((int8x16_t)__rev0, __p1, 2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vrshrn_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrn_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 0); \
+  __ret; \
+})
+#else
+#define vrshrn_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vrshrn_n_v((int8x16_t)__rev0, __p1, 0); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vrshrn_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vrsqrteq_u32(uint32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vrsqrteq_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vrsqrteq_u32(uint32x4_t __p0) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vrsqrteq_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vrsqrteq_f32(float32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrsqrteq_v((int8x16_t)__p0, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vrsqrteq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrsqrteq_v((int8x16_t)__rev0, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vrsqrte_u32(uint32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vrsqrte_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vrsqrte_u32(uint32x2_t __p0) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vrsqrte_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vrsqrte_f32(float32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrsqrte_v((int8x8_t)__p0, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vrsqrte_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrsqrte_v((int8x8_t)__rev0, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vrsqrtsq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrsqrtsq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vrsqrtsq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrsqrtsq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vrsqrts_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrsqrts_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vrsqrts_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrsqrts_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrsraq_n_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __s1 = __p1; \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 48); \
+  __ret; \
+})
+#else
+#define vrsraq_n_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __s1 = __p1; \
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 48); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrsraq_n_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 50); \
+  __ret; \
+})
+#else
+#define vrsraq_n_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 50); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrsraq_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __s1 = __p1; \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 51); \
+  __ret; \
+})
+#else
+#define vrsraq_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __s1 = __p1; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 51); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrsraq_n_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 49); \
+  __ret; \
+})
+#else
+#define vrsraq_n_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 49); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrsraq_n_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __s1 = __p1; \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 32); \
+  __ret; \
+})
+#else
+#define vrsraq_n_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __s1 = __p1; \
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 32); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrsraq_n_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 34); \
+  __ret; \
+})
+#else
+#define vrsraq_n_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 34); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrsraq_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __s1 = __p1; \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 35); \
+  __ret; \
+})
+#else
+#define vrsraq_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __s1 = __p1; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 35); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrsraq_n_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 33); \
+  __ret; \
+})
+#else
+#define vrsraq_n_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 33); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrsra_n_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __s1 = __p1; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 16); \
+  __ret; \
+})
+#else
+#define vrsra_n_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __s1 = __p1; \
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vrsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 16); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrsra_n_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 18); \
+  __ret; \
+})
+#else
+#define vrsra_n_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vrsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrsra_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x1_t __s1 = __p1; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
+  __ret; \
+})
+#else
+#define vrsra_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x1_t __s1 = __p1; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrsra_n_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 17); \
+  __ret; \
+})
+#else
+#define vrsra_n_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vrsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 17); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrsra_n_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __s1 = __p1; \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 0); \
+  __ret; \
+})
+#else
+#define vrsra_n_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __s1 = __p1; \
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vrsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 0); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrsra_n_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 2); \
+  __ret; \
+})
+#else
+#define vrsra_n_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vrsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrsra_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x1_t __s1 = __p1; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
+  __ret; \
+})
+#else
+#define vrsra_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x1_t __s1 = __p1; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrsra_n_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 1); \
+  __ret; \
+})
+#else
+#define vrsra_n_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vrsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vrsubhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vrsubhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vrsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint16x4_t __noswap_vrsubhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vrsubhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vrsubhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vrsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai uint32x2_t __noswap_vrsubhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vrsubhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vrsubhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vrsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint8x8_t __noswap_vrsubhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vrsubhn_s32(int32x4_t __p0, int32x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vrsubhn_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vrsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int16x4_t __noswap_vrsubhn_s32(int32x4_t __p0, int32x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vrsubhn_s64(int64x2_t __p0, int64x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vrsubhn_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vrsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int32x2_t __noswap_vrsubhn_s64(int64x2_t __p0, int64x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vrsubhn_s16(int16x8_t __p0, int16x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vrsubhn_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vrsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int8x8_t __noswap_vrsubhn_s16(int16x8_t __p0, int16x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vset_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8_t __s0 = __p0; \
+  poly8x8_t __s1 = __p1; \
+  poly8x8_t __ret; \
+  __ret = (poly8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vset_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8_t __s0 = __p0; \
+  poly8x8_t __s1 = __p1; \
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x8_t __ret; \
+  __ret = (poly8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vset_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8_t __s0 = __p0; \
+  poly8x8_t __s1 = __p1; \
+  poly8x8_t __ret; \
+  __ret = (poly8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vset_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16_t __s0 = __p0; \
+  poly16x4_t __s1 = __p1; \
+  poly16x4_t __ret; \
+  __ret = (poly16x4_t) __builtin_neon_vset_lane_i16(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vset_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16_t __s0 = __p0; \
+  poly16x4_t __s1 = __p1; \
+  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  poly16x4_t __ret; \
+  __ret = (poly16x4_t) __builtin_neon_vset_lane_i16(__s0, (int8x8_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vset_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16_t __s0 = __p0; \
+  poly16x4_t __s1 = __p1; \
+  poly16x4_t __ret; \
+  __ret = (poly16x4_t) __builtin_neon_vset_lane_i16(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsetq_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8_t __s0 = __p0; \
+  poly8x16_t __s1 = __p1; \
+  poly8x16_t __ret; \
+  __ret = (poly8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vsetq_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8_t __s0 = __p0; \
+  poly8x16_t __s1 = __p1; \
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16_t __ret; \
+  __ret = (poly8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vsetq_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8_t __s0 = __p0; \
+  poly8x16_t __s1 = __p1; \
+  poly8x16_t __ret; \
+  __ret = (poly8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsetq_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16_t __s0 = __p0; \
+  poly16x8_t __s1 = __p1; \
+  poly16x8_t __ret; \
+  __ret = (poly16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vsetq_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16_t __s0 = __p0; \
+  poly16x8_t __s1 = __p1; \
+  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x8_t __ret; \
+  __ret = (poly16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int8x16_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vsetq_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16_t __s0 = __p0; \
+  poly16x8_t __s1 = __p1; \
+  poly16x8_t __ret; \
+  __ret = (poly16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsetq_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8_t __s0 = __p0; \
+  uint8x16_t __s1 = __p1; \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vsetq_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8_t __s0 = __p0; \
+  uint8x16_t __s1 = __p1; \
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vsetq_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8_t __s0 = __p0; \
+  uint8x16_t __s1 = __p1; \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsetq_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vsetq_lane_i32(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vsetq_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vsetq_lane_i32(__s0, (int8x16_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vsetq_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vsetq_lane_i32(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsetq_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64x2_t __s1 = __p1; \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vsetq_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64x2_t __s1 = __p1; \
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int8x16_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vsetq_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64x2_t __s1 = __p1; \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsetq_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vsetq_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int8x16_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vsetq_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsetq_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8_t __s0 = __p0; \
+  int8x16_t __s1 = __p1; \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vsetq_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8_t __s0 = __p0; \
+  int8x16_t __s1 = __p1; \
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vsetq_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8_t __s0 = __p0; \
+  int8x16_t __s1 = __p1; \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsetq_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vsetq_lane_f32(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vsetq_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vsetq_lane_f32(__s0, (int8x16_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vsetq_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vsetq_lane_f32(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsetq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vsetq_lane_i32(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vsetq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vsetq_lane_i32(__s0, (int8x16_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vsetq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vsetq_lane_i32(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsetq_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64x2_t __s1 = __p1; \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vsetq_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64x2_t __s1 = __p1; \
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int8x16_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vsetq_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64x2_t __s1 = __p1; \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsetq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vsetq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int8x16_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vsetq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vset_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8_t __s0 = __p0; \
+  uint8x8_t __s1 = __p1; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vset_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8_t __s0 = __p0; \
+  uint8x8_t __s1 = __p1; \
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vset_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8_t __s0 = __p0; \
+  uint8x8_t __s1 = __p1; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vset_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vset_lane_i32(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vset_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vset_lane_i32(__s0, (int8x8_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vset_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vset_lane_i32(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vset_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64x1_t __s1 = __p1; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vset_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64x1_t __s1 = __p1; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#define __noswap_vset_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64x1_t __s1 = __p1; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vset_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vset_lane_i16(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vset_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vset_lane_i16(__s0, (int8x8_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vset_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vset_lane_i16(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vset_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8_t __s0 = __p0; \
+  int8x8_t __s1 = __p1; \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vset_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8_t __s0 = __p0; \
+  int8x8_t __s1 = __p1; \
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vset_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8_t __s0 = __p0; \
+  int8x8_t __s1 = __p1; \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vset_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vset_lane_f32(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vset_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vset_lane_f32(__s0, (int8x8_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vset_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vset_lane_f32(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vset_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vset_lane_i32(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vset_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vset_lane_i32(__s0, (int8x8_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vset_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vset_lane_i32(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vset_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64x1_t __s1 = __p1; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vset_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64x1_t __s1 = __p1; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#define __noswap_vset_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64x1_t __s1 = __p1; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vset_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vset_lane_i16(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vset_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vset_lane_i16(__s0, (int8x8_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vset_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vset_lane_i16(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vshlq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vshlq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vshlq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vshlq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vshlq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vshlq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vshlq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vshlq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vshl_u8(uint8x8_t __p0, int8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vshl_u8(uint8x8_t __p0, int8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vshl_u32(uint32x2_t __p0, int32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vshl_u32(uint32x2_t __p0, int32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vshl_u64(uint64x1_t __p0, int64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vshl_u64(uint64x1_t __p0, int64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vshl_u16(uint16x4_t __p0, int16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vshl_u16(uint16x4_t __p0, int16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vshl_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vshl_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vshl_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vshl_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vshl_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
+  return __ret;
+}
+#else
+__ai int64x1_t vshl_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vshl_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vshl_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshlq_n_u8(__p0, __p1) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 48); \
+  __ret; \
+})
+#else
+#define vshlq_n_u8(__p0, __p1) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 48); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshlq_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 50); \
+  __ret; \
+})
+#else
+#define vshlq_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 50); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshlq_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 51); \
+  __ret; \
+})
+#else
+#define vshlq_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 51); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshlq_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 49); \
+  __ret; \
+})
+#else
+#define vshlq_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 49); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshlq_n_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 32); \
+  __ret; \
+})
+#else
+#define vshlq_n_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 32); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshlq_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 34); \
+  __ret; \
+})
+#else
+#define vshlq_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 34); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshlq_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 35); \
+  __ret; \
+})
+#else
+#define vshlq_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 35); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshlq_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 33); \
+  __ret; \
+})
+#else
+#define vshlq_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 33); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshl_n_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 16); \
+  __ret; \
+})
+#else
+#define vshl_n_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vshl_n_v((int8x8_t)__rev0, __p1, 16); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshl_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 18); \
+  __ret; \
+})
+#else
+#define vshl_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vshl_n_v((int8x8_t)__rev0, __p1, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshl_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 19); \
+  __ret; \
+})
+#else
+#define vshl_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshl_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 17); \
+  __ret; \
+})
+#else
+#define vshl_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vshl_n_v((int8x8_t)__rev0, __p1, 17); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshl_n_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 0); \
+  __ret; \
+})
+#else
+#define vshl_n_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vshl_n_v((int8x8_t)__rev0, __p1, 0); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshl_n_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 2); \
+  __ret; \
+})
+#else
+#define vshl_n_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vshl_n_v((int8x8_t)__rev0, __p1, 2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshl_n_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 3); \
+  __ret; \
+})
+#else
+#define vshl_n_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshl_n_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 1); \
+  __ret; \
+})
+#else
+#define vshl_n_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vshl_n_v((int8x8_t)__rev0, __p1, 1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshll_n_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 49); \
+  __ret; \
+})
+#else
+#define vshll_n_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vshll_n_v((int8x8_t)__rev0, __p1, 49); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vshll_n_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 49); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshll_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 51); \
+  __ret; \
+})
+#else
+#define vshll_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vshll_n_v((int8x8_t)__rev0, __p1, 51); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vshll_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 51); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshll_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 50); \
+  __ret; \
+})
+#else
+#define vshll_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vshll_n_v((int8x8_t)__rev0, __p1, 50); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vshll_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 50); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshll_n_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 33); \
+  __ret; \
+})
+#else
+#define vshll_n_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vshll_n_v((int8x8_t)__rev0, __p1, 33); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vshll_n_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 33); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshll_n_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 35); \
+  __ret; \
+})
+#else
+#define vshll_n_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vshll_n_v((int8x8_t)__rev0, __p1, 35); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vshll_n_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 35); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshll_n_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 34); \
+  __ret; \
+})
+#else
+#define vshll_n_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vshll_n_v((int8x8_t)__rev0, __p1, 34); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vshll_n_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 34); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrq_n_u8(__p0, __p1) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 48); \
+  __ret; \
+})
+#else
+#define vshrq_n_u8(__p0, __p1) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 48); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrq_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 50); \
+  __ret; \
+})
+#else
+#define vshrq_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 50); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrq_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 51); \
+  __ret; \
+})
+#else
+#define vshrq_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 51); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrq_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 49); \
+  __ret; \
+})
+#else
+#define vshrq_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 49); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrq_n_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 32); \
+  __ret; \
+})
+#else
+#define vshrq_n_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 32); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrq_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 34); \
+  __ret; \
+})
+#else
+#define vshrq_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 34); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrq_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 35); \
+  __ret; \
+})
+#else
+#define vshrq_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 35); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrq_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 33); \
+  __ret; \
+})
+#else
+#define vshrq_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 33); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshr_n_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 16); \
+  __ret; \
+})
+#else
+#define vshr_n_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vshr_n_v((int8x8_t)__rev0, __p1, 16); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshr_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 18); \
+  __ret; \
+})
+#else
+#define vshr_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vshr_n_v((int8x8_t)__rev0, __p1, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshr_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 19); \
+  __ret; \
+})
+#else
+#define vshr_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshr_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 17); \
+  __ret; \
+})
+#else
+#define vshr_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vshr_n_v((int8x8_t)__rev0, __p1, 17); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshr_n_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 0); \
+  __ret; \
+})
+#else
+#define vshr_n_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vshr_n_v((int8x8_t)__rev0, __p1, 0); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshr_n_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 2); \
+  __ret; \
+})
+#else
+#define vshr_n_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vshr_n_v((int8x8_t)__rev0, __p1, 2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshr_n_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 3); \
+  __ret; \
+})
+#else
+#define vshr_n_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshr_n_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 1); \
+  __ret; \
+})
+#else
+#define vshr_n_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vshr_n_v((int8x8_t)__rev0, __p1, 1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrn_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 17); \
+  __ret; \
+})
+#else
+#define vshrn_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vshrn_n_v((int8x16_t)__rev0, __p1, 17); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vshrn_n_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 17); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrn_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 18); \
+  __ret; \
+})
+#else
+#define vshrn_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vshrn_n_v((int8x16_t)__rev0, __p1, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vshrn_n_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 18); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrn_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 16); \
+  __ret; \
+})
+#else
+#define vshrn_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vshrn_n_v((int8x16_t)__rev0, __p1, 16); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vshrn_n_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 16); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrn_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 1); \
+  __ret; \
+})
+#else
+#define vshrn_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vshrn_n_v((int8x16_t)__rev0, __p1, 1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vshrn_n_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrn_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 2); \
+  __ret; \
+})
+#else
+#define vshrn_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vshrn_n_v((int8x16_t)__rev0, __p1, 2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vshrn_n_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrn_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 0); \
+  __ret; \
+})
+#else
+#define vshrn_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vshrn_n_v((int8x16_t)__rev0, __p1, 0); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vshrn_n_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsli_n_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8_t __s0 = __p0; \
+  poly8x8_t __s1 = __p1; \
+  poly8x8_t __ret; \
+  __ret = (poly8x8_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 4); \
+  __ret; \
+})
+#else
+#define vsli_n_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8_t __s0 = __p0; \
+  poly8x8_t __s1 = __p1; \
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x8_t __ret; \
+  __ret = (poly8x8_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 4); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsli_n_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4_t __s0 = __p0; \
+  poly16x4_t __s1 = __p1; \
+  poly16x4_t __ret; \
+  __ret = (poly16x4_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 5); \
+  __ret; \
+})
+#else
+#define vsli_n_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4_t __s0 = __p0; \
+  poly16x4_t __s1 = __p1; \
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  poly16x4_t __ret; \
+  __ret = (poly16x4_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 5); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsliq_n_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16_t __s0 = __p0; \
+  poly8x16_t __s1 = __p1; \
+  poly8x16_t __ret; \
+  __ret = (poly8x16_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 36); \
+  __ret; \
+})
+#else
+#define vsliq_n_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16_t __s0 = __p0; \
+  poly8x16_t __s1 = __p1; \
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16_t __ret; \
+  __ret = (poly8x16_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 36); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsliq_n_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8_t __s0 = __p0; \
+  poly16x8_t __s1 = __p1; \
+  poly16x8_t __ret; \
+  __ret = (poly16x8_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 37); \
+  __ret; \
+})
+#else
+#define vsliq_n_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8_t __s0 = __p0; \
+  poly16x8_t __s1 = __p1; \
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x8_t __ret; \
+  __ret = (poly16x8_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 37); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsliq_n_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __s1 = __p1; \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 48); \
+  __ret; \
+})
+#else
+#define vsliq_n_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __s1 = __p1; \
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 48); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsliq_n_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 50); \
+  __ret; \
+})
+#else
+#define vsliq_n_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 50); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsliq_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __s1 = __p1; \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 51); \
+  __ret; \
+})
+#else
+#define vsliq_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __s1 = __p1; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 51); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsliq_n_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 49); \
+  __ret; \
+})
+#else
+#define vsliq_n_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 49); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsliq_n_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __s1 = __p1; \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 32); \
+  __ret; \
+})
+#else
+#define vsliq_n_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __s1 = __p1; \
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 32); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsliq_n_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 34); \
+  __ret; \
+})
+#else
+#define vsliq_n_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 34); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsliq_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __s1 = __p1; \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 35); \
+  __ret; \
+})
+#else
+#define vsliq_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __s1 = __p1; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 35); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsliq_n_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 33); \
+  __ret; \
+})
+#else
+#define vsliq_n_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 33); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsli_n_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __s1 = __p1; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 16); \
+  __ret; \
+})
+#else
+#define vsli_n_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __s1 = __p1; \
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 16); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsli_n_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 18); \
+  __ret; \
+})
+#else
+#define vsli_n_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsli_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x1_t __s1 = __p1; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
+  __ret; \
+})
+#else
+#define vsli_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x1_t __s1 = __p1; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsli_n_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 17); \
+  __ret; \
+})
+#else
+#define vsli_n_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 17); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsli_n_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __s1 = __p1; \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 0); \
+  __ret; \
+})
+#else
+#define vsli_n_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __s1 = __p1; \
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 0); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsli_n_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 2); \
+  __ret; \
+})
+#else
+#define vsli_n_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsli_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x1_t __s1 = __p1; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
+  __ret; \
+})
+#else
+#define vsli_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x1_t __s1 = __p1; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsli_n_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 1); \
+  __ret; \
+})
+#else
+#define vsli_n_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsraq_n_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __s1 = __p1; \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 48); \
+  __ret; \
+})
+#else
+#define vsraq_n_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __s1 = __p1; \
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 48); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsraq_n_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 50); \
+  __ret; \
+})
+#else
+#define vsraq_n_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 50); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsraq_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __s1 = __p1; \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 51); \
+  __ret; \
+})
+#else
+#define vsraq_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __s1 = __p1; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 51); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsraq_n_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 49); \
+  __ret; \
+})
+#else
+#define vsraq_n_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 49); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsraq_n_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __s1 = __p1; \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 32); \
+  __ret; \
+})
+#else
+#define vsraq_n_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __s1 = __p1; \
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 32); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsraq_n_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 34); \
+  __ret; \
+})
+#else
+#define vsraq_n_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 34); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsraq_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __s1 = __p1; \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 35); \
+  __ret; \
+})
+#else
+#define vsraq_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __s1 = __p1; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 35); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsraq_n_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 33); \
+  __ret; \
+})
+#else
+#define vsraq_n_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 33); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsra_n_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __s1 = __p1; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 16); \
+  __ret; \
+})
+#else
+#define vsra_n_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __s1 = __p1; \
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 16); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsra_n_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 18); \
+  __ret; \
+})
+#else
+#define vsra_n_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsra_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x1_t __s1 = __p1; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
+  __ret; \
+})
+#else
+#define vsra_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x1_t __s1 = __p1; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsra_n_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 17); \
+  __ret; \
+})
+#else
+#define vsra_n_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 17); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsra_n_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __s1 = __p1; \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 0); \
+  __ret; \
+})
+#else
+#define vsra_n_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __s1 = __p1; \
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 0); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsra_n_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 2); \
+  __ret; \
+})
+#else
+#define vsra_n_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsra_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x1_t __s1 = __p1; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
+  __ret; \
+})
+#else
+#define vsra_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x1_t __s1 = __p1; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsra_n_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 1); \
+  __ret; \
+})
+#else
+#define vsra_n_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsri_n_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8_t __s0 = __p0; \
+  poly8x8_t __s1 = __p1; \
+  poly8x8_t __ret; \
+  __ret = (poly8x8_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 4); \
+  __ret; \
+})
+#else
+#define vsri_n_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8_t __s0 = __p0; \
+  poly8x8_t __s1 = __p1; \
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x8_t __ret; \
+  __ret = (poly8x8_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 4); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsri_n_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4_t __s0 = __p0; \
+  poly16x4_t __s1 = __p1; \
+  poly16x4_t __ret; \
+  __ret = (poly16x4_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 5); \
+  __ret; \
+})
+#else
+#define vsri_n_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4_t __s0 = __p0; \
+  poly16x4_t __s1 = __p1; \
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  poly16x4_t __ret; \
+  __ret = (poly16x4_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 5); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsriq_n_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16_t __s0 = __p0; \
+  poly8x16_t __s1 = __p1; \
+  poly8x16_t __ret; \
+  __ret = (poly8x16_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 36); \
+  __ret; \
+})
+#else
+#define vsriq_n_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16_t __s0 = __p0; \
+  poly8x16_t __s1 = __p1; \
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16_t __ret; \
+  __ret = (poly8x16_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 36); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsriq_n_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8_t __s0 = __p0; \
+  poly16x8_t __s1 = __p1; \
+  poly16x8_t __ret; \
+  __ret = (poly16x8_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 37); \
+  __ret; \
+})
+#else
+#define vsriq_n_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8_t __s0 = __p0; \
+  poly16x8_t __s1 = __p1; \
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x8_t __ret; \
+  __ret = (poly16x8_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 37); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsriq_n_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __s1 = __p1; \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 48); \
+  __ret; \
+})
+#else
+#define vsriq_n_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __s1 = __p1; \
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret; \
+  __ret = (uint8x16_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 48); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsriq_n_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 50); \
+  __ret; \
+})
+#else
+#define vsriq_n_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = (uint32x4_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 50); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsriq_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __s1 = __p1; \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 51); \
+  __ret; \
+})
+#else
+#define vsriq_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __s1 = __p1; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 51); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsriq_n_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 49); \
+  __ret; \
+})
+#else
+#define vsriq_n_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = (uint16x8_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 49); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsriq_n_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __s1 = __p1; \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 32); \
+  __ret; \
+})
+#else
+#define vsriq_n_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __s1 = __p1; \
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret; \
+  __ret = (int8x16_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 32); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsriq_n_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 34); \
+  __ret; \
+})
+#else
+#define vsriq_n_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = (int32x4_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 34); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsriq_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __s1 = __p1; \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 35); \
+  __ret; \
+})
+#else
+#define vsriq_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __s1 = __p1; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 35); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsriq_n_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 33); \
+  __ret; \
+})
+#else
+#define vsriq_n_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = (int16x8_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 33); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsri_n_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __s1 = __p1; \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 16); \
+  __ret; \
+})
+#else
+#define vsri_n_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __s1 = __p1; \
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret; \
+  __ret = (uint8x8_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 16); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsri_n_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 18); \
+  __ret; \
+})
+#else
+#define vsri_n_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = (uint32x2_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 18); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsri_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x1_t __s1 = __p1; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
+  __ret; \
+})
+#else
+#define vsri_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64x1_t __s1 = __p1; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsri_n_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 17); \
+  __ret; \
+})
+#else
+#define vsri_n_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = (uint16x4_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 17); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsri_n_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __s1 = __p1; \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 0); \
+  __ret; \
+})
+#else
+#define vsri_n_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __s1 = __p1; \
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret; \
+  __ret = (int8x8_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 0); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsri_n_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 2); \
+  __ret; \
+})
+#else
+#define vsri_n_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x2_t __ret; \
+  __ret = (int32x2_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsri_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x1_t __s1 = __p1; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
+  __ret; \
+})
+#else
+#define vsri_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64x1_t __s1 = __p1; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsri_n_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 1); \
+  __ret; \
+})
+#else
+#define vsri_n_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = (int16x4_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_p8(__p0, __p1) __extension__ ({ \
+  poly8x8_t __s1 = __p1; \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 4); \
+})
+#else
+#define vst1_p8(__p0, __p1) __extension__ ({ \
+  poly8x8_t __s1 = __p1; \
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 4); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_p16(__p0, __p1) __extension__ ({ \
+  poly16x4_t __s1 = __p1; \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 5); \
+})
+#else
+#define vst1_p16(__p0, __p1) __extension__ ({ \
+  poly16x4_t __s1 = __p1; \
+  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 5); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_p8(__p0, __p1) __extension__ ({ \
+  poly8x16_t __s1 = __p1; \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 36); \
+})
+#else
+#define vst1q_p8(__p0, __p1) __extension__ ({ \
+  poly8x16_t __s1 = __p1; \
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 36); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_p16(__p0, __p1) __extension__ ({ \
+  poly16x8_t __s1 = __p1; \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 37); \
+})
+#else
+#define vst1q_p16(__p0, __p1) __extension__ ({ \
+  poly16x8_t __s1 = __p1; \
+  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 37); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_u8(__p0, __p1) __extension__ ({ \
+  uint8x16_t __s1 = __p1; \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 48); \
+})
+#else
+#define vst1q_u8(__p0, __p1) __extension__ ({ \
+  uint8x16_t __s1 = __p1; \
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 48); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s1 = __p1; \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 50); \
+})
+#else
+#define vst1q_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 50); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 51); \
+})
+#else
+#define vst1q_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s1 = __p1; \
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 51); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s1 = __p1; \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 49); \
+})
+#else
+#define vst1q_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 49); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s1 = __p1; \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 32); \
+})
+#else
+#define vst1q_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s1 = __p1; \
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 32); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_f32(__p0, __p1) __extension__ ({ \
+  float32x4_t __s1 = __p1; \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 41); \
+})
+#else
+#define vst1q_f32(__p0, __p1) __extension__ ({ \
+  float32x4_t __s1 = __p1; \
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 41); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_f16(__p0, __p1) __extension__ ({ \
+  float16x8_t __s1 = __p1; \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 40); \
+})
+#else
+#define vst1q_f16(__p0, __p1) __extension__ ({ \
+  float16x8_t __s1 = __p1; \
+  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 40); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s1 = __p1; \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 34); \
+})
+#else
+#define vst1q_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 34); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 35); \
+})
+#else
+#define vst1q_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s1 = __p1; \
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 35); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s1 = __p1; \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 33); \
+})
+#else
+#define vst1q_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 33); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s1 = __p1; \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 16); \
+})
+#else
+#define vst1_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s1 = __p1; \
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 16); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s1 = __p1; \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 18); \
+})
+#else
+#define vst1_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 18); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s1 = __p1; \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 19); \
+})
+#else
+#define vst1_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s1 = __p1; \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 19); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s1 = __p1; \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 17); \
+})
+#else
+#define vst1_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 17); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s1 = __p1; \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 0); \
+})
+#else
+#define vst1_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s1 = __p1; \
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 0); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_f32(__p0, __p1) __extension__ ({ \
+  float32x2_t __s1 = __p1; \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 9); \
+})
+#else
+#define vst1_f32(__p0, __p1) __extension__ ({ \
+  float32x2_t __s1 = __p1; \
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 9); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_f16(__p0, __p1) __extension__ ({ \
+  float16x4_t __s1 = __p1; \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 8); \
+})
+#else
+#define vst1_f16(__p0, __p1) __extension__ ({ \
+  float16x4_t __s1 = __p1; \
+  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 8); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s1 = __p1; \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 2); \
+})
+#else
+#define vst1_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 2); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s1 = __p1; \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 3); \
+})
+#else
+#define vst1_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s1 = __p1; \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 3); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s1 = __p1; \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 1); \
+})
+#else
+#define vst1_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 1); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8_t __s1 = __p1; \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 4); \
+})
+#else
+#define vst1_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8_t __s1 = __p1; \
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 4); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4_t __s1 = __p1; \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 5); \
+})
+#else
+#define vst1_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4_t __s1 = __p1; \
+  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 5); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16_t __s1 = __p1; \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 36); \
+})
+#else
+#define vst1q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16_t __s1 = __p1; \
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 36); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8_t __s1 = __p1; \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 37); \
+})
+#else
+#define vst1q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8_t __s1 = __p1; \
+  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 37); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16_t __s1 = __p1; \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 48); \
+})
+#else
+#define vst1q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16_t __s1 = __p1; \
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 48); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s1 = __p1; \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 50); \
+})
+#else
+#define vst1q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 50); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 51); \
+})
+#else
+#define vst1q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2_t __s1 = __p1; \
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 51); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s1 = __p1; \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 49); \
+})
+#else
+#define vst1q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 49); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16_t __s1 = __p1; \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 32); \
+})
+#else
+#define vst1q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16_t __s1 = __p1; \
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 32); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4_t __s1 = __p1; \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 41); \
+})
+#else
+#define vst1q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4_t __s1 = __p1; \
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 41); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x8_t __s1 = __p1; \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 40); \
+})
+#else
+#define vst1q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x8_t __s1 = __p1; \
+  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 40); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s1 = __p1; \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 34); \
+})
+#else
+#define vst1q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 34); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 35); \
+})
+#else
+#define vst1q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2_t __s1 = __p1; \
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 35); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s1 = __p1; \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 33); \
+})
+#else
+#define vst1q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 33); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8_t __s1 = __p1; \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 16); \
+})
+#else
+#define vst1_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8_t __s1 = __p1; \
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 16); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s1 = __p1; \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 18); \
+})
+#else
+#define vst1_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 18); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1_t __s1 = __p1; \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 19); \
+})
+#else
+#define vst1_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1_t __s1 = __p1; \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 19); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s1 = __p1; \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 17); \
+})
+#else
+#define vst1_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 17); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8_t __s1 = __p1; \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 0); \
+})
+#else
+#define vst1_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8_t __s1 = __p1; \
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 0); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2_t __s1 = __p1; \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 9); \
+})
+#else
+#define vst1_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2_t __s1 = __p1; \
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 9); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x4_t __s1 = __p1; \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 8); \
+})
+#else
+#define vst1_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x4_t __s1 = __p1; \
+  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 8); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s1 = __p1; \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 2); \
+})
+#else
+#define vst1_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 2); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1_t __s1 = __p1; \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 3); \
+})
+#else
+#define vst1_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1_t __s1 = __p1; \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 3); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s1 = __p1; \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 1); \
+})
+#else
+#define vst1_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 1); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_p8(__p0, __p1) __extension__ ({ \
+  poly8x8x2_t __s1 = __p1; \
+  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 4); \
+})
+#else
+#define vst2_p8(__p0, __p1) __extension__ ({ \
+  poly8x8x2_t __s1 = __p1; \
+  poly8x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 4); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_p16(__p0, __p1) __extension__ ({ \
+  poly16x4x2_t __s1 = __p1; \
+  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 5); \
+})
+#else
+#define vst2_p16(__p0, __p1) __extension__ ({ \
+  poly16x4x2_t __s1 = __p1; \
+  poly16x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 5); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_p8(__p0, __p1) __extension__ ({ \
+  poly8x16x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 36); \
+})
+#else
+#define vst2q_p8(__p0, __p1) __extension__ ({ \
+  poly8x16x2_t __s1 = __p1; \
+  poly8x16x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 36); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_p16(__p0, __p1) __extension__ ({ \
+  poly16x8x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 37); \
+})
+#else
+#define vst2q_p16(__p0, __p1) __extension__ ({ \
+  poly16x8x2_t __s1 = __p1; \
+  poly16x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 37); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_u8(__p0, __p1) __extension__ ({ \
+  uint8x16x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 48); \
+})
+#else
+#define vst2q_u8(__p0, __p1) __extension__ ({ \
+  uint8x16x2_t __s1 = __p1; \
+  uint8x16x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 48); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_u32(__p0, __p1) __extension__ ({ \
+  uint32x4x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 50); \
+})
+#else
+#define vst2q_u32(__p0, __p1) __extension__ ({ \
+  uint32x4x2_t __s1 = __p1; \
+  uint32x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 50); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_u16(__p0, __p1) __extension__ ({ \
+  uint16x8x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 49); \
+})
+#else
+#define vst2q_u16(__p0, __p1) __extension__ ({ \
+  uint16x8x2_t __s1 = __p1; \
+  uint16x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 49); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_s8(__p0, __p1) __extension__ ({ \
+  int8x16x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 32); \
+})
+#else
+#define vst2q_s8(__p0, __p1) __extension__ ({ \
+  int8x16x2_t __s1 = __p1; \
+  int8x16x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 32); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_f32(__p0, __p1) __extension__ ({ \
+  float32x4x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_v(__p0, __s1.val[0], __s1.val[1], 41); \
+})
+#else
+#define vst2q_f32(__p0, __p1) __extension__ ({ \
+  float32x4x2_t __s1 = __p1; \
+  float32x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst2q_v(__p0, __rev1.val[0], __rev1.val[1], 41); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_f16(__p0, __p1) __extension__ ({ \
+  float16x8x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_v(__p0, __s1.val[0], __s1.val[1], 40); \
+})
+#else
+#define vst2q_f16(__p0, __p1) __extension__ ({ \
+  float16x8x2_t __s1 = __p1; \
+  float16x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst2q_v(__p0, __rev1.val[0], __rev1.val[1], 40); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_s32(__p0, __p1) __extension__ ({ \
+  int32x4x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_v(__p0, __s1.val[0], __s1.val[1], 34); \
+})
+#else
+#define vst2q_s32(__p0, __p1) __extension__ ({ \
+  int32x4x2_t __s1 = __p1; \
+  int32x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst2q_v(__p0, __rev1.val[0], __rev1.val[1], 34); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_s16(__p0, __p1) __extension__ ({ \
+  int16x8x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_v(__p0, __s1.val[0], __s1.val[1], 33); \
+})
+#else
+#define vst2q_s16(__p0, __p1) __extension__ ({ \
+  int16x8x2_t __s1 = __p1; \
+  int16x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst2q_v(__p0, __rev1.val[0], __rev1.val[1], 33); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_u8(__p0, __p1) __extension__ ({ \
+  uint8x8x2_t __s1 = __p1; \
+  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 16); \
+})
+#else
+#define vst2_u8(__p0, __p1) __extension__ ({ \
+  uint8x8x2_t __s1 = __p1; \
+  uint8x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 16); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_u32(__p0, __p1) __extension__ ({ \
+  uint32x2x2_t __s1 = __p1; \
+  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 18); \
+})
+#else
+#define vst2_u32(__p0, __p1) __extension__ ({ \
+  uint32x2x2_t __s1 = __p1; \
+  uint32x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 18); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_u64(__p0, __p1) __extension__ ({ \
+  uint64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 19); \
+})
+#else
+#define vst2_u64(__p0, __p1) __extension__ ({ \
+  uint64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 19); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_u16(__p0, __p1) __extension__ ({ \
+  uint16x4x2_t __s1 = __p1; \
+  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 17); \
+})
+#else
+#define vst2_u16(__p0, __p1) __extension__ ({ \
+  uint16x4x2_t __s1 = __p1; \
+  uint16x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 17); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_s8(__p0, __p1) __extension__ ({ \
+  int8x8x2_t __s1 = __p1; \
+  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 0); \
+})
+#else
+#define vst2_s8(__p0, __p1) __extension__ ({ \
+  int8x8x2_t __s1 = __p1; \
+  int8x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 0); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_f32(__p0, __p1) __extension__ ({ \
+  float32x2x2_t __s1 = __p1; \
+  __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 9); \
+})
+#else
+#define vst2_f32(__p0, __p1) __extension__ ({ \
+  float32x2x2_t __s1 = __p1; \
+  float32x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst2_v(__p0, __rev1.val[0], __rev1.val[1], 9); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_f16(__p0, __p1) __extension__ ({ \
+  float16x4x2_t __s1 = __p1; \
+  __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 8); \
+})
+#else
+#define vst2_f16(__p0, __p1) __extension__ ({ \
+  float16x4x2_t __s1 = __p1; \
+  float16x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst2_v(__p0, __rev1.val[0], __rev1.val[1], 8); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_s32(__p0, __p1) __extension__ ({ \
+  int32x2x2_t __s1 = __p1; \
+  __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 2); \
+})
+#else
+#define vst2_s32(__p0, __p1) __extension__ ({ \
+  int32x2x2_t __s1 = __p1; \
+  int32x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst2_v(__p0, __rev1.val[0], __rev1.val[1], 2); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_s64(__p0, __p1) __extension__ ({ \
+  int64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 3); \
+})
+#else
+#define vst2_s64(__p0, __p1) __extension__ ({ \
+  int64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 3); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_s16(__p0, __p1) __extension__ ({ \
+  int16x4x2_t __s1 = __p1; \
+  __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 1); \
+})
+#else
+#define vst2_s16(__p0, __p1) __extension__ ({ \
+  int16x4x2_t __s1 = __p1; \
+  int16x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst2_v(__p0, __rev1.val[0], __rev1.val[1], 1); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8x2_t __s1 = __p1; \
+  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 4); \
+})
+#else
+#define vst2_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8x2_t __s1 = __p1; \
+  poly8x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 4); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4x2_t __s1 = __p1; \
+  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 5); \
+})
+#else
+#define vst2_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4x2_t __s1 = __p1; \
+  poly16x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 5); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 37); \
+})
+#else
+#define vst2q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8x2_t __s1 = __p1; \
+  poly16x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 37); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 50); \
+})
+#else
+#define vst2q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4x2_t __s1 = __p1; \
+  uint32x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 50); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 49); \
+})
+#else
+#define vst2q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8x2_t __s1 = __p1; \
+  uint16x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 49); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 41); \
+})
+#else
+#define vst2q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4x2_t __s1 = __p1; \
+  float32x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst2q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __p2, 41); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x8x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 40); \
+})
+#else
+#define vst2q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x8x2_t __s1 = __p1; \
+  float16x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst2q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __p2, 40); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 34); \
+})
+#else
+#define vst2q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4x2_t __s1 = __p1; \
+  int32x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst2q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __p2, 34); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 33); \
+})
+#else
+#define vst2q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8x2_t __s1 = __p1; \
+  int16x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst2q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __p2, 33); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8x2_t __s1 = __p1; \
+  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 16); \
+})
+#else
+#define vst2_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8x2_t __s1 = __p1; \
+  uint8x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 16); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2x2_t __s1 = __p1; \
+  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 18); \
+})
+#else
+#define vst2_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2x2_t __s1 = __p1; \
+  uint32x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 18); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4x2_t __s1 = __p1; \
+  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 17); \
+})
+#else
+#define vst2_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4x2_t __s1 = __p1; \
+  uint16x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 17); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8x2_t __s1 = __p1; \
+  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 0); \
+})
+#else
+#define vst2_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8x2_t __s1 = __p1; \
+  int8x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 0); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2x2_t __s1 = __p1; \
+  __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 9); \
+})
+#else
+#define vst2_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2x2_t __s1 = __p1; \
+  float32x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst2_lane_v(__p0, __rev1.val[0], __rev1.val[1], __p2, 9); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x4x2_t __s1 = __p1; \
+  __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 8); \
+})
+#else
+#define vst2_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x4x2_t __s1 = __p1; \
+  float16x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst2_lane_v(__p0, __rev1.val[0], __rev1.val[1], __p2, 8); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2x2_t __s1 = __p1; \
+  __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 2); \
+})
+#else
+#define vst2_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2x2_t __s1 = __p1; \
+  int32x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst2_lane_v(__p0, __rev1.val[0], __rev1.val[1], __p2, 2); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4x2_t __s1 = __p1; \
+  __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 1); \
+})
+#else
+#define vst2_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4x2_t __s1 = __p1; \
+  int16x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst2_lane_v(__p0, __rev1.val[0], __rev1.val[1], __p2, 1); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_p8(__p0, __p1) __extension__ ({ \
+  poly8x8x3_t __s1 = __p1; \
+  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 4); \
+})
+#else
+#define vst3_p8(__p0, __p1) __extension__ ({ \
+  poly8x8x3_t __s1 = __p1; \
+  poly8x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 4); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_p16(__p0, __p1) __extension__ ({ \
+  poly16x4x3_t __s1 = __p1; \
+  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 5); \
+})
+#else
+#define vst3_p16(__p0, __p1) __extension__ ({ \
+  poly16x4x3_t __s1 = __p1; \
+  poly16x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 5); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_p8(__p0, __p1) __extension__ ({ \
+  poly8x16x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 36); \
+})
+#else
+#define vst3q_p8(__p0, __p1) __extension__ ({ \
+  poly8x16x3_t __s1 = __p1; \
+  poly8x16x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 36); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_p16(__p0, __p1) __extension__ ({ \
+  poly16x8x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 37); \
+})
+#else
+#define vst3q_p16(__p0, __p1) __extension__ ({ \
+  poly16x8x3_t __s1 = __p1; \
+  poly16x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 37); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_u8(__p0, __p1) __extension__ ({ \
+  uint8x16x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 48); \
+})
+#else
+#define vst3q_u8(__p0, __p1) __extension__ ({ \
+  uint8x16x3_t __s1 = __p1; \
+  uint8x16x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 48); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_u32(__p0, __p1) __extension__ ({ \
+  uint32x4x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 50); \
+})
+#else
+#define vst3q_u32(__p0, __p1) __extension__ ({ \
+  uint32x4x3_t __s1 = __p1; \
+  uint32x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 50); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_u16(__p0, __p1) __extension__ ({ \
+  uint16x8x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 49); \
+})
+#else
+#define vst3q_u16(__p0, __p1) __extension__ ({ \
+  uint16x8x3_t __s1 = __p1; \
+  uint16x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 49); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_s8(__p0, __p1) __extension__ ({ \
+  int8x16x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 32); \
+})
+#else
+#define vst3q_s8(__p0, __p1) __extension__ ({ \
+  int8x16x3_t __s1 = __p1; \
+  int8x16x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 32); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_f32(__p0, __p1) __extension__ ({ \
+  float32x4x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 41); \
+})
+#else
+#define vst3q_f32(__p0, __p1) __extension__ ({ \
+  float32x4x3_t __s1 = __p1; \
+  float32x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst3q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 41); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_f16(__p0, __p1) __extension__ ({ \
+  float16x8x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 40); \
+})
+#else
+#define vst3q_f16(__p0, __p1) __extension__ ({ \
+  float16x8x3_t __s1 = __p1; \
+  float16x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst3q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 40); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_s32(__p0, __p1) __extension__ ({ \
+  int32x4x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 34); \
+})
+#else
+#define vst3q_s32(__p0, __p1) __extension__ ({ \
+  int32x4x3_t __s1 = __p1; \
+  int32x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst3q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 34); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_s16(__p0, __p1) __extension__ ({ \
+  int16x8x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 33); \
+})
+#else
+#define vst3q_s16(__p0, __p1) __extension__ ({ \
+  int16x8x3_t __s1 = __p1; \
+  int16x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst3q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 33); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_u8(__p0, __p1) __extension__ ({ \
+  uint8x8x3_t __s1 = __p1; \
+  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 16); \
+})
+#else
+#define vst3_u8(__p0, __p1) __extension__ ({ \
+  uint8x8x3_t __s1 = __p1; \
+  uint8x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 16); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_u32(__p0, __p1) __extension__ ({ \
+  uint32x2x3_t __s1 = __p1; \
+  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 18); \
+})
+#else
+#define vst3_u32(__p0, __p1) __extension__ ({ \
+  uint32x2x3_t __s1 = __p1; \
+  uint32x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 18); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_u64(__p0, __p1) __extension__ ({ \
+  uint64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 19); \
+})
+#else
+#define vst3_u64(__p0, __p1) __extension__ ({ \
+  uint64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 19); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_u16(__p0, __p1) __extension__ ({ \
+  uint16x4x3_t __s1 = __p1; \
+  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 17); \
+})
+#else
+#define vst3_u16(__p0, __p1) __extension__ ({ \
+  uint16x4x3_t __s1 = __p1; \
+  uint16x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 17); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_s8(__p0, __p1) __extension__ ({ \
+  int8x8x3_t __s1 = __p1; \
+  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 0); \
+})
+#else
+#define vst3_s8(__p0, __p1) __extension__ ({ \
+  int8x8x3_t __s1 = __p1; \
+  int8x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 0); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_f32(__p0, __p1) __extension__ ({ \
+  float32x2x3_t __s1 = __p1; \
+  __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 9); \
+})
+#else
+#define vst3_f32(__p0, __p1) __extension__ ({ \
+  float32x2x3_t __s1 = __p1; \
+  float32x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 9); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_f16(__p0, __p1) __extension__ ({ \
+  float16x4x3_t __s1 = __p1; \
+  __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 8); \
+})
+#else
+#define vst3_f16(__p0, __p1) __extension__ ({ \
+  float16x4x3_t __s1 = __p1; \
+  float16x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 8); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_s32(__p0, __p1) __extension__ ({ \
+  int32x2x3_t __s1 = __p1; \
+  __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 2); \
+})
+#else
+#define vst3_s32(__p0, __p1) __extension__ ({ \
+  int32x2x3_t __s1 = __p1; \
+  int32x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 2); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_s64(__p0, __p1) __extension__ ({ \
+  int64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 3); \
+})
+#else
+#define vst3_s64(__p0, __p1) __extension__ ({ \
+  int64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 3); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_s16(__p0, __p1) __extension__ ({ \
+  int16x4x3_t __s1 = __p1; \
+  __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 1); \
+})
+#else
+#define vst3_s16(__p0, __p1) __extension__ ({ \
+  int16x4x3_t __s1 = __p1; \
+  int16x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 1); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8x3_t __s1 = __p1; \
+  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 4); \
+})
+#else
+#define vst3_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8x3_t __s1 = __p1; \
+  poly8x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 4); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4x3_t __s1 = __p1; \
+  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 5); \
+})
+#else
+#define vst3_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4x3_t __s1 = __p1; \
+  poly16x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 5); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 37); \
+})
+#else
+#define vst3q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8x3_t __s1 = __p1; \
+  poly16x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 37); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 50); \
+})
+#else
+#define vst3q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4x3_t __s1 = __p1; \
+  uint32x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 50); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 49); \
+})
+#else
+#define vst3q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8x3_t __s1 = __p1; \
+  uint16x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 49); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 41); \
+})
+#else
+#define vst3q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4x3_t __s1 = __p1; \
+  float32x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst3q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 41); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x8x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 40); \
+})
+#else
+#define vst3q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x8x3_t __s1 = __p1; \
+  float16x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst3q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 40); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 34); \
+})
+#else
+#define vst3q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4x3_t __s1 = __p1; \
+  int32x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst3q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 34); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 33); \
+})
+#else
+#define vst3q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8x3_t __s1 = __p1; \
+  int16x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst3q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 33); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8x3_t __s1 = __p1; \
+  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 16); \
+})
+#else
+#define vst3_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8x3_t __s1 = __p1; \
+  uint8x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 16); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2x3_t __s1 = __p1; \
+  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 18); \
+})
+#else
+#define vst3_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2x3_t __s1 = __p1; \
+  uint32x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 18); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4x3_t __s1 = __p1; \
+  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 17); \
+})
+#else
+#define vst3_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4x3_t __s1 = __p1; \
+  uint16x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 17); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8x3_t __s1 = __p1; \
+  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 0); \
+})
+#else
+#define vst3_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8x3_t __s1 = __p1; \
+  int8x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 0); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2x3_t __s1 = __p1; \
+  __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 9); \
+})
+#else
+#define vst3_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2x3_t __s1 = __p1; \
+  float32x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst3_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 9); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x4x3_t __s1 = __p1; \
+  __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 8); \
+})
+#else
+#define vst3_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x4x3_t __s1 = __p1; \
+  float16x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst3_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 8); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2x3_t __s1 = __p1; \
+  __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 2); \
+})
+#else
+#define vst3_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2x3_t __s1 = __p1; \
+  int32x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst3_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 2); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4x3_t __s1 = __p1; \
+  __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 1); \
+})
+#else
+#define vst3_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4x3_t __s1 = __p1; \
+  int16x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst3_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 1); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_p8(__p0, __p1) __extension__ ({ \
+  poly8x8x4_t __s1 = __p1; \
+  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 4); \
+})
+#else
+#define vst4_p8(__p0, __p1) __extension__ ({ \
+  poly8x8x4_t __s1 = __p1; \
+  poly8x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 4); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_p16(__p0, __p1) __extension__ ({ \
+  poly16x4x4_t __s1 = __p1; \
+  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 5); \
+})
+#else
+#define vst4_p16(__p0, __p1) __extension__ ({ \
+  poly16x4x4_t __s1 = __p1; \
+  poly16x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 5); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_p8(__p0, __p1) __extension__ ({ \
+  poly8x16x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 36); \
+})
+#else
+#define vst4q_p8(__p0, __p1) __extension__ ({ \
+  poly8x16x4_t __s1 = __p1; \
+  poly8x16x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 36); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_p16(__p0, __p1) __extension__ ({ \
+  poly16x8x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 37); \
+})
+#else
+#define vst4q_p16(__p0, __p1) __extension__ ({ \
+  poly16x8x4_t __s1 = __p1; \
+  poly16x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 37); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_u8(__p0, __p1) __extension__ ({ \
+  uint8x16x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 48); \
+})
+#else
+#define vst4q_u8(__p0, __p1) __extension__ ({ \
+  uint8x16x4_t __s1 = __p1; \
+  uint8x16x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 48); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_u32(__p0, __p1) __extension__ ({ \
+  uint32x4x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 50); \
+})
+#else
+#define vst4q_u32(__p0, __p1) __extension__ ({ \
+  uint32x4x4_t __s1 = __p1; \
+  uint32x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 50); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_u16(__p0, __p1) __extension__ ({ \
+  uint16x8x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 49); \
+})
+#else
+#define vst4q_u16(__p0, __p1) __extension__ ({ \
+  uint16x8x4_t __s1 = __p1; \
+  uint16x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 49); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_s8(__p0, __p1) __extension__ ({ \
+  int8x16x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 32); \
+})
+#else
+#define vst4q_s8(__p0, __p1) __extension__ ({ \
+  int8x16x4_t __s1 = __p1; \
+  int8x16x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 32); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_f32(__p0, __p1) __extension__ ({ \
+  float32x4x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 41); \
+})
+#else
+#define vst4q_f32(__p0, __p1) __extension__ ({ \
+  float32x4x4_t __s1 = __p1; \
+  float32x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst4q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 41); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_f16(__p0, __p1) __extension__ ({ \
+  float16x8x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 40); \
+})
+#else
+#define vst4q_f16(__p0, __p1) __extension__ ({ \
+  float16x8x4_t __s1 = __p1; \
+  float16x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst4q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 40); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_s32(__p0, __p1) __extension__ ({ \
+  int32x4x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 34); \
+})
+#else
+#define vst4q_s32(__p0, __p1) __extension__ ({ \
+  int32x4x4_t __s1 = __p1; \
+  int32x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst4q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 34); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_s16(__p0, __p1) __extension__ ({ \
+  int16x8x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 33); \
+})
+#else
+#define vst4q_s16(__p0, __p1) __extension__ ({ \
+  int16x8x4_t __s1 = __p1; \
+  int16x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst4q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 33); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_u8(__p0, __p1) __extension__ ({ \
+  uint8x8x4_t __s1 = __p1; \
+  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 16); \
+})
+#else
+#define vst4_u8(__p0, __p1) __extension__ ({ \
+  uint8x8x4_t __s1 = __p1; \
+  uint8x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 16); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_u32(__p0, __p1) __extension__ ({ \
+  uint32x2x4_t __s1 = __p1; \
+  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 18); \
+})
+#else
+#define vst4_u32(__p0, __p1) __extension__ ({ \
+  uint32x2x4_t __s1 = __p1; \
+  uint32x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 18); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_u64(__p0, __p1) __extension__ ({ \
+  uint64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 19); \
+})
+#else
+#define vst4_u64(__p0, __p1) __extension__ ({ \
+  uint64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 19); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_u16(__p0, __p1) __extension__ ({ \
+  uint16x4x4_t __s1 = __p1; \
+  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 17); \
+})
+#else
+#define vst4_u16(__p0, __p1) __extension__ ({ \
+  uint16x4x4_t __s1 = __p1; \
+  uint16x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 17); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_s8(__p0, __p1) __extension__ ({ \
+  int8x8x4_t __s1 = __p1; \
+  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 0); \
+})
+#else
+#define vst4_s8(__p0, __p1) __extension__ ({ \
+  int8x8x4_t __s1 = __p1; \
+  int8x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 0); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_f32(__p0, __p1) __extension__ ({ \
+  float32x2x4_t __s1 = __p1; \
+  __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 9); \
+})
+#else
+#define vst4_f32(__p0, __p1) __extension__ ({ \
+  float32x2x4_t __s1 = __p1; \
+  float32x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 9); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_f16(__p0, __p1) __extension__ ({ \
+  float16x4x4_t __s1 = __p1; \
+  __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 8); \
+})
+#else
+#define vst4_f16(__p0, __p1) __extension__ ({ \
+  float16x4x4_t __s1 = __p1; \
+  float16x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 8); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_s32(__p0, __p1) __extension__ ({ \
+  int32x2x4_t __s1 = __p1; \
+  __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 2); \
+})
+#else
+#define vst4_s32(__p0, __p1) __extension__ ({ \
+  int32x2x4_t __s1 = __p1; \
+  int32x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 2); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_s64(__p0, __p1) __extension__ ({ \
+  int64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 3); \
+})
+#else
+#define vst4_s64(__p0, __p1) __extension__ ({ \
+  int64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 3); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_s16(__p0, __p1) __extension__ ({ \
+  int16x4x4_t __s1 = __p1; \
+  __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 1); \
+})
+#else
+#define vst4_s16(__p0, __p1) __extension__ ({ \
+  int16x4x4_t __s1 = __p1; \
+  int16x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 1); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8x4_t __s1 = __p1; \
+  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 4); \
+})
+#else
+#define vst4_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x8x4_t __s1 = __p1; \
+  poly8x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 4); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4x4_t __s1 = __p1; \
+  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 5); \
+})
+#else
+#define vst4_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x4x4_t __s1 = __p1; \
+  poly16x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 5); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 37); \
+})
+#else
+#define vst4q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
+  poly16x8x4_t __s1 = __p1; \
+  poly16x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 37); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 50); \
+})
+#else
+#define vst4q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4x4_t __s1 = __p1; \
+  uint32x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 50); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 49); \
+})
+#else
+#define vst4q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8x4_t __s1 = __p1; \
+  uint16x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 49); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 41); \
+})
+#else
+#define vst4q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4x4_t __s1 = __p1; \
+  float32x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst4q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 41); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x8x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 40); \
+})
+#else
+#define vst4q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x8x4_t __s1 = __p1; \
+  float16x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst4q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 40); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 34); \
+})
+#else
+#define vst4q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4x4_t __s1 = __p1; \
+  int32x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst4q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 34); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 33); \
+})
+#else
+#define vst4q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8x4_t __s1 = __p1; \
+  int16x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst4q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 33); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8x4_t __s1 = __p1; \
+  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 16); \
+})
+#else
+#define vst4_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x8x4_t __s1 = __p1; \
+  uint8x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 16); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2x4_t __s1 = __p1; \
+  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 18); \
+})
+#else
+#define vst4_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2x4_t __s1 = __p1; \
+  uint32x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 18); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4x4_t __s1 = __p1; \
+  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 17); \
+})
+#else
+#define vst4_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4x4_t __s1 = __p1; \
+  uint16x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 17); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8x4_t __s1 = __p1; \
+  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 0); \
+})
+#else
+#define vst4_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x8x4_t __s1 = __p1; \
+  int8x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 0); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2x4_t __s1 = __p1; \
+  __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 9); \
+})
+#else
+#define vst4_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2x4_t __s1 = __p1; \
+  float32x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst4_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 9); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x4x4_t __s1 = __p1; \
+  __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 8); \
+})
+#else
+#define vst4_lane_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16x4x4_t __s1 = __p1; \
+  float16x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst4_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 8); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2x4_t __s1 = __p1; \
+  __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 2); \
+})
+#else
+#define vst4_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2x4_t __s1 = __p1; \
+  int32x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst4_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 2); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4x4_t __s1 = __p1; \
+  __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 1); \
+})
+#else
+#define vst4_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4x4_t __s1 = __p1; \
+  int16x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst4_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 1); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vsubq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#else
+__ai uint8x16_t vsubq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __rev0 - __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vsubq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#else
+__ai uint32x4_t vsubq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 - __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vsubq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#else
+__ai uint64x2_t vsubq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = __rev0 - __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vsubq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#else
+__ai uint16x8_t vsubq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 - __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vsubq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#else
+__ai int8x16_t vsubq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __rev0 - __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vsubq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#else
+__ai float32x4_t vsubq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = __rev0 - __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vsubq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#else
+__ai int32x4_t vsubq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 - __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vsubq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#else
+__ai int64x2_t vsubq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = __rev0 - __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vsubq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#else
+__ai int16x8_t vsubq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 - __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vsub_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#else
+__ai uint8x8_t vsub_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __rev0 - __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vsub_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#else
+__ai uint32x2_t vsub_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = __rev0 - __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vsub_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#else
+__ai uint64x1_t vsub_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#else
+__ai uint16x4_t vsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __rev0 - __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vsub_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#else
+__ai int8x8_t vsub_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __rev0 - __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vsub_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#else
+__ai float32x2_t vsub_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = __rev0 - __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vsub_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#else
+__ai int32x2_t vsub_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = __rev0 - __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vsub_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#else
+__ai int64x1_t vsub_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vsub_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#else
+__ai int16x4_t vsub_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __rev0 - __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vsubhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vsubhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint16x4_t __noswap_vsubhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vsubhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vsubhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai uint32x2_t __noswap_vsubhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vsubhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vsubhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint8x8_t __noswap_vsubhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vsubhn_s32(int32x4_t __p0, int32x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vsubhn_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int16x4_t __noswap_vsubhn_s32(int32x4_t __p0, int32x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vsubhn_s64(int64x2_t __p0, int64x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vsubhn_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int32x2_t __noswap_vsubhn_s64(int64x2_t __p0, int64x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vsubhn_s16(int16x8_t __p0, int16x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vsubhn_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int8x8_t __noswap_vsubhn_s16(int16x8_t __p0, int16x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vsubl_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = vmovl_u8(__p0) - vmovl_u8(__p1);
+  return __ret;
+}
+#else
+__ai uint16x8_t vsubl_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __noswap_vmovl_u8(__rev0) - __noswap_vmovl_u8(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vsubl_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = vmovl_u32(__p0) - vmovl_u32(__p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vsubl_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = __noswap_vmovl_u32(__rev0) - __noswap_vmovl_u32(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vsubl_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = vmovl_u16(__p0) - vmovl_u16(__p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vsubl_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __noswap_vmovl_u16(__rev0) - __noswap_vmovl_u16(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vsubl_s8(int8x8_t __p0, int8x8_t __p1) {
+  int16x8_t __ret;
+  __ret = vmovl_s8(__p0) - vmovl_s8(__p1);
+  return __ret;
+}
+#else
+__ai int16x8_t vsubl_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __noswap_vmovl_s8(__rev0) - __noswap_vmovl_s8(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vsubl_s32(int32x2_t __p0, int32x2_t __p1) {
+  int64x2_t __ret;
+  __ret = vmovl_s32(__p0) - vmovl_s32(__p1);
+  return __ret;
+}
+#else
+__ai int64x2_t vsubl_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = __noswap_vmovl_s32(__rev0) - __noswap_vmovl_s32(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vsubl_s16(int16x4_t __p0, int16x4_t __p1) {
+  int32x4_t __ret;
+  __ret = vmovl_s16(__p0) - vmovl_s16(__p1);
+  return __ret;
+}
+#else
+__ai int32x4_t vsubl_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vmovl_s16(__rev0) - __noswap_vmovl_s16(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vsubw_u8(uint16x8_t __p0, uint8x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = __p0 - vmovl_u8(__p1);
+  return __ret;
+}
+#else
+__ai uint16x8_t vsubw_u8(uint16x8_t __p0, uint8x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 - __noswap_vmovl_u8(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vsubw_u32(uint64x2_t __p0, uint32x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = __p0 - vmovl_u32(__p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vsubw_u32(uint64x2_t __p0, uint32x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = __rev0 - __noswap_vmovl_u32(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vsubw_u16(uint32x4_t __p0, uint16x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = __p0 - vmovl_u16(__p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vsubw_u16(uint32x4_t __p0, uint16x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 - __noswap_vmovl_u16(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vsubw_s8(int16x8_t __p0, int8x8_t __p1) {
+  int16x8_t __ret;
+  __ret = __p0 - vmovl_s8(__p1);
+  return __ret;
+}
+#else
+__ai int16x8_t vsubw_s8(int16x8_t __p0, int8x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 - __noswap_vmovl_s8(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vsubw_s32(int64x2_t __p0, int32x2_t __p1) {
+  int64x2_t __ret;
+  __ret = __p0 - vmovl_s32(__p1);
+  return __ret;
+}
+#else
+__ai int64x2_t vsubw_s32(int64x2_t __p0, int32x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = __rev0 - __noswap_vmovl_s32(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vsubw_s16(int32x4_t __p0, int16x4_t __p1) {
+  int32x4_t __ret;
+  __ret = __p0 - vmovl_s16(__p1);
+  return __ret;
+}
+#else
+__ai int32x4_t vsubw_s16(int32x4_t __p0, int16x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 - __noswap_vmovl_s16(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vtbl1_p8(poly8x8_t __p0, uint8x8_t __p1) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vtbl1_v((int8x8_t)__p0, (int8x8_t)__p1, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vtbl1_p8(poly8x8_t __p0, uint8x8_t __p1) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vtbl1_v((int8x8_t)__rev0, (int8x8_t)__rev1, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vtbl1_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtbl1_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vtbl1_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtbl1_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vtbl1_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vtbl1_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vtbl1_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vtbl1_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vtbl2_p8(poly8x8x2_t __p0, uint8x8_t __p1) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vtbl2_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p1, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vtbl2_p8(poly8x8x2_t __p0, uint8x8_t __p1) {
+  poly8x8x2_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vtbl2_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev1, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vtbl2_u8(uint8x8x2_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtbl2_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vtbl2_u8(uint8x8x2_t __p0, uint8x8_t __p1) {
+  uint8x8x2_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtbl2_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vtbl2_s8(int8x8x2_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vtbl2_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vtbl2_s8(int8x8x2_t __p0, int8x8_t __p1) {
+  int8x8x2_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vtbl2_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vtbl3_p8(poly8x8x3_t __p0, uint8x8_t __p1) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vtbl3_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p0.val[2], (int8x8_t)__p1, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vtbl3_p8(poly8x8x3_t __p0, uint8x8_t __p1) {
+  poly8x8x3_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vtbl3_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev0.val[2], (int8x8_t)__rev1, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vtbl3_u8(uint8x8x3_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtbl3_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p0.val[2], (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vtbl3_u8(uint8x8x3_t __p0, uint8x8_t __p1) {
+  uint8x8x3_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtbl3_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev0.val[2], (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vtbl3_s8(int8x8x3_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vtbl3_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p0.val[2], (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vtbl3_s8(int8x8x3_t __p0, int8x8_t __p1) {
+  int8x8x3_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vtbl3_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev0.val[2], (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vtbl4_p8(poly8x8x4_t __p0, uint8x8_t __p1) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vtbl4_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p0.val[2], (int8x8_t)__p0.val[3], (int8x8_t)__p1, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vtbl4_p8(poly8x8x4_t __p0, uint8x8_t __p1) {
+  poly8x8x4_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vtbl4_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev0.val[2], (int8x8_t)__rev0.val[3], (int8x8_t)__rev1, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vtbl4_u8(uint8x8x4_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtbl4_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p0.val[2], (int8x8_t)__p0.val[3], (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vtbl4_u8(uint8x8x4_t __p0, uint8x8_t __p1) {
+  uint8x8x4_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtbl4_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev0.val[2], (int8x8_t)__rev0.val[3], (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vtbl4_s8(int8x8x4_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vtbl4_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p0.val[2], (int8x8_t)__p0.val[3], (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vtbl4_s8(int8x8x4_t __p0, int8x8_t __p1) {
+  int8x8x4_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vtbl4_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev0.val[2], (int8x8_t)__rev0.val[3], (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vtbx1_p8(poly8x8_t __p0, poly8x8_t __p1, uint8x8_t __p2) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vtbx1_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vtbx1_p8(poly8x8_t __p0, poly8x8_t __p1, uint8x8_t __p2) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vtbx1_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vtbx1_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtbx1_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vtbx1_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtbx1_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vtbx1_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vtbx1_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vtbx1_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vtbx1_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vtbx2_p8(poly8x8_t __p0, poly8x8x2_t __p1, uint8x8_t __p2) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vtbx2_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p2, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vtbx2_p8(poly8x8_t __p0, poly8x8x2_t __p1, uint8x8_t __p2) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8x2_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vtbx2_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev2, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vtbx2_u8(uint8x8_t __p0, uint8x8x2_t __p1, uint8x8_t __p2) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtbx2_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p2, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vtbx2_u8(uint8x8_t __p0, uint8x8x2_t __p1, uint8x8_t __p2) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8x2_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtbx2_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev2, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vtbx2_s8(int8x8_t __p0, int8x8x2_t __p1, int8x8_t __p2) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vtbx2_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p2, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vtbx2_s8(int8x8_t __p0, int8x8x2_t __p1, int8x8_t __p2) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8x2_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vtbx2_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev2, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vtbx3_p8(poly8x8_t __p0, poly8x8x3_t __p1, uint8x8_t __p2) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vtbx3_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p1.val[2], (int8x8_t)__p2, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vtbx3_p8(poly8x8_t __p0, poly8x8x3_t __p1, uint8x8_t __p2) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8x3_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vtbx3_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev2, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vtbx3_u8(uint8x8_t __p0, uint8x8x3_t __p1, uint8x8_t __p2) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtbx3_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p1.val[2], (int8x8_t)__p2, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vtbx3_u8(uint8x8_t __p0, uint8x8x3_t __p1, uint8x8_t __p2) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8x3_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtbx3_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev2, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vtbx3_s8(int8x8_t __p0, int8x8x3_t __p1, int8x8_t __p2) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vtbx3_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p1.val[2], (int8x8_t)__p2, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vtbx3_s8(int8x8_t __p0, int8x8x3_t __p1, int8x8_t __p2) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8x3_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vtbx3_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev2, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vtbx4_p8(poly8x8_t __p0, poly8x8x4_t __p1, uint8x8_t __p2) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vtbx4_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p1.val[2], (int8x8_t)__p1.val[3], (int8x8_t)__p2, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vtbx4_p8(poly8x8_t __p0, poly8x8x4_t __p1, uint8x8_t __p2) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8x4_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vtbx4_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], (int8x8_t)__rev2, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vtbx4_u8(uint8x8_t __p0, uint8x8x4_t __p1, uint8x8_t __p2) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtbx4_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p1.val[2], (int8x8_t)__p1.val[3], (int8x8_t)__p2, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vtbx4_u8(uint8x8_t __p0, uint8x8x4_t __p1, uint8x8_t __p2) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8x4_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtbx4_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], (int8x8_t)__rev2, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vtbx4_s8(int8x8_t __p0, int8x8x4_t __p1, int8x8_t __p2) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vtbx4_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p1.val[2], (int8x8_t)__p1.val[3], (int8x8_t)__p2, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vtbx4_s8(int8x8_t __p0, int8x8x4_t __p1, int8x8_t __p2) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8x4_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vtbx4_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], (int8x8_t)__rev2, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8x2_t vtrn_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8x2_t __ret;
+  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 4);
+  return __ret;
+}
+#else
+__ai poly8x8x2_t vtrn_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8x2_t __ret;
+  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 4);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4x2_t vtrn_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x4x2_t __ret;
+  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 5);
+  return __ret;
+}
+#else
+__ai poly16x4x2_t vtrn_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  poly16x4x2_t __ret;
+  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 5);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16x2_t vtrnq_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16x2_t __ret;
+  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 36);
+  return __ret;
+}
+#else
+__ai poly8x16x2_t vtrnq_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16x2_t __ret;
+  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 36);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8x2_t vtrnq_p16(poly16x8_t __p0, poly16x8_t __p1) {
+  poly16x8x2_t __ret;
+  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 37);
+  return __ret;
+}
+#else
+__ai poly16x8x2_t vtrnq_p16(poly16x8_t __p0, poly16x8_t __p1) {
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8x2_t __ret;
+  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 37);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16x2_t vtrnq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16x2_t __ret;
+  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16x2_t vtrnq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16x2_t __ret;
+  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4x2_t vtrnq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4x2_t __ret;
+  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4x2_t vtrnq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4x2_t __ret;
+  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8x2_t vtrnq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8x2_t __ret;
+  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8x2_t vtrnq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8x2_t __ret;
+  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16x2_t vtrnq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16x2_t __ret;
+  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16x2_t vtrnq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16x2_t __ret;
+  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 32);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4x2_t vtrnq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4x2_t __ret;
+  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 41);
+  return __ret;
+}
+#else
+__ai float32x4x2_t vtrnq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4x2_t __ret;
+  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 41);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4x2_t vtrnq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4x2_t __ret;
+  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4x2_t vtrnq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4x2_t __ret;
+  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8x2_t vtrnq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8x2_t __ret;
+  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8x2_t vtrnq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8x2_t __ret;
+  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8x2_t vtrn_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8x2_t __ret;
+  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8x2_t vtrn_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8x2_t __ret;
+  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2x2_t vtrn_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2x2_t __ret;
+  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2x2_t vtrn_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2x2_t __ret;
+  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4x2_t vtrn_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4x2_t __ret;
+  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4x2_t vtrn_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4x2_t __ret;
+  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8x2_t vtrn_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8x2_t __ret;
+  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8x2_t vtrn_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8x2_t __ret;
+  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 0);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2x2_t vtrn_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2x2_t __ret;
+  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 9);
+  return __ret;
+}
+#else
+__ai float32x2x2_t vtrn_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2x2_t __ret;
+  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 9);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2x2_t vtrn_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2x2_t __ret;
+  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2x2_t vtrn_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2x2_t __ret;
+  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4x2_t vtrn_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4x2_t __ret;
+  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4x2_t vtrn_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4x2_t __ret;
+  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vtst_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vtst_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vtst_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vtst_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vtstq_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vtstq_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vtstq_p16(poly16x8_t __p0, poly16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vtstq_p16(poly16x8_t __p0, poly16x8_t __p1) {
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vtstq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vtstq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vtstq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vtstq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vtstq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vtstq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vtstq_s8(int8x16_t __p0, int8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vtstq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vtstq_s32(int32x4_t __p0, int32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vtstq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vtstq_s16(int16x8_t __p0, int16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vtstq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vtst_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vtst_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vtst_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vtst_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vtst_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vtst_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vtst_s8(int8x8_t __p0, int8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vtst_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vtst_s32(int32x2_t __p0, int32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vtst_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vtst_s16(int16x4_t __p0, int16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vtst_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8x2_t vuzp_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8x2_t __ret;
+  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 4);
+  return __ret;
+}
+#else
+__ai poly8x8x2_t vuzp_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8x2_t __ret;
+  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 4);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4x2_t vuzp_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x4x2_t __ret;
+  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 5);
+  return __ret;
+}
+#else
+__ai poly16x4x2_t vuzp_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  poly16x4x2_t __ret;
+  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 5);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16x2_t vuzpq_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16x2_t __ret;
+  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 36);
+  return __ret;
+}
+#else
+__ai poly8x16x2_t vuzpq_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16x2_t __ret;
+  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 36);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8x2_t vuzpq_p16(poly16x8_t __p0, poly16x8_t __p1) {
+  poly16x8x2_t __ret;
+  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 37);
+  return __ret;
+}
+#else
+__ai poly16x8x2_t vuzpq_p16(poly16x8_t __p0, poly16x8_t __p1) {
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8x2_t __ret;
+  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 37);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16x2_t vuzpq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16x2_t __ret;
+  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16x2_t vuzpq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16x2_t __ret;
+  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4x2_t vuzpq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4x2_t __ret;
+  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4x2_t vuzpq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4x2_t __ret;
+  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8x2_t vuzpq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8x2_t __ret;
+  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8x2_t vuzpq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8x2_t __ret;
+  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16x2_t vuzpq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16x2_t __ret;
+  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16x2_t vuzpq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16x2_t __ret;
+  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 32);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4x2_t vuzpq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4x2_t __ret;
+  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 41);
+  return __ret;
+}
+#else
+__ai float32x4x2_t vuzpq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4x2_t __ret;
+  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 41);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4x2_t vuzpq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4x2_t __ret;
+  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4x2_t vuzpq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4x2_t __ret;
+  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8x2_t vuzpq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8x2_t __ret;
+  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8x2_t vuzpq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8x2_t __ret;
+  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8x2_t vuzp_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8x2_t __ret;
+  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8x2_t vuzp_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8x2_t __ret;
+  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2x2_t vuzp_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2x2_t __ret;
+  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2x2_t vuzp_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2x2_t __ret;
+  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4x2_t vuzp_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4x2_t __ret;
+  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4x2_t vuzp_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4x2_t __ret;
+  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8x2_t vuzp_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8x2_t __ret;
+  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8x2_t vuzp_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8x2_t __ret;
+  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 0);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2x2_t vuzp_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2x2_t __ret;
+  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 9);
+  return __ret;
+}
+#else
+__ai float32x2x2_t vuzp_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2x2_t __ret;
+  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 9);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2x2_t vuzp_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2x2_t __ret;
+  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2x2_t vuzp_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2x2_t __ret;
+  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4x2_t vuzp_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4x2_t __ret;
+  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4x2_t vuzp_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4x2_t __ret;
+  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8x2_t vzip_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8x2_t __ret;
+  __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 4);
+  return __ret;
+}
+#else
+__ai poly8x8x2_t vzip_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8x2_t __ret;
+  __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 4);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4x2_t vzip_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x4x2_t __ret;
+  __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 5);
+  return __ret;
+}
+#else
+__ai poly16x4x2_t vzip_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  poly16x4x2_t __ret;
+  __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 5);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16x2_t vzipq_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16x2_t __ret;
+  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 36);
+  return __ret;
+}
+#else
+__ai poly8x16x2_t vzipq_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16x2_t __ret;
+  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 36);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8x2_t vzipq_p16(poly16x8_t __p0, poly16x8_t __p1) {
+  poly16x8x2_t __ret;
+  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 37);
+  return __ret;
+}
+#else
+__ai poly16x8x2_t vzipq_p16(poly16x8_t __p0, poly16x8_t __p1) {
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8x2_t __ret;
+  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 37);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16x2_t vzipq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16x2_t __ret;
+  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16x2_t vzipq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16x2_t __ret;
+  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4x2_t vzipq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4x2_t __ret;
+  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4x2_t vzipq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4x2_t __ret;
+  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8x2_t vzipq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8x2_t __ret;
+  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8x2_t vzipq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8x2_t __ret;
+  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16x2_t vzipq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16x2_t __ret;
+  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16x2_t vzipq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16x2_t __ret;
+  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 32);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4x2_t vzipq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4x2_t __ret;
+  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 41);
+  return __ret;
+}
+#else
+__ai float32x4x2_t vzipq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4x2_t __ret;
+  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 41);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4x2_t vzipq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4x2_t __ret;
+  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4x2_t vzipq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4x2_t __ret;
+  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8x2_t vzipq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8x2_t __ret;
+  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8x2_t vzipq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8x2_t __ret;
+  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8x2_t vzip_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8x2_t __ret;
+  __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8x2_t vzip_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8x2_t __ret;
+  __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2x2_t vzip_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2x2_t __ret;
+  __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2x2_t vzip_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2x2_t __ret;
+  __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4x2_t vzip_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4x2_t __ret;
+  __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4x2_t vzip_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4x2_t __ret;
+  __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8x2_t vzip_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8x2_t __ret;
+  __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8x2_t vzip_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8x2_t __ret;
+  __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 0);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2x2_t vzip_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2x2_t __ret;
+  __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 9);
+  return __ret;
+}
+#else
+__ai float32x2x2_t vzip_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2x2_t __ret;
+  __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 9);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2x2_t vzip_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2x2_t __ret;
+  __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2x2_t vzip_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2x2_t __ret;
+  __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4x2_t vzip_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4x2_t __ret;
+  __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4x2_t vzip_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4x2_t __ret;
+  __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#if !defined(__aarch64__)
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_p16(poly16x4_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_p16(poly16x4_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_u8(uint8x8_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_u8(uint8x8_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_u32(uint32x2_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_u32(uint32x2_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_u64(uint64x1_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_u64(uint64x1_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_u16(uint16x4_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_u16(uint16x4_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_s8(int8x8_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_s8(int8x8_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_f32(float32x2_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_f32(float32x2_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_f16(float16x4_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_f16(float16x4_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_s32(int32x2_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_s32(int32x2_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_s64(int64x1_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_s64(int64x1_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_s16(int16x4_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_s16(int16x4_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_p8(poly8x8_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_p8(poly8x8_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_u8(uint8x8_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_u8(uint8x8_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_u32(uint32x2_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_u32(uint32x2_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_u64(uint64x1_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_u64(uint64x1_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_u16(uint16x4_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_u16(uint16x4_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_s8(int8x8_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_s8(int8x8_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_f32(float32x2_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_f32(float32x2_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_f16(float16x4_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_f16(float16x4_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_s32(int32x2_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_s32(int32x2_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_s64(int64x1_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_s64(int64x1_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_s16(int16x4_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_s16(int16x4_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_p16(poly16x8_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_p16(poly16x8_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_u8(uint8x16_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_u8(uint8x16_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_u32(uint32x4_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_u32(uint32x4_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_u64(uint64x2_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_u64(uint64x2_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_u16(uint16x8_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_u16(uint16x8_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_s8(int8x16_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_s8(int8x16_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_f32(float32x4_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_f32(float32x4_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_f16(float16x8_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_f16(float16x8_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_s32(int32x4_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_s32(int32x4_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_s64(int64x2_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_s64(int64x2_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_s16(int16x8_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_s16(int16x8_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_p8(poly8x16_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_p8(poly8x16_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_u8(uint8x16_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_u8(uint8x16_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_u32(uint32x4_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_u32(uint32x4_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_u64(uint64x2_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_u64(uint64x2_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_u16(uint16x8_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_u16(uint16x8_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_s8(int8x16_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_s8(int8x16_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_f32(float32x4_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_f32(float32x4_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_f16(float16x8_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_f16(float16x8_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_s32(int32x4_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_s32(int32x4_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_s64(int64x2_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_s64(int64x2_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_s16(int16x8_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_s16(int16x8_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_p8(poly8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_p8(poly8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_p16(poly16x8_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_p16(poly16x8_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_u32(uint32x4_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_u32(uint32x4_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_u64(uint64x2_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_u64(uint64x2_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_u16(uint16x8_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_u16(uint16x8_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_s8(int8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_s8(int8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_f32(float32x4_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_f32(float32x4_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_f16(float16x8_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_f16(float16x8_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_s32(int32x4_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_s32(int32x4_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_s64(int64x2_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_s64(int64x2_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_s16(int16x8_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_s16(int16x8_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_p8(poly8x16_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_p8(poly8x16_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_p16(poly16x8_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_p16(poly16x8_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_u8(uint8x16_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_u8(uint8x16_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_u64(uint64x2_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_u64(uint64x2_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_u16(uint16x8_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_u16(uint16x8_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_s8(int8x16_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_s8(int8x16_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_f32(float32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_f32(float32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_f16(float16x8_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_f16(float16x8_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_s32(int32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_s32(int32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_s64(int64x2_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_s64(int64x2_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_s16(int16x8_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_s16(int16x8_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_p8(poly8x16_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_p8(poly8x16_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_p16(poly16x8_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_p16(poly16x8_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_u8(uint8x16_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_u8(uint8x16_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_u32(uint32x4_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_u32(uint32x4_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_u16(uint16x8_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_u16(uint16x8_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_s8(int8x16_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_s8(int8x16_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_f32(float32x4_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_f32(float32x4_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_f16(float16x8_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_f16(float16x8_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_s32(int32x4_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_s32(int32x4_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_s64(int64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_s64(int64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_s16(int16x8_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_s16(int16x8_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_p8(poly8x16_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_p8(poly8x16_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_p16(poly16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_p16(poly16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_u8(uint8x16_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_u8(uint8x16_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_u32(uint32x4_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_u32(uint32x4_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_u64(uint64x2_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_u64(uint64x2_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_s8(int8x16_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_s8(int8x16_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_f32(float32x4_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_f32(float32x4_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_f16(float16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_f16(float16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_s32(int32x4_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_s32(int32x4_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_s64(int64x2_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_s64(int64x2_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_s16(int16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_s16(int16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_p8(poly8x16_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_p8(poly8x16_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_p16(poly16x8_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_p16(poly16x8_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_u8(uint8x16_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_u8(uint8x16_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_u32(uint32x4_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_u32(uint32x4_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_u64(uint64x2_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_u64(uint64x2_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_u16(uint16x8_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_u16(uint16x8_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_f32(float32x4_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_f32(float32x4_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_f16(float16x8_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_f16(float16x8_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_s32(int32x4_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_s32(int32x4_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_s64(int64x2_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_s64(int64x2_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_s16(int16x8_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_s16(int16x8_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_p8(poly8x16_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_p8(poly8x16_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_p16(poly16x8_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_p16(poly16x8_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_u8(uint8x16_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_u8(uint8x16_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_u32(uint32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_u32(uint32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_u64(uint64x2_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_u64(uint64x2_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_u16(uint16x8_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_u16(uint16x8_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_s8(int8x16_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_s8(int8x16_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_f16(float16x8_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_f16(float16x8_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_s32(int32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_s32(int32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_s64(int64x2_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_s64(int64x2_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_s16(int16x8_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_s16(int16x8_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_p8(poly8x16_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_p8(poly8x16_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_p16(poly16x8_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_p16(poly16x8_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_u8(uint8x16_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_u8(uint8x16_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_u32(uint32x4_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_u32(uint32x4_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_u64(uint64x2_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_u64(uint64x2_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_u16(uint16x8_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_u16(uint16x8_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_s8(int8x16_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_s8(int8x16_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_f32(float32x4_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_f32(float32x4_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_s32(int32x4_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_s32(int32x4_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_s64(int64x2_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_s64(int64x2_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_s16(int16x8_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_s16(int16x8_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_p8(poly8x16_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_p8(poly8x16_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_p16(poly16x8_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_p16(poly16x8_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_u8(uint8x16_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_u8(uint8x16_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_u32(uint32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_u32(uint32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_u64(uint64x2_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_u64(uint64x2_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_u16(uint16x8_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_u16(uint16x8_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_s8(int8x16_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_s8(int8x16_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_f32(float32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_f32(float32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_f16(float16x8_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_f16(float16x8_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_s64(int64x2_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_s64(int64x2_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_s16(int16x8_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_s16(int16x8_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_p8(poly8x16_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_p8(poly8x16_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_p16(poly16x8_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_p16(poly16x8_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_u8(uint8x16_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_u8(uint8x16_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_u32(uint32x4_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_u32(uint32x4_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_u64(uint64x2_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_u64(uint64x2_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_u16(uint16x8_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_u16(uint16x8_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_s8(int8x16_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_s8(int8x16_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_f32(float32x4_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_f32(float32x4_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_f16(float16x8_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_f16(float16x8_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_s32(int32x4_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_s32(int32x4_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_s16(int16x8_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_s16(int16x8_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_p8(poly8x16_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_p8(poly8x16_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_p16(poly16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_p16(poly16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_u8(uint8x16_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_u8(uint8x16_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_u32(uint32x4_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_u32(uint32x4_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_u64(uint64x2_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_u64(uint64x2_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_u16(uint16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_u16(uint16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_s8(int8x16_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_s8(int8x16_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_f32(float32x4_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_f32(float32x4_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_f16(float16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_f16(float16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_s32(int32x4_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_s32(int32x4_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_s64(int64x2_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_s64(int64x2_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_p8(poly8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_p8(poly8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_p16(poly16x4_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_p16(poly16x4_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_u32(uint32x2_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_u32(uint32x2_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_u64(uint64x1_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_u64(uint64x1_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_u16(uint16x4_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_u16(uint16x4_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_s8(int8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_s8(int8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_f32(float32x2_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_f32(float32x2_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_f16(float16x4_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_f16(float16x4_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_s32(int32x2_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_s32(int32x2_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_s64(int64x1_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_s64(int64x1_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_s16(int16x4_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_s16(int16x4_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_p8(poly8x8_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_p8(poly8x8_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_p16(poly16x4_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_p16(poly16x4_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_u8(uint8x8_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_u8(uint8x8_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_u64(uint64x1_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_u64(uint64x1_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_u16(uint16x4_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_u16(uint16x4_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_s8(int8x8_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_s8(int8x8_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_f32(float32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_f32(float32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_f16(float16x4_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_f16(float16x4_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_s32(int32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_s32(int32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_s64(int64x1_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_s64(int64x1_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_s16(int16x4_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_s16(int16x4_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_p8(poly8x8_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_p8(poly8x8_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_p16(poly16x4_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_p16(poly16x4_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_u8(uint8x8_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_u8(uint8x8_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_u32(uint32x2_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_u32(uint32x2_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_u16(uint16x4_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_u16(uint16x4_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_s8(int8x8_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_s8(int8x8_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_f32(float32x2_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_f32(float32x2_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_f16(float16x4_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_f16(float16x4_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_s32(int32x2_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_s32(int32x2_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_s64(int64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_s64(int64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_s16(int16x4_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_s16(int16x4_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_p8(poly8x8_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_p8(poly8x8_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_p16(poly16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_p16(poly16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_u8(uint8x8_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_u8(uint8x8_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_u32(uint32x2_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_u32(uint32x2_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_u64(uint64x1_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_u64(uint64x1_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_s8(int8x8_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_s8(int8x8_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_f32(float32x2_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_f32(float32x2_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_f16(float16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_f16(float16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_s32(int32x2_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_s32(int32x2_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_s64(int64x1_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_s64(int64x1_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_s16(int16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_s16(int16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_p8(poly8x8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_p8(poly8x8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_p16(poly16x4_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_p16(poly16x4_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_u8(uint8x8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_u8(uint8x8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_u32(uint32x2_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_u32(uint32x2_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_u64(uint64x1_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_u64(uint64x1_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_u16(uint16x4_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_u16(uint16x4_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_f32(float32x2_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_f32(float32x2_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_f16(float16x4_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_f16(float16x4_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_s32(int32x2_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_s32(int32x2_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_s64(int64x1_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_s64(int64x1_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_s16(int16x4_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_s16(int16x4_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_p8(poly8x8_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_p8(poly8x8_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_p16(poly16x4_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_p16(poly16x4_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_u8(uint8x8_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_u8(uint8x8_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_u32(uint32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_u32(uint32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_u64(uint64x1_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_u64(uint64x1_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_u16(uint16x4_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_u16(uint16x4_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_s8(int8x8_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_s8(int8x8_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_f16(float16x4_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_f16(float16x4_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_s32(int32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_s32(int32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_s64(int64x1_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_s64(int64x1_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_s16(int16x4_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_s16(int16x4_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_p8(poly8x8_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_p8(poly8x8_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_p16(poly16x4_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_p16(poly16x4_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_u8(uint8x8_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_u8(uint8x8_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_u32(uint32x2_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_u32(uint32x2_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_u64(uint64x1_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_u64(uint64x1_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_u16(uint16x4_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_u16(uint16x4_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_s8(int8x8_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_s8(int8x8_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_f32(float32x2_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_f32(float32x2_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_s32(int32x2_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_s32(int32x2_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_s64(int64x1_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_s64(int64x1_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_s16(int16x4_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_s16(int16x4_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_p8(poly8x8_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_p8(poly8x8_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_p16(poly16x4_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_p16(poly16x4_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_u8(uint8x8_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_u8(uint8x8_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_u32(uint32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_u32(uint32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_u64(uint64x1_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_u64(uint64x1_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_u16(uint16x4_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_u16(uint16x4_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_s8(int8x8_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_s8(int8x8_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_f32(float32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_f32(float32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_f16(float16x4_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_f16(float16x4_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_s64(int64x1_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_s64(int64x1_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_s16(int16x4_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_s16(int16x4_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_p8(poly8x8_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_p8(poly8x8_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_p16(poly16x4_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_p16(poly16x4_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_u8(uint8x8_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_u8(uint8x8_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_u32(uint32x2_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_u32(uint32x2_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_u64(uint64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_u64(uint64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_u16(uint16x4_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_u16(uint16x4_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_s8(int8x8_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_s8(int8x8_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_f32(float32x2_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_f32(float32x2_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_f16(float16x4_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_f16(float16x4_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_s32(int32x2_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_s32(int32x2_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_s16(int16x4_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_s16(int16x4_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_p8(poly8x8_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_p8(poly8x8_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_p16(poly16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_p16(poly16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_u8(uint8x8_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_u8(uint8x8_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_u32(uint32x2_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_u32(uint32x2_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_u64(uint64x1_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_u64(uint64x1_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_u16(uint16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_u16(uint16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_s8(int8x8_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_s8(int8x8_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_f32(float32x2_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_f32(float32x2_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_f16(float16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_f16(float16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_s32(int32x2_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_s32(int32x2_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_s64(int64x1_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_s64(int64x1_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#endif
+#if __ARM_ARCH >= 8
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vcvtaq_s32_f32(float32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vcvtaq_s32_v((int8x16_t)__p0, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vcvtaq_s32_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vcvtaq_s32_v((int8x16_t)__rev0, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vcvta_s32_f32(float32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vcvta_s32_v((int8x8_t)__p0, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vcvta_s32_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vcvta_s32_v((int8x8_t)__rev0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcvtaq_u32_f32(float32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcvtaq_u32_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcvtaq_u32_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcvtaq_u32_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcvta_u32_f32(float32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcvta_u32_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcvta_u32_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcvta_u32_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vcvtmq_s32_f32(float32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vcvtmq_s32_v((int8x16_t)__p0, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vcvtmq_s32_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vcvtmq_s32_v((int8x16_t)__rev0, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vcvtm_s32_f32(float32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vcvtm_s32_v((int8x8_t)__p0, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vcvtm_s32_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vcvtm_s32_v((int8x8_t)__rev0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcvtmq_u32_f32(float32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcvtmq_u32_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcvtmq_u32_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcvtmq_u32_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcvtm_u32_f32(float32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcvtm_u32_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcvtm_u32_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcvtm_u32_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vcvtnq_s32_f32(float32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vcvtnq_s32_v((int8x16_t)__p0, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vcvtnq_s32_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vcvtnq_s32_v((int8x16_t)__rev0, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vcvtn_s32_f32(float32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vcvtn_s32_v((int8x8_t)__p0, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vcvtn_s32_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vcvtn_s32_v((int8x8_t)__rev0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcvtnq_u32_f32(float32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcvtnq_u32_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcvtnq_u32_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcvtnq_u32_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcvtn_u32_f32(float32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcvtn_u32_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcvtn_u32_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcvtn_u32_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vcvtpq_s32_f32(float32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vcvtpq_s32_v((int8x16_t)__p0, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vcvtpq_s32_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vcvtpq_s32_v((int8x16_t)__rev0, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vcvtp_s32_f32(float32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vcvtp_s32_v((int8x8_t)__p0, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vcvtp_s32_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vcvtp_s32_v((int8x8_t)__rev0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcvtpq_u32_f32(float32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcvtpq_u32_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcvtpq_u32_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcvtpq_u32_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcvtp_u32_f32(float32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcvtp_u32_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcvtp_u32_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcvtp_u32_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#endif
+#if __ARM_ARCH >= 8 && defined(__ARM_FEATURE_DIRECTED_ROUNDING)
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vrndq_f32(float32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrndq_v((int8x16_t)__p0, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vrndq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrndq_v((int8x16_t)__rev0, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vrnd_f32(float32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrnd_v((int8x8_t)__p0, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vrnd_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrnd_v((int8x8_t)__rev0, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vrndaq_f32(float32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrndaq_v((int8x16_t)__p0, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vrndaq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrndaq_v((int8x16_t)__rev0, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vrnda_f32(float32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrnda_v((int8x8_t)__p0, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vrnda_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrnda_v((int8x8_t)__rev0, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vrndmq_f32(float32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrndmq_v((int8x16_t)__p0, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vrndmq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrndmq_v((int8x16_t)__rev0, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vrndm_f32(float32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrndm_v((int8x8_t)__p0, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vrndm_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrndm_v((int8x8_t)__rev0, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vrndnq_f32(float32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrndnq_v((int8x16_t)__p0, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vrndnq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrndnq_v((int8x16_t)__rev0, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vrndn_f32(float32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrndn_v((int8x8_t)__p0, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vrndn_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrndn_v((int8x8_t)__rev0, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vrndpq_f32(float32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrndpq_v((int8x16_t)__p0, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vrndpq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrndpq_v((int8x16_t)__rev0, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vrndp_f32(float32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrndp_v((int8x8_t)__p0, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vrndp_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrndp_v((int8x8_t)__rev0, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vrndxq_f32(float32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrndxq_v((int8x16_t)__p0, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vrndxq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrndxq_v((int8x16_t)__rev0, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vrndx_f32(float32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrndx_v((int8x8_t)__p0, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vrndx_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrndx_v((int8x8_t)__rev0, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#endif
+#if __ARM_ARCH >= 8 && defined(__ARM_FEATURE_NUMERIC_MAXMIN)
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vmaxnmq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vmaxnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vmaxnmq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vmaxnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vmaxnm_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vmaxnm_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vmaxnm_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vmaxnm_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vminnmq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vminnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vminnmq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vminnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vminnm_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vminnm_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vminnm_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vminnm_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#endif
+#if __ARM_ARCH >= 8 && defined(__aarch64__)
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vcvtaq_s64_f64(float64x2_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vcvtaq_s64_v((int8x16_t)__p0, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vcvtaq_s64_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vcvtaq_s64_v((int8x16_t)__rev0, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vcvta_s64_f64(float64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vcvta_s64_v((int8x8_t)__p0, 3);
+  return __ret;
+}
+#else
+__ai int64x1_t vcvta_s64_f64(float64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vcvta_s64_v((int8x8_t)__p0, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcvtaq_u64_f64(float64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcvtaq_u64_v((int8x16_t)__p0, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcvtaq_u64_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcvtaq_u64_v((int8x16_t)__rev0, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcvta_u64_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcvta_u64_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcvta_u64_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcvta_u64_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vcvtmq_s64_f64(float64x2_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vcvtmq_s64_v((int8x16_t)__p0, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vcvtmq_s64_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vcvtmq_s64_v((int8x16_t)__rev0, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vcvtm_s64_f64(float64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vcvtm_s64_v((int8x8_t)__p0, 3);
+  return __ret;
+}
+#else
+__ai int64x1_t vcvtm_s64_f64(float64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vcvtm_s64_v((int8x8_t)__p0, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcvtmq_u64_f64(float64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcvtmq_u64_v((int8x16_t)__p0, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcvtmq_u64_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcvtmq_u64_v((int8x16_t)__rev0, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcvtm_u64_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcvtm_u64_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcvtm_u64_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcvtm_u64_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vcvtnq_s64_f64(float64x2_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vcvtnq_s64_v((int8x16_t)__p0, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vcvtnq_s64_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vcvtnq_s64_v((int8x16_t)__rev0, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vcvtn_s64_f64(float64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vcvtn_s64_v((int8x8_t)__p0, 3);
+  return __ret;
+}
+#else
+__ai int64x1_t vcvtn_s64_f64(float64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vcvtn_s64_v((int8x8_t)__p0, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcvtnq_u64_f64(float64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcvtnq_u64_v((int8x16_t)__p0, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcvtnq_u64_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcvtnq_u64_v((int8x16_t)__rev0, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcvtn_u64_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcvtn_u64_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcvtn_u64_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcvtn_u64_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vcvtpq_s64_f64(float64x2_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vcvtpq_s64_v((int8x16_t)__p0, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vcvtpq_s64_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vcvtpq_s64_v((int8x16_t)__rev0, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vcvtp_s64_f64(float64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vcvtp_s64_v((int8x8_t)__p0, 3);
+  return __ret;
+}
+#else
+__ai int64x1_t vcvtp_s64_f64(float64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vcvtp_s64_v((int8x8_t)__p0, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcvtpq_u64_f64(float64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcvtpq_u64_v((int8x16_t)__p0, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcvtpq_u64_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcvtpq_u64_v((int8x16_t)__rev0, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcvtp_u64_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcvtp_u64_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcvtp_u64_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcvtp_u64_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_p64(poly64x1_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_p64(poly64x1_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_p16(poly16x4_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_p16(poly16x4_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_u8(uint8x8_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_u8(uint8x8_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_u32(uint32x2_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_u32(uint32x2_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_u64(uint64x1_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_u64(uint64x1_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_u16(uint16x4_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_u16(uint16x4_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_s8(int8x8_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_s8(int8x8_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_f64(float64x1_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_f64(float64x1_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_f32(float32x2_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_f32(float32x2_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_f16(float16x4_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_f16(float16x4_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_s32(int32x2_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_s32(int32x2_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_s64(int64x1_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_s64(int64x1_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vreinterpret_p8_s16(int16x4_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x8_t vreinterpret_p8_s16(int16x4_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x1_t vreinterpret_p64_p8(poly8x8_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x1_t vreinterpret_p64_p8(poly8x8_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x1_t vreinterpret_p64_p16(poly16x4_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x1_t vreinterpret_p64_p16(poly16x4_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x1_t vreinterpret_p64_u8(uint8x8_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x1_t vreinterpret_p64_u8(uint8x8_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x1_t vreinterpret_p64_u32(uint32x2_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x1_t vreinterpret_p64_u32(uint32x2_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x1_t vreinterpret_p64_u64(uint64x1_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x1_t vreinterpret_p64_u64(uint64x1_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x1_t vreinterpret_p64_u16(uint16x4_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x1_t vreinterpret_p64_u16(uint16x4_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x1_t vreinterpret_p64_s8(int8x8_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x1_t vreinterpret_p64_s8(int8x8_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x1_t vreinterpret_p64_f64(float64x1_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x1_t vreinterpret_p64_f64(float64x1_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x1_t vreinterpret_p64_f32(float32x2_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x1_t vreinterpret_p64_f32(float32x2_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x1_t vreinterpret_p64_f16(float16x4_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x1_t vreinterpret_p64_f16(float16x4_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x1_t vreinterpret_p64_s32(int32x2_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x1_t vreinterpret_p64_s32(int32x2_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x1_t vreinterpret_p64_s64(int64x1_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x1_t vreinterpret_p64_s64(int64x1_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x1_t vreinterpret_p64_s16(int16x4_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x1_t vreinterpret_p64_s16(int16x4_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_p8(poly8x8_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_p8(poly8x8_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_p64(poly64x1_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_p64(poly64x1_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_u8(uint8x8_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_u8(uint8x8_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_u32(uint32x2_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_u32(uint32x2_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_u64(uint64x1_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_u64(uint64x1_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_u16(uint16x4_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_u16(uint16x4_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_s8(int8x8_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_s8(int8x8_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_f64(float64x1_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_f64(float64x1_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_f32(float32x2_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_f32(float32x2_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_f16(float16x4_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_f16(float16x4_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_s32(int32x2_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_s32(int32x2_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_s64(int64x1_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_s64(int64x1_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vreinterpret_p16_s16(int16x4_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x4_t vreinterpret_p16_s16(int16x4_t __p0) {
+  poly16x4_t __ret;
+  __ret = (poly16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_p128(poly128_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_p128(poly128_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_p64(poly64x2_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_p64(poly64x2_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_p16(poly16x8_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_p16(poly16x8_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_u8(uint8x16_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_u8(uint8x16_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_u32(uint32x4_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_u32(uint32x4_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_u64(uint64x2_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_u64(uint64x2_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_u16(uint16x8_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_u16(uint16x8_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_s8(int8x16_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_s8(int8x16_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_f64(float64x2_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_f64(float64x2_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_f32(float32x4_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_f32(float32x4_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_f16(float16x8_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_f16(float16x8_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_s32(int32x4_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_s32(int32x4_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_s64(int64x2_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_s64(int64x2_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vreinterpretq_p8_s16(int16x8_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly8x16_t vreinterpretq_p8_s16(int16x8_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly128_t vreinterpretq_p128_p8(poly8x16_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly128_t vreinterpretq_p128_p8(poly8x16_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly128_t vreinterpretq_p128_p64(poly64x2_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly128_t vreinterpretq_p128_p64(poly64x2_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly128_t vreinterpretq_p128_p16(poly16x8_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly128_t vreinterpretq_p128_p16(poly16x8_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly128_t vreinterpretq_p128_u8(uint8x16_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly128_t vreinterpretq_p128_u8(uint8x16_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly128_t vreinterpretq_p128_u32(uint32x4_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly128_t vreinterpretq_p128_u32(uint32x4_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly128_t vreinterpretq_p128_u64(uint64x2_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly128_t vreinterpretq_p128_u64(uint64x2_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly128_t vreinterpretq_p128_u16(uint16x8_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly128_t vreinterpretq_p128_u16(uint16x8_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly128_t vreinterpretq_p128_s8(int8x16_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly128_t vreinterpretq_p128_s8(int8x16_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly128_t vreinterpretq_p128_f64(float64x2_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly128_t vreinterpretq_p128_f64(float64x2_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly128_t vreinterpretq_p128_f32(float32x4_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly128_t vreinterpretq_p128_f32(float32x4_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly128_t vreinterpretq_p128_f16(float16x8_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly128_t vreinterpretq_p128_f16(float16x8_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly128_t vreinterpretq_p128_s32(int32x4_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly128_t vreinterpretq_p128_s32(int32x4_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly128_t vreinterpretq_p128_s64(int64x2_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly128_t vreinterpretq_p128_s64(int64x2_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly128_t vreinterpretq_p128_s16(int16x8_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly128_t vreinterpretq_p128_s16(int16x8_t __p0) {
+  poly128_t __ret;
+  __ret = (poly128_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vreinterpretq_p64_p8(poly8x16_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x2_t vreinterpretq_p64_p8(poly8x16_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vreinterpretq_p64_p128(poly128_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x2_t vreinterpretq_p64_p128(poly128_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vreinterpretq_p64_p16(poly16x8_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x2_t vreinterpretq_p64_p16(poly16x8_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vreinterpretq_p64_u8(uint8x16_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x2_t vreinterpretq_p64_u8(uint8x16_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vreinterpretq_p64_u32(uint32x4_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x2_t vreinterpretq_p64_u32(uint32x4_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vreinterpretq_p64_u64(uint64x2_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x2_t vreinterpretq_p64_u64(uint64x2_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vreinterpretq_p64_u16(uint16x8_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x2_t vreinterpretq_p64_u16(uint16x8_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vreinterpretq_p64_s8(int8x16_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x2_t vreinterpretq_p64_s8(int8x16_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vreinterpretq_p64_f64(float64x2_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x2_t vreinterpretq_p64_f64(float64x2_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vreinterpretq_p64_f32(float32x4_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x2_t vreinterpretq_p64_f32(float32x4_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vreinterpretq_p64_f16(float16x8_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x2_t vreinterpretq_p64_f16(float16x8_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vreinterpretq_p64_s32(int32x4_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x2_t vreinterpretq_p64_s32(int32x4_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vreinterpretq_p64_s64(int64x2_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x2_t vreinterpretq_p64_s64(int64x2_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vreinterpretq_p64_s16(int16x8_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x2_t vreinterpretq_p64_s16(int16x8_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_p8(poly8x16_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_p8(poly8x16_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_p128(poly128_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_p128(poly128_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_p64(poly64x2_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_p64(poly64x2_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_u8(uint8x16_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_u8(uint8x16_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_u32(uint32x4_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_u32(uint32x4_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_u64(uint64x2_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_u64(uint64x2_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_u16(uint16x8_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_u16(uint16x8_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_s8(int8x16_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_s8(int8x16_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_f64(float64x2_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_f64(float64x2_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_f32(float32x4_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_f32(float32x4_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_f16(float16x8_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_f16(float16x8_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_s32(int32x4_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_s32(int32x4_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_s64(int64x2_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_s64(int64x2_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vreinterpretq_p16_s16(int16x8_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly16x8_t vreinterpretq_p16_s16(int16x8_t __p0) {
+  poly16x8_t __ret;
+  __ret = (poly16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_p8(poly8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_p8(poly8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_p128(poly128_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_p128(poly128_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_p64(poly64x2_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_p64(poly64x2_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_p16(poly16x8_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_p16(poly16x8_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_u32(uint32x4_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_u32(uint32x4_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_u64(uint64x2_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_u64(uint64x2_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_u16(uint16x8_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_u16(uint16x8_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_s8(int8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_s8(int8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_f64(float64x2_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_f64(float64x2_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_f32(float32x4_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_f32(float32x4_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_f16(float16x8_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_f16(float16x8_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_s32(int32x4_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_s32(int32x4_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_s64(int64x2_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_s64(int64x2_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vreinterpretq_u8_s16(int16x8_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x16_t vreinterpretq_u8_s16(int16x8_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_p8(poly8x16_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_p8(poly8x16_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_p128(poly128_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_p128(poly128_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_p64(poly64x2_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_p64(poly64x2_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_p16(poly16x8_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_p16(poly16x8_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_u8(uint8x16_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_u8(uint8x16_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_u64(uint64x2_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_u64(uint64x2_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_u16(uint16x8_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_u16(uint16x8_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_s8(int8x16_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_s8(int8x16_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_f64(float64x2_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_f64(float64x2_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_f32(float32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_f32(float32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_f16(float16x8_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_f16(float16x8_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_s32(int32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_s32(int32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_s64(int64x2_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_s64(int64x2_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vreinterpretq_u32_s16(int16x8_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x4_t vreinterpretq_u32_s16(int16x8_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_p8(poly8x16_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_p8(poly8x16_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_p128(poly128_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_p128(poly128_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_p64(poly64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_p64(poly64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_p16(poly16x8_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_p16(poly16x8_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_u8(uint8x16_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_u8(uint8x16_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_u32(uint32x4_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_u32(uint32x4_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_u16(uint16x8_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_u16(uint16x8_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_s8(int8x16_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_s8(int8x16_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_f64(float64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_f64(float64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_f32(float32x4_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_f32(float32x4_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_f16(float16x8_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_f16(float16x8_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_s32(int32x4_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_s32(int32x4_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_s64(int64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_s64(int64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vreinterpretq_u64_s16(int16x8_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x2_t vreinterpretq_u64_s16(int16x8_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_p8(poly8x16_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_p8(poly8x16_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_p128(poly128_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_p128(poly128_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_p64(poly64x2_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_p64(poly64x2_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_p16(poly16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_p16(poly16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_u8(uint8x16_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_u8(uint8x16_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_u32(uint32x4_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_u32(uint32x4_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_u64(uint64x2_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_u64(uint64x2_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_s8(int8x16_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_s8(int8x16_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_f64(float64x2_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_f64(float64x2_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_f32(float32x4_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_f32(float32x4_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_f16(float16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_f16(float16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_s32(int32x4_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_s32(int32x4_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_s64(int64x2_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_s64(int64x2_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vreinterpretq_u16_s16(int16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x8_t vreinterpretq_u16_s16(int16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_p8(poly8x16_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_p8(poly8x16_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_p128(poly128_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_p128(poly128_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_p64(poly64x2_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_p64(poly64x2_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_p16(poly16x8_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_p16(poly16x8_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_u8(uint8x16_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_u8(uint8x16_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_u32(uint32x4_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_u32(uint32x4_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_u64(uint64x2_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_u64(uint64x2_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_u16(uint16x8_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_u16(uint16x8_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_f64(float64x2_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_f64(float64x2_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_f32(float32x4_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_f32(float32x4_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_f16(float16x8_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_f16(float16x8_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_s32(int32x4_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_s32(int32x4_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_s64(int64x2_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_s64(int64x2_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vreinterpretq_s8_s16(int16x8_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x16_t vreinterpretq_s8_s16(int16x8_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vreinterpretq_f64_p8(poly8x16_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x2_t vreinterpretq_f64_p8(poly8x16_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vreinterpretq_f64_p128(poly128_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x2_t vreinterpretq_f64_p128(poly128_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vreinterpretq_f64_p64(poly64x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x2_t vreinterpretq_f64_p64(poly64x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vreinterpretq_f64_p16(poly16x8_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x2_t vreinterpretq_f64_p16(poly16x8_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vreinterpretq_f64_u8(uint8x16_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x2_t vreinterpretq_f64_u8(uint8x16_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vreinterpretq_f64_u32(uint32x4_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x2_t vreinterpretq_f64_u32(uint32x4_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vreinterpretq_f64_u64(uint64x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x2_t vreinterpretq_f64_u64(uint64x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vreinterpretq_f64_u16(uint16x8_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x2_t vreinterpretq_f64_u16(uint16x8_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vreinterpretq_f64_s8(int8x16_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x2_t vreinterpretq_f64_s8(int8x16_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vreinterpretq_f64_f32(float32x4_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x2_t vreinterpretq_f64_f32(float32x4_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vreinterpretq_f64_f16(float16x8_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x2_t vreinterpretq_f64_f16(float16x8_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vreinterpretq_f64_s32(int32x4_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x2_t vreinterpretq_f64_s32(int32x4_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vreinterpretq_f64_s64(int64x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x2_t vreinterpretq_f64_s64(int64x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vreinterpretq_f64_s16(int16x8_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x2_t vreinterpretq_f64_s16(int16x8_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_p8(poly8x16_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_p8(poly8x16_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_p128(poly128_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_p128(poly128_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_p64(poly64x2_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_p64(poly64x2_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_p16(poly16x8_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_p16(poly16x8_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_u8(uint8x16_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_u8(uint8x16_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_u32(uint32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_u32(uint32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_u64(uint64x2_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_u64(uint64x2_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_u16(uint16x8_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_u16(uint16x8_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_s8(int8x16_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_s8(int8x16_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_f64(float64x2_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_f64(float64x2_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_f16(float16x8_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_f16(float16x8_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_s32(int32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_s32(int32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_s64(int64x2_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_s64(int64x2_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vreinterpretq_f32_s16(int16x8_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x4_t vreinterpretq_f32_s16(int16x8_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_p8(poly8x16_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_p8(poly8x16_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_p128(poly128_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_p128(poly128_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_p64(poly64x2_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_p64(poly64x2_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_p16(poly16x8_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_p16(poly16x8_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_u8(uint8x16_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_u8(uint8x16_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_u32(uint32x4_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_u32(uint32x4_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_u64(uint64x2_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_u64(uint64x2_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_u16(uint16x8_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_u16(uint16x8_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_s8(int8x16_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_s8(int8x16_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_f64(float64x2_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_f64(float64x2_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_f32(float32x4_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_f32(float32x4_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_s32(int32x4_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_s32(int32x4_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_s64(int64x2_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_s64(int64x2_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vreinterpretq_f16_s16(int16x8_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x8_t vreinterpretq_f16_s16(int16x8_t __p0) {
+  float16x8_t __ret;
+  __ret = (float16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_p8(poly8x16_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_p8(poly8x16_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_p128(poly128_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_p128(poly128_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_p64(poly64x2_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_p64(poly64x2_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_p16(poly16x8_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_p16(poly16x8_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_u8(uint8x16_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_u8(uint8x16_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_u32(uint32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_u32(uint32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_u64(uint64x2_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_u64(uint64x2_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_u16(uint16x8_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_u16(uint16x8_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_s8(int8x16_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_s8(int8x16_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_f64(float64x2_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_f64(float64x2_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_f32(float32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_f32(float32x4_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_f16(float16x8_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_f16(float16x8_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_s64(int64x2_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_s64(int64x2_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vreinterpretq_s32_s16(int16x8_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x4_t vreinterpretq_s32_s16(int16x8_t __p0) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_p8(poly8x16_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_p8(poly8x16_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_p128(poly128_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_p128(poly128_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_p64(poly64x2_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_p64(poly64x2_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_p16(poly16x8_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_p16(poly16x8_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_u8(uint8x16_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_u8(uint8x16_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_u32(uint32x4_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_u32(uint32x4_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_u64(uint64x2_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_u64(uint64x2_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_u16(uint16x8_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_u16(uint16x8_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_s8(int8x16_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_s8(int8x16_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_f64(float64x2_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_f64(float64x2_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_f32(float32x4_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_f32(float32x4_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_f16(float16x8_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_f16(float16x8_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_s32(int32x4_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_s32(int32x4_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vreinterpretq_s64_s16(int16x8_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x2_t vreinterpretq_s64_s16(int16x8_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_p8(poly8x16_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_p8(poly8x16_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_p128(poly128_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_p128(poly128_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_p64(poly64x2_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_p64(poly64x2_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_p16(poly16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_p16(poly16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_u8(uint8x16_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_u8(uint8x16_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_u32(uint32x4_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_u32(uint32x4_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_u64(uint64x2_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_u64(uint64x2_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_u16(uint16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_u16(uint16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_s8(int8x16_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_s8(int8x16_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_f64(float64x2_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_f64(float64x2_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_f32(float32x4_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_f32(float32x4_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_f16(float16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_f16(float16x8_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_s32(int32x4_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_s32(int32x4_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vreinterpretq_s16_s64(int64x2_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x8_t vreinterpretq_s16_s64(int64x2_t __p0) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_p8(poly8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_p8(poly8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_p64(poly64x1_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_p64(poly64x1_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_p16(poly16x4_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_p16(poly16x4_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_u32(uint32x2_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_u32(uint32x2_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_u64(uint64x1_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_u64(uint64x1_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_u16(uint16x4_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_u16(uint16x4_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_s8(int8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_s8(int8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_f64(float64x1_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_f64(float64x1_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_f32(float32x2_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_f32(float32x2_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_f16(float16x4_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_f16(float16x4_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_s32(int32x2_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_s32(int32x2_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_s64(int64x1_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_s64(int64x1_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vreinterpret_u8_s16(int16x4_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint8x8_t vreinterpret_u8_s16(int16x4_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_p8(poly8x8_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_p8(poly8x8_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_p64(poly64x1_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_p64(poly64x1_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_p16(poly16x4_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_p16(poly16x4_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_u8(uint8x8_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_u8(uint8x8_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_u64(uint64x1_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_u64(uint64x1_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_u16(uint16x4_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_u16(uint16x4_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_s8(int8x8_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_s8(int8x8_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_f64(float64x1_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_f64(float64x1_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_f32(float32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_f32(float32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_f16(float16x4_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_f16(float16x4_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_s32(int32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_s32(int32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_s64(int64x1_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_s64(int64x1_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vreinterpret_u32_s16(int16x4_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint32x2_t vreinterpret_u32_s16(int16x4_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_p8(poly8x8_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_p8(poly8x8_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_p64(poly64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_p64(poly64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_p16(poly16x4_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_p16(poly16x4_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_u8(uint8x8_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_u8(uint8x8_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_u32(uint32x2_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_u32(uint32x2_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_u16(uint16x4_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_u16(uint16x4_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_s8(int8x8_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_s8(int8x8_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_f32(float32x2_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_f32(float32x2_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_f16(float16x4_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_f16(float16x4_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_s32(int32x2_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_s32(int32x2_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_s64(int64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_s64(int64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vreinterpret_u64_s16(int16x4_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint64x1_t vreinterpret_u64_s16(int16x4_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_p8(poly8x8_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_p8(poly8x8_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_p64(poly64x1_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_p64(poly64x1_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_p16(poly16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_p16(poly16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_u8(uint8x8_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_u8(uint8x8_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_u32(uint32x2_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_u32(uint32x2_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_u64(uint64x1_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_u64(uint64x1_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_s8(int8x8_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_s8(int8x8_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_f64(float64x1_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_f64(float64x1_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_f32(float32x2_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_f32(float32x2_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_f16(float16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_f16(float16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_s32(int32x2_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_s32(int32x2_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_s64(int64x1_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_s64(int64x1_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vreinterpret_u16_s16(int16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai uint16x4_t vreinterpret_u16_s16(int16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_p8(poly8x8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_p8(poly8x8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_p64(poly64x1_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_p64(poly64x1_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_p16(poly16x4_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_p16(poly16x4_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_u8(uint8x8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_u8(uint8x8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_u32(uint32x2_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_u32(uint32x2_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_u64(uint64x1_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_u64(uint64x1_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_u16(uint16x4_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_u16(uint16x4_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_f64(float64x1_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_f64(float64x1_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_f32(float32x2_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_f32(float32x2_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_f16(float16x4_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_f16(float16x4_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_s32(int32x2_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_s32(int32x2_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_s64(int64x1_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_s64(int64x1_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vreinterpret_s8_s16(int16x4_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#else
+__ai int8x8_t vreinterpret_s8_s16(int16x4_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vreinterpret_f64_p8(poly8x8_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x1_t vreinterpret_f64_p8(poly8x8_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vreinterpret_f64_p64(poly64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x1_t vreinterpret_f64_p64(poly64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vreinterpret_f64_p16(poly16x4_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x1_t vreinterpret_f64_p16(poly16x4_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vreinterpret_f64_u8(uint8x8_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x1_t vreinterpret_f64_u8(uint8x8_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vreinterpret_f64_u32(uint32x2_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x1_t vreinterpret_f64_u32(uint32x2_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vreinterpret_f64_u64(uint64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x1_t vreinterpret_f64_u64(uint64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vreinterpret_f64_u16(uint16x4_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x1_t vreinterpret_f64_u16(uint16x4_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vreinterpret_f64_s8(int8x8_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x1_t vreinterpret_f64_s8(int8x8_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vreinterpret_f64_f32(float32x2_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x1_t vreinterpret_f64_f32(float32x2_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vreinterpret_f64_f16(float16x4_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x1_t vreinterpret_f64_f16(float16x4_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vreinterpret_f64_s32(int32x2_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x1_t vreinterpret_f64_s32(int32x2_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vreinterpret_f64_s64(int64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x1_t vreinterpret_f64_s64(int64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vreinterpret_f64_s16(int16x4_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x1_t vreinterpret_f64_s16(int16x4_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_p8(poly8x8_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_p8(poly8x8_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_p64(poly64x1_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_p64(poly64x1_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_p16(poly16x4_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_p16(poly16x4_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_u8(uint8x8_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_u8(uint8x8_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_u32(uint32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_u32(uint32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_u64(uint64x1_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_u64(uint64x1_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_u16(uint16x4_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_u16(uint16x4_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_s8(int8x8_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_s8(int8x8_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_f64(float64x1_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_f64(float64x1_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_f16(float16x4_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_f16(float16x4_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_s32(int32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_s32(int32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_s64(int64x1_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_s64(int64x1_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vreinterpret_f32_s16(int16x4_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai float32x2_t vreinterpret_f32_s16(int16x4_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_p8(poly8x8_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_p8(poly8x8_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_p64(poly64x1_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_p64(poly64x1_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_p16(poly16x4_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_p16(poly16x4_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_u8(uint8x8_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_u8(uint8x8_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_u32(uint32x2_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_u32(uint32x2_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_u64(uint64x1_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_u64(uint64x1_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_u16(uint16x4_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_u16(uint16x4_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_s8(int8x8_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_s8(int8x8_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_f64(float64x1_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_f64(float64x1_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_f32(float32x2_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_f32(float32x2_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_s32(int32x2_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_s32(int32x2_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_s64(int64x1_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_s64(int64x1_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x4_t vreinterpret_f16_s16(int16x4_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai float16x4_t vreinterpret_f16_s16(int16x4_t __p0) {
+  float16x4_t __ret;
+  __ret = (float16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_p8(poly8x8_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_p8(poly8x8_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_p64(poly64x1_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_p64(poly64x1_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_p16(poly16x4_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_p16(poly16x4_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_u8(uint8x8_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_u8(uint8x8_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_u32(uint32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_u32(uint32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_u64(uint64x1_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_u64(uint64x1_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_u16(uint16x4_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_u16(uint16x4_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_s8(int8x8_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_s8(int8x8_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_f64(float64x1_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_f64(float64x1_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_f32(float32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_f32(float32x2_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_f16(float16x4_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_f16(float16x4_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_s64(int64x1_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_s64(int64x1_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vreinterpret_s32_s16(int16x4_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#else
+__ai int32x2_t vreinterpret_s32_s16(int16x4_t __p0) {
+  int32x2_t __ret;
+  __ret = (int32x2_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_p8(poly8x8_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_p8(poly8x8_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_p64(poly64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_p64(poly64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_p16(poly16x4_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_p16(poly16x4_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_u8(uint8x8_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_u8(uint8x8_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_u32(uint32x2_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_u32(uint32x2_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_u64(uint64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_u64(uint64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_u16(uint16x4_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_u16(uint16x4_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_s8(int8x8_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_s8(int8x8_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_f64(float64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_f64(float64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_f32(float32x2_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_f32(float32x2_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_f16(float16x4_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_f16(float16x4_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_s32(int32x2_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_s32(int32x2_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vreinterpret_s64_s16(int16x4_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai int64x1_t vreinterpret_s64_s16(int16x4_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_p8(poly8x8_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_p8(poly8x8_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_p64(poly64x1_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_p64(poly64x1_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_p16(poly16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_p16(poly16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_u8(uint8x8_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_u8(uint8x8_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_u32(uint32x2_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_u32(uint32x2_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_u64(uint64x1_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_u64(uint64x1_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_u16(uint16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_u16(uint16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_s8(int8x8_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_s8(int8x8_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_f64(float64x1_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_f64(float64x1_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_f32(float32x2_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_f32(float32x2_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_f16(float16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_f16(float16x4_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_s32(int32x2_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_s32(int32x2_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vreinterpret_s16_s64(int64x1_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#else
+__ai int16x4_t vreinterpret_s16_s64(int64x1_t __p0) {
+  int16x4_t __ret;
+  __ret = (int16x4_t)(__p0);
+  return __ret;
+}
+#endif
+
+#endif
+#if __ARM_ARCH >= 8 && defined(__aarch64__) && defined(__ARM_FEATURE_DIRECTED_ROUNDING)
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vrndq_f64(float64x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrndq_v((int8x16_t)__p0, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vrndq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrndq_v((int8x16_t)__rev0, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vrnd_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrnd_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vrnd_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrnd_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vrndaq_f64(float64x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrndaq_v((int8x16_t)__p0, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vrndaq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrndaq_v((int8x16_t)__rev0, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vrnda_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrnda_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vrnda_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrnda_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vrndiq_f64(float64x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrndiq_v((int8x16_t)__p0, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vrndiq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrndiq_v((int8x16_t)__rev0, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vrndiq_f32(float32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrndiq_v((int8x16_t)__p0, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vrndiq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vrndiq_v((int8x16_t)__rev0, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vrndi_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrndi_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vrndi_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrndi_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vrndi_f32(float32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrndi_v((int8x8_t)__p0, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vrndi_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vrndi_v((int8x8_t)__rev0, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vrndmq_f64(float64x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrndmq_v((int8x16_t)__p0, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vrndmq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrndmq_v((int8x16_t)__rev0, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vrndm_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrndm_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vrndm_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrndm_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vrndnq_f64(float64x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrndnq_v((int8x16_t)__p0, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vrndnq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrndnq_v((int8x16_t)__rev0, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vrndn_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrndn_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vrndn_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrndn_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vrndpq_f64(float64x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrndpq_v((int8x16_t)__p0, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vrndpq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrndpq_v((int8x16_t)__rev0, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vrndp_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrndp_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vrndp_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrndp_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vrndxq_f64(float64x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrndxq_v((int8x16_t)__p0, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vrndxq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrndxq_v((int8x16_t)__rev0, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vrndx_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrndx_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vrndx_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrndx_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#endif
+
+#endif
+#if __ARM_ARCH >= 8 && defined(__aarch64__) && defined(__ARM_FEATURE_NUMERIC_MAXMIN)
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vmaxnmq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vmaxnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vmaxnmq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vmaxnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vmaxnm_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vmaxnm_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vmaxnm_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vmaxnm_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vminnmq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vminnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vminnmq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vminnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vminnm_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vminnm_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vminnm_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vminnm_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
+  return __ret;
+}
+#endif
+
+#endif
+#if __ARM_FEATURE_CRYPTO
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vaesdq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vaesdq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vaesdq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vaesdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vaeseq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vaeseq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vaeseq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vaeseq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vaesimcq_u8(uint8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vaesimcq_v((int8x16_t)__p0, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vaesimcq_u8(uint8x16_t __p0) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vaesimcq_v((int8x16_t)__rev0, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vaesmcq_u8(uint8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vaesmcq_v((int8x16_t)__p0, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vaesmcq_u8(uint8x16_t __p0) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vaesmcq_v((int8x16_t)__rev0, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vsha1cq_u32(uint32x4_t __p0, uint32_t __p1, uint32x4_t __p2) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vsha1cq_u32((int8x16_t)__p0, __p1, (int8x16_t)__p2);
+  return __ret;
+}
+#else
+__ai uint32x4_t vsha1cq_u32(uint32x4_t __p0, uint32_t __p1, uint32x4_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vsha1cq_u32((int8x16_t)__rev0, __p1, (int8x16_t)__rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vsha1h_u32(uint32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vsha1h_u32(__p0);
+  return __ret;
+}
+#else
+__ai uint32_t vsha1h_u32(uint32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vsha1h_u32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vsha1mq_u32(uint32x4_t __p0, uint32_t __p1, uint32x4_t __p2) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vsha1mq_u32((int8x16_t)__p0, __p1, (int8x16_t)__p2);
+  return __ret;
+}
+#else
+__ai uint32x4_t vsha1mq_u32(uint32x4_t __p0, uint32_t __p1, uint32x4_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vsha1mq_u32((int8x16_t)__rev0, __p1, (int8x16_t)__rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vsha1pq_u32(uint32x4_t __p0, uint32_t __p1, uint32x4_t __p2) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vsha1pq_u32((int8x16_t)__p0, __p1, (int8x16_t)__p2);
+  return __ret;
+}
+#else
+__ai uint32x4_t vsha1pq_u32(uint32x4_t __p0, uint32_t __p1, uint32x4_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vsha1pq_u32((int8x16_t)__rev0, __p1, (int8x16_t)__rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vsha1su0q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vsha1su0q_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vsha1su0q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vsha1su0q_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vsha1su1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vsha1su1q_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vsha1su1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vsha1su1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vsha256hq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vsha256hq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vsha256hq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vsha256hq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vsha256h2q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vsha256h2q_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vsha256h2q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vsha256h2q_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vsha256su0q_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vsha256su0q_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vsha256su0q_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vsha256su0q_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vsha256su1q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vsha256su1q_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vsha256su1q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vsha256su1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#endif
+#if defined(__aarch64__)
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vabdq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vabdq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vabd_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vabd_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64_t vabdd_f64(float64_t __p0, float64_t __p1) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vabdd_f64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai float64_t vabdd_f64(float64_t __p0, float64_t __p1) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vabdd_f64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vabds_f32(float32_t __p0, float32_t __p1) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vabds_f32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai float32_t vabds_f32(float32_t __p0, float32_t __p1) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vabds_f32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vabsq_f64(float64x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vabsq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vabsq_s64(int64x2_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vabsq_s64(int64x2_t __p0) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vabs_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vabs_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vabs_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vabs_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vabs_s64(int64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vabs_v((int8x8_t)__p0, 3);
+  return __ret;
+}
+#else
+__ai int64x1_t vabs_s64(int64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vabs_v((int8x8_t)__p0, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vabsd_s64(int64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vabsd_s64(__p0);
+  return __ret;
+}
+#else
+__ai int64_t vabsd_s64(int64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vabsd_s64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vaddq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#else
+__ai float64x2_t vaddq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = __rev0 + __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vadd_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#else
+__ai float64x1_t vadd_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = __p0 + __p1;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vaddd_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vaddd_u64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vaddd_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vaddd_u64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vaddd_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vaddd_s64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int64_t vaddd_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vaddd_s64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vaddhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint16x8_t __ret;
+  __ret = vcombine_u16(__p0, vaddhn_u32(__p1, __p2));
+  return __ret;
+}
+#else
+__ai uint16x8_t vaddhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __noswap_vcombine_u16(__rev0, __noswap_vaddhn_u32(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vaddhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
+  uint32x4_t __ret;
+  __ret = vcombine_u32(__p0, vaddhn_u64(__p1, __p2));
+  return __ret;
+}
+#else
+__ai uint32x4_t vaddhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __noswap_vcombine_u32(__rev0, __noswap_vaddhn_u64(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vaddhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint8x16_t __ret;
+  __ret = vcombine_u8(__p0, vaddhn_u16(__p1, __p2));
+  return __ret;
+}
+#else
+__ai uint8x16_t vaddhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __noswap_vcombine_u8(__rev0, __noswap_vaddhn_u16(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vaddhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int16x8_t __ret;
+  __ret = vcombine_s16(__p0, vaddhn_s32(__p1, __p2));
+  return __ret;
+}
+#else
+__ai int16x8_t vaddhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __noswap_vcombine_s16(__rev0, __noswap_vaddhn_s32(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vaddhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
+  int32x4_t __ret;
+  __ret = vcombine_s32(__p0, vaddhn_s64(__p1, __p2));
+  return __ret;
+}
+#else
+__ai int32x4_t vaddhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vcombine_s32(__rev0, __noswap_vaddhn_s64(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vaddhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int8x16_t __ret;
+  __ret = vcombine_s8(__p0, vaddhn_s16(__p1, __p2));
+  return __ret;
+}
+#else
+__ai int8x16_t vaddhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __noswap_vcombine_s8(__rev0, __noswap_vaddhn_s16(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16_t vaddlvq_u8(uint8x16_t __p0) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vaddlvq_u8((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai uint16_t vaddlvq_u8(uint8x16_t __p0) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vaddlvq_u8((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vaddlvq_u32(uint32x4_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vaddlvq_u32((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai uint64_t vaddlvq_u32(uint32x4_t __p0) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vaddlvq_u32((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vaddlvq_u16(uint16x8_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vaddlvq_u16((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai uint32_t vaddlvq_u16(uint16x8_t __p0) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vaddlvq_u16((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16_t vaddlvq_s8(int8x16_t __p0) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vaddlvq_s8((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai int16_t vaddlvq_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vaddlvq_s8((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vaddlvq_s32(int32x4_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vaddlvq_s32((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai int64_t vaddlvq_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vaddlvq_s32((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vaddlvq_s16(int16x8_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vaddlvq_s16((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai int32_t vaddlvq_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vaddlvq_s16((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16_t vaddlv_u8(uint8x8_t __p0) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vaddlv_u8((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai uint16_t vaddlv_u8(uint8x8_t __p0) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vaddlv_u8((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vaddlv_u32(uint32x2_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vaddlv_u32((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai uint64_t vaddlv_u32(uint32x2_t __p0) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vaddlv_u32((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vaddlv_u16(uint16x4_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vaddlv_u16((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai uint32_t vaddlv_u16(uint16x4_t __p0) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vaddlv_u16((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16_t vaddlv_s8(int8x8_t __p0) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vaddlv_s8((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai int16_t vaddlv_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vaddlv_s8((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vaddlv_s32(int32x2_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vaddlv_s32((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai int64_t vaddlv_s32(int32x2_t __p0) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vaddlv_s32((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vaddlv_s16(int16x4_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vaddlv_s16((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai int32_t vaddlv_s16(int16x4_t __p0) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vaddlv_s16((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8_t vaddvq_u8(uint8x16_t __p0) {
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vaddvq_u8((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai uint8_t vaddvq_u8(uint8x16_t __p0) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vaddvq_u8((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vaddvq_u32(uint32x4_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vaddvq_u32((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai uint32_t vaddvq_u32(uint32x4_t __p0) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vaddvq_u32((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vaddvq_u64(uint64x2_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vaddvq_u64((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai uint64_t vaddvq_u64(uint64x2_t __p0) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vaddvq_u64((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16_t vaddvq_u16(uint16x8_t __p0) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vaddvq_u16((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai uint16_t vaddvq_u16(uint16x8_t __p0) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vaddvq_u16((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8_t vaddvq_s8(int8x16_t __p0) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vaddvq_s8((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai int8_t vaddvq_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vaddvq_s8((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64_t vaddvq_f64(float64x2_t __p0) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vaddvq_f64((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai float64_t vaddvq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vaddvq_f64((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vaddvq_f32(float32x4_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vaddvq_f32((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai float32_t vaddvq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vaddvq_f32((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vaddvq_s32(int32x4_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vaddvq_s32((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai int32_t vaddvq_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vaddvq_s32((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vaddvq_s64(int64x2_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vaddvq_s64((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai int64_t vaddvq_s64(int64x2_t __p0) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vaddvq_s64((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16_t vaddvq_s16(int16x8_t __p0) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vaddvq_s16((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai int16_t vaddvq_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vaddvq_s16((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8_t vaddv_u8(uint8x8_t __p0) {
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vaddv_u8((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai uint8_t vaddv_u8(uint8x8_t __p0) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vaddv_u8((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vaddv_u32(uint32x2_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vaddv_u32((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai uint32_t vaddv_u32(uint32x2_t __p0) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vaddv_u32((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16_t vaddv_u16(uint16x4_t __p0) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vaddv_u16((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai uint16_t vaddv_u16(uint16x4_t __p0) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vaddv_u16((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8_t vaddv_s8(int8x8_t __p0) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vaddv_s8((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai int8_t vaddv_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vaddv_s8((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vaddv_f32(float32x2_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vaddv_f32((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai float32_t vaddv_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vaddv_f32((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vaddv_s32(int32x2_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vaddv_s32((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai int32_t vaddv_s32(int32x2_t __p0) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vaddv_s32((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16_t vaddv_s16(int16x4_t __p0) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vaddv_s16((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai int16_t vaddv_s16(int16x4_t __p0) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vaddv_s16((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x1_t vbsl_p64(uint64x1_t __p0, poly64x1_t __p1, poly64x1_t __p2) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 6);
+  return __ret;
+}
+#else
+__ai poly64x1_t vbsl_p64(uint64x1_t __p0, poly64x1_t __p1, poly64x1_t __p2) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 6);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vbslq_p64(uint64x2_t __p0, poly64x2_t __p1, poly64x2_t __p2) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 38);
+  return __ret;
+}
+#else
+__ai poly64x2_t vbslq_p64(uint64x2_t __p0, poly64x2_t __p1, poly64x2_t __p2) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  poly64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  poly64x2_t __ret;
+  __ret = (poly64x2_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 38);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vbslq_f64(uint64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vbslq_f64(uint64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vbsl_f64(uint64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vbsl_f64(uint64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcageq_f64(float64x2_t __p0, float64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcageq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcageq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcageq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcage_f64(float64x1_t __p0, float64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcage_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcage_f64(float64x1_t __p0, float64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcage_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vcaged_f64(float64_t __p0, float64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcaged_f64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vcaged_f64(float64_t __p0, float64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcaged_f64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vcages_f32(float32_t __p0, float32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcages_f32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint32_t vcages_f32(float32_t __p0, float32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcages_f32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcagtq_f64(float64x2_t __p0, float64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcagtq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcagtq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcagtq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcagt_f64(float64x1_t __p0, float64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcagt_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcagt_f64(float64x1_t __p0, float64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcagt_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vcagtd_f64(float64_t __p0, float64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcagtd_f64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vcagtd_f64(float64_t __p0, float64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcagtd_f64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vcagts_f32(float32_t __p0, float32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcagts_f32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint32_t vcagts_f32(float32_t __p0, float32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcagts_f32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcaleq_f64(float64x2_t __p0, float64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcaleq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcaleq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcaleq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcale_f64(float64x1_t __p0, float64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcale_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcale_f64(float64x1_t __p0, float64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcale_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vcaled_f64(float64_t __p0, float64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcaled_f64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vcaled_f64(float64_t __p0, float64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcaled_f64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vcales_f32(float32_t __p0, float32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcales_f32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint32_t vcales_f32(float32_t __p0, float32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcales_f32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcaltq_f64(float64x2_t __p0, float64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcaltq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcaltq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcaltq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcalt_f64(float64x1_t __p0, float64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcalt_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcalt_f64(float64x1_t __p0, float64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcalt_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vcaltd_f64(float64_t __p0, float64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcaltd_f64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vcaltd_f64(float64_t __p0, float64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcaltd_f64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vcalts_f32(float32_t __p0, float32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcalts_f32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint32_t vcalts_f32(float32_t __p0, float32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcalts_f32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vceq_p64(poly64x1_t __p0, poly64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint64x1_t vceq_p64(poly64x1_t __p0, poly64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 == __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vceqq_p64(poly64x2_t __p0, poly64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vceqq_p64(poly64x2_t __p0, poly64x2_t __p1) {
+  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__rev0 == __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vceqq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vceqq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__rev0 == __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vceqq_f64(float64x2_t __p0, float64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vceqq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__rev0 == __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vceqq_s64(int64x2_t __p0, int64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vceqq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__rev0 == __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vceq_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint64x1_t vceq_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 == __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vceq_f64(float64x1_t __p0, float64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint64x1_t vceq_f64(float64x1_t __p0, float64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 == __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vceq_s64(int64x1_t __p0, int64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 == __p1);
+  return __ret;
+}
+#else
+__ai uint64x1_t vceq_s64(int64x1_t __p0, int64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 == __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vceqd_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vceqd_u64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vceqd_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vceqd_u64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vceqd_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vceqd_s64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int64_t vceqd_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vceqd_s64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vceqd_f64(float64_t __p0, float64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vceqd_f64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vceqd_f64(float64_t __p0, float64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vceqd_f64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vceqs_f32(float32_t __p0, float32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vceqs_f32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint32_t vceqs_f32(float32_t __p0, float32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vceqs_f32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vceqz_p8(poly8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vceqz_p8(poly8x8_t __p0) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vceqz_p64(poly64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vceqz_p64(poly64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vceqz_p16(poly16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vceqz_p16(poly16x4_t __p0) {
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vceqzq_p8(poly8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vceqzq_p8(poly8x16_t __p0) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vceqzq_p64(poly64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vceqzq_p64(poly64x2_t __p0) {
+  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vceqzq_p16(poly16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vceqzq_p16(poly16x8_t __p0) {
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vceqzq_u8(uint8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vceqzq_u8(uint8x16_t __p0) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vceqzq_u32(uint32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vceqzq_u32(uint32x4_t __p0) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vceqzq_u64(uint64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vceqzq_u64(uint64x2_t __p0) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vceqzq_u16(uint16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vceqzq_u16(uint16x8_t __p0) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vceqzq_s8(int8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vceqzq_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vceqzq_f64(float64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vceqzq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vceqzq_f32(float32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vceqzq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vceqzq_s32(int32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vceqzq_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vceqzq_s64(int64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vceqzq_s64(int64x2_t __p0) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vceqzq_s16(int16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vceqzq_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vceqz_u8(uint8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vceqz_u8(uint8x8_t __p0) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vceqz_u32(uint32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vceqz_u32(uint32x2_t __p0) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vceqz_u64(uint64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vceqz_u64(uint64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vceqz_u16(uint16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vceqz_u16(uint16x4_t __p0) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vceqz_s8(int8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vceqz_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vceqz_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vceqz_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vceqz_f32(float32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vceqz_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vceqz_s32(int32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vceqz_s32(int32x2_t __p0) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vceqz_s64(int64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vceqz_s64(int64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vceqz_s16(int16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vceqz_s16(int16x4_t __p0) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vceqzd_u64(uint64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vceqzd_u64(__p0);
+  return __ret;
+}
+#else
+__ai uint64_t vceqzd_u64(uint64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vceqzd_u64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vceqzd_s64(int64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vceqzd_s64(__p0);
+  return __ret;
+}
+#else
+__ai int64_t vceqzd_s64(int64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vceqzd_s64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vceqzd_f64(float64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vceqzd_f64(__p0);
+  return __ret;
+}
+#else
+__ai uint64_t vceqzd_f64(float64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vceqzd_f64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vceqzs_f32(float32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vceqzs_f32(__p0);
+  return __ret;
+}
+#else
+__ai uint32_t vceqzs_f32(float32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vceqzs_f32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcgeq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0 >= __p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcgeq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__rev0 >= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcgeq_f64(float64x2_t __p0, float64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0 >= __p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcgeq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__rev0 >= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcgeq_s64(int64x2_t __p0, int64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0 >= __p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcgeq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__rev0 >= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcge_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 >= __p1);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcge_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 >= __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcge_f64(float64x1_t __p0, float64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 >= __p1);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcge_f64(float64x1_t __p0, float64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 >= __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcge_s64(int64x1_t __p0, int64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 >= __p1);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcge_s64(int64x1_t __p0, int64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 >= __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vcged_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcged_s64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int64_t vcged_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcged_s64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vcged_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcged_u64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vcged_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcged_u64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vcged_f64(float64_t __p0, float64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcged_f64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vcged_f64(float64_t __p0, float64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcged_f64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vcges_f32(float32_t __p0, float32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcges_f32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint32_t vcges_f32(float32_t __p0, float32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcges_f32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vcgezq_s8(int8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vcgezq_v((int8x16_t)__p0, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vcgezq_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vcgezq_v((int8x16_t)__rev0, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcgezq_f64(float64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcgezq_v((int8x16_t)__p0, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcgezq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcgezq_v((int8x16_t)__rev0, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcgezq_f32(float32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcgezq_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcgezq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcgezq_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcgezq_s32(int32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcgezq_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcgezq_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcgezq_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcgezq_s64(int64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcgezq_v((int8x16_t)__p0, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcgezq_s64(int64x2_t __p0) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcgezq_v((int8x16_t)__rev0, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vcgezq_s16(int16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vcgezq_v((int8x16_t)__p0, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vcgezq_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vcgezq_v((int8x16_t)__rev0, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vcgez_s8(int8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vcgez_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vcgez_v((int8x8_t)__rev0, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcgez_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcgez_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcgez_f32(float32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcgez_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcgez_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcgez_s32(int32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcgez_s32(int32x2_t __p0) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcgez_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcgez_s64(int64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcgez_s64(int64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vcgez_s16(int16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vcgez_s16(int16x4_t __p0) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vcgez_v((int8x8_t)__rev0, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vcgezd_s64(int64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcgezd_s64(__p0);
+  return __ret;
+}
+#else
+__ai int64_t vcgezd_s64(int64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcgezd_s64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vcgezd_f64(float64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcgezd_f64(__p0);
+  return __ret;
+}
+#else
+__ai uint64_t vcgezd_f64(float64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcgezd_f64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vcgezs_f32(float32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcgezs_f32(__p0);
+  return __ret;
+}
+#else
+__ai uint32_t vcgezs_f32(float32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcgezs_f32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcgtq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0 > __p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcgtq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__rev0 > __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcgtq_f64(float64x2_t __p0, float64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0 > __p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcgtq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__rev0 > __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcgtq_s64(int64x2_t __p0, int64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0 > __p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcgtq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__rev0 > __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcgt_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 > __p1);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcgt_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 > __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcgt_f64(float64x1_t __p0, float64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 > __p1);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcgt_f64(float64x1_t __p0, float64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 > __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcgt_s64(int64x1_t __p0, int64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 > __p1);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcgt_s64(int64x1_t __p0, int64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 > __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vcgtd_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcgtd_s64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int64_t vcgtd_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcgtd_s64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vcgtd_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcgtd_u64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vcgtd_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcgtd_u64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vcgtd_f64(float64_t __p0, float64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcgtd_f64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vcgtd_f64(float64_t __p0, float64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcgtd_f64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vcgts_f32(float32_t __p0, float32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcgts_f32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint32_t vcgts_f32(float32_t __p0, float32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcgts_f32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vcgtzq_s8(int8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vcgtzq_v((int8x16_t)__p0, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vcgtzq_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vcgtzq_v((int8x16_t)__rev0, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcgtzq_f64(float64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcgtzq_v((int8x16_t)__p0, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcgtzq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcgtzq_v((int8x16_t)__rev0, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcgtzq_f32(float32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcgtzq_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcgtzq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcgtzq_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcgtzq_s32(int32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcgtzq_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcgtzq_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcgtzq_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcgtzq_s64(int64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcgtzq_v((int8x16_t)__p0, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcgtzq_s64(int64x2_t __p0) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcgtzq_v((int8x16_t)__rev0, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vcgtzq_s16(int16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vcgtzq_v((int8x16_t)__p0, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vcgtzq_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vcgtzq_v((int8x16_t)__rev0, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vcgtz_s8(int8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vcgtz_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vcgtz_v((int8x8_t)__rev0, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcgtz_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcgtz_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcgtz_f32(float32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcgtz_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcgtz_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcgtz_s32(int32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcgtz_s32(int32x2_t __p0) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcgtz_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcgtz_s64(int64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcgtz_s64(int64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vcgtz_s16(int16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vcgtz_s16(int16x4_t __p0) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vcgtz_v((int8x8_t)__rev0, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vcgtzd_s64(int64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcgtzd_s64(__p0);
+  return __ret;
+}
+#else
+__ai int64_t vcgtzd_s64(int64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcgtzd_s64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vcgtzd_f64(float64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcgtzd_f64(__p0);
+  return __ret;
+}
+#else
+__ai uint64_t vcgtzd_f64(float64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcgtzd_f64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vcgtzs_f32(float32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcgtzs_f32(__p0);
+  return __ret;
+}
+#else
+__ai uint32_t vcgtzs_f32(float32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcgtzs_f32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcleq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0 <= __p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcleq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__rev0 <= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcleq_f64(float64x2_t __p0, float64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0 <= __p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcleq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__rev0 <= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcleq_s64(int64x2_t __p0, int64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0 <= __p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcleq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__rev0 <= __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcle_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 <= __p1);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcle_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 <= __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcle_f64(float64x1_t __p0, float64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 <= __p1);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcle_f64(float64x1_t __p0, float64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 <= __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcle_s64(int64x1_t __p0, int64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 <= __p1);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcle_s64(int64x1_t __p0, int64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 <= __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vcled_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcled_u64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vcled_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcled_u64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vcled_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcled_s64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int64_t vcled_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcled_s64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vcled_f64(float64_t __p0, float64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcled_f64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vcled_f64(float64_t __p0, float64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcled_f64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vcles_f32(float32_t __p0, float32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcles_f32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint32_t vcles_f32(float32_t __p0, float32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcles_f32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vclezq_s8(int8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vclezq_v((int8x16_t)__p0, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vclezq_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vclezq_v((int8x16_t)__rev0, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vclezq_f64(float64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vclezq_v((int8x16_t)__p0, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vclezq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vclezq_v((int8x16_t)__rev0, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vclezq_f32(float32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vclezq_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vclezq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vclezq_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vclezq_s32(int32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vclezq_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vclezq_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vclezq_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vclezq_s64(int64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vclezq_v((int8x16_t)__p0, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vclezq_s64(int64x2_t __p0) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vclezq_v((int8x16_t)__rev0, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vclezq_s16(int16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vclezq_v((int8x16_t)__p0, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vclezq_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vclezq_v((int8x16_t)__rev0, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vclez_s8(int8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vclez_v((int8x8_t)__p0, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vclez_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vclez_v((int8x8_t)__rev0, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vclez_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vclez_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vclez_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vclez_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vclez_f32(float32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vclez_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vclez_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vclez_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vclez_s32(int32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vclez_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vclez_s32(int32x2_t __p0) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vclez_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vclez_s64(int64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vclez_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vclez_s64(int64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vclez_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vclez_s16(int16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vclez_v((int8x8_t)__p0, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vclez_s16(int16x4_t __p0) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vclez_v((int8x8_t)__rev0, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vclezd_s64(int64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vclezd_s64(__p0);
+  return __ret;
+}
+#else
+__ai int64_t vclezd_s64(int64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vclezd_s64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vclezd_f64(float64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vclezd_f64(__p0);
+  return __ret;
+}
+#else
+__ai uint64_t vclezd_f64(float64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vclezd_f64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vclezs_f32(float32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vclezs_f32(__p0);
+  return __ret;
+}
+#else
+__ai uint32_t vclezs_f32(float32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vclezs_f32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcltq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0 < __p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcltq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__rev0 < __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcltq_f64(float64x2_t __p0, float64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0 < __p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcltq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__rev0 < __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcltq_s64(int64x2_t __p0, int64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__p0 < __p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcltq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__rev0 < __rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vclt_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 < __p1);
+  return __ret;
+}
+#else
+__ai uint64x1_t vclt_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 < __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vclt_f64(float64x1_t __p0, float64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 < __p1);
+  return __ret;
+}
+#else
+__ai uint64x1_t vclt_f64(float64x1_t __p0, float64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 < __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vclt_s64(int64x1_t __p0, int64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 < __p1);
+  return __ret;
+}
+#else
+__ai uint64x1_t vclt_s64(int64x1_t __p0, int64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t)(__p0 < __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vcltd_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcltd_u64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vcltd_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcltd_u64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vcltd_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcltd_s64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int64_t vcltd_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcltd_s64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vcltd_f64(float64_t __p0, float64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcltd_f64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vcltd_f64(float64_t __p0, float64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcltd_f64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vclts_f32(float32_t __p0, float32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vclts_f32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint32_t vclts_f32(float32_t __p0, float32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vclts_f32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vcltzq_s8(int8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vcltzq_v((int8x16_t)__p0, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vcltzq_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vcltzq_v((int8x16_t)__rev0, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcltzq_f64(float64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcltzq_v((int8x16_t)__p0, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcltzq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcltzq_v((int8x16_t)__rev0, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcltzq_f32(float32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcltzq_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcltzq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcltzq_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vcltzq_s32(int32x4_t __p0) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcltzq_v((int8x16_t)__p0, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vcltzq_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vcltzq_v((int8x16_t)__rev0, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcltzq_s64(int64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcltzq_v((int8x16_t)__p0, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcltzq_s64(int64x2_t __p0) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcltzq_v((int8x16_t)__rev0, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vcltzq_s16(int16x8_t __p0) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vcltzq_v((int8x16_t)__p0, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vcltzq_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vcltzq_v((int8x16_t)__rev0, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vcltz_s8(int8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vcltz_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vcltz_v((int8x8_t)__rev0, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcltz_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcltz_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcltz_f32(float32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcltz_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcltz_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vcltz_s32(int32x2_t __p0) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vcltz_s32(int32x2_t __p0) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vcltz_v((int8x8_t)__rev0, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcltz_s64(int64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcltz_s64(int64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vcltz_s16(int16x4_t __p0) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vcltz_s16(int16x4_t __p0) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vcltz_v((int8x8_t)__rev0, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vcltzd_s64(int64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcltzd_s64(__p0);
+  return __ret;
+}
+#else
+__ai int64_t vcltzd_s64(int64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcltzd_s64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vcltzd_f64(float64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcltzd_f64(__p0);
+  return __ret;
+}
+#else
+__ai uint64_t vcltzd_f64(float64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcltzd_f64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vcltzs_f32(float32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcltzs_f32(__p0);
+  return __ret;
+}
+#else
+__ai uint32_t vcltzs_f32(float32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcltzs_f32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vcombine_p64(poly64x1_t __p0, poly64x1_t __p1) {
+  poly64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
+  return __ret;
+}
+#else
+__ai poly64x2_t vcombine_p64(poly64x1_t __p0, poly64x1_t __p1) {
+  poly64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vcombine_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
+  return __ret;
+}
+#else
+__ai float64x2_t vcombine_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_lane_p8(__p0_0, __p1_0, __p2_0, __p3_0) __extension__ ({ \
+  poly8x16_t __s0_0 = __p0_0; \
+  poly8x8_t __s2_0 = __p2_0; \
+  poly8x16_t __ret_0; \
+  __ret_0 = vsetq_lane_p8(vget_lane_p8(__s2_0, __p3_0), __s0_0, __p1_0); \
+  __ret_0; \
+})
+#else
+#define vcopyq_lane_p8(__p0_1, __p1_1, __p2_1, __p3_1) __extension__ ({ \
+  poly8x16_t __s0_1 = __p0_1; \
+  poly8x8_t __s2_1 = __p2_1; \
+  poly8x16_t __rev0_1;  __rev0_1 = __builtin_shufflevector(__s0_1, __s0_1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x8_t __rev2_1;  __rev2_1 = __builtin_shufflevector(__s2_1, __s2_1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16_t __ret_1; \
+  __ret_1 = __noswap_vsetq_lane_p8(__noswap_vget_lane_p8(__rev2_1, __p3_1), __rev0_1, __p1_1); \
+  __ret_1 = __builtin_shufflevector(__ret_1, __ret_1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_1; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_lane_p16(__p0_2, __p1_2, __p2_2, __p3_2) __extension__ ({ \
+  poly16x8_t __s0_2 = __p0_2; \
+  poly16x4_t __s2_2 = __p2_2; \
+  poly16x8_t __ret_2; \
+  __ret_2 = vsetq_lane_p16(vget_lane_p16(__s2_2, __p3_2), __s0_2, __p1_2); \
+  __ret_2; \
+})
+#else
+#define vcopyq_lane_p16(__p0_3, __p1_3, __p2_3, __p3_3) __extension__ ({ \
+  poly16x8_t __s0_3 = __p0_3; \
+  poly16x4_t __s2_3 = __p2_3; \
+  poly16x8_t __rev0_3;  __rev0_3 = __builtin_shufflevector(__s0_3, __s0_3, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x4_t __rev2_3;  __rev2_3 = __builtin_shufflevector(__s2_3, __s2_3, 3, 2, 1, 0); \
+  poly16x8_t __ret_3; \
+  __ret_3 = __noswap_vsetq_lane_p16(__noswap_vget_lane_p16(__rev2_3, __p3_3), __rev0_3, __p1_3); \
+  __ret_3 = __builtin_shufflevector(__ret_3, __ret_3, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_3; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_lane_u8(__p0_4, __p1_4, __p2_4, __p3_4) __extension__ ({ \
+  uint8x16_t __s0_4 = __p0_4; \
+  uint8x8_t __s2_4 = __p2_4; \
+  uint8x16_t __ret_4; \
+  __ret_4 = vsetq_lane_u8(vget_lane_u8(__s2_4, __p3_4), __s0_4, __p1_4); \
+  __ret_4; \
+})
+#else
+#define vcopyq_lane_u8(__p0_5, __p1_5, __p2_5, __p3_5) __extension__ ({ \
+  uint8x16_t __s0_5 = __p0_5; \
+  uint8x8_t __s2_5 = __p2_5; \
+  uint8x16_t __rev0_5;  __rev0_5 = __builtin_shufflevector(__s0_5, __s0_5, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __rev2_5;  __rev2_5 = __builtin_shufflevector(__s2_5, __s2_5, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret_5; \
+  __ret_5 = __noswap_vsetq_lane_u8(__noswap_vget_lane_u8(__rev2_5, __p3_5), __rev0_5, __p1_5); \
+  __ret_5 = __builtin_shufflevector(__ret_5, __ret_5, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_5; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_lane_u32(__p0_6, __p1_6, __p2_6, __p3_6) __extension__ ({ \
+  uint32x4_t __s0_6 = __p0_6; \
+  uint32x2_t __s2_6 = __p2_6; \
+  uint32x4_t __ret_6; \
+  __ret_6 = vsetq_lane_u32(vget_lane_u32(__s2_6, __p3_6), __s0_6, __p1_6); \
+  __ret_6; \
+})
+#else
+#define vcopyq_lane_u32(__p0_7, __p1_7, __p2_7, __p3_7) __extension__ ({ \
+  uint32x4_t __s0_7 = __p0_7; \
+  uint32x2_t __s2_7 = __p2_7; \
+  uint32x4_t __rev0_7;  __rev0_7 = __builtin_shufflevector(__s0_7, __s0_7, 3, 2, 1, 0); \
+  uint32x2_t __rev2_7;  __rev2_7 = __builtin_shufflevector(__s2_7, __s2_7, 1, 0); \
+  uint32x4_t __ret_7; \
+  __ret_7 = __noswap_vsetq_lane_u32(__noswap_vget_lane_u32(__rev2_7, __p3_7), __rev0_7, __p1_7); \
+  __ret_7 = __builtin_shufflevector(__ret_7, __ret_7, 3, 2, 1, 0); \
+  __ret_7; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_lane_u64(__p0_8, __p1_8, __p2_8, __p3_8) __extension__ ({ \
+  uint64x2_t __s0_8 = __p0_8; \
+  uint64x1_t __s2_8 = __p2_8; \
+  uint64x2_t __ret_8; \
+  __ret_8 = vsetq_lane_u64(vget_lane_u64(__s2_8, __p3_8), __s0_8, __p1_8); \
+  __ret_8; \
+})
+#else
+#define vcopyq_lane_u64(__p0_9, __p1_9, __p2_9, __p3_9) __extension__ ({ \
+  uint64x2_t __s0_9 = __p0_9; \
+  uint64x1_t __s2_9 = __p2_9; \
+  uint64x2_t __rev0_9;  __rev0_9 = __builtin_shufflevector(__s0_9, __s0_9, 1, 0); \
+  uint64x2_t __ret_9; \
+  __ret_9 = __noswap_vsetq_lane_u64(__noswap_vget_lane_u64(__s2_9, __p3_9), __rev0_9, __p1_9); \
+  __ret_9 = __builtin_shufflevector(__ret_9, __ret_9, 1, 0); \
+  __ret_9; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_lane_u16(__p0_10, __p1_10, __p2_10, __p3_10) __extension__ ({ \
+  uint16x8_t __s0_10 = __p0_10; \
+  uint16x4_t __s2_10 = __p2_10; \
+  uint16x8_t __ret_10; \
+  __ret_10 = vsetq_lane_u16(vget_lane_u16(__s2_10, __p3_10), __s0_10, __p1_10); \
+  __ret_10; \
+})
+#else
+#define vcopyq_lane_u16(__p0_11, __p1_11, __p2_11, __p3_11) __extension__ ({ \
+  uint16x8_t __s0_11 = __p0_11; \
+  uint16x4_t __s2_11 = __p2_11; \
+  uint16x8_t __rev0_11;  __rev0_11 = __builtin_shufflevector(__s0_11, __s0_11, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x4_t __rev2_11;  __rev2_11 = __builtin_shufflevector(__s2_11, __s2_11, 3, 2, 1, 0); \
+  uint16x8_t __ret_11; \
+  __ret_11 = __noswap_vsetq_lane_u16(__noswap_vget_lane_u16(__rev2_11, __p3_11), __rev0_11, __p1_11); \
+  __ret_11 = __builtin_shufflevector(__ret_11, __ret_11, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_11; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_lane_s8(__p0_12, __p1_12, __p2_12, __p3_12) __extension__ ({ \
+  int8x16_t __s0_12 = __p0_12; \
+  int8x8_t __s2_12 = __p2_12; \
+  int8x16_t __ret_12; \
+  __ret_12 = vsetq_lane_s8(vget_lane_s8(__s2_12, __p3_12), __s0_12, __p1_12); \
+  __ret_12; \
+})
+#else
+#define vcopyq_lane_s8(__p0_13, __p1_13, __p2_13, __p3_13) __extension__ ({ \
+  int8x16_t __s0_13 = __p0_13; \
+  int8x8_t __s2_13 = __p2_13; \
+  int8x16_t __rev0_13;  __rev0_13 = __builtin_shufflevector(__s0_13, __s0_13, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __rev2_13;  __rev2_13 = __builtin_shufflevector(__s2_13, __s2_13, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret_13; \
+  __ret_13 = __noswap_vsetq_lane_s8(__noswap_vget_lane_s8(__rev2_13, __p3_13), __rev0_13, __p1_13); \
+  __ret_13 = __builtin_shufflevector(__ret_13, __ret_13, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_13; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_lane_f32(__p0_14, __p1_14, __p2_14, __p3_14) __extension__ ({ \
+  float32x4_t __s0_14 = __p0_14; \
+  float32x2_t __s2_14 = __p2_14; \
+  float32x4_t __ret_14; \
+  __ret_14 = vsetq_lane_f32(vget_lane_f32(__s2_14, __p3_14), __s0_14, __p1_14); \
+  __ret_14; \
+})
+#else
+#define vcopyq_lane_f32(__p0_15, __p1_15, __p2_15, __p3_15) __extension__ ({ \
+  float32x4_t __s0_15 = __p0_15; \
+  float32x2_t __s2_15 = __p2_15; \
+  float32x4_t __rev0_15;  __rev0_15 = __builtin_shufflevector(__s0_15, __s0_15, 3, 2, 1, 0); \
+  float32x2_t __rev2_15;  __rev2_15 = __builtin_shufflevector(__s2_15, __s2_15, 1, 0); \
+  float32x4_t __ret_15; \
+  __ret_15 = __noswap_vsetq_lane_f32(__noswap_vget_lane_f32(__rev2_15, __p3_15), __rev0_15, __p1_15); \
+  __ret_15 = __builtin_shufflevector(__ret_15, __ret_15, 3, 2, 1, 0); \
+  __ret_15; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_lane_s32(__p0_16, __p1_16, __p2_16, __p3_16) __extension__ ({ \
+  int32x4_t __s0_16 = __p0_16; \
+  int32x2_t __s2_16 = __p2_16; \
+  int32x4_t __ret_16; \
+  __ret_16 = vsetq_lane_s32(vget_lane_s32(__s2_16, __p3_16), __s0_16, __p1_16); \
+  __ret_16; \
+})
+#else
+#define vcopyq_lane_s32(__p0_17, __p1_17, __p2_17, __p3_17) __extension__ ({ \
+  int32x4_t __s0_17 = __p0_17; \
+  int32x2_t __s2_17 = __p2_17; \
+  int32x4_t __rev0_17;  __rev0_17 = __builtin_shufflevector(__s0_17, __s0_17, 3, 2, 1, 0); \
+  int32x2_t __rev2_17;  __rev2_17 = __builtin_shufflevector(__s2_17, __s2_17, 1, 0); \
+  int32x4_t __ret_17; \
+  __ret_17 = __noswap_vsetq_lane_s32(__noswap_vget_lane_s32(__rev2_17, __p3_17), __rev0_17, __p1_17); \
+  __ret_17 = __builtin_shufflevector(__ret_17, __ret_17, 3, 2, 1, 0); \
+  __ret_17; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_lane_s64(__p0_18, __p1_18, __p2_18, __p3_18) __extension__ ({ \
+  int64x2_t __s0_18 = __p0_18; \
+  int64x1_t __s2_18 = __p2_18; \
+  int64x2_t __ret_18; \
+  __ret_18 = vsetq_lane_s64(vget_lane_s64(__s2_18, __p3_18), __s0_18, __p1_18); \
+  __ret_18; \
+})
+#else
+#define vcopyq_lane_s64(__p0_19, __p1_19, __p2_19, __p3_19) __extension__ ({ \
+  int64x2_t __s0_19 = __p0_19; \
+  int64x1_t __s2_19 = __p2_19; \
+  int64x2_t __rev0_19;  __rev0_19 = __builtin_shufflevector(__s0_19, __s0_19, 1, 0); \
+  int64x2_t __ret_19; \
+  __ret_19 = __noswap_vsetq_lane_s64(__noswap_vget_lane_s64(__s2_19, __p3_19), __rev0_19, __p1_19); \
+  __ret_19 = __builtin_shufflevector(__ret_19, __ret_19, 1, 0); \
+  __ret_19; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_lane_s16(__p0_20, __p1_20, __p2_20, __p3_20) __extension__ ({ \
+  int16x8_t __s0_20 = __p0_20; \
+  int16x4_t __s2_20 = __p2_20; \
+  int16x8_t __ret_20; \
+  __ret_20 = vsetq_lane_s16(vget_lane_s16(__s2_20, __p3_20), __s0_20, __p1_20); \
+  __ret_20; \
+})
+#else
+#define vcopyq_lane_s16(__p0_21, __p1_21, __p2_21, __p3_21) __extension__ ({ \
+  int16x8_t __s0_21 = __p0_21; \
+  int16x4_t __s2_21 = __p2_21; \
+  int16x8_t __rev0_21;  __rev0_21 = __builtin_shufflevector(__s0_21, __s0_21, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __rev2_21;  __rev2_21 = __builtin_shufflevector(__s2_21, __s2_21, 3, 2, 1, 0); \
+  int16x8_t __ret_21; \
+  __ret_21 = __noswap_vsetq_lane_s16(__noswap_vget_lane_s16(__rev2_21, __p3_21), __rev0_21, __p1_21); \
+  __ret_21 = __builtin_shufflevector(__ret_21, __ret_21, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_21; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_p8(__p0_22, __p1_22, __p2_22, __p3_22) __extension__ ({ \
+  poly8x8_t __s0_22 = __p0_22; \
+  poly8x8_t __s2_22 = __p2_22; \
+  poly8x8_t __ret_22; \
+  __ret_22 = vset_lane_p8(vget_lane_p8(__s2_22, __p3_22), __s0_22, __p1_22); \
+  __ret_22; \
+})
+#else
+#define vcopy_lane_p8(__p0_23, __p1_23, __p2_23, __p3_23) __extension__ ({ \
+  poly8x8_t __s0_23 = __p0_23; \
+  poly8x8_t __s2_23 = __p2_23; \
+  poly8x8_t __rev0_23;  __rev0_23 = __builtin_shufflevector(__s0_23, __s0_23, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x8_t __rev2_23;  __rev2_23 = __builtin_shufflevector(__s2_23, __s2_23, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x8_t __ret_23; \
+  __ret_23 = __noswap_vset_lane_p8(__noswap_vget_lane_p8(__rev2_23, __p3_23), __rev0_23, __p1_23); \
+  __ret_23 = __builtin_shufflevector(__ret_23, __ret_23, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_23; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_p16(__p0_24, __p1_24, __p2_24, __p3_24) __extension__ ({ \
+  poly16x4_t __s0_24 = __p0_24; \
+  poly16x4_t __s2_24 = __p2_24; \
+  poly16x4_t __ret_24; \
+  __ret_24 = vset_lane_p16(vget_lane_p16(__s2_24, __p3_24), __s0_24, __p1_24); \
+  __ret_24; \
+})
+#else
+#define vcopy_lane_p16(__p0_25, __p1_25, __p2_25, __p3_25) __extension__ ({ \
+  poly16x4_t __s0_25 = __p0_25; \
+  poly16x4_t __s2_25 = __p2_25; \
+  poly16x4_t __rev0_25;  __rev0_25 = __builtin_shufflevector(__s0_25, __s0_25, 3, 2, 1, 0); \
+  poly16x4_t __rev2_25;  __rev2_25 = __builtin_shufflevector(__s2_25, __s2_25, 3, 2, 1, 0); \
+  poly16x4_t __ret_25; \
+  __ret_25 = __noswap_vset_lane_p16(__noswap_vget_lane_p16(__rev2_25, __p3_25), __rev0_25, __p1_25); \
+  __ret_25 = __builtin_shufflevector(__ret_25, __ret_25, 3, 2, 1, 0); \
+  __ret_25; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_u8(__p0_26, __p1_26, __p2_26, __p3_26) __extension__ ({ \
+  uint8x8_t __s0_26 = __p0_26; \
+  uint8x8_t __s2_26 = __p2_26; \
+  uint8x8_t __ret_26; \
+  __ret_26 = vset_lane_u8(vget_lane_u8(__s2_26, __p3_26), __s0_26, __p1_26); \
+  __ret_26; \
+})
+#else
+#define vcopy_lane_u8(__p0_27, __p1_27, __p2_27, __p3_27) __extension__ ({ \
+  uint8x8_t __s0_27 = __p0_27; \
+  uint8x8_t __s2_27 = __p2_27; \
+  uint8x8_t __rev0_27;  __rev0_27 = __builtin_shufflevector(__s0_27, __s0_27, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __rev2_27;  __rev2_27 = __builtin_shufflevector(__s2_27, __s2_27, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret_27; \
+  __ret_27 = __noswap_vset_lane_u8(__noswap_vget_lane_u8(__rev2_27, __p3_27), __rev0_27, __p1_27); \
+  __ret_27 = __builtin_shufflevector(__ret_27, __ret_27, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_27; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_u32(__p0_28, __p1_28, __p2_28, __p3_28) __extension__ ({ \
+  uint32x2_t __s0_28 = __p0_28; \
+  uint32x2_t __s2_28 = __p2_28; \
+  uint32x2_t __ret_28; \
+  __ret_28 = vset_lane_u32(vget_lane_u32(__s2_28, __p3_28), __s0_28, __p1_28); \
+  __ret_28; \
+})
+#else
+#define vcopy_lane_u32(__p0_29, __p1_29, __p2_29, __p3_29) __extension__ ({ \
+  uint32x2_t __s0_29 = __p0_29; \
+  uint32x2_t __s2_29 = __p2_29; \
+  uint32x2_t __rev0_29;  __rev0_29 = __builtin_shufflevector(__s0_29, __s0_29, 1, 0); \
+  uint32x2_t __rev2_29;  __rev2_29 = __builtin_shufflevector(__s2_29, __s2_29, 1, 0); \
+  uint32x2_t __ret_29; \
+  __ret_29 = __noswap_vset_lane_u32(__noswap_vget_lane_u32(__rev2_29, __p3_29), __rev0_29, __p1_29); \
+  __ret_29 = __builtin_shufflevector(__ret_29, __ret_29, 1, 0); \
+  __ret_29; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_u64(__p0_30, __p1_30, __p2_30, __p3_30) __extension__ ({ \
+  uint64x1_t __s0_30 = __p0_30; \
+  uint64x1_t __s2_30 = __p2_30; \
+  uint64x1_t __ret_30; \
+  __ret_30 = vset_lane_u64(vget_lane_u64(__s2_30, __p3_30), __s0_30, __p1_30); \
+  __ret_30; \
+})
+#else
+#define vcopy_lane_u64(__p0_31, __p1_31, __p2_31, __p3_31) __extension__ ({ \
+  uint64x1_t __s0_31 = __p0_31; \
+  uint64x1_t __s2_31 = __p2_31; \
+  uint64x1_t __ret_31; \
+  __ret_31 = __noswap_vset_lane_u64(__noswap_vget_lane_u64(__s2_31, __p3_31), __s0_31, __p1_31); \
+  __ret_31; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_u16(__p0_32, __p1_32, __p2_32, __p3_32) __extension__ ({ \
+  uint16x4_t __s0_32 = __p0_32; \
+  uint16x4_t __s2_32 = __p2_32; \
+  uint16x4_t __ret_32; \
+  __ret_32 = vset_lane_u16(vget_lane_u16(__s2_32, __p3_32), __s0_32, __p1_32); \
+  __ret_32; \
+})
+#else
+#define vcopy_lane_u16(__p0_33, __p1_33, __p2_33, __p3_33) __extension__ ({ \
+  uint16x4_t __s0_33 = __p0_33; \
+  uint16x4_t __s2_33 = __p2_33; \
+  uint16x4_t __rev0_33;  __rev0_33 = __builtin_shufflevector(__s0_33, __s0_33, 3, 2, 1, 0); \
+  uint16x4_t __rev2_33;  __rev2_33 = __builtin_shufflevector(__s2_33, __s2_33, 3, 2, 1, 0); \
+  uint16x4_t __ret_33; \
+  __ret_33 = __noswap_vset_lane_u16(__noswap_vget_lane_u16(__rev2_33, __p3_33), __rev0_33, __p1_33); \
+  __ret_33 = __builtin_shufflevector(__ret_33, __ret_33, 3, 2, 1, 0); \
+  __ret_33; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_s8(__p0_34, __p1_34, __p2_34, __p3_34) __extension__ ({ \
+  int8x8_t __s0_34 = __p0_34; \
+  int8x8_t __s2_34 = __p2_34; \
+  int8x8_t __ret_34; \
+  __ret_34 = vset_lane_s8(vget_lane_s8(__s2_34, __p3_34), __s0_34, __p1_34); \
+  __ret_34; \
+})
+#else
+#define vcopy_lane_s8(__p0_35, __p1_35, __p2_35, __p3_35) __extension__ ({ \
+  int8x8_t __s0_35 = __p0_35; \
+  int8x8_t __s2_35 = __p2_35; \
+  int8x8_t __rev0_35;  __rev0_35 = __builtin_shufflevector(__s0_35, __s0_35, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __rev2_35;  __rev2_35 = __builtin_shufflevector(__s2_35, __s2_35, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret_35; \
+  __ret_35 = __noswap_vset_lane_s8(__noswap_vget_lane_s8(__rev2_35, __p3_35), __rev0_35, __p1_35); \
+  __ret_35 = __builtin_shufflevector(__ret_35, __ret_35, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_35; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_f32(__p0_36, __p1_36, __p2_36, __p3_36) __extension__ ({ \
+  float32x2_t __s0_36 = __p0_36; \
+  float32x2_t __s2_36 = __p2_36; \
+  float32x2_t __ret_36; \
+  __ret_36 = vset_lane_f32(vget_lane_f32(__s2_36, __p3_36), __s0_36, __p1_36); \
+  __ret_36; \
+})
+#else
+#define vcopy_lane_f32(__p0_37, __p1_37, __p2_37, __p3_37) __extension__ ({ \
+  float32x2_t __s0_37 = __p0_37; \
+  float32x2_t __s2_37 = __p2_37; \
+  float32x2_t __rev0_37;  __rev0_37 = __builtin_shufflevector(__s0_37, __s0_37, 1, 0); \
+  float32x2_t __rev2_37;  __rev2_37 = __builtin_shufflevector(__s2_37, __s2_37, 1, 0); \
+  float32x2_t __ret_37; \
+  __ret_37 = __noswap_vset_lane_f32(__noswap_vget_lane_f32(__rev2_37, __p3_37), __rev0_37, __p1_37); \
+  __ret_37 = __builtin_shufflevector(__ret_37, __ret_37, 1, 0); \
+  __ret_37; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_s32(__p0_38, __p1_38, __p2_38, __p3_38) __extension__ ({ \
+  int32x2_t __s0_38 = __p0_38; \
+  int32x2_t __s2_38 = __p2_38; \
+  int32x2_t __ret_38; \
+  __ret_38 = vset_lane_s32(vget_lane_s32(__s2_38, __p3_38), __s0_38, __p1_38); \
+  __ret_38; \
+})
+#else
+#define vcopy_lane_s32(__p0_39, __p1_39, __p2_39, __p3_39) __extension__ ({ \
+  int32x2_t __s0_39 = __p0_39; \
+  int32x2_t __s2_39 = __p2_39; \
+  int32x2_t __rev0_39;  __rev0_39 = __builtin_shufflevector(__s0_39, __s0_39, 1, 0); \
+  int32x2_t __rev2_39;  __rev2_39 = __builtin_shufflevector(__s2_39, __s2_39, 1, 0); \
+  int32x2_t __ret_39; \
+  __ret_39 = __noswap_vset_lane_s32(__noswap_vget_lane_s32(__rev2_39, __p3_39), __rev0_39, __p1_39); \
+  __ret_39 = __builtin_shufflevector(__ret_39, __ret_39, 1, 0); \
+  __ret_39; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_s64(__p0_40, __p1_40, __p2_40, __p3_40) __extension__ ({ \
+  int64x1_t __s0_40 = __p0_40; \
+  int64x1_t __s2_40 = __p2_40; \
+  int64x1_t __ret_40; \
+  __ret_40 = vset_lane_s64(vget_lane_s64(__s2_40, __p3_40), __s0_40, __p1_40); \
+  __ret_40; \
+})
+#else
+#define vcopy_lane_s64(__p0_41, __p1_41, __p2_41, __p3_41) __extension__ ({ \
+  int64x1_t __s0_41 = __p0_41; \
+  int64x1_t __s2_41 = __p2_41; \
+  int64x1_t __ret_41; \
+  __ret_41 = __noswap_vset_lane_s64(__noswap_vget_lane_s64(__s2_41, __p3_41), __s0_41, __p1_41); \
+  __ret_41; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_s16(__p0_42, __p1_42, __p2_42, __p3_42) __extension__ ({ \
+  int16x4_t __s0_42 = __p0_42; \
+  int16x4_t __s2_42 = __p2_42; \
+  int16x4_t __ret_42; \
+  __ret_42 = vset_lane_s16(vget_lane_s16(__s2_42, __p3_42), __s0_42, __p1_42); \
+  __ret_42; \
+})
+#else
+#define vcopy_lane_s16(__p0_43, __p1_43, __p2_43, __p3_43) __extension__ ({ \
+  int16x4_t __s0_43 = __p0_43; \
+  int16x4_t __s2_43 = __p2_43; \
+  int16x4_t __rev0_43;  __rev0_43 = __builtin_shufflevector(__s0_43, __s0_43, 3, 2, 1, 0); \
+  int16x4_t __rev2_43;  __rev2_43 = __builtin_shufflevector(__s2_43, __s2_43, 3, 2, 1, 0); \
+  int16x4_t __ret_43; \
+  __ret_43 = __noswap_vset_lane_s16(__noswap_vget_lane_s16(__rev2_43, __p3_43), __rev0_43, __p1_43); \
+  __ret_43 = __builtin_shufflevector(__ret_43, __ret_43, 3, 2, 1, 0); \
+  __ret_43; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_laneq_p8(__p0_44, __p1_44, __p2_44, __p3_44) __extension__ ({ \
+  poly8x16_t __s0_44 = __p0_44; \
+  poly8x16_t __s2_44 = __p2_44; \
+  poly8x16_t __ret_44; \
+  __ret_44 = vsetq_lane_p8(vgetq_lane_p8(__s2_44, __p3_44), __s0_44, __p1_44); \
+  __ret_44; \
+})
+#else
+#define vcopyq_laneq_p8(__p0_45, __p1_45, __p2_45, __p3_45) __extension__ ({ \
+  poly8x16_t __s0_45 = __p0_45; \
+  poly8x16_t __s2_45 = __p2_45; \
+  poly8x16_t __rev0_45;  __rev0_45 = __builtin_shufflevector(__s0_45, __s0_45, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16_t __rev2_45;  __rev2_45 = __builtin_shufflevector(__s2_45, __s2_45, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16_t __ret_45; \
+  __ret_45 = __noswap_vsetq_lane_p8(__noswap_vgetq_lane_p8(__rev2_45, __p3_45), __rev0_45, __p1_45); \
+  __ret_45 = __builtin_shufflevector(__ret_45, __ret_45, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_45; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_laneq_p16(__p0_46, __p1_46, __p2_46, __p3_46) __extension__ ({ \
+  poly16x8_t __s0_46 = __p0_46; \
+  poly16x8_t __s2_46 = __p2_46; \
+  poly16x8_t __ret_46; \
+  __ret_46 = vsetq_lane_p16(vgetq_lane_p16(__s2_46, __p3_46), __s0_46, __p1_46); \
+  __ret_46; \
+})
+#else
+#define vcopyq_laneq_p16(__p0_47, __p1_47, __p2_47, __p3_47) __extension__ ({ \
+  poly16x8_t __s0_47 = __p0_47; \
+  poly16x8_t __s2_47 = __p2_47; \
+  poly16x8_t __rev0_47;  __rev0_47 = __builtin_shufflevector(__s0_47, __s0_47, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x8_t __rev2_47;  __rev2_47 = __builtin_shufflevector(__s2_47, __s2_47, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x8_t __ret_47; \
+  __ret_47 = __noswap_vsetq_lane_p16(__noswap_vgetq_lane_p16(__rev2_47, __p3_47), __rev0_47, __p1_47); \
+  __ret_47 = __builtin_shufflevector(__ret_47, __ret_47, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_47; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_laneq_u8(__p0_48, __p1_48, __p2_48, __p3_48) __extension__ ({ \
+  uint8x16_t __s0_48 = __p0_48; \
+  uint8x16_t __s2_48 = __p2_48; \
+  uint8x16_t __ret_48; \
+  __ret_48 = vsetq_lane_u8(vgetq_lane_u8(__s2_48, __p3_48), __s0_48, __p1_48); \
+  __ret_48; \
+})
+#else
+#define vcopyq_laneq_u8(__p0_49, __p1_49, __p2_49, __p3_49) __extension__ ({ \
+  uint8x16_t __s0_49 = __p0_49; \
+  uint8x16_t __s2_49 = __p2_49; \
+  uint8x16_t __rev0_49;  __rev0_49 = __builtin_shufflevector(__s0_49, __s0_49, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __rev2_49;  __rev2_49 = __builtin_shufflevector(__s2_49, __s2_49, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret_49; \
+  __ret_49 = __noswap_vsetq_lane_u8(__noswap_vgetq_lane_u8(__rev2_49, __p3_49), __rev0_49, __p1_49); \
+  __ret_49 = __builtin_shufflevector(__ret_49, __ret_49, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_49; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_laneq_u32(__p0_50, __p1_50, __p2_50, __p3_50) __extension__ ({ \
+  uint32x4_t __s0_50 = __p0_50; \
+  uint32x4_t __s2_50 = __p2_50; \
+  uint32x4_t __ret_50; \
+  __ret_50 = vsetq_lane_u32(vgetq_lane_u32(__s2_50, __p3_50), __s0_50, __p1_50); \
+  __ret_50; \
+})
+#else
+#define vcopyq_laneq_u32(__p0_51, __p1_51, __p2_51, __p3_51) __extension__ ({ \
+  uint32x4_t __s0_51 = __p0_51; \
+  uint32x4_t __s2_51 = __p2_51; \
+  uint32x4_t __rev0_51;  __rev0_51 = __builtin_shufflevector(__s0_51, __s0_51, 3, 2, 1, 0); \
+  uint32x4_t __rev2_51;  __rev2_51 = __builtin_shufflevector(__s2_51, __s2_51, 3, 2, 1, 0); \
+  uint32x4_t __ret_51; \
+  __ret_51 = __noswap_vsetq_lane_u32(__noswap_vgetq_lane_u32(__rev2_51, __p3_51), __rev0_51, __p1_51); \
+  __ret_51 = __builtin_shufflevector(__ret_51, __ret_51, 3, 2, 1, 0); \
+  __ret_51; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_laneq_u64(__p0_52, __p1_52, __p2_52, __p3_52) __extension__ ({ \
+  uint64x2_t __s0_52 = __p0_52; \
+  uint64x2_t __s2_52 = __p2_52; \
+  uint64x2_t __ret_52; \
+  __ret_52 = vsetq_lane_u64(vgetq_lane_u64(__s2_52, __p3_52), __s0_52, __p1_52); \
+  __ret_52; \
+})
+#else
+#define vcopyq_laneq_u64(__p0_53, __p1_53, __p2_53, __p3_53) __extension__ ({ \
+  uint64x2_t __s0_53 = __p0_53; \
+  uint64x2_t __s2_53 = __p2_53; \
+  uint64x2_t __rev0_53;  __rev0_53 = __builtin_shufflevector(__s0_53, __s0_53, 1, 0); \
+  uint64x2_t __rev2_53;  __rev2_53 = __builtin_shufflevector(__s2_53, __s2_53, 1, 0); \
+  uint64x2_t __ret_53; \
+  __ret_53 = __noswap_vsetq_lane_u64(__noswap_vgetq_lane_u64(__rev2_53, __p3_53), __rev0_53, __p1_53); \
+  __ret_53 = __builtin_shufflevector(__ret_53, __ret_53, 1, 0); \
+  __ret_53; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_laneq_u16(__p0_54, __p1_54, __p2_54, __p3_54) __extension__ ({ \
+  uint16x8_t __s0_54 = __p0_54; \
+  uint16x8_t __s2_54 = __p2_54; \
+  uint16x8_t __ret_54; \
+  __ret_54 = vsetq_lane_u16(vgetq_lane_u16(__s2_54, __p3_54), __s0_54, __p1_54); \
+  __ret_54; \
+})
+#else
+#define vcopyq_laneq_u16(__p0_55, __p1_55, __p2_55, __p3_55) __extension__ ({ \
+  uint16x8_t __s0_55 = __p0_55; \
+  uint16x8_t __s2_55 = __p2_55; \
+  uint16x8_t __rev0_55;  __rev0_55 = __builtin_shufflevector(__s0_55, __s0_55, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev2_55;  __rev2_55 = __builtin_shufflevector(__s2_55, __s2_55, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret_55; \
+  __ret_55 = __noswap_vsetq_lane_u16(__noswap_vgetq_lane_u16(__rev2_55, __p3_55), __rev0_55, __p1_55); \
+  __ret_55 = __builtin_shufflevector(__ret_55, __ret_55, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_55; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_laneq_s8(__p0_56, __p1_56, __p2_56, __p3_56) __extension__ ({ \
+  int8x16_t __s0_56 = __p0_56; \
+  int8x16_t __s2_56 = __p2_56; \
+  int8x16_t __ret_56; \
+  __ret_56 = vsetq_lane_s8(vgetq_lane_s8(__s2_56, __p3_56), __s0_56, __p1_56); \
+  __ret_56; \
+})
+#else
+#define vcopyq_laneq_s8(__p0_57, __p1_57, __p2_57, __p3_57) __extension__ ({ \
+  int8x16_t __s0_57 = __p0_57; \
+  int8x16_t __s2_57 = __p2_57; \
+  int8x16_t __rev0_57;  __rev0_57 = __builtin_shufflevector(__s0_57, __s0_57, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __rev2_57;  __rev2_57 = __builtin_shufflevector(__s2_57, __s2_57, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret_57; \
+  __ret_57 = __noswap_vsetq_lane_s8(__noswap_vgetq_lane_s8(__rev2_57, __p3_57), __rev0_57, __p1_57); \
+  __ret_57 = __builtin_shufflevector(__ret_57, __ret_57, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_57; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_laneq_f32(__p0_58, __p1_58, __p2_58, __p3_58) __extension__ ({ \
+  float32x4_t __s0_58 = __p0_58; \
+  float32x4_t __s2_58 = __p2_58; \
+  float32x4_t __ret_58; \
+  __ret_58 = vsetq_lane_f32(vgetq_lane_f32(__s2_58, __p3_58), __s0_58, __p1_58); \
+  __ret_58; \
+})
+#else
+#define vcopyq_laneq_f32(__p0_59, __p1_59, __p2_59, __p3_59) __extension__ ({ \
+  float32x4_t __s0_59 = __p0_59; \
+  float32x4_t __s2_59 = __p2_59; \
+  float32x4_t __rev0_59;  __rev0_59 = __builtin_shufflevector(__s0_59, __s0_59, 3, 2, 1, 0); \
+  float32x4_t __rev2_59;  __rev2_59 = __builtin_shufflevector(__s2_59, __s2_59, 3, 2, 1, 0); \
+  float32x4_t __ret_59; \
+  __ret_59 = __noswap_vsetq_lane_f32(__noswap_vgetq_lane_f32(__rev2_59, __p3_59), __rev0_59, __p1_59); \
+  __ret_59 = __builtin_shufflevector(__ret_59, __ret_59, 3, 2, 1, 0); \
+  __ret_59; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_laneq_s32(__p0_60, __p1_60, __p2_60, __p3_60) __extension__ ({ \
+  int32x4_t __s0_60 = __p0_60; \
+  int32x4_t __s2_60 = __p2_60; \
+  int32x4_t __ret_60; \
+  __ret_60 = vsetq_lane_s32(vgetq_lane_s32(__s2_60, __p3_60), __s0_60, __p1_60); \
+  __ret_60; \
+})
+#else
+#define vcopyq_laneq_s32(__p0_61, __p1_61, __p2_61, __p3_61) __extension__ ({ \
+  int32x4_t __s0_61 = __p0_61; \
+  int32x4_t __s2_61 = __p2_61; \
+  int32x4_t __rev0_61;  __rev0_61 = __builtin_shufflevector(__s0_61, __s0_61, 3, 2, 1, 0); \
+  int32x4_t __rev2_61;  __rev2_61 = __builtin_shufflevector(__s2_61, __s2_61, 3, 2, 1, 0); \
+  int32x4_t __ret_61; \
+  __ret_61 = __noswap_vsetq_lane_s32(__noswap_vgetq_lane_s32(__rev2_61, __p3_61), __rev0_61, __p1_61); \
+  __ret_61 = __builtin_shufflevector(__ret_61, __ret_61, 3, 2, 1, 0); \
+  __ret_61; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_laneq_s64(__p0_62, __p1_62, __p2_62, __p3_62) __extension__ ({ \
+  int64x2_t __s0_62 = __p0_62; \
+  int64x2_t __s2_62 = __p2_62; \
+  int64x2_t __ret_62; \
+  __ret_62 = vsetq_lane_s64(vgetq_lane_s64(__s2_62, __p3_62), __s0_62, __p1_62); \
+  __ret_62; \
+})
+#else
+#define vcopyq_laneq_s64(__p0_63, __p1_63, __p2_63, __p3_63) __extension__ ({ \
+  int64x2_t __s0_63 = __p0_63; \
+  int64x2_t __s2_63 = __p2_63; \
+  int64x2_t __rev0_63;  __rev0_63 = __builtin_shufflevector(__s0_63, __s0_63, 1, 0); \
+  int64x2_t __rev2_63;  __rev2_63 = __builtin_shufflevector(__s2_63, __s2_63, 1, 0); \
+  int64x2_t __ret_63; \
+  __ret_63 = __noswap_vsetq_lane_s64(__noswap_vgetq_lane_s64(__rev2_63, __p3_63), __rev0_63, __p1_63); \
+  __ret_63 = __builtin_shufflevector(__ret_63, __ret_63, 1, 0); \
+  __ret_63; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_laneq_s16(__p0_64, __p1_64, __p2_64, __p3_64) __extension__ ({ \
+  int16x8_t __s0_64 = __p0_64; \
+  int16x8_t __s2_64 = __p2_64; \
+  int16x8_t __ret_64; \
+  __ret_64 = vsetq_lane_s16(vgetq_lane_s16(__s2_64, __p3_64), __s0_64, __p1_64); \
+  __ret_64; \
+})
+#else
+#define vcopyq_laneq_s16(__p0_65, __p1_65, __p2_65, __p3_65) __extension__ ({ \
+  int16x8_t __s0_65 = __p0_65; \
+  int16x8_t __s2_65 = __p2_65; \
+  int16x8_t __rev0_65;  __rev0_65 = __builtin_shufflevector(__s0_65, __s0_65, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev2_65;  __rev2_65 = __builtin_shufflevector(__s2_65, __s2_65, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret_65; \
+  __ret_65 = __noswap_vsetq_lane_s16(__noswap_vgetq_lane_s16(__rev2_65, __p3_65), __rev0_65, __p1_65); \
+  __ret_65 = __builtin_shufflevector(__ret_65, __ret_65, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_65; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_laneq_p8(__p0_66, __p1_66, __p2_66, __p3_66) __extension__ ({ \
+  poly8x8_t __s0_66 = __p0_66; \
+  poly8x16_t __s2_66 = __p2_66; \
+  poly8x8_t __ret_66; \
+  __ret_66 = vset_lane_p8(vgetq_lane_p8(__s2_66, __p3_66), __s0_66, __p1_66); \
+  __ret_66; \
+})
+#else
+#define vcopy_laneq_p8(__p0_67, __p1_67, __p2_67, __p3_67) __extension__ ({ \
+  poly8x8_t __s0_67 = __p0_67; \
+  poly8x16_t __s2_67 = __p2_67; \
+  poly8x8_t __rev0_67;  __rev0_67 = __builtin_shufflevector(__s0_67, __s0_67, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16_t __rev2_67;  __rev2_67 = __builtin_shufflevector(__s2_67, __s2_67, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x8_t __ret_67; \
+  __ret_67 = __noswap_vset_lane_p8(__noswap_vgetq_lane_p8(__rev2_67, __p3_67), __rev0_67, __p1_67); \
+  __ret_67 = __builtin_shufflevector(__ret_67, __ret_67, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_67; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_laneq_p16(__p0_68, __p1_68, __p2_68, __p3_68) __extension__ ({ \
+  poly16x4_t __s0_68 = __p0_68; \
+  poly16x8_t __s2_68 = __p2_68; \
+  poly16x4_t __ret_68; \
+  __ret_68 = vset_lane_p16(vgetq_lane_p16(__s2_68, __p3_68), __s0_68, __p1_68); \
+  __ret_68; \
+})
+#else
+#define vcopy_laneq_p16(__p0_69, __p1_69, __p2_69, __p3_69) __extension__ ({ \
+  poly16x4_t __s0_69 = __p0_69; \
+  poly16x8_t __s2_69 = __p2_69; \
+  poly16x4_t __rev0_69;  __rev0_69 = __builtin_shufflevector(__s0_69, __s0_69, 3, 2, 1, 0); \
+  poly16x8_t __rev2_69;  __rev2_69 = __builtin_shufflevector(__s2_69, __s2_69, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x4_t __ret_69; \
+  __ret_69 = __noswap_vset_lane_p16(__noswap_vgetq_lane_p16(__rev2_69, __p3_69), __rev0_69, __p1_69); \
+  __ret_69 = __builtin_shufflevector(__ret_69, __ret_69, 3, 2, 1, 0); \
+  __ret_69; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_laneq_u8(__p0_70, __p1_70, __p2_70, __p3_70) __extension__ ({ \
+  uint8x8_t __s0_70 = __p0_70; \
+  uint8x16_t __s2_70 = __p2_70; \
+  uint8x8_t __ret_70; \
+  __ret_70 = vset_lane_u8(vgetq_lane_u8(__s2_70, __p3_70), __s0_70, __p1_70); \
+  __ret_70; \
+})
+#else
+#define vcopy_laneq_u8(__p0_71, __p1_71, __p2_71, __p3_71) __extension__ ({ \
+  uint8x8_t __s0_71 = __p0_71; \
+  uint8x16_t __s2_71 = __p2_71; \
+  uint8x8_t __rev0_71;  __rev0_71 = __builtin_shufflevector(__s0_71, __s0_71, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __rev2_71;  __rev2_71 = __builtin_shufflevector(__s2_71, __s2_71, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret_71; \
+  __ret_71 = __noswap_vset_lane_u8(__noswap_vgetq_lane_u8(__rev2_71, __p3_71), __rev0_71, __p1_71); \
+  __ret_71 = __builtin_shufflevector(__ret_71, __ret_71, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_71; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_laneq_u32(__p0_72, __p1_72, __p2_72, __p3_72) __extension__ ({ \
+  uint32x2_t __s0_72 = __p0_72; \
+  uint32x4_t __s2_72 = __p2_72; \
+  uint32x2_t __ret_72; \
+  __ret_72 = vset_lane_u32(vgetq_lane_u32(__s2_72, __p3_72), __s0_72, __p1_72); \
+  __ret_72; \
+})
+#else
+#define vcopy_laneq_u32(__p0_73, __p1_73, __p2_73, __p3_73) __extension__ ({ \
+  uint32x2_t __s0_73 = __p0_73; \
+  uint32x4_t __s2_73 = __p2_73; \
+  uint32x2_t __rev0_73;  __rev0_73 = __builtin_shufflevector(__s0_73, __s0_73, 1, 0); \
+  uint32x4_t __rev2_73;  __rev2_73 = __builtin_shufflevector(__s2_73, __s2_73, 3, 2, 1, 0); \
+  uint32x2_t __ret_73; \
+  __ret_73 = __noswap_vset_lane_u32(__noswap_vgetq_lane_u32(__rev2_73, __p3_73), __rev0_73, __p1_73); \
+  __ret_73 = __builtin_shufflevector(__ret_73, __ret_73, 1, 0); \
+  __ret_73; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_laneq_u64(__p0_74, __p1_74, __p2_74, __p3_74) __extension__ ({ \
+  uint64x1_t __s0_74 = __p0_74; \
+  uint64x2_t __s2_74 = __p2_74; \
+  uint64x1_t __ret_74; \
+  __ret_74 = vset_lane_u64(vgetq_lane_u64(__s2_74, __p3_74), __s0_74, __p1_74); \
+  __ret_74; \
+})
+#else
+#define vcopy_laneq_u64(__p0_75, __p1_75, __p2_75, __p3_75) __extension__ ({ \
+  uint64x1_t __s0_75 = __p0_75; \
+  uint64x2_t __s2_75 = __p2_75; \
+  uint64x2_t __rev2_75;  __rev2_75 = __builtin_shufflevector(__s2_75, __s2_75, 1, 0); \
+  uint64x1_t __ret_75; \
+  __ret_75 = __noswap_vset_lane_u64(__noswap_vgetq_lane_u64(__rev2_75, __p3_75), __s0_75, __p1_75); \
+  __ret_75; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_laneq_u16(__p0_76, __p1_76, __p2_76, __p3_76) __extension__ ({ \
+  uint16x4_t __s0_76 = __p0_76; \
+  uint16x8_t __s2_76 = __p2_76; \
+  uint16x4_t __ret_76; \
+  __ret_76 = vset_lane_u16(vgetq_lane_u16(__s2_76, __p3_76), __s0_76, __p1_76); \
+  __ret_76; \
+})
+#else
+#define vcopy_laneq_u16(__p0_77, __p1_77, __p2_77, __p3_77) __extension__ ({ \
+  uint16x4_t __s0_77 = __p0_77; \
+  uint16x8_t __s2_77 = __p2_77; \
+  uint16x4_t __rev0_77;  __rev0_77 = __builtin_shufflevector(__s0_77, __s0_77, 3, 2, 1, 0); \
+  uint16x8_t __rev2_77;  __rev2_77 = __builtin_shufflevector(__s2_77, __s2_77, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x4_t __ret_77; \
+  __ret_77 = __noswap_vset_lane_u16(__noswap_vgetq_lane_u16(__rev2_77, __p3_77), __rev0_77, __p1_77); \
+  __ret_77 = __builtin_shufflevector(__ret_77, __ret_77, 3, 2, 1, 0); \
+  __ret_77; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_laneq_s8(__p0_78, __p1_78, __p2_78, __p3_78) __extension__ ({ \
+  int8x8_t __s0_78 = __p0_78; \
+  int8x16_t __s2_78 = __p2_78; \
+  int8x8_t __ret_78; \
+  __ret_78 = vset_lane_s8(vgetq_lane_s8(__s2_78, __p3_78), __s0_78, __p1_78); \
+  __ret_78; \
+})
+#else
+#define vcopy_laneq_s8(__p0_79, __p1_79, __p2_79, __p3_79) __extension__ ({ \
+  int8x8_t __s0_79 = __p0_79; \
+  int8x16_t __s2_79 = __p2_79; \
+  int8x8_t __rev0_79;  __rev0_79 = __builtin_shufflevector(__s0_79, __s0_79, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __rev2_79;  __rev2_79 = __builtin_shufflevector(__s2_79, __s2_79, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret_79; \
+  __ret_79 = __noswap_vset_lane_s8(__noswap_vgetq_lane_s8(__rev2_79, __p3_79), __rev0_79, __p1_79); \
+  __ret_79 = __builtin_shufflevector(__ret_79, __ret_79, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_79; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_laneq_f32(__p0_80, __p1_80, __p2_80, __p3_80) __extension__ ({ \
+  float32x2_t __s0_80 = __p0_80; \
+  float32x4_t __s2_80 = __p2_80; \
+  float32x2_t __ret_80; \
+  __ret_80 = vset_lane_f32(vgetq_lane_f32(__s2_80, __p3_80), __s0_80, __p1_80); \
+  __ret_80; \
+})
+#else
+#define vcopy_laneq_f32(__p0_81, __p1_81, __p2_81, __p3_81) __extension__ ({ \
+  float32x2_t __s0_81 = __p0_81; \
+  float32x4_t __s2_81 = __p2_81; \
+  float32x2_t __rev0_81;  __rev0_81 = __builtin_shufflevector(__s0_81, __s0_81, 1, 0); \
+  float32x4_t __rev2_81;  __rev2_81 = __builtin_shufflevector(__s2_81, __s2_81, 3, 2, 1, 0); \
+  float32x2_t __ret_81; \
+  __ret_81 = __noswap_vset_lane_f32(__noswap_vgetq_lane_f32(__rev2_81, __p3_81), __rev0_81, __p1_81); \
+  __ret_81 = __builtin_shufflevector(__ret_81, __ret_81, 1, 0); \
+  __ret_81; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_laneq_s32(__p0_82, __p1_82, __p2_82, __p3_82) __extension__ ({ \
+  int32x2_t __s0_82 = __p0_82; \
+  int32x4_t __s2_82 = __p2_82; \
+  int32x2_t __ret_82; \
+  __ret_82 = vset_lane_s32(vgetq_lane_s32(__s2_82, __p3_82), __s0_82, __p1_82); \
+  __ret_82; \
+})
+#else
+#define vcopy_laneq_s32(__p0_83, __p1_83, __p2_83, __p3_83) __extension__ ({ \
+  int32x2_t __s0_83 = __p0_83; \
+  int32x4_t __s2_83 = __p2_83; \
+  int32x2_t __rev0_83;  __rev0_83 = __builtin_shufflevector(__s0_83, __s0_83, 1, 0); \
+  int32x4_t __rev2_83;  __rev2_83 = __builtin_shufflevector(__s2_83, __s2_83, 3, 2, 1, 0); \
+  int32x2_t __ret_83; \
+  __ret_83 = __noswap_vset_lane_s32(__noswap_vgetq_lane_s32(__rev2_83, __p3_83), __rev0_83, __p1_83); \
+  __ret_83 = __builtin_shufflevector(__ret_83, __ret_83, 1, 0); \
+  __ret_83; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_laneq_s64(__p0_84, __p1_84, __p2_84, __p3_84) __extension__ ({ \
+  int64x1_t __s0_84 = __p0_84; \
+  int64x2_t __s2_84 = __p2_84; \
+  int64x1_t __ret_84; \
+  __ret_84 = vset_lane_s64(vgetq_lane_s64(__s2_84, __p3_84), __s0_84, __p1_84); \
+  __ret_84; \
+})
+#else
+#define vcopy_laneq_s64(__p0_85, __p1_85, __p2_85, __p3_85) __extension__ ({ \
+  int64x1_t __s0_85 = __p0_85; \
+  int64x2_t __s2_85 = __p2_85; \
+  int64x2_t __rev2_85;  __rev2_85 = __builtin_shufflevector(__s2_85, __s2_85, 1, 0); \
+  int64x1_t __ret_85; \
+  __ret_85 = __noswap_vset_lane_s64(__noswap_vgetq_lane_s64(__rev2_85, __p3_85), __s0_85, __p1_85); \
+  __ret_85; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_laneq_s16(__p0_86, __p1_86, __p2_86, __p3_86) __extension__ ({ \
+  int16x4_t __s0_86 = __p0_86; \
+  int16x8_t __s2_86 = __p2_86; \
+  int16x4_t __ret_86; \
+  __ret_86 = vset_lane_s16(vgetq_lane_s16(__s2_86, __p3_86), __s0_86, __p1_86); \
+  __ret_86; \
+})
+#else
+#define vcopy_laneq_s16(__p0_87, __p1_87, __p2_87, __p3_87) __extension__ ({ \
+  int16x4_t __s0_87 = __p0_87; \
+  int16x8_t __s2_87 = __p2_87; \
+  int16x4_t __rev0_87;  __rev0_87 = __builtin_shufflevector(__s0_87, __s0_87, 3, 2, 1, 0); \
+  int16x8_t __rev2_87;  __rev2_87 = __builtin_shufflevector(__s2_87, __s2_87, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __ret_87; \
+  __ret_87 = __noswap_vset_lane_s16(__noswap_vgetq_lane_s16(__rev2_87, __p3_87), __rev0_87, __p1_87); \
+  __ret_87 = __builtin_shufflevector(__ret_87, __ret_87, 3, 2, 1, 0); \
+  __ret_87; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x1_t vcreate_p64(uint64_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai poly64x1_t vcreate_p64(uint64_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vcreate_f64(uint64_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#else
+__ai float64x1_t vcreate_f64(uint64_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t)(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vcvts_f32_s32(int32_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vcvts_f32_s32(__p0);
+  return __ret;
+}
+#else
+__ai float32_t vcvts_f32_s32(int32_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vcvts_f32_s32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vcvts_f32_u32(uint32_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vcvts_f32_u32(__p0);
+  return __ret;
+}
+#else
+__ai float32_t vcvts_f32_u32(uint32_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vcvts_f32_u32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vcvt_f32_f64(float64x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vcvt_f32_f64((int8x16_t)__p0, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vcvt_f32_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vcvt_f32_f64((int8x16_t)__rev0, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai float32x2_t __noswap_vcvt_f32_f64(float64x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vcvt_f32_f64((int8x16_t)__p0, 9);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64_t vcvtd_f64_s64(int64_t __p0) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vcvtd_f64_s64(__p0);
+  return __ret;
+}
+#else
+__ai float64_t vcvtd_f64_s64(int64_t __p0) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vcvtd_f64_s64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64_t vcvtd_f64_u64(uint64_t __p0) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vcvtd_f64_u64(__p0);
+  return __ret;
+}
+#else
+__ai float64_t vcvtd_f64_u64(uint64_t __p0) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vcvtd_f64_u64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vcvtq_f64_u64(uint64x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vcvtq_f64_v((int8x16_t)__p0, 51);
+  return __ret;
+}
+#else
+__ai float64x2_t vcvtq_f64_u64(uint64x2_t __p0) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vcvtq_f64_v((int8x16_t)__rev0, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vcvtq_f64_s64(int64x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vcvtq_f64_v((int8x16_t)__p0, 35);
+  return __ret;
+}
+#else
+__ai float64x2_t vcvtq_f64_s64(int64x2_t __p0) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vcvtq_f64_v((int8x16_t)__rev0, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vcvt_f64_u64(uint64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vcvt_f64_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#else
+__ai float64x1_t vcvt_f64_u64(uint64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vcvt_f64_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vcvt_f64_s64(int64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vcvt_f64_v((int8x8_t)__p0, 3);
+  return __ret;
+}
+#else
+__ai float64x1_t vcvt_f64_s64(int64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vcvt_f64_v((int8x8_t)__p0, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vcvt_f64_f32(float32x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vcvt_f64_f32((int8x8_t)__p0, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vcvt_f64_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vcvt_f64_f32((int8x8_t)__rev0, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai float64x2_t __noswap_vcvt_f64_f32(float32x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vcvt_f64_f32((int8x8_t)__p0, 42);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float16x8_t vcvt_high_f16_f32(float16x4_t __p0, float32x4_t __p1) {
+  float16x8_t __ret;
+  __ret = vcombine_f16(__p0, vcvt_f16_f32(__p1));
+  return __ret;
+}
+#else
+__ai float16x8_t vcvt_high_f16_f32(float16x4_t __p0, float32x4_t __p1) {
+  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float16x8_t __ret;
+  __ret = __noswap_vcombine_f16(__rev0, __noswap_vcvt_f16_f32(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vcvt_high_f32_f16(float16x8_t __p0) {
+  float32x4_t __ret;
+  __ret = vcvt_f32_f16(vget_high_f16(__p0));
+  return __ret;
+}
+#else
+__ai float32x4_t vcvt_high_f32_f16(float16x8_t __p0) {
+  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = __noswap_vcvt_f32_f16(__noswap_vget_high_f16(__rev0));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vcvt_high_f32_f64(float32x2_t __p0, float64x2_t __p1) {
+  float32x4_t __ret;
+  __ret = vcombine_f32(__p0, vcvt_f32_f64(__p1));
+  return __ret;
+}
+#else
+__ai float32x4_t vcvt_high_f32_f64(float32x2_t __p0, float64x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x4_t __ret;
+  __ret = __noswap_vcombine_f32(__rev0, __noswap_vcvt_f32_f64(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vcvt_high_f64_f32(float32x4_t __p0) {
+  float64x2_t __ret;
+  __ret = vcvt_f64_f32(vget_high_f32(__p0));
+  return __ret;
+}
+#else
+__ai float64x2_t vcvt_high_f64_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float64x2_t __ret;
+  __ret = __noswap_vcvt_f64_f32(__noswap_vget_high_f32(__rev0));
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvts_n_f32_u32(__p0, __p1) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  float32_t __ret; \
+  __ret = (float32_t) __builtin_neon_vcvts_n_f32_u32(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vcvts_n_f32_u32(__p0, __p1) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  float32_t __ret; \
+  __ret = (float32_t) __builtin_neon_vcvts_n_f32_u32(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvts_n_f32_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  float32_t __ret; \
+  __ret = (float32_t) __builtin_neon_vcvts_n_f32_s32(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vcvts_n_f32_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  float32_t __ret; \
+  __ret = (float32_t) __builtin_neon_vcvts_n_f32_s32(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvtq_n_f64_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vcvtq_n_f64_v((int8x16_t)__s0, __p1, 51); \
+  __ret; \
+})
+#else
+#define vcvtq_n_f64_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vcvtq_n_f64_v((int8x16_t)__rev0, __p1, 51); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvtq_n_f64_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vcvtq_n_f64_v((int8x16_t)__s0, __p1, 35); \
+  __ret; \
+})
+#else
+#define vcvtq_n_f64_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vcvtq_n_f64_v((int8x16_t)__rev0, __p1, 35); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvt_n_f64_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vcvt_n_f64_v((int8x8_t)__s0, __p1, 19); \
+  __ret; \
+})
+#else
+#define vcvt_n_f64_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vcvt_n_f64_v((int8x8_t)__s0, __p1, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvt_n_f64_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vcvt_n_f64_v((int8x8_t)__s0, __p1, 3); \
+  __ret; \
+})
+#else
+#define vcvt_n_f64_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vcvt_n_f64_v((int8x8_t)__s0, __p1, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvtd_n_f64_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  float64_t __ret; \
+  __ret = (float64_t) __builtin_neon_vcvtd_n_f64_u64(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vcvtd_n_f64_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  float64_t __ret; \
+  __ret = (float64_t) __builtin_neon_vcvtd_n_f64_u64(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvtd_n_f64_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  float64_t __ret; \
+  __ret = (float64_t) __builtin_neon_vcvtd_n_f64_s64(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vcvtd_n_f64_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  float64_t __ret; \
+  __ret = (float64_t) __builtin_neon_vcvtd_n_f64_s64(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvts_n_s32_f32(__p0, __p1) __extension__ ({ \
+  float32_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vcvts_n_s32_f32(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vcvts_n_s32_f32(__p0, __p1) __extension__ ({ \
+  float32_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vcvts_n_s32_f32(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvtq_n_s64_f64(__p0, __p1) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vcvtq_n_s64_v((int8x16_t)__s0, __p1, 35); \
+  __ret; \
+})
+#else
+#define vcvtq_n_s64_f64(__p0, __p1) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int64x2_t __ret; \
+  __ret = (int64x2_t) __builtin_neon_vcvtq_n_s64_v((int8x16_t)__rev0, __p1, 35); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvt_n_s64_f64(__p0, __p1) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vcvt_n_s64_v((int8x8_t)__s0, __p1, 3); \
+  __ret; \
+})
+#else
+#define vcvt_n_s64_f64(__p0, __p1) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  int64x1_t __ret; \
+  __ret = (int64x1_t) __builtin_neon_vcvt_n_s64_v((int8x8_t)__s0, __p1, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvtd_n_s64_f64(__p0, __p1) __extension__ ({ \
+  float64_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vcvtd_n_s64_f64(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vcvtd_n_s64_f64(__p0, __p1) __extension__ ({ \
+  float64_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vcvtd_n_s64_f64(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvts_n_u32_f32(__p0, __p1) __extension__ ({ \
+  float32_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vcvts_n_u32_f32(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vcvts_n_u32_f32(__p0, __p1) __extension__ ({ \
+  float32_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vcvts_n_u32_f32(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvtq_n_u64_f64(__p0, __p1) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vcvtq_n_u64_v((int8x16_t)__s0, __p1, 51); \
+  __ret; \
+})
+#else
+#define vcvtq_n_u64_f64(__p0, __p1) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = (uint64x2_t) __builtin_neon_vcvtq_n_u64_v((int8x16_t)__rev0, __p1, 51); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvt_n_u64_f64(__p0, __p1) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vcvt_n_u64_v((int8x8_t)__s0, __p1, 19); \
+  __ret; \
+})
+#else
+#define vcvt_n_u64_f64(__p0, __p1) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  uint64x1_t __ret; \
+  __ret = (uint64x1_t) __builtin_neon_vcvt_n_u64_v((int8x8_t)__s0, __p1, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcvtd_n_u64_f64(__p0, __p1) __extension__ ({ \
+  float64_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vcvtd_n_u64_f64(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vcvtd_n_u64_f64(__p0, __p1) __extension__ ({ \
+  float64_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vcvtd_n_u64_f64(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vcvts_s32_f32(float32_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vcvts_s32_f32(__p0);
+  return __ret;
+}
+#else
+__ai int32_t vcvts_s32_f32(float32_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vcvts_s32_f32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vcvtd_s64_f64(float64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcvtd_s64_f64(__p0);
+  return __ret;
+}
+#else
+__ai int64_t vcvtd_s64_f64(float64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcvtd_s64_f64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vcvtq_s64_f64(float64x2_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vcvtq_s64_v((int8x16_t)__p0, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vcvtq_s64_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vcvtq_s64_v((int8x16_t)__rev0, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vcvt_s64_f64(float64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vcvt_s64_v((int8x8_t)__p0, 3);
+  return __ret;
+}
+#else
+__ai int64x1_t vcvt_s64_f64(float64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vcvt_s64_v((int8x8_t)__p0, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vcvts_u32_f32(float32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcvts_u32_f32(__p0);
+  return __ret;
+}
+#else
+__ai uint32_t vcvts_u32_f32(float32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcvts_u32_f32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vcvtd_u64_f64(float64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcvtd_u64_f64(__p0);
+  return __ret;
+}
+#else
+__ai uint64_t vcvtd_u64_f64(float64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcvtd_u64_f64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vcvtq_u64_f64(float64x2_t __p0) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcvtq_u64_v((int8x16_t)__p0, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vcvtq_u64_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vcvtq_u64_v((int8x16_t)__rev0, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vcvt_u64_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcvt_u64_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vcvt_u64_f64(float64x1_t __p0) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vcvt_u64_v((int8x8_t)__p0, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vcvtas_s32_f32(float32_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vcvtas_s32_f32(__p0);
+  return __ret;
+}
+#else
+__ai int32_t vcvtas_s32_f32(float32_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vcvtas_s32_f32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vcvtad_s64_f64(float64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcvtad_s64_f64(__p0);
+  return __ret;
+}
+#else
+__ai int64_t vcvtad_s64_f64(float64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcvtad_s64_f64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vcvtas_u32_f32(float32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcvtas_u32_f32(__p0);
+  return __ret;
+}
+#else
+__ai uint32_t vcvtas_u32_f32(float32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcvtas_u32_f32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vcvtad_u64_f64(float64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcvtad_u64_f64(__p0);
+  return __ret;
+}
+#else
+__ai uint64_t vcvtad_u64_f64(float64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcvtad_u64_f64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vcvtms_s32_f32(float32_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vcvtms_s32_f32(__p0);
+  return __ret;
+}
+#else
+__ai int32_t vcvtms_s32_f32(float32_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vcvtms_s32_f32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vcvtmd_s64_f64(float64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcvtmd_s64_f64(__p0);
+  return __ret;
+}
+#else
+__ai int64_t vcvtmd_s64_f64(float64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcvtmd_s64_f64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vcvtms_u32_f32(float32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcvtms_u32_f32(__p0);
+  return __ret;
+}
+#else
+__ai uint32_t vcvtms_u32_f32(float32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcvtms_u32_f32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vcvtmd_u64_f64(float64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcvtmd_u64_f64(__p0);
+  return __ret;
+}
+#else
+__ai uint64_t vcvtmd_u64_f64(float64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcvtmd_u64_f64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vcvtns_s32_f32(float32_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vcvtns_s32_f32(__p0);
+  return __ret;
+}
+#else
+__ai int32_t vcvtns_s32_f32(float32_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vcvtns_s32_f32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vcvtnd_s64_f64(float64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcvtnd_s64_f64(__p0);
+  return __ret;
+}
+#else
+__ai int64_t vcvtnd_s64_f64(float64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcvtnd_s64_f64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vcvtns_u32_f32(float32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcvtns_u32_f32(__p0);
+  return __ret;
+}
+#else
+__ai uint32_t vcvtns_u32_f32(float32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcvtns_u32_f32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vcvtnd_u64_f64(float64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcvtnd_u64_f64(__p0);
+  return __ret;
+}
+#else
+__ai uint64_t vcvtnd_u64_f64(float64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcvtnd_u64_f64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vcvtps_s32_f32(float32_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vcvtps_s32_f32(__p0);
+  return __ret;
+}
+#else
+__ai int32_t vcvtps_s32_f32(float32_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vcvtps_s32_f32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vcvtpd_s64_f64(float64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcvtpd_s64_f64(__p0);
+  return __ret;
+}
+#else
+__ai int64_t vcvtpd_s64_f64(float64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vcvtpd_s64_f64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vcvtps_u32_f32(float32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcvtps_u32_f32(__p0);
+  return __ret;
+}
+#else
+__ai uint32_t vcvtps_u32_f32(float32_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vcvtps_u32_f32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vcvtpd_u64_f64(float64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcvtpd_u64_f64(__p0);
+  return __ret;
+}
+#else
+__ai uint64_t vcvtpd_u64_f64(float64_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vcvtpd_u64_f64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vcvtxd_f32_f64(float64_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vcvtxd_f32_f64(__p0);
+  return __ret;
+}
+#else
+__ai float32_t vcvtxd_f32_f64(float64_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vcvtxd_f32_f64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vcvtx_f32_f64(float64x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vcvtx_f32_v((int8x16_t)__p0, 42);
+  return __ret;
+}
+#else
+__ai float32x2_t vcvtx_f32_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vcvtx_f32_v((int8x16_t)__rev0, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai float32x2_t __noswap_vcvtx_f32_f64(float64x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vcvtx_f32_v((int8x16_t)__p0, 42);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vcvtx_high_f32_f64(float32x2_t __p0, float64x2_t __p1) {
+  float32x4_t __ret;
+  __ret = vcombine_f32(__p0, vcvtx_f32_f64(__p1));
+  return __ret;
+}
+#else
+__ai float32x4_t vcvtx_high_f32_f64(float32x2_t __p0, float64x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x4_t __ret;
+  __ret = __noswap_vcombine_f32(__rev0, __noswap_vcvtx_f32_f64(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vdivq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = __p0 / __p1;
+  return __ret;
+}
+#else
+__ai float64x2_t vdivq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = __rev0 / __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vdivq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = __p0 / __p1;
+  return __ret;
+}
+#else
+__ai float32x4_t vdivq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = __rev0 / __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vdiv_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = __p0 / __p1;
+  return __ret;
+}
+#else
+__ai float64x1_t vdiv_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = __p0 / __p1;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vdiv_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = __p0 / __p1;
+  return __ret;
+}
+#else
+__ai float32x2_t vdiv_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = __rev0 / __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupb_lane_p8(__p0, __p1) __extension__ ({ \
+  poly8x8_t __s0 = __p0; \
+  poly8_t __ret; \
+  __ret = (poly8_t) __builtin_neon_vdupb_lane_i8((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vdupb_lane_p8(__p0, __p1) __extension__ ({ \
+  poly8x8_t __s0 = __p0; \
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8_t __ret; \
+  __ret = (poly8_t) __builtin_neon_vdupb_lane_i8((int8x8_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vduph_lane_p16(__p0, __p1) __extension__ ({ \
+  poly16x4_t __s0 = __p0; \
+  poly16_t __ret; \
+  __ret = (poly16_t) __builtin_neon_vduph_lane_i16((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vduph_lane_p16(__p0, __p1) __extension__ ({ \
+  poly16x4_t __s0 = __p0; \
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  poly16_t __ret; \
+  __ret = (poly16_t) __builtin_neon_vduph_lane_i16((int8x8_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupb_lane_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vdupb_lane_i8((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vdupb_lane_u8(__p0, __p1) __extension__ ({ \
+  uint8x8_t __s0 = __p0; \
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vdupb_lane_i8((int8x8_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdups_lane_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vdups_lane_i32((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vdups_lane_u32(__p0, __p1) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vdups_lane_i32((int8x8_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupd_lane_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vdupd_lane_i64((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vdupd_lane_u64(__p0, __p1) __extension__ ({ \
+  uint64x1_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vdupd_lane_i64((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vduph_lane_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vduph_lane_i16((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vduph_lane_u16(__p0, __p1) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vduph_lane_i16((int8x8_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupb_lane_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vdupb_lane_i8((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vdupb_lane_s8(__p0, __p1) __extension__ ({ \
+  int8x8_t __s0 = __p0; \
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vdupb_lane_i8((int8x8_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupd_lane_f64(__p0, __p1) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64_t __ret; \
+  __ret = (float64_t) __builtin_neon_vdupd_lane_f64((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vdupd_lane_f64(__p0, __p1) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64_t __ret; \
+  __ret = (float64_t) __builtin_neon_vdupd_lane_f64((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdups_lane_f32(__p0, __p1) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32_t __ret; \
+  __ret = (float32_t) __builtin_neon_vdups_lane_f32((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vdups_lane_f32(__p0, __p1) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float32_t __ret; \
+  __ret = (float32_t) __builtin_neon_vdups_lane_f32((int8x8_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdups_lane_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vdups_lane_i32((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vdups_lane_s32(__p0, __p1) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vdups_lane_i32((int8x8_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupd_lane_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vdupd_lane_i64((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vdupd_lane_s64(__p0, __p1) __extension__ ({ \
+  int64x1_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vdupd_lane_i64((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vduph_lane_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vduph_lane_i16((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vduph_lane_s16(__p0, __p1) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vduph_lane_i16((int8x8_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_lane_p64(__p0, __p1) __extension__ ({ \
+  poly64x1_t __s0 = __p0; \
+  poly64x1_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
+  __ret; \
+})
+#else
+#define vdup_lane_p64(__p0, __p1) __extension__ ({ \
+  poly64x1_t __s0 = __p0; \
+  poly64x1_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_lane_p64(__p0, __p1) __extension__ ({ \
+  poly64x1_t __s0 = __p0; \
+  poly64x2_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_lane_p64(__p0, __p1) __extension__ ({ \
+  poly64x1_t __s0 = __p0; \
+  poly64x2_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_lane_f64(__p0, __p1) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64x2_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_lane_f64(__p0, __p1) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64x2_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_lane_f16(__p0, __p1) __extension__ ({ \
+  float16x4_t __s0 = __p0; \
+  float16x8_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_lane_f16(__p0, __p1) __extension__ ({ \
+  float16x4_t __s0 = __p0; \
+  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  float16x8_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_lane_f64(__p0, __p1) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64x1_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
+  __ret; \
+})
+#else
+#define vdup_lane_f64(__p0, __p1) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64x1_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_lane_f16(__p0, __p1) __extension__ ({ \
+  float16x4_t __s0 = __p0; \
+  float16x4_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdup_lane_f16(__p0, __p1) __extension__ ({ \
+  float16x4_t __s0 = __p0; \
+  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  float16x4_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupb_laneq_p8(__p0, __p1) __extension__ ({ \
+  poly8x16_t __s0 = __p0; \
+  poly8_t __ret; \
+  __ret = (poly8_t) __builtin_neon_vdupb_laneq_i8((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vdupb_laneq_p8(__p0, __p1) __extension__ ({ \
+  poly8x16_t __s0 = __p0; \
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8_t __ret; \
+  __ret = (poly8_t) __builtin_neon_vdupb_laneq_i8((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vduph_laneq_p16(__p0, __p1) __extension__ ({ \
+  poly16x8_t __s0 = __p0; \
+  poly16_t __ret; \
+  __ret = (poly16_t) __builtin_neon_vduph_laneq_i16((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vduph_laneq_p16(__p0, __p1) __extension__ ({ \
+  poly16x8_t __s0 = __p0; \
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16_t __ret; \
+  __ret = (poly16_t) __builtin_neon_vduph_laneq_i16((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupb_laneq_u8(__p0, __p1) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vdupb_laneq_i8((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vdupb_laneq_u8(__p0, __p1) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vdupb_laneq_i8((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdups_laneq_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vdups_laneq_i32((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vdups_laneq_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vdups_laneq_i32((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupd_laneq_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vdupd_laneq_i64((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vdupd_laneq_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vdupd_laneq_i64((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vduph_laneq_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vduph_laneq_i16((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vduph_laneq_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vduph_laneq_i16((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupb_laneq_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vdupb_laneq_i8((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vdupb_laneq_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vdupb_laneq_i8((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupd_laneq_f64(__p0, __p1) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64_t __ret; \
+  __ret = (float64_t) __builtin_neon_vdupd_laneq_f64((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vdupd_laneq_f64(__p0, __p1) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float64_t __ret; \
+  __ret = (float64_t) __builtin_neon_vdupd_laneq_f64((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdups_laneq_f32(__p0, __p1) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32_t __ret; \
+  __ret = (float32_t) __builtin_neon_vdups_laneq_f32((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vdups_laneq_f32(__p0, __p1) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  float32_t __ret; \
+  __ret = (float32_t) __builtin_neon_vdups_laneq_f32((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdups_laneq_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vdups_laneq_i32((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vdups_laneq_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vdups_laneq_i32((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupd_laneq_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vdupd_laneq_i64((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vdupd_laneq_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vdupd_laneq_i64((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vduph_laneq_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vduph_laneq_i16((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vduph_laneq_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vduph_laneq_i16((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_laneq_p8(__p0, __p1) __extension__ ({ \
+  poly8x16_t __s0 = __p0; \
+  poly8x8_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdup_laneq_p8(__p0, __p1) __extension__ ({ \
+  poly8x16_t __s0 = __p0; \
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x8_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_laneq_p64(__p0, __p1) __extension__ ({ \
+  poly64x2_t __s0 = __p0; \
+  poly64x1_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
+  __ret; \
+})
+#else
+#define vdup_laneq_p64(__p0, __p1) __extension__ ({ \
+  poly64x2_t __s0 = __p0; \
+  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  poly64x1_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_laneq_p16(__p0, __p1) __extension__ ({ \
+  poly16x8_t __s0 = __p0; \
+  poly16x4_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdup_laneq_p16(__p0, __p1) __extension__ ({ \
+  poly16x8_t __s0 = __p0; \
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x4_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_laneq_p8(__p0, __p1) __extension__ ({ \
+  poly8x16_t __s0 = __p0; \
+  poly8x16_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_laneq_p8(__p0, __p1) __extension__ ({ \
+  poly8x16_t __s0 = __p0; \
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_laneq_p64(__p0, __p1) __extension__ ({ \
+  poly64x2_t __s0 = __p0; \
+  poly64x2_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_laneq_p64(__p0, __p1) __extension__ ({ \
+  poly64x2_t __s0 = __p0; \
+  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  poly64x2_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_laneq_p16(__p0, __p1) __extension__ ({ \
+  poly16x8_t __s0 = __p0; \
+  poly16x8_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_laneq_p16(__p0, __p1) __extension__ ({ \
+  poly16x8_t __s0 = __p0; \
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x8_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_laneq_u8(__p0, __p1) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_laneq_u8(__p0, __p1) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_laneq_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_laneq_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_laneq_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_laneq_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_laneq_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_laneq_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_laneq_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_laneq_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_laneq_f64(__p0, __p1) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x2_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_laneq_f64(__p0, __p1) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float64x2_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_laneq_f32(__p0, __p1) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_laneq_f32(__p0, __p1) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  float32x4_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_laneq_f16(__p0, __p1) __extension__ ({ \
+  float16x8_t __s0 = __p0; \
+  float16x8_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_laneq_f16(__p0, __p1) __extension__ ({ \
+  float16x8_t __s0 = __p0; \
+  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x8_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_laneq_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_laneq_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_laneq_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_laneq_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdupq_laneq_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdupq_laneq_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_laneq_u8(__p0, __p1) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x8_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdup_laneq_u8(__p0, __p1) __extension__ ({ \
+  uint8x16_t __s0 = __p0; \
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_laneq_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x2_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdup_laneq_u32(__p0, __p1) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_laneq_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x1_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
+  __ret; \
+})
+#else
+#define vdup_laneq_u64(__p0, __p1) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint64x1_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_laneq_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x4_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdup_laneq_u16(__p0, __p1) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_laneq_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x8_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdup_laneq_s8(__p0, __p1) __extension__ ({ \
+  int8x16_t __s0 = __p0; \
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_laneq_f64(__p0, __p1) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x1_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
+  __ret; \
+})
+#else
+#define vdup_laneq_f64(__p0, __p1) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float64x1_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_laneq_f32(__p0, __p1) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x2_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdup_laneq_f32(__p0, __p1) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  float32x2_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_laneq_f16(__p0, __p1) __extension__ ({ \
+  float16x8_t __s0 = __p0; \
+  float16x4_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdup_laneq_f16(__p0, __p1) __extension__ ({ \
+  float16x8_t __s0 = __p0; \
+  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x4_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_laneq_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x2_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdup_laneq_s32(__p0, __p1) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x2_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_laneq_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x1_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
+  __ret; \
+})
+#else
+#define vdup_laneq_s64(__p0, __p1) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int64x1_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vdup_laneq_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x4_t __ret; \
+  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
+  __ret; \
+})
+#else
+#define vdup_laneq_s16(__p0, __p1) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x1_t vdup_n_p64(poly64_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t) {__p0};
+  return __ret;
+}
+#else
+__ai poly64x1_t vdup_n_p64(poly64_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t) {__p0};
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vdupq_n_p64(poly64_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t) {__p0, __p0};
+  return __ret;
+}
+#else
+__ai poly64x2_t vdupq_n_p64(poly64_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t) {__p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vdupq_n_f64(float64_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) {__p0, __p0};
+  return __ret;
+}
+#else
+__ai float64x2_t vdupq_n_f64(float64_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) {__p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vdup_n_f64(float64_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) {__p0};
+  return __ret;
+}
+#else
+__ai float64x1_t vdup_n_f64(float64_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) {__p0};
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vext_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1_t __s0 = __p0; \
+  poly64x1_t __s1 = __p1; \
+  poly64x1_t __ret; \
+  __ret = (poly64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
+  __ret; \
+})
+#else
+#define vext_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1_t __s0 = __p0; \
+  poly64x1_t __s1 = __p1; \
+  poly64x1_t __ret; \
+  __ret = (poly64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vextq_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2_t __s0 = __p0; \
+  poly64x2_t __s1 = __p1; \
+  poly64x2_t __ret; \
+  __ret = (poly64x2_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 38); \
+  __ret; \
+})
+#else
+#define vextq_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2_t __s0 = __p0; \
+  poly64x2_t __s1 = __p1; \
+  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  poly64x2_t __ret; \
+  __ret = (poly64x2_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 38); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vextq_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x2_t __s1 = __p1; \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 42); \
+  __ret; \
+})
+#else
+#define vextq_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x2_t __s1 = __p1; \
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 42); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vext_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64x1_t __s1 = __p1; \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 10); \
+  __ret; \
+})
+#else
+#define vext_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64x1_t __s1 = __p1; \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 10); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vfmaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vfmaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vfmaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vfmaq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai float64x2_t __noswap_vfmaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vfmaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vfma_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vfma_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmad_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float64_t __s0 = __p0; \
+  float64_t __s1 = __p1; \
+  float64x1_t __s2 = __p2; \
+  float64_t __ret; \
+  __ret = (float64_t) __builtin_neon_vfmad_lane_f64(__s0, __s1, (int8x8_t)__s2, __p3); \
+  __ret; \
+})
+#else
+#define vfmad_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float64_t __s0 = __p0; \
+  float64_t __s1 = __p1; \
+  float64x1_t __s2 = __p2; \
+  float64_t __ret; \
+  __ret = (float64_t) __builtin_neon_vfmad_lane_f64(__s0, __s1, (int8x8_t)__s2, __p3); \
+  __ret; \
+})
+#define __noswap_vfmad_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float64_t __s0 = __p0; \
+  float64_t __s1 = __p1; \
+  float64x1_t __s2 = __p2; \
+  float64_t __ret; \
+  __ret = (float64_t) __builtin_neon_vfmad_lane_f64(__s0, __s1, (int8x8_t)__s2, __p3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmas_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32_t __s0 = __p0; \
+  float32_t __s1 = __p1; \
+  float32x2_t __s2 = __p2; \
+  float32_t __ret; \
+  __ret = (float32_t) __builtin_neon_vfmas_lane_f32(__s0, __s1, (int8x8_t)__s2, __p3); \
+  __ret; \
+})
+#else
+#define vfmas_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32_t __s0 = __p0; \
+  float32_t __s1 = __p1; \
+  float32x2_t __s2 = __p2; \
+  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  float32_t __ret; \
+  __ret = (float32_t) __builtin_neon_vfmas_lane_f32(__s0, __s1, (int8x8_t)__rev2, __p3); \
+  __ret; \
+})
+#define __noswap_vfmas_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32_t __s0 = __p0; \
+  float32_t __s1 = __p1; \
+  float32x2_t __s2 = __p2; \
+  float32_t __ret; \
+  __ret = (float32_t) __builtin_neon_vfmas_lane_f32(__s0, __s1, (int8x8_t)__s2, __p3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmaq_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x2_t __s1 = __p1; \
+  float64x1_t __s2 = __p2; \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x8_t)__s2, __p3, 42); \
+  __ret; \
+})
+#else
+#define vfmaq_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x2_t __s1 = __p1; \
+  float64x1_t __s2 = __p2; \
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x8_t)__s2, __p3, 42); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vfmaq_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x2_t __s1 = __p1; \
+  float64x1_t __s2 = __p2; \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x8_t)__s2, __p3, 42); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmaq_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x2_t __s2 = __p2; \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x8_t)__s2, __p3, 41); \
+  __ret; \
+})
+#else
+#define vfmaq_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x2_t __s2 = __p2; \
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x8_t)__rev2, __p3, 41); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vfmaq_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x2_t __s2 = __p2; \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x8_t)__s2, __p3, 41); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfma_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64x1_t __s1 = __p1; \
+  float64x1_t __s2 = __p2; \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 10); \
+  __ret; \
+})
+#else
+#define vfma_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64x1_t __s1 = __p1; \
+  float64x1_t __s2 = __p2; \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 10); \
+  __ret; \
+})
+#define __noswap_vfma_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64x1_t __s1 = __p1; \
+  float64x1_t __s2 = __p2; \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 10); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfma_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x2_t __s2 = __p2; \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 9); \
+  __ret; \
+})
+#else
+#define vfma_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x2_t __s2 = __p2; \
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vfma_lane_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, __p3, 9); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vfma_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x2_t __s2 = __p2; \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 9); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmad_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float64_t __s0 = __p0; \
+  float64_t __s1 = __p1; \
+  float64x2_t __s2 = __p2; \
+  float64_t __ret; \
+  __ret = (float64_t) __builtin_neon_vfmad_laneq_f64(__s0, __s1, (int8x16_t)__s2, __p3); \
+  __ret; \
+})
+#else
+#define vfmad_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float64_t __s0 = __p0; \
+  float64_t __s1 = __p1; \
+  float64x2_t __s2 = __p2; \
+  float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  float64_t __ret; \
+  __ret = (float64_t) __builtin_neon_vfmad_laneq_f64(__s0, __s1, (int8x16_t)__rev2, __p3); \
+  __ret; \
+})
+#define __noswap_vfmad_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float64_t __s0 = __p0; \
+  float64_t __s1 = __p1; \
+  float64x2_t __s2 = __p2; \
+  float64_t __ret; \
+  __ret = (float64_t) __builtin_neon_vfmad_laneq_f64(__s0, __s1, (int8x16_t)__s2, __p3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmas_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32_t __s0 = __p0; \
+  float32_t __s1 = __p1; \
+  float32x4_t __s2 = __p2; \
+  float32_t __ret; \
+  __ret = (float32_t) __builtin_neon_vfmas_laneq_f32(__s0, __s1, (int8x16_t)__s2, __p3); \
+  __ret; \
+})
+#else
+#define vfmas_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32_t __s0 = __p0; \
+  float32_t __s1 = __p1; \
+  float32x4_t __s2 = __p2; \
+  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  float32_t __ret; \
+  __ret = (float32_t) __builtin_neon_vfmas_laneq_f32(__s0, __s1, (int8x16_t)__rev2, __p3); \
+  __ret; \
+})
+#define __noswap_vfmas_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32_t __s0 = __p0; \
+  float32_t __s1 = __p1; \
+  float32x4_t __s2 = __p2; \
+  float32_t __ret; \
+  __ret = (float32_t) __builtin_neon_vfmas_laneq_f32(__s0, __s1, (int8x16_t)__s2, __p3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmaq_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x2_t __s1 = __p1; \
+  float64x2_t __s2 = __p2; \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x16_t)__s2, __p3, 42); \
+  __ret; \
+})
+#else
+#define vfmaq_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x2_t __s1 = __p1; \
+  float64x2_t __s2 = __p2; \
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, __p3, 42); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vfmaq_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x2_t __s1 = __p1; \
+  float64x2_t __s2 = __p2; \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x16_t)__s2, __p3, 42); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmaq_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x4_t __s2 = __p2; \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x16_t)__s2, __p3, 41); \
+  __ret; \
+})
+#else
+#define vfmaq_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x4_t __s2 = __p2; \
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, __p3, 41); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#define __noswap_vfmaq_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x4_t __s2 = __p2; \
+  float32x4_t __ret; \
+  __ret = (float32x4_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x16_t)__s2, __p3, 41); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfma_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64x1_t __s1 = __p1; \
+  float64x2_t __s2 = __p2; \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vfma_laneq_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x16_t)__s2, __p3, 10); \
+  __ret; \
+})
+#else
+#define vfma_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64x1_t __s1 = __p1; \
+  float64x2_t __s2 = __p2; \
+  float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vfma_laneq_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x16_t)__rev2, __p3, 10); \
+  __ret; \
+})
+#define __noswap_vfma_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64x1_t __s1 = __p1; \
+  float64x2_t __s2 = __p2; \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vfma_laneq_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x16_t)__s2, __p3, 10); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfma_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x4_t __s2 = __p2; \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vfma_laneq_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x16_t)__s2, __p3, 9); \
+  __ret; \
+})
+#else
+#define vfma_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x4_t __s2 = __p2; \
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vfma_laneq_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x16_t)__rev2, __p3, 9); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vfma_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x4_t __s2 = __p2; \
+  float32x2_t __ret; \
+  __ret = (float32x2_t) __builtin_neon_vfma_laneq_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x16_t)__s2, __p3, 9); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vfmaq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
+  float64x2_t __ret;
+  __ret = vfmaq_f64(__p0, __p1, (float64x2_t) {__p2, __p2});
+  return __ret;
+}
+#else
+__ai float64x2_t vfmaq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = __noswap_vfmaq_f64(__rev0, __rev1, (float64x2_t) {__p2, __p2});
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vfmaq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
+  float32x4_t __ret;
+  __ret = vfmaq_f32(__p0, __p1, (float32x4_t) {__p2, __p2, __p2, __p2});
+  return __ret;
+}
+#else
+__ai float32x4_t vfmaq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = __noswap_vfmaq_f32(__rev0, __rev1, (float32x4_t) {__p2, __p2, __p2, __p2});
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vfma_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
+  float32x2_t __ret;
+  __ret = vfma_f32(__p0, __p1, (float32x2_t) {__p2, __p2});
+  return __ret;
+}
+#else
+__ai float32x2_t vfma_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = __noswap_vfma_f32(__rev0, __rev1, (float32x2_t) {__p2, __p2});
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vfmsq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vfmsq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vfmsq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vfmsq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai float64x2_t __noswap_vfmsq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vfmsq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vfmsq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vfmsq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vfmsq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vfmsq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai float32x4_t __noswap_vfmsq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vfmsq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vfms_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vfms_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vfms_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vfms_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vfms_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vfms_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vfms_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vfms_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai float32x2_t __noswap_vfms_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vfms_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmsd_lane_f64(__p0_88, __p1_88, __p2_88, __p3_88) __extension__ ({ \
+  float64_t __s0_88 = __p0_88; \
+  float64_t __s1_88 = __p1_88; \
+  float64x1_t __s2_88 = __p2_88; \
+  float64_t __ret_88; \
+  __ret_88 = vfmad_lane_f64(__s0_88, __s1_88, -__s2_88, __p3_88); \
+  __ret_88; \
+})
+#else
+#define vfmsd_lane_f64(__p0_89, __p1_89, __p2_89, __p3_89) __extension__ ({ \
+  float64_t __s0_89 = __p0_89; \
+  float64_t __s1_89 = __p1_89; \
+  float64x1_t __s2_89 = __p2_89; \
+  float64_t __ret_89; \
+  __ret_89 = __noswap_vfmad_lane_f64(__s0_89, __s1_89, -__s2_89, __p3_89); \
+  __ret_89; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmss_lane_f32(__p0_90, __p1_90, __p2_90, __p3_90) __extension__ ({ \
+  float32_t __s0_90 = __p0_90; \
+  float32_t __s1_90 = __p1_90; \
+  float32x2_t __s2_90 = __p2_90; \
+  float32_t __ret_90; \
+  __ret_90 = vfmas_lane_f32(__s0_90, __s1_90, -__s2_90, __p3_90); \
+  __ret_90; \
+})
+#else
+#define vfmss_lane_f32(__p0_91, __p1_91, __p2_91, __p3_91) __extension__ ({ \
+  float32_t __s0_91 = __p0_91; \
+  float32_t __s1_91 = __p1_91; \
+  float32x2_t __s2_91 = __p2_91; \
+  float32x2_t __rev2_91;  __rev2_91 = __builtin_shufflevector(__s2_91, __s2_91, 1, 0); \
+  float32_t __ret_91; \
+  __ret_91 = __noswap_vfmas_lane_f32(__s0_91, __s1_91, -__rev2_91, __p3_91); \
+  __ret_91; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmsq_lane_f64(__p0_92, __p1_92, __p2_92, __p3_92) __extension__ ({ \
+  float64x2_t __s0_92 = __p0_92; \
+  float64x2_t __s1_92 = __p1_92; \
+  float64x1_t __s2_92 = __p2_92; \
+  float64x2_t __ret_92; \
+  __ret_92 = vfmaq_lane_f64(__s0_92, __s1_92, -__s2_92, __p3_92); \
+  __ret_92; \
+})
+#else
+#define vfmsq_lane_f64(__p0_93, __p1_93, __p2_93, __p3_93) __extension__ ({ \
+  float64x2_t __s0_93 = __p0_93; \
+  float64x2_t __s1_93 = __p1_93; \
+  float64x1_t __s2_93 = __p2_93; \
+  float64x2_t __rev0_93;  __rev0_93 = __builtin_shufflevector(__s0_93, __s0_93, 1, 0); \
+  float64x2_t __rev1_93;  __rev1_93 = __builtin_shufflevector(__s1_93, __s1_93, 1, 0); \
+  float64x2_t __ret_93; \
+  __ret_93 = __noswap_vfmaq_lane_f64(__rev0_93, __rev1_93, -__s2_93, __p3_93); \
+  __ret_93 = __builtin_shufflevector(__ret_93, __ret_93, 1, 0); \
+  __ret_93; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmsq_lane_f32(__p0_94, __p1_94, __p2_94, __p3_94) __extension__ ({ \
+  float32x4_t __s0_94 = __p0_94; \
+  float32x4_t __s1_94 = __p1_94; \
+  float32x2_t __s2_94 = __p2_94; \
+  float32x4_t __ret_94; \
+  __ret_94 = vfmaq_lane_f32(__s0_94, __s1_94, -__s2_94, __p3_94); \
+  __ret_94; \
+})
+#else
+#define vfmsq_lane_f32(__p0_95, __p1_95, __p2_95, __p3_95) __extension__ ({ \
+  float32x4_t __s0_95 = __p0_95; \
+  float32x4_t __s1_95 = __p1_95; \
+  float32x2_t __s2_95 = __p2_95; \
+  float32x4_t __rev0_95;  __rev0_95 = __builtin_shufflevector(__s0_95, __s0_95, 3, 2, 1, 0); \
+  float32x4_t __rev1_95;  __rev1_95 = __builtin_shufflevector(__s1_95, __s1_95, 3, 2, 1, 0); \
+  float32x2_t __rev2_95;  __rev2_95 = __builtin_shufflevector(__s2_95, __s2_95, 1, 0); \
+  float32x4_t __ret_95; \
+  __ret_95 = __noswap_vfmaq_lane_f32(__rev0_95, __rev1_95, -__rev2_95, __p3_95); \
+  __ret_95 = __builtin_shufflevector(__ret_95, __ret_95, 3, 2, 1, 0); \
+  __ret_95; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfms_lane_f64(__p0_96, __p1_96, __p2_96, __p3_96) __extension__ ({ \
+  float64x1_t __s0_96 = __p0_96; \
+  float64x1_t __s1_96 = __p1_96; \
+  float64x1_t __s2_96 = __p2_96; \
+  float64x1_t __ret_96; \
+  __ret_96 = vfma_lane_f64(__s0_96, __s1_96, -__s2_96, __p3_96); \
+  __ret_96; \
+})
+#else
+#define vfms_lane_f64(__p0_97, __p1_97, __p2_97, __p3_97) __extension__ ({ \
+  float64x1_t __s0_97 = __p0_97; \
+  float64x1_t __s1_97 = __p1_97; \
+  float64x1_t __s2_97 = __p2_97; \
+  float64x1_t __ret_97; \
+  __ret_97 = __noswap_vfma_lane_f64(__s0_97, __s1_97, -__s2_97, __p3_97); \
+  __ret_97; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfms_lane_f32(__p0_98, __p1_98, __p2_98, __p3_98) __extension__ ({ \
+  float32x2_t __s0_98 = __p0_98; \
+  float32x2_t __s1_98 = __p1_98; \
+  float32x2_t __s2_98 = __p2_98; \
+  float32x2_t __ret_98; \
+  __ret_98 = vfma_lane_f32(__s0_98, __s1_98, -__s2_98, __p3_98); \
+  __ret_98; \
+})
+#else
+#define vfms_lane_f32(__p0_99, __p1_99, __p2_99, __p3_99) __extension__ ({ \
+  float32x2_t __s0_99 = __p0_99; \
+  float32x2_t __s1_99 = __p1_99; \
+  float32x2_t __s2_99 = __p2_99; \
+  float32x2_t __rev0_99;  __rev0_99 = __builtin_shufflevector(__s0_99, __s0_99, 1, 0); \
+  float32x2_t __rev1_99;  __rev1_99 = __builtin_shufflevector(__s1_99, __s1_99, 1, 0); \
+  float32x2_t __rev2_99;  __rev2_99 = __builtin_shufflevector(__s2_99, __s2_99, 1, 0); \
+  float32x2_t __ret_99; \
+  __ret_99 = __noswap_vfma_lane_f32(__rev0_99, __rev1_99, -__rev2_99, __p3_99); \
+  __ret_99 = __builtin_shufflevector(__ret_99, __ret_99, 1, 0); \
+  __ret_99; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmsd_laneq_f64(__p0_100, __p1_100, __p2_100, __p3_100) __extension__ ({ \
+  float64_t __s0_100 = __p0_100; \
+  float64_t __s1_100 = __p1_100; \
+  float64x2_t __s2_100 = __p2_100; \
+  float64_t __ret_100; \
+  __ret_100 = vfmad_laneq_f64(__s0_100, __s1_100, -__s2_100, __p3_100); \
+  __ret_100; \
+})
+#else
+#define vfmsd_laneq_f64(__p0_101, __p1_101, __p2_101, __p3_101) __extension__ ({ \
+  float64_t __s0_101 = __p0_101; \
+  float64_t __s1_101 = __p1_101; \
+  float64x2_t __s2_101 = __p2_101; \
+  float64x2_t __rev2_101;  __rev2_101 = __builtin_shufflevector(__s2_101, __s2_101, 1, 0); \
+  float64_t __ret_101; \
+  __ret_101 = __noswap_vfmad_laneq_f64(__s0_101, __s1_101, -__rev2_101, __p3_101); \
+  __ret_101; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmss_laneq_f32(__p0_102, __p1_102, __p2_102, __p3_102) __extension__ ({ \
+  float32_t __s0_102 = __p0_102; \
+  float32_t __s1_102 = __p1_102; \
+  float32x4_t __s2_102 = __p2_102; \
+  float32_t __ret_102; \
+  __ret_102 = vfmas_laneq_f32(__s0_102, __s1_102, -__s2_102, __p3_102); \
+  __ret_102; \
+})
+#else
+#define vfmss_laneq_f32(__p0_103, __p1_103, __p2_103, __p3_103) __extension__ ({ \
+  float32_t __s0_103 = __p0_103; \
+  float32_t __s1_103 = __p1_103; \
+  float32x4_t __s2_103 = __p2_103; \
+  float32x4_t __rev2_103;  __rev2_103 = __builtin_shufflevector(__s2_103, __s2_103, 3, 2, 1, 0); \
+  float32_t __ret_103; \
+  __ret_103 = __noswap_vfmas_laneq_f32(__s0_103, __s1_103, -__rev2_103, __p3_103); \
+  __ret_103; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmsq_laneq_f64(__p0_104, __p1_104, __p2_104, __p3_104) __extension__ ({ \
+  float64x2_t __s0_104 = __p0_104; \
+  float64x2_t __s1_104 = __p1_104; \
+  float64x2_t __s2_104 = __p2_104; \
+  float64x2_t __ret_104; \
+  __ret_104 = vfmaq_laneq_f64(__s0_104, __s1_104, -__s2_104, __p3_104); \
+  __ret_104; \
+})
+#else
+#define vfmsq_laneq_f64(__p0_105, __p1_105, __p2_105, __p3_105) __extension__ ({ \
+  float64x2_t __s0_105 = __p0_105; \
+  float64x2_t __s1_105 = __p1_105; \
+  float64x2_t __s2_105 = __p2_105; \
+  float64x2_t __rev0_105;  __rev0_105 = __builtin_shufflevector(__s0_105, __s0_105, 1, 0); \
+  float64x2_t __rev1_105;  __rev1_105 = __builtin_shufflevector(__s1_105, __s1_105, 1, 0); \
+  float64x2_t __rev2_105;  __rev2_105 = __builtin_shufflevector(__s2_105, __s2_105, 1, 0); \
+  float64x2_t __ret_105; \
+  __ret_105 = __noswap_vfmaq_laneq_f64(__rev0_105, __rev1_105, -__rev2_105, __p3_105); \
+  __ret_105 = __builtin_shufflevector(__ret_105, __ret_105, 1, 0); \
+  __ret_105; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmsq_laneq_f32(__p0_106, __p1_106, __p2_106, __p3_106) __extension__ ({ \
+  float32x4_t __s0_106 = __p0_106; \
+  float32x4_t __s1_106 = __p1_106; \
+  float32x4_t __s2_106 = __p2_106; \
+  float32x4_t __ret_106; \
+  __ret_106 = vfmaq_laneq_f32(__s0_106, __s1_106, -__s2_106, __p3_106); \
+  __ret_106; \
+})
+#else
+#define vfmsq_laneq_f32(__p0_107, __p1_107, __p2_107, __p3_107) __extension__ ({ \
+  float32x4_t __s0_107 = __p0_107; \
+  float32x4_t __s1_107 = __p1_107; \
+  float32x4_t __s2_107 = __p2_107; \
+  float32x4_t __rev0_107;  __rev0_107 = __builtin_shufflevector(__s0_107, __s0_107, 3, 2, 1, 0); \
+  float32x4_t __rev1_107;  __rev1_107 = __builtin_shufflevector(__s1_107, __s1_107, 3, 2, 1, 0); \
+  float32x4_t __rev2_107;  __rev2_107 = __builtin_shufflevector(__s2_107, __s2_107, 3, 2, 1, 0); \
+  float32x4_t __ret_107; \
+  __ret_107 = __noswap_vfmaq_laneq_f32(__rev0_107, __rev1_107, -__rev2_107, __p3_107); \
+  __ret_107 = __builtin_shufflevector(__ret_107, __ret_107, 3, 2, 1, 0); \
+  __ret_107; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfms_laneq_f64(__p0_108, __p1_108, __p2_108, __p3_108) __extension__ ({ \
+  float64x1_t __s0_108 = __p0_108; \
+  float64x1_t __s1_108 = __p1_108; \
+  float64x2_t __s2_108 = __p2_108; \
+  float64x1_t __ret_108; \
+  __ret_108 = vfma_laneq_f64(__s0_108, __s1_108, -__s2_108, __p3_108); \
+  __ret_108; \
+})
+#else
+#define vfms_laneq_f64(__p0_109, __p1_109, __p2_109, __p3_109) __extension__ ({ \
+  float64x1_t __s0_109 = __p0_109; \
+  float64x1_t __s1_109 = __p1_109; \
+  float64x2_t __s2_109 = __p2_109; \
+  float64x2_t __rev2_109;  __rev2_109 = __builtin_shufflevector(__s2_109, __s2_109, 1, 0); \
+  float64x1_t __ret_109; \
+  __ret_109 = __noswap_vfma_laneq_f64(__s0_109, __s1_109, -__rev2_109, __p3_109); \
+  __ret_109; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfms_laneq_f32(__p0_110, __p1_110, __p2_110, __p3_110) __extension__ ({ \
+  float32x2_t __s0_110 = __p0_110; \
+  float32x2_t __s1_110 = __p1_110; \
+  float32x4_t __s2_110 = __p2_110; \
+  float32x2_t __ret_110; \
+  __ret_110 = vfma_laneq_f32(__s0_110, __s1_110, -__s2_110, __p3_110); \
+  __ret_110; \
+})
+#else
+#define vfms_laneq_f32(__p0_111, __p1_111, __p2_111, __p3_111) __extension__ ({ \
+  float32x2_t __s0_111 = __p0_111; \
+  float32x2_t __s1_111 = __p1_111; \
+  float32x4_t __s2_111 = __p2_111; \
+  float32x2_t __rev0_111;  __rev0_111 = __builtin_shufflevector(__s0_111, __s0_111, 1, 0); \
+  float32x2_t __rev1_111;  __rev1_111 = __builtin_shufflevector(__s1_111, __s1_111, 1, 0); \
+  float32x4_t __rev2_111;  __rev2_111 = __builtin_shufflevector(__s2_111, __s2_111, 3, 2, 1, 0); \
+  float32x2_t __ret_111; \
+  __ret_111 = __noswap_vfma_laneq_f32(__rev0_111, __rev1_111, -__rev2_111, __p3_111); \
+  __ret_111 = __builtin_shufflevector(__ret_111, __ret_111, 1, 0); \
+  __ret_111; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vfmsq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
+  float64x2_t __ret;
+  __ret = vfmsq_f64(__p0, __p1, (float64x2_t) {__p2, __p2});
+  return __ret;
+}
+#else
+__ai float64x2_t vfmsq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = __noswap_vfmsq_f64(__rev0, __rev1, (float64x2_t) {__p2, __p2});
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vfmsq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
+  float32x4_t __ret;
+  __ret = vfmsq_f32(__p0, __p1, (float32x4_t) {__p2, __p2, __p2, __p2});
+  return __ret;
+}
+#else
+__ai float32x4_t vfmsq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = __noswap_vfmsq_f32(__rev0, __rev1, (float32x4_t) {__p2, __p2, __p2, __p2});
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vfms_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
+  float32x2_t __ret;
+  __ret = vfms_f32(__p0, __p1, (float32x2_t) {__p2, __p2});
+  return __ret;
+}
+#else
+__ai float32x2_t vfms_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = __noswap_vfms_f32(__rev0, __rev1, (float32x2_t) {__p2, __p2});
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x1_t vget_high_p64(poly64x2_t __p0) {
+  poly64x1_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1);
+  return __ret;
+}
+#else
+__ai poly64x1_t vget_high_p64(poly64x2_t __p0) {
+  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  poly64x1_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1);
+  return __ret;
+}
+__ai poly64x1_t __noswap_vget_high_p64(poly64x2_t __p0) {
+  poly64x1_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vget_high_f64(float64x2_t __p0) {
+  float64x1_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 1);
+  return __ret;
+}
+#else
+__ai float64x1_t vget_high_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x1_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vget_lane_p64(__p0, __p1) __extension__ ({ \
+  poly64x1_t __s0 = __p0; \
+  poly64_t __ret; \
+  __ret = (poly64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vget_lane_p64(__p0, __p1) __extension__ ({ \
+  poly64x1_t __s0 = __p0; \
+  poly64_t __ret; \
+  __ret = (poly64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#define __noswap_vget_lane_p64(__p0, __p1) __extension__ ({ \
+  poly64x1_t __s0 = __p0; \
+  poly64_t __ret; \
+  __ret = (poly64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vgetq_lane_p64(__p0, __p1) __extension__ ({ \
+  poly64x2_t __s0 = __p0; \
+  poly64_t __ret; \
+  __ret = (poly64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vgetq_lane_p64(__p0, __p1) __extension__ ({ \
+  poly64x2_t __s0 = __p0; \
+  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  poly64_t __ret; \
+  __ret = (poly64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vgetq_lane_p64(__p0, __p1) __extension__ ({ \
+  poly64x2_t __s0 = __p0; \
+  poly64_t __ret; \
+  __ret = (poly64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vgetq_lane_f64(__p0, __p1) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64_t __ret; \
+  __ret = (float64_t) __builtin_neon_vgetq_lane_f64((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vgetq_lane_f64(__p0, __p1) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float64_t __ret; \
+  __ret = (float64_t) __builtin_neon_vgetq_lane_f64((int8x16_t)__rev0, __p1); \
+  __ret; \
+})
+#define __noswap_vgetq_lane_f64(__p0, __p1) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64_t __ret; \
+  __ret = (float64_t) __builtin_neon_vgetq_lane_f64((int8x16_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vget_lane_f64(__p0, __p1) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64_t __ret; \
+  __ret = (float64_t) __builtin_neon_vget_lane_f64((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#else
+#define vget_lane_f64(__p0, __p1) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64_t __ret; \
+  __ret = (float64_t) __builtin_neon_vget_lane_f64((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#define __noswap_vget_lane_f64(__p0, __p1) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64_t __ret; \
+  __ret = (float64_t) __builtin_neon_vget_lane_f64((int8x8_t)__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vget_lane_f16(__p0_112, __p1_112) __extension__ ({ \
+  float16x4_t __s0_112 = __p0_112; \
+  float16_t __ret_112; \
+float16x4_t __reint_112 = __s0_112; \
+int16_t __reint1_112 = vget_lane_s16(*(int16x4_t *) &__reint_112, __p1_112); \
+  __ret_112 = *(float16_t *) &__reint1_112; \
+  __ret_112; \
+})
+#else
+#define vget_lane_f16(__p0_113, __p1_113) __extension__ ({ \
+  float16x4_t __s0_113 = __p0_113; \
+  float16x4_t __rev0_113;  __rev0_113 = __builtin_shufflevector(__s0_113, __s0_113, 3, 2, 1, 0); \
+  float16_t __ret_113; \
+float16x4_t __reint_113 = __rev0_113; \
+int16_t __reint1_113 = __noswap_vget_lane_s16(*(int16x4_t *) &__reint_113, __p1_113); \
+  __ret_113 = *(float16_t *) &__reint1_113; \
+  __ret_113; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vgetq_lane_f16(__p0_114, __p1_114) __extension__ ({ \
+  float16x8_t __s0_114 = __p0_114; \
+  float16_t __ret_114; \
+float16x8_t __reint_114 = __s0_114; \
+int16_t __reint1_114 = vgetq_lane_s16(*(int16x8_t *) &__reint_114, __p1_114); \
+  __ret_114 = *(float16_t *) &__reint1_114; \
+  __ret_114; \
+})
+#else
+#define vgetq_lane_f16(__p0_115, __p1_115) __extension__ ({ \
+  float16x8_t __s0_115 = __p0_115; \
+  float16x8_t __rev0_115;  __rev0_115 = __builtin_shufflevector(__s0_115, __s0_115, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16_t __ret_115; \
+float16x8_t __reint_115 = __rev0_115; \
+int16_t __reint1_115 = __noswap_vgetq_lane_s16(*(int16x8_t *) &__reint_115, __p1_115); \
+  __ret_115 = *(float16_t *) &__reint1_115; \
+  __ret_115; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x1_t vget_low_p64(poly64x2_t __p0) {
+  poly64x1_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 0);
+  return __ret;
+}
+#else
+__ai poly64x1_t vget_low_p64(poly64x2_t __p0) {
+  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  poly64x1_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vget_low_f64(float64x2_t __p0) {
+  float64x1_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p0, 0);
+  return __ret;
+}
+#else
+__ai float64x1_t vget_low_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x1_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev0, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_p64(__p0) __extension__ ({ \
+  poly64x1_t __ret; \
+  __ret = (poly64x1_t) __builtin_neon_vld1_v(__p0, 6); \
+  __ret; \
+})
+#else
+#define vld1_p64(__p0) __extension__ ({ \
+  poly64x1_t __ret; \
+  __ret = (poly64x1_t) __builtin_neon_vld1_v(__p0, 6); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_p64(__p0) __extension__ ({ \
+  poly64x2_t __ret; \
+  __ret = (poly64x2_t) __builtin_neon_vld1q_v(__p0, 38); \
+  __ret; \
+})
+#else
+#define vld1q_p64(__p0) __extension__ ({ \
+  poly64x2_t __ret; \
+  __ret = (poly64x2_t) __builtin_neon_vld1q_v(__p0, 38); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_f64(__p0) __extension__ ({ \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vld1q_v(__p0, 42); \
+  __ret; \
+})
+#else
+#define vld1q_f64(__p0) __extension__ ({ \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vld1q_v(__p0, 42); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_f64(__p0) __extension__ ({ \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vld1_v(__p0, 10); \
+  __ret; \
+})
+#else
+#define vld1_f64(__p0) __extension__ ({ \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vld1_v(__p0, 10); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_dup_p64(__p0) __extension__ ({ \
+  poly64x1_t __ret; \
+  __ret = (poly64x1_t) __builtin_neon_vld1_dup_v(__p0, 6); \
+  __ret; \
+})
+#else
+#define vld1_dup_p64(__p0) __extension__ ({ \
+  poly64x1_t __ret; \
+  __ret = (poly64x1_t) __builtin_neon_vld1_dup_v(__p0, 6); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_dup_p64(__p0) __extension__ ({ \
+  poly64x2_t __ret; \
+  __ret = (poly64x2_t) __builtin_neon_vld1q_dup_v(__p0, 38); \
+  __ret; \
+})
+#else
+#define vld1q_dup_p64(__p0) __extension__ ({ \
+  poly64x2_t __ret; \
+  __ret = (poly64x2_t) __builtin_neon_vld1q_dup_v(__p0, 38); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_dup_f64(__p0) __extension__ ({ \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vld1q_dup_v(__p0, 42); \
+  __ret; \
+})
+#else
+#define vld1q_dup_f64(__p0) __extension__ ({ \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vld1q_dup_v(__p0, 42); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_dup_f64(__p0) __extension__ ({ \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vld1_dup_v(__p0, 10); \
+  __ret; \
+})
+#else
+#define vld1_dup_f64(__p0) __extension__ ({ \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vld1_dup_v(__p0, 10); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1_t __s1 = __p1; \
+  poly64x1_t __ret; \
+  __ret = (poly64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 6); \
+  __ret; \
+})
+#else
+#define vld1_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1_t __s1 = __p1; \
+  poly64x1_t __ret; \
+  __ret = (poly64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 6); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2_t __s1 = __p1; \
+  poly64x2_t __ret; \
+  __ret = (poly64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 38); \
+  __ret; \
+})
+#else
+#define vld1q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2_t __s1 = __p1; \
+  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  poly64x2_t __ret; \
+  __ret = (poly64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 38); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2_t __s1 = __p1; \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 42); \
+  __ret; \
+})
+#else
+#define vld1q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2_t __s1 = __p1; \
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 42); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1_t __s1 = __p1; \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 10); \
+  __ret; \
+})
+#else
+#define vld1_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1_t __s1 = __p1; \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 10); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_p8_x2(__p0) __extension__ ({ \
+  poly8x8x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 4); \
+  __ret; \
+})
+#else
+#define vld1_p8_x2(__p0) __extension__ ({ \
+  poly8x8x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 4); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_p64_x2(__p0) __extension__ ({ \
+  poly64x1x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 6); \
+  __ret; \
+})
+#else
+#define vld1_p64_x2(__p0) __extension__ ({ \
+  poly64x1x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 6); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_p16_x2(__p0) __extension__ ({ \
+  poly16x4x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 5); \
+  __ret; \
+})
+#else
+#define vld1_p16_x2(__p0) __extension__ ({ \
+  poly16x4x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 5); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_p8_x2(__p0) __extension__ ({ \
+  poly8x16x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 36); \
+  __ret; \
+})
+#else
+#define vld1q_p8_x2(__p0) __extension__ ({ \
+  poly8x16x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 36); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_p64_x2(__p0) __extension__ ({ \
+  poly64x2x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 38); \
+  __ret; \
+})
+#else
+#define vld1q_p64_x2(__p0) __extension__ ({ \
+  poly64x2x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 38); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_p16_x2(__p0) __extension__ ({ \
+  poly16x8x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 37); \
+  __ret; \
+})
+#else
+#define vld1q_p16_x2(__p0) __extension__ ({ \
+  poly16x8x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 37); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_u8_x2(__p0) __extension__ ({ \
+  uint8x16x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 48); \
+  __ret; \
+})
+#else
+#define vld1q_u8_x2(__p0) __extension__ ({ \
+  uint8x16x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 48); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_u32_x2(__p0) __extension__ ({ \
+  uint32x4x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 50); \
+  __ret; \
+})
+#else
+#define vld1q_u32_x2(__p0) __extension__ ({ \
+  uint32x4x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 50); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_u64_x2(__p0) __extension__ ({ \
+  uint64x2x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 51); \
+  __ret; \
+})
+#else
+#define vld1q_u64_x2(__p0) __extension__ ({ \
+  uint64x2x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 51); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_u16_x2(__p0) __extension__ ({ \
+  uint16x8x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 49); \
+  __ret; \
+})
+#else
+#define vld1q_u16_x2(__p0) __extension__ ({ \
+  uint16x8x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 49); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_s8_x2(__p0) __extension__ ({ \
+  int8x16x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 32); \
+  __ret; \
+})
+#else
+#define vld1q_s8_x2(__p0) __extension__ ({ \
+  int8x16x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 32); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_f64_x2(__p0) __extension__ ({ \
+  float64x2x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 42); \
+  __ret; \
+})
+#else
+#define vld1q_f64_x2(__p0) __extension__ ({ \
+  float64x2x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 42); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_f32_x2(__p0) __extension__ ({ \
+  float32x4x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 41); \
+  __ret; \
+})
+#else
+#define vld1q_f32_x2(__p0) __extension__ ({ \
+  float32x4x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 41); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_f16_x2(__p0) __extension__ ({ \
+  float16x8x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 40); \
+  __ret; \
+})
+#else
+#define vld1q_f16_x2(__p0) __extension__ ({ \
+  float16x8x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 40); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_s32_x2(__p0) __extension__ ({ \
+  int32x4x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 34); \
+  __ret; \
+})
+#else
+#define vld1q_s32_x2(__p0) __extension__ ({ \
+  int32x4x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 34); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_s64_x2(__p0) __extension__ ({ \
+  int64x2x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 35); \
+  __ret; \
+})
+#else
+#define vld1q_s64_x2(__p0) __extension__ ({ \
+  int64x2x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 35); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_s16_x2(__p0) __extension__ ({ \
+  int16x8x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 33); \
+  __ret; \
+})
+#else
+#define vld1q_s16_x2(__p0) __extension__ ({ \
+  int16x8x2_t __ret; \
+  __builtin_neon_vld1q_x2_v(&__ret, __p0, 33); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_u8_x2(__p0) __extension__ ({ \
+  uint8x8x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 16); \
+  __ret; \
+})
+#else
+#define vld1_u8_x2(__p0) __extension__ ({ \
+  uint8x8x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 16); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_u32_x2(__p0) __extension__ ({ \
+  uint32x2x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 18); \
+  __ret; \
+})
+#else
+#define vld1_u32_x2(__p0) __extension__ ({ \
+  uint32x2x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 18); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_u64_x2(__p0) __extension__ ({ \
+  uint64x1x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 19); \
+  __ret; \
+})
+#else
+#define vld1_u64_x2(__p0) __extension__ ({ \
+  uint64x1x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_u16_x2(__p0) __extension__ ({ \
+  uint16x4x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 17); \
+  __ret; \
+})
+#else
+#define vld1_u16_x2(__p0) __extension__ ({ \
+  uint16x4x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 17); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_s8_x2(__p0) __extension__ ({ \
+  int8x8x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 0); \
+  __ret; \
+})
+#else
+#define vld1_s8_x2(__p0) __extension__ ({ \
+  int8x8x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 0); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_f64_x2(__p0) __extension__ ({ \
+  float64x1x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 10); \
+  __ret; \
+})
+#else
+#define vld1_f64_x2(__p0) __extension__ ({ \
+  float64x1x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 10); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_f32_x2(__p0) __extension__ ({ \
+  float32x2x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 9); \
+  __ret; \
+})
+#else
+#define vld1_f32_x2(__p0) __extension__ ({ \
+  float32x2x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 9); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_f16_x2(__p0) __extension__ ({ \
+  float16x4x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 8); \
+  __ret; \
+})
+#else
+#define vld1_f16_x2(__p0) __extension__ ({ \
+  float16x4x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 8); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_s32_x2(__p0) __extension__ ({ \
+  int32x2x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 2); \
+  __ret; \
+})
+#else
+#define vld1_s32_x2(__p0) __extension__ ({ \
+  int32x2x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 2); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_s64_x2(__p0) __extension__ ({ \
+  int64x1x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 3); \
+  __ret; \
+})
+#else
+#define vld1_s64_x2(__p0) __extension__ ({ \
+  int64x1x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_s16_x2(__p0) __extension__ ({ \
+  int16x4x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 1); \
+  __ret; \
+})
+#else
+#define vld1_s16_x2(__p0) __extension__ ({ \
+  int16x4x2_t __ret; \
+  __builtin_neon_vld1_x2_v(&__ret, __p0, 1); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_p8_x3(__p0) __extension__ ({ \
+  poly8x8x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 4); \
+  __ret; \
+})
+#else
+#define vld1_p8_x3(__p0) __extension__ ({ \
+  poly8x8x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 4); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_p64_x3(__p0) __extension__ ({ \
+  poly64x1x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 6); \
+  __ret; \
+})
+#else
+#define vld1_p64_x3(__p0) __extension__ ({ \
+  poly64x1x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 6); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_p16_x3(__p0) __extension__ ({ \
+  poly16x4x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 5); \
+  __ret; \
+})
+#else
+#define vld1_p16_x3(__p0) __extension__ ({ \
+  poly16x4x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 5); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_p8_x3(__p0) __extension__ ({ \
+  poly8x16x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 36); \
+  __ret; \
+})
+#else
+#define vld1q_p8_x3(__p0) __extension__ ({ \
+  poly8x16x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 36); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_p64_x3(__p0) __extension__ ({ \
+  poly64x2x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 38); \
+  __ret; \
+})
+#else
+#define vld1q_p64_x3(__p0) __extension__ ({ \
+  poly64x2x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 38); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_p16_x3(__p0) __extension__ ({ \
+  poly16x8x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 37); \
+  __ret; \
+})
+#else
+#define vld1q_p16_x3(__p0) __extension__ ({ \
+  poly16x8x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 37); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_u8_x3(__p0) __extension__ ({ \
+  uint8x16x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 48); \
+  __ret; \
+})
+#else
+#define vld1q_u8_x3(__p0) __extension__ ({ \
+  uint8x16x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 48); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_u32_x3(__p0) __extension__ ({ \
+  uint32x4x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 50); \
+  __ret; \
+})
+#else
+#define vld1q_u32_x3(__p0) __extension__ ({ \
+  uint32x4x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 50); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_u64_x3(__p0) __extension__ ({ \
+  uint64x2x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 51); \
+  __ret; \
+})
+#else
+#define vld1q_u64_x3(__p0) __extension__ ({ \
+  uint64x2x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 51); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_u16_x3(__p0) __extension__ ({ \
+  uint16x8x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 49); \
+  __ret; \
+})
+#else
+#define vld1q_u16_x3(__p0) __extension__ ({ \
+  uint16x8x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 49); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_s8_x3(__p0) __extension__ ({ \
+  int8x16x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 32); \
+  __ret; \
+})
+#else
+#define vld1q_s8_x3(__p0) __extension__ ({ \
+  int8x16x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 32); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_f64_x3(__p0) __extension__ ({ \
+  float64x2x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 42); \
+  __ret; \
+})
+#else
+#define vld1q_f64_x3(__p0) __extension__ ({ \
+  float64x2x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 42); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_f32_x3(__p0) __extension__ ({ \
+  float32x4x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 41); \
+  __ret; \
+})
+#else
+#define vld1q_f32_x3(__p0) __extension__ ({ \
+  float32x4x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 41); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_f16_x3(__p0) __extension__ ({ \
+  float16x8x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 40); \
+  __ret; \
+})
+#else
+#define vld1q_f16_x3(__p0) __extension__ ({ \
+  float16x8x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 40); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_s32_x3(__p0) __extension__ ({ \
+  int32x4x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 34); \
+  __ret; \
+})
+#else
+#define vld1q_s32_x3(__p0) __extension__ ({ \
+  int32x4x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 34); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_s64_x3(__p0) __extension__ ({ \
+  int64x2x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 35); \
+  __ret; \
+})
+#else
+#define vld1q_s64_x3(__p0) __extension__ ({ \
+  int64x2x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 35); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_s16_x3(__p0) __extension__ ({ \
+  int16x8x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 33); \
+  __ret; \
+})
+#else
+#define vld1q_s16_x3(__p0) __extension__ ({ \
+  int16x8x3_t __ret; \
+  __builtin_neon_vld1q_x3_v(&__ret, __p0, 33); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_u8_x3(__p0) __extension__ ({ \
+  uint8x8x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 16); \
+  __ret; \
+})
+#else
+#define vld1_u8_x3(__p0) __extension__ ({ \
+  uint8x8x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 16); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_u32_x3(__p0) __extension__ ({ \
+  uint32x2x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 18); \
+  __ret; \
+})
+#else
+#define vld1_u32_x3(__p0) __extension__ ({ \
+  uint32x2x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 18); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_u64_x3(__p0) __extension__ ({ \
+  uint64x1x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 19); \
+  __ret; \
+})
+#else
+#define vld1_u64_x3(__p0) __extension__ ({ \
+  uint64x1x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_u16_x3(__p0) __extension__ ({ \
+  uint16x4x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 17); \
+  __ret; \
+})
+#else
+#define vld1_u16_x3(__p0) __extension__ ({ \
+  uint16x4x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 17); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_s8_x3(__p0) __extension__ ({ \
+  int8x8x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 0); \
+  __ret; \
+})
+#else
+#define vld1_s8_x3(__p0) __extension__ ({ \
+  int8x8x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 0); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_f64_x3(__p0) __extension__ ({ \
+  float64x1x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 10); \
+  __ret; \
+})
+#else
+#define vld1_f64_x3(__p0) __extension__ ({ \
+  float64x1x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 10); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_f32_x3(__p0) __extension__ ({ \
+  float32x2x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 9); \
+  __ret; \
+})
+#else
+#define vld1_f32_x3(__p0) __extension__ ({ \
+  float32x2x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 9); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_f16_x3(__p0) __extension__ ({ \
+  float16x4x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 8); \
+  __ret; \
+})
+#else
+#define vld1_f16_x3(__p0) __extension__ ({ \
+  float16x4x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 8); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_s32_x3(__p0) __extension__ ({ \
+  int32x2x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 2); \
+  __ret; \
+})
+#else
+#define vld1_s32_x3(__p0) __extension__ ({ \
+  int32x2x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 2); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_s64_x3(__p0) __extension__ ({ \
+  int64x1x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 3); \
+  __ret; \
+})
+#else
+#define vld1_s64_x3(__p0) __extension__ ({ \
+  int64x1x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_s16_x3(__p0) __extension__ ({ \
+  int16x4x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 1); \
+  __ret; \
+})
+#else
+#define vld1_s16_x3(__p0) __extension__ ({ \
+  int16x4x3_t __ret; \
+  __builtin_neon_vld1_x3_v(&__ret, __p0, 1); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_p8_x4(__p0) __extension__ ({ \
+  poly8x8x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 4); \
+  __ret; \
+})
+#else
+#define vld1_p8_x4(__p0) __extension__ ({ \
+  poly8x8x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 4); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_p64_x4(__p0) __extension__ ({ \
+  poly64x1x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 6); \
+  __ret; \
+})
+#else
+#define vld1_p64_x4(__p0) __extension__ ({ \
+  poly64x1x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 6); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_p16_x4(__p0) __extension__ ({ \
+  poly16x4x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 5); \
+  __ret; \
+})
+#else
+#define vld1_p16_x4(__p0) __extension__ ({ \
+  poly16x4x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 5); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_p8_x4(__p0) __extension__ ({ \
+  poly8x16x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 36); \
+  __ret; \
+})
+#else
+#define vld1q_p8_x4(__p0) __extension__ ({ \
+  poly8x16x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 36); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_p64_x4(__p0) __extension__ ({ \
+  poly64x2x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 38); \
+  __ret; \
+})
+#else
+#define vld1q_p64_x4(__p0) __extension__ ({ \
+  poly64x2x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 38); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_p16_x4(__p0) __extension__ ({ \
+  poly16x8x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 37); \
+  __ret; \
+})
+#else
+#define vld1q_p16_x4(__p0) __extension__ ({ \
+  poly16x8x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 37); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_u8_x4(__p0) __extension__ ({ \
+  uint8x16x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 48); \
+  __ret; \
+})
+#else
+#define vld1q_u8_x4(__p0) __extension__ ({ \
+  uint8x16x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 48); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_u32_x4(__p0) __extension__ ({ \
+  uint32x4x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 50); \
+  __ret; \
+})
+#else
+#define vld1q_u32_x4(__p0) __extension__ ({ \
+  uint32x4x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 50); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_u64_x4(__p0) __extension__ ({ \
+  uint64x2x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 51); \
+  __ret; \
+})
+#else
+#define vld1q_u64_x4(__p0) __extension__ ({ \
+  uint64x2x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 51); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_u16_x4(__p0) __extension__ ({ \
+  uint16x8x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 49); \
+  __ret; \
+})
+#else
+#define vld1q_u16_x4(__p0) __extension__ ({ \
+  uint16x8x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 49); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_s8_x4(__p0) __extension__ ({ \
+  int8x16x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 32); \
+  __ret; \
+})
+#else
+#define vld1q_s8_x4(__p0) __extension__ ({ \
+  int8x16x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 32); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_f64_x4(__p0) __extension__ ({ \
+  float64x2x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 42); \
+  __ret; \
+})
+#else
+#define vld1q_f64_x4(__p0) __extension__ ({ \
+  float64x2x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 42); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_f32_x4(__p0) __extension__ ({ \
+  float32x4x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 41); \
+  __ret; \
+})
+#else
+#define vld1q_f32_x4(__p0) __extension__ ({ \
+  float32x4x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 41); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_f16_x4(__p0) __extension__ ({ \
+  float16x8x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 40); \
+  __ret; \
+})
+#else
+#define vld1q_f16_x4(__p0) __extension__ ({ \
+  float16x8x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 40); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_s32_x4(__p0) __extension__ ({ \
+  int32x4x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 34); \
+  __ret; \
+})
+#else
+#define vld1q_s32_x4(__p0) __extension__ ({ \
+  int32x4x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 34); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_s64_x4(__p0) __extension__ ({ \
+  int64x2x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 35); \
+  __ret; \
+})
+#else
+#define vld1q_s64_x4(__p0) __extension__ ({ \
+  int64x2x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 35); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1q_s16_x4(__p0) __extension__ ({ \
+  int16x8x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 33); \
+  __ret; \
+})
+#else
+#define vld1q_s16_x4(__p0) __extension__ ({ \
+  int16x8x4_t __ret; \
+  __builtin_neon_vld1q_x4_v(&__ret, __p0, 33); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_u8_x4(__p0) __extension__ ({ \
+  uint8x8x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 16); \
+  __ret; \
+})
+#else
+#define vld1_u8_x4(__p0) __extension__ ({ \
+  uint8x8x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 16); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_u32_x4(__p0) __extension__ ({ \
+  uint32x2x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 18); \
+  __ret; \
+})
+#else
+#define vld1_u32_x4(__p0) __extension__ ({ \
+  uint32x2x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 18); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_u64_x4(__p0) __extension__ ({ \
+  uint64x1x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 19); \
+  __ret; \
+})
+#else
+#define vld1_u64_x4(__p0) __extension__ ({ \
+  uint64x1x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_u16_x4(__p0) __extension__ ({ \
+  uint16x4x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 17); \
+  __ret; \
+})
+#else
+#define vld1_u16_x4(__p0) __extension__ ({ \
+  uint16x4x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 17); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_s8_x4(__p0) __extension__ ({ \
+  int8x8x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 0); \
+  __ret; \
+})
+#else
+#define vld1_s8_x4(__p0) __extension__ ({ \
+  int8x8x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 0); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_f64_x4(__p0) __extension__ ({ \
+  float64x1x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 10); \
+  __ret; \
+})
+#else
+#define vld1_f64_x4(__p0) __extension__ ({ \
+  float64x1x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 10); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_f32_x4(__p0) __extension__ ({ \
+  float32x2x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 9); \
+  __ret; \
+})
+#else
+#define vld1_f32_x4(__p0) __extension__ ({ \
+  float32x2x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 9); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_f16_x4(__p0) __extension__ ({ \
+  float16x4x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 8); \
+  __ret; \
+})
+#else
+#define vld1_f16_x4(__p0) __extension__ ({ \
+  float16x4x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 8); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_s32_x4(__p0) __extension__ ({ \
+  int32x2x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 2); \
+  __ret; \
+})
+#else
+#define vld1_s32_x4(__p0) __extension__ ({ \
+  int32x2x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 2); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_s64_x4(__p0) __extension__ ({ \
+  int64x1x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 3); \
+  __ret; \
+})
+#else
+#define vld1_s64_x4(__p0) __extension__ ({ \
+  int64x1x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld1_s16_x4(__p0) __extension__ ({ \
+  int16x4x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 1); \
+  __ret; \
+})
+#else
+#define vld1_s16_x4(__p0) __extension__ ({ \
+  int16x4x4_t __ret; \
+  __builtin_neon_vld1_x4_v(&__ret, __p0, 1); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_p64(__p0) __extension__ ({ \
+  poly64x1x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 6); \
+  __ret; \
+})
+#else
+#define vld2_p64(__p0) __extension__ ({ \
+  poly64x1x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 6); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_p64(__p0) __extension__ ({ \
+  poly64x2x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 38); \
+  __ret; \
+})
+#else
+#define vld2q_p64(__p0) __extension__ ({ \
+  poly64x2x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 38); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_u64(__p0) __extension__ ({ \
+  uint64x2x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 51); \
+  __ret; \
+})
+#else
+#define vld2q_u64(__p0) __extension__ ({ \
+  uint64x2x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 51); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_f64(__p0) __extension__ ({ \
+  float64x2x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 42); \
+  __ret; \
+})
+#else
+#define vld2q_f64(__p0) __extension__ ({ \
+  float64x2x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 42); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_s64(__p0) __extension__ ({ \
+  int64x2x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 35); \
+  __ret; \
+})
+#else
+#define vld2q_s64(__p0) __extension__ ({ \
+  int64x2x2_t __ret; \
+  __builtin_neon_vld2q_v(&__ret, __p0, 35); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_f64(__p0) __extension__ ({ \
+  float64x1x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 10); \
+  __ret; \
+})
+#else
+#define vld2_f64(__p0) __extension__ ({ \
+  float64x1x2_t __ret; \
+  __builtin_neon_vld2_v(&__ret, __p0, 10); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_dup_p64(__p0) __extension__ ({ \
+  poly64x1x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 6); \
+  __ret; \
+})
+#else
+#define vld2_dup_p64(__p0) __extension__ ({ \
+  poly64x1x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 6); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_dup_p8(__p0) __extension__ ({ \
+  poly8x16x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 36); \
+  __ret; \
+})
+#else
+#define vld2q_dup_p8(__p0) __extension__ ({ \
+  poly8x16x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 36); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_dup_p64(__p0) __extension__ ({ \
+  poly64x2x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 38); \
+  __ret; \
+})
+#else
+#define vld2q_dup_p64(__p0) __extension__ ({ \
+  poly64x2x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 38); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_dup_p16(__p0) __extension__ ({ \
+  poly16x8x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 37); \
+  __ret; \
+})
+#else
+#define vld2q_dup_p16(__p0) __extension__ ({ \
+  poly16x8x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 37); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_dup_u8(__p0) __extension__ ({ \
+  uint8x16x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 48); \
+  __ret; \
+})
+#else
+#define vld2q_dup_u8(__p0) __extension__ ({ \
+  uint8x16x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 48); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_dup_u32(__p0) __extension__ ({ \
+  uint32x4x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 50); \
+  __ret; \
+})
+#else
+#define vld2q_dup_u32(__p0) __extension__ ({ \
+  uint32x4x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 50); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_dup_u64(__p0) __extension__ ({ \
+  uint64x2x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 51); \
+  __ret; \
+})
+#else
+#define vld2q_dup_u64(__p0) __extension__ ({ \
+  uint64x2x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 51); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_dup_u16(__p0) __extension__ ({ \
+  uint16x8x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 49); \
+  __ret; \
+})
+#else
+#define vld2q_dup_u16(__p0) __extension__ ({ \
+  uint16x8x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 49); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_dup_s8(__p0) __extension__ ({ \
+  int8x16x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 32); \
+  __ret; \
+})
+#else
+#define vld2q_dup_s8(__p0) __extension__ ({ \
+  int8x16x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 32); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_dup_f64(__p0) __extension__ ({ \
+  float64x2x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 42); \
+  __ret; \
+})
+#else
+#define vld2q_dup_f64(__p0) __extension__ ({ \
+  float64x2x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 42); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_dup_f32(__p0) __extension__ ({ \
+  float32x4x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 41); \
+  __ret; \
+})
+#else
+#define vld2q_dup_f32(__p0) __extension__ ({ \
+  float32x4x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 41); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_dup_f16(__p0) __extension__ ({ \
+  float16x8x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 40); \
+  __ret; \
+})
+#else
+#define vld2q_dup_f16(__p0) __extension__ ({ \
+  float16x8x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 40); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_dup_s32(__p0) __extension__ ({ \
+  int32x4x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 34); \
+  __ret; \
+})
+#else
+#define vld2q_dup_s32(__p0) __extension__ ({ \
+  int32x4x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 34); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_dup_s64(__p0) __extension__ ({ \
+  int64x2x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 35); \
+  __ret; \
+})
+#else
+#define vld2q_dup_s64(__p0) __extension__ ({ \
+  int64x2x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 35); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_dup_s16(__p0) __extension__ ({ \
+  int16x8x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 33); \
+  __ret; \
+})
+#else
+#define vld2q_dup_s16(__p0) __extension__ ({ \
+  int16x8x2_t __ret; \
+  __builtin_neon_vld2q_dup_v(&__ret, __p0, 33); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_dup_f64(__p0) __extension__ ({ \
+  float64x1x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 10); \
+  __ret; \
+})
+#else
+#define vld2_dup_f64(__p0) __extension__ ({ \
+  float64x1x2_t __ret; \
+  __builtin_neon_vld2_dup_v(&__ret, __p0, 10); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1x2_t __s1 = __p1; \
+  poly64x1x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 6); \
+  __ret; \
+})
+#else
+#define vld2_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1x2_t __s1 = __p1; \
+  poly64x1x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 6); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16x2_t __s1 = __p1; \
+  poly8x16x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 36); \
+  __ret; \
+})
+#else
+#define vld2q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16x2_t __s1 = __p1; \
+  poly8x16x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 36); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2x2_t __s1 = __p1; \
+  poly64x2x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 38); \
+  __ret; \
+})
+#else
+#define vld2q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2x2_t __s1 = __p1; \
+  poly64x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  poly64x2x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 38); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16x2_t __s1 = __p1; \
+  uint8x16x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 48); \
+  __ret; \
+})
+#else
+#define vld2q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16x2_t __s1 = __p1; \
+  uint8x16x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 48); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2x2_t __s1 = __p1; \
+  uint64x2x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 51); \
+  __ret; \
+})
+#else
+#define vld2q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2x2_t __s1 = __p1; \
+  uint64x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  uint64x2x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 51); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16x2_t __s1 = __p1; \
+  int8x16x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 32); \
+  __ret; \
+})
+#else
+#define vld2q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16x2_t __s1 = __p1; \
+  int8x16x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 32); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2x2_t __s1 = __p1; \
+  float64x2x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 42); \
+  __ret; \
+})
+#else
+#define vld2q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2x2_t __s1 = __p1; \
+  float64x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  float64x2x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 42); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2x2_t __s1 = __p1; \
+  int64x2x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 35); \
+  __ret; \
+})
+#else
+#define vld2q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2x2_t __s1 = __p1; \
+  int64x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  int64x2x2_t __ret; \
+  __builtin_neon_vld2q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 35); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1x2_t __s1 = __p1; \
+  uint64x1x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 19); \
+  __ret; \
+})
+#else
+#define vld2_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1x2_t __s1 = __p1; \
+  uint64x1x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1x2_t __s1 = __p1; \
+  float64x1x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 10); \
+  __ret; \
+})
+#else
+#define vld2_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1x2_t __s1 = __p1; \
+  float64x1x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 10); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld2_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1x2_t __s1 = __p1; \
+  int64x1x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 3); \
+  __ret; \
+})
+#else
+#define vld2_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1x2_t __s1 = __p1; \
+  int64x1x2_t __ret; \
+  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_p64(__p0) __extension__ ({ \
+  poly64x1x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 6); \
+  __ret; \
+})
+#else
+#define vld3_p64(__p0) __extension__ ({ \
+  poly64x1x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 6); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_p64(__p0) __extension__ ({ \
+  poly64x2x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 38); \
+  __ret; \
+})
+#else
+#define vld3q_p64(__p0) __extension__ ({ \
+  poly64x2x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 38); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_u64(__p0) __extension__ ({ \
+  uint64x2x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 51); \
+  __ret; \
+})
+#else
+#define vld3q_u64(__p0) __extension__ ({ \
+  uint64x2x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 51); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_f64(__p0) __extension__ ({ \
+  float64x2x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 42); \
+  __ret; \
+})
+#else
+#define vld3q_f64(__p0) __extension__ ({ \
+  float64x2x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 42); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_s64(__p0) __extension__ ({ \
+  int64x2x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 35); \
+  __ret; \
+})
+#else
+#define vld3q_s64(__p0) __extension__ ({ \
+  int64x2x3_t __ret; \
+  __builtin_neon_vld3q_v(&__ret, __p0, 35); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_f64(__p0) __extension__ ({ \
+  float64x1x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 10); \
+  __ret; \
+})
+#else
+#define vld3_f64(__p0) __extension__ ({ \
+  float64x1x3_t __ret; \
+  __builtin_neon_vld3_v(&__ret, __p0, 10); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_dup_p64(__p0) __extension__ ({ \
+  poly64x1x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 6); \
+  __ret; \
+})
+#else
+#define vld3_dup_p64(__p0) __extension__ ({ \
+  poly64x1x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 6); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_dup_p8(__p0) __extension__ ({ \
+  poly8x16x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 36); \
+  __ret; \
+})
+#else
+#define vld3q_dup_p8(__p0) __extension__ ({ \
+  poly8x16x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 36); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_dup_p64(__p0) __extension__ ({ \
+  poly64x2x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 38); \
+  __ret; \
+})
+#else
+#define vld3q_dup_p64(__p0) __extension__ ({ \
+  poly64x2x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 38); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_dup_p16(__p0) __extension__ ({ \
+  poly16x8x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 37); \
+  __ret; \
+})
+#else
+#define vld3q_dup_p16(__p0) __extension__ ({ \
+  poly16x8x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 37); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_dup_u8(__p0) __extension__ ({ \
+  uint8x16x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 48); \
+  __ret; \
+})
+#else
+#define vld3q_dup_u8(__p0) __extension__ ({ \
+  uint8x16x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 48); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_dup_u32(__p0) __extension__ ({ \
+  uint32x4x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 50); \
+  __ret; \
+})
+#else
+#define vld3q_dup_u32(__p0) __extension__ ({ \
+  uint32x4x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 50); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_dup_u64(__p0) __extension__ ({ \
+  uint64x2x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 51); \
+  __ret; \
+})
+#else
+#define vld3q_dup_u64(__p0) __extension__ ({ \
+  uint64x2x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 51); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_dup_u16(__p0) __extension__ ({ \
+  uint16x8x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 49); \
+  __ret; \
+})
+#else
+#define vld3q_dup_u16(__p0) __extension__ ({ \
+  uint16x8x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 49); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_dup_s8(__p0) __extension__ ({ \
+  int8x16x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 32); \
+  __ret; \
+})
+#else
+#define vld3q_dup_s8(__p0) __extension__ ({ \
+  int8x16x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 32); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_dup_f64(__p0) __extension__ ({ \
+  float64x2x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 42); \
+  __ret; \
+})
+#else
+#define vld3q_dup_f64(__p0) __extension__ ({ \
+  float64x2x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 42); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_dup_f32(__p0) __extension__ ({ \
+  float32x4x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 41); \
+  __ret; \
+})
+#else
+#define vld3q_dup_f32(__p0) __extension__ ({ \
+  float32x4x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 41); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_dup_f16(__p0) __extension__ ({ \
+  float16x8x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 40); \
+  __ret; \
+})
+#else
+#define vld3q_dup_f16(__p0) __extension__ ({ \
+  float16x8x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 40); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_dup_s32(__p0) __extension__ ({ \
+  int32x4x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 34); \
+  __ret; \
+})
+#else
+#define vld3q_dup_s32(__p0) __extension__ ({ \
+  int32x4x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 34); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_dup_s64(__p0) __extension__ ({ \
+  int64x2x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 35); \
+  __ret; \
+})
+#else
+#define vld3q_dup_s64(__p0) __extension__ ({ \
+  int64x2x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 35); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_dup_s16(__p0) __extension__ ({ \
+  int16x8x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 33); \
+  __ret; \
+})
+#else
+#define vld3q_dup_s16(__p0) __extension__ ({ \
+  int16x8x3_t __ret; \
+  __builtin_neon_vld3q_dup_v(&__ret, __p0, 33); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_dup_f64(__p0) __extension__ ({ \
+  float64x1x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 10); \
+  __ret; \
+})
+#else
+#define vld3_dup_f64(__p0) __extension__ ({ \
+  float64x1x3_t __ret; \
+  __builtin_neon_vld3_dup_v(&__ret, __p0, 10); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1x3_t __s1 = __p1; \
+  poly64x1x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 6); \
+  __ret; \
+})
+#else
+#define vld3_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1x3_t __s1 = __p1; \
+  poly64x1x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 6); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16x3_t __s1 = __p1; \
+  poly8x16x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 36); \
+  __ret; \
+})
+#else
+#define vld3q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16x3_t __s1 = __p1; \
+  poly8x16x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 36); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2x3_t __s1 = __p1; \
+  poly64x2x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 38); \
+  __ret; \
+})
+#else
+#define vld3q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2x3_t __s1 = __p1; \
+  poly64x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  poly64x2x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 38); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16x3_t __s1 = __p1; \
+  uint8x16x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 48); \
+  __ret; \
+})
+#else
+#define vld3q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16x3_t __s1 = __p1; \
+  uint8x16x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 48); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2x3_t __s1 = __p1; \
+  uint64x2x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 51); \
+  __ret; \
+})
+#else
+#define vld3q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2x3_t __s1 = __p1; \
+  uint64x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  uint64x2x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 51); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16x3_t __s1 = __p1; \
+  int8x16x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 32); \
+  __ret; \
+})
+#else
+#define vld3q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16x3_t __s1 = __p1; \
+  int8x16x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 32); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2x3_t __s1 = __p1; \
+  float64x2x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 42); \
+  __ret; \
+})
+#else
+#define vld3q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2x3_t __s1 = __p1; \
+  float64x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  float64x2x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 42); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2x3_t __s1 = __p1; \
+  int64x2x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 35); \
+  __ret; \
+})
+#else
+#define vld3q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2x3_t __s1 = __p1; \
+  int64x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  int64x2x3_t __ret; \
+  __builtin_neon_vld3q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 35); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1x3_t __s1 = __p1; \
+  uint64x1x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 19); \
+  __ret; \
+})
+#else
+#define vld3_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1x3_t __s1 = __p1; \
+  uint64x1x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1x3_t __s1 = __p1; \
+  float64x1x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 10); \
+  __ret; \
+})
+#else
+#define vld3_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1x3_t __s1 = __p1; \
+  float64x1x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 10); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld3_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1x3_t __s1 = __p1; \
+  int64x1x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 3); \
+  __ret; \
+})
+#else
+#define vld3_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1x3_t __s1 = __p1; \
+  int64x1x3_t __ret; \
+  __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_p64(__p0) __extension__ ({ \
+  poly64x1x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 6); \
+  __ret; \
+})
+#else
+#define vld4_p64(__p0) __extension__ ({ \
+  poly64x1x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 6); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_p64(__p0) __extension__ ({ \
+  poly64x2x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 38); \
+  __ret; \
+})
+#else
+#define vld4q_p64(__p0) __extension__ ({ \
+  poly64x2x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 38); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_u64(__p0) __extension__ ({ \
+  uint64x2x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 51); \
+  __ret; \
+})
+#else
+#define vld4q_u64(__p0) __extension__ ({ \
+  uint64x2x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 51); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_f64(__p0) __extension__ ({ \
+  float64x2x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 42); \
+  __ret; \
+})
+#else
+#define vld4q_f64(__p0) __extension__ ({ \
+  float64x2x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 42); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_s64(__p0) __extension__ ({ \
+  int64x2x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 35); \
+  __ret; \
+})
+#else
+#define vld4q_s64(__p0) __extension__ ({ \
+  int64x2x4_t __ret; \
+  __builtin_neon_vld4q_v(&__ret, __p0, 35); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_f64(__p0) __extension__ ({ \
+  float64x1x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 10); \
+  __ret; \
+})
+#else
+#define vld4_f64(__p0) __extension__ ({ \
+  float64x1x4_t __ret; \
+  __builtin_neon_vld4_v(&__ret, __p0, 10); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_dup_p64(__p0) __extension__ ({ \
+  poly64x1x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 6); \
+  __ret; \
+})
+#else
+#define vld4_dup_p64(__p0) __extension__ ({ \
+  poly64x1x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 6); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_dup_p8(__p0) __extension__ ({ \
+  poly8x16x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 36); \
+  __ret; \
+})
+#else
+#define vld4q_dup_p8(__p0) __extension__ ({ \
+  poly8x16x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 36); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_dup_p64(__p0) __extension__ ({ \
+  poly64x2x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 38); \
+  __ret; \
+})
+#else
+#define vld4q_dup_p64(__p0) __extension__ ({ \
+  poly64x2x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 38); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_dup_p16(__p0) __extension__ ({ \
+  poly16x8x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 37); \
+  __ret; \
+})
+#else
+#define vld4q_dup_p16(__p0) __extension__ ({ \
+  poly16x8x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 37); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_dup_u8(__p0) __extension__ ({ \
+  uint8x16x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 48); \
+  __ret; \
+})
+#else
+#define vld4q_dup_u8(__p0) __extension__ ({ \
+  uint8x16x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 48); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_dup_u32(__p0) __extension__ ({ \
+  uint32x4x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 50); \
+  __ret; \
+})
+#else
+#define vld4q_dup_u32(__p0) __extension__ ({ \
+  uint32x4x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 50); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_dup_u64(__p0) __extension__ ({ \
+  uint64x2x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 51); \
+  __ret; \
+})
+#else
+#define vld4q_dup_u64(__p0) __extension__ ({ \
+  uint64x2x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 51); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_dup_u16(__p0) __extension__ ({ \
+  uint16x8x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 49); \
+  __ret; \
+})
+#else
+#define vld4q_dup_u16(__p0) __extension__ ({ \
+  uint16x8x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 49); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_dup_s8(__p0) __extension__ ({ \
+  int8x16x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 32); \
+  __ret; \
+})
+#else
+#define vld4q_dup_s8(__p0) __extension__ ({ \
+  int8x16x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 32); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_dup_f64(__p0) __extension__ ({ \
+  float64x2x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 42); \
+  __ret; \
+})
+#else
+#define vld4q_dup_f64(__p0) __extension__ ({ \
+  float64x2x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 42); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_dup_f32(__p0) __extension__ ({ \
+  float32x4x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 41); \
+  __ret; \
+})
+#else
+#define vld4q_dup_f32(__p0) __extension__ ({ \
+  float32x4x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 41); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_dup_f16(__p0) __extension__ ({ \
+  float16x8x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 40); \
+  __ret; \
+})
+#else
+#define vld4q_dup_f16(__p0) __extension__ ({ \
+  float16x8x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 40); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_dup_s32(__p0) __extension__ ({ \
+  int32x4x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 34); \
+  __ret; \
+})
+#else
+#define vld4q_dup_s32(__p0) __extension__ ({ \
+  int32x4x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 34); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_dup_s64(__p0) __extension__ ({ \
+  int64x2x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 35); \
+  __ret; \
+})
+#else
+#define vld4q_dup_s64(__p0) __extension__ ({ \
+  int64x2x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 35); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_dup_s16(__p0) __extension__ ({ \
+  int16x8x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 33); \
+  __ret; \
+})
+#else
+#define vld4q_dup_s16(__p0) __extension__ ({ \
+  int16x8x4_t __ret; \
+  __builtin_neon_vld4q_dup_v(&__ret, __p0, 33); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_dup_f64(__p0) __extension__ ({ \
+  float64x1x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 10); \
+  __ret; \
+})
+#else
+#define vld4_dup_f64(__p0) __extension__ ({ \
+  float64x1x4_t __ret; \
+  __builtin_neon_vld4_dup_v(&__ret, __p0, 10); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1x4_t __s1 = __p1; \
+  poly64x1x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 6); \
+  __ret; \
+})
+#else
+#define vld4_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1x4_t __s1 = __p1; \
+  poly64x1x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 6); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16x4_t __s1 = __p1; \
+  poly8x16x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 36); \
+  __ret; \
+})
+#else
+#define vld4q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16x4_t __s1 = __p1; \
+  poly8x16x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 36); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2x4_t __s1 = __p1; \
+  poly64x2x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 38); \
+  __ret; \
+})
+#else
+#define vld4q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2x4_t __s1 = __p1; \
+  poly64x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  poly64x2x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 38); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16x4_t __s1 = __p1; \
+  uint8x16x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 48); \
+  __ret; \
+})
+#else
+#define vld4q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16x4_t __s1 = __p1; \
+  uint8x16x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 48); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2x4_t __s1 = __p1; \
+  uint64x2x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 51); \
+  __ret; \
+})
+#else
+#define vld4q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2x4_t __s1 = __p1; \
+  uint64x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  uint64x2x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 51); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16x4_t __s1 = __p1; \
+  int8x16x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 32); \
+  __ret; \
+})
+#else
+#define vld4q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16x4_t __s1 = __p1; \
+  int8x16x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 32); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2x4_t __s1 = __p1; \
+  float64x2x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 42); \
+  __ret; \
+})
+#else
+#define vld4q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2x4_t __s1 = __p1; \
+  float64x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  float64x2x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 42); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2x4_t __s1 = __p1; \
+  int64x2x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 35); \
+  __ret; \
+})
+#else
+#define vld4q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2x4_t __s1 = __p1; \
+  int64x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  int64x2x4_t __ret; \
+  __builtin_neon_vld4q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 35); \
+ \
+  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
+  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
+  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
+  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1x4_t __s1 = __p1; \
+  uint64x1x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 19); \
+  __ret; \
+})
+#else
+#define vld4_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1x4_t __s1 = __p1; \
+  uint64x1x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 19); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1x4_t __s1 = __p1; \
+  float64x1x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 10); \
+  __ret; \
+})
+#else
+#define vld4_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1x4_t __s1 = __p1; \
+  float64x1x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 10); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vld4_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1x4_t __s1 = __p1; \
+  int64x1x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 3); \
+  __ret; \
+})
+#else
+#define vld4_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1x4_t __s1 = __p1; \
+  int64x1x4_t __ret; \
+  __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vldrq_p128(__p0) __extension__ ({ \
+  poly128_t __ret; \
+  __ret = (poly128_t) __builtin_neon_vldrq_p128(__p0); \
+  __ret; \
+})
+#else
+#define vldrq_p128(__p0) __extension__ ({ \
+  poly128_t __ret; \
+  __ret = (poly128_t) __builtin_neon_vldrq_p128(__p0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vmaxq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vmaxq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vmax_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vmax_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64_t vmaxnmvq_f64(float64x2_t __p0) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vmaxnmvq_f64((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai float64_t vmaxnmvq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vmaxnmvq_f64((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vmaxnmvq_f32(float32x4_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vmaxnmvq_f32((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai float32_t vmaxnmvq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vmaxnmvq_f32((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vmaxnmv_f32(float32x2_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vmaxnmv_f32((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai float32_t vmaxnmv_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vmaxnmv_f32((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8_t vmaxvq_u8(uint8x16_t __p0) {
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vmaxvq_u8((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai uint8_t vmaxvq_u8(uint8x16_t __p0) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vmaxvq_u8((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vmaxvq_u32(uint32x4_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vmaxvq_u32((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai uint32_t vmaxvq_u32(uint32x4_t __p0) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vmaxvq_u32((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16_t vmaxvq_u16(uint16x8_t __p0) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vmaxvq_u16((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai uint16_t vmaxvq_u16(uint16x8_t __p0) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vmaxvq_u16((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8_t vmaxvq_s8(int8x16_t __p0) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vmaxvq_s8((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai int8_t vmaxvq_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vmaxvq_s8((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64_t vmaxvq_f64(float64x2_t __p0) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vmaxvq_f64((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai float64_t vmaxvq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vmaxvq_f64((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vmaxvq_f32(float32x4_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vmaxvq_f32((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai float32_t vmaxvq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vmaxvq_f32((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vmaxvq_s32(int32x4_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vmaxvq_s32((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai int32_t vmaxvq_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vmaxvq_s32((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16_t vmaxvq_s16(int16x8_t __p0) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vmaxvq_s16((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai int16_t vmaxvq_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vmaxvq_s16((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8_t vmaxv_u8(uint8x8_t __p0) {
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vmaxv_u8((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai uint8_t vmaxv_u8(uint8x8_t __p0) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vmaxv_u8((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vmaxv_u32(uint32x2_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vmaxv_u32((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai uint32_t vmaxv_u32(uint32x2_t __p0) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vmaxv_u32((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16_t vmaxv_u16(uint16x4_t __p0) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vmaxv_u16((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai uint16_t vmaxv_u16(uint16x4_t __p0) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vmaxv_u16((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8_t vmaxv_s8(int8x8_t __p0) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vmaxv_s8((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai int8_t vmaxv_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vmaxv_s8((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vmaxv_f32(float32x2_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vmaxv_f32((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai float32_t vmaxv_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vmaxv_f32((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vmaxv_s32(int32x2_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vmaxv_s32((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai int32_t vmaxv_s32(int32x2_t __p0) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vmaxv_s32((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16_t vmaxv_s16(int16x4_t __p0) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vmaxv_s16((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai int16_t vmaxv_s16(int16x4_t __p0) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vmaxv_s16((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vminq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vminq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vmin_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vmin_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64_t vminnmvq_f64(float64x2_t __p0) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vminnmvq_f64((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai float64_t vminnmvq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vminnmvq_f64((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vminnmvq_f32(float32x4_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vminnmvq_f32((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai float32_t vminnmvq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vminnmvq_f32((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vminnmv_f32(float32x2_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vminnmv_f32((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai float32_t vminnmv_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vminnmv_f32((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8_t vminvq_u8(uint8x16_t __p0) {
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vminvq_u8((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai uint8_t vminvq_u8(uint8x16_t __p0) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vminvq_u8((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vminvq_u32(uint32x4_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vminvq_u32((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai uint32_t vminvq_u32(uint32x4_t __p0) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vminvq_u32((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16_t vminvq_u16(uint16x8_t __p0) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vminvq_u16((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai uint16_t vminvq_u16(uint16x8_t __p0) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vminvq_u16((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8_t vminvq_s8(int8x16_t __p0) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vminvq_s8((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai int8_t vminvq_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vminvq_s8((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64_t vminvq_f64(float64x2_t __p0) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vminvq_f64((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai float64_t vminvq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vminvq_f64((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vminvq_f32(float32x4_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vminvq_f32((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai float32_t vminvq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vminvq_f32((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vminvq_s32(int32x4_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vminvq_s32((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai int32_t vminvq_s32(int32x4_t __p0) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vminvq_s32((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16_t vminvq_s16(int16x8_t __p0) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vminvq_s16((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai int16_t vminvq_s16(int16x8_t __p0) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vminvq_s16((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8_t vminv_u8(uint8x8_t __p0) {
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vminv_u8((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai uint8_t vminv_u8(uint8x8_t __p0) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vminv_u8((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vminv_u32(uint32x2_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vminv_u32((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai uint32_t vminv_u32(uint32x2_t __p0) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vminv_u32((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16_t vminv_u16(uint16x4_t __p0) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vminv_u16((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai uint16_t vminv_u16(uint16x4_t __p0) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vminv_u16((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8_t vminv_s8(int8x8_t __p0) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vminv_s8((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai int8_t vminv_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vminv_s8((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vminv_f32(float32x2_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vminv_f32((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai float32_t vminv_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vminv_f32((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vminv_s32(int32x2_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vminv_s32((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai int32_t vminv_s32(int32x2_t __p0) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vminv_s32((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16_t vminv_s16(int16x4_t __p0) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vminv_s16((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai int16_t vminv_s16(int16x4_t __p0) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vminv_s16((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vmlaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
+  float64x2_t __ret;
+  __ret = __p0 + __p1 * __p2;
+  return __ret;
+}
+#else
+__ai float64x2_t vmlaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  float64x2_t __ret;
+  __ret = __rev0 + __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vmla_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
+  float64x1_t __ret;
+  __ret = __p0 + __p1 * __p2;
+  return __ret;
+}
+#else
+__ai float64x1_t vmla_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
+  float64x1_t __ret;
+  __ret = __p0 + __p1 * __p2;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlaq_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __s2 = __p2; \
+  uint32x4_t __ret; \
+  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmlaq_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __s2 = __p2; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlaq_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __s2 = __p2; \
+  uint16x8_t __ret; \
+  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmlaq_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __s2 = __p2; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlaq_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x4_t __s2 = __p2; \
+  float32x4_t __ret; \
+  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmlaq_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x4_t __s2 = __p2; \
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  float32x4_t __ret; \
+  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlaq_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int32x4_t __ret; \
+  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmlaq_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlaq_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int16x8_t __ret; \
+  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmlaq_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmla_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x4_t __s2 = __p2; \
+  uint32x2_t __ret; \
+  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmla_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x4_t __s2 = __p2; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmla_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x8_t __s2 = __p2; \
+  uint16x4_t __ret; \
+  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmla_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x8_t __s2 = __p2; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmla_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x4_t __s2 = __p2; \
+  float32x2_t __ret; \
+  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmla_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x4_t __s2 = __p2; \
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  float32x2_t __ret; \
+  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmla_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int32x2_t __ret; \
+  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmla_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int32x2_t __ret; \
+  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmla_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int16x4_t __ret; \
+  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmla_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vmlaq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
+  float64x2_t __ret;
+  __ret = __p0 + __p1 * (float64x2_t) {__p2, __p2};
+  return __ret;
+}
+#else
+__ai float64x2_t vmlaq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = __rev0 + __rev1 * (float64x2_t) {__p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlal_high_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x2_t __s2 = __p2; \
+  uint64x2_t __ret; \
+  __ret = __s0 + vmull_u32(vget_high_u32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlal_high_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x2_t __s2 = __p2; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = __rev0 + __noswap_vmull_u32(__noswap_vget_high_u32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlal_high_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x4_t __s2 = __p2; \
+  uint32x4_t __ret; \
+  __ret = __s0 + vmull_u16(vget_high_u16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlal_high_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x4_t __s2 = __p2; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = __rev0 + __noswap_vmull_u16(__noswap_vget_high_u16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlal_high_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int64x2_t __ret; \
+  __ret = __s0 + vmull_s32(vget_high_s32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlal_high_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __rev0 + __noswap_vmull_s32(__noswap_vget_high_s32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlal_high_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int32x4_t __ret; \
+  __ret = __s0 + vmull_s16(vget_high_s16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlal_high_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __rev0 + __noswap_vmull_s16(__noswap_vget_high_s16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlal_high_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __s2 = __p2; \
+  uint64x2_t __ret; \
+  __ret = __s0 + vmull_u32(vget_high_u32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlal_high_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __s2 = __p2; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = __rev0 + __noswap_vmull_u32(__noswap_vget_high_u32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlal_high_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __s2 = __p2; \
+  uint32x4_t __ret; \
+  __ret = __s0 + vmull_u16(vget_high_u16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlal_high_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __s2 = __p2; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = __rev0 + __noswap_vmull_u16(__noswap_vget_high_u16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlal_high_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int64x2_t __ret; \
+  __ret = __s0 + vmull_s32(vget_high_s32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlal_high_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __rev0 + __noswap_vmull_s32(__noswap_vget_high_s32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlal_high_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int32x4_t __ret; \
+  __ret = __s0 + vmull_s16(vget_high_s16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlal_high_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __rev0 + __noswap_vmull_s16(__noswap_vget_high_s16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlal_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x4_t __s2 = __p2; \
+  uint64x2_t __ret; \
+  __ret = __s0 + vmull_u32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlal_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x4_t __s2 = __p2; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = __rev0 + __noswap_vmull_u32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlal_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x8_t __s2 = __p2; \
+  uint32x4_t __ret; \
+  __ret = __s0 + vmull_u16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlal_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x8_t __s2 = __p2; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = __rev0 + __noswap_vmull_u16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlal_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int64x2_t __ret; \
+  __ret = __s0 + vmull_s32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlal_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __rev0 + __noswap_vmull_s32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlal_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int32x4_t __ret; \
+  __ret = __s0 + vmull_s16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlal_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __rev0 + __noswap_vmull_s16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vmlsq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
+  float64x2_t __ret;
+  __ret = __p0 - __p1 * __p2;
+  return __ret;
+}
+#else
+__ai float64x2_t vmlsq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  float64x2_t __ret;
+  __ret = __rev0 - __rev1 * __rev2;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vmls_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
+  float64x1_t __ret;
+  __ret = __p0 - __p1 * __p2;
+  return __ret;
+}
+#else
+__ai float64x1_t vmls_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
+  float64x1_t __ret;
+  __ret = __p0 - __p1 * __p2;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsq_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __s2 = __p2; \
+  uint32x4_t __ret; \
+  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmlsq_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __s2 = __p2; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsq_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __s2 = __p2; \
+  uint16x8_t __ret; \
+  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmlsq_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __s2 = __p2; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsq_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x4_t __s2 = __p2; \
+  float32x4_t __ret; \
+  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmlsq_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x4_t __s2 = __p2; \
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  float32x4_t __ret; \
+  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsq_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int32x4_t __ret; \
+  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmlsq_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsq_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int16x8_t __ret; \
+  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmlsq_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmls_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x4_t __s2 = __p2; \
+  uint32x2_t __ret; \
+  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmls_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x4_t __s2 = __p2; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmls_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x8_t __s2 = __p2; \
+  uint16x4_t __ret; \
+  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmls_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x8_t __s2 = __p2; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmls_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x4_t __s2 = __p2; \
+  float32x2_t __ret; \
+  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmls_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x4_t __s2 = __p2; \
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  float32x2_t __ret; \
+  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmls_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int32x2_t __ret; \
+  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmls_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int32x2_t __ret; \
+  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmls_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int16x4_t __ret; \
+  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
+  __ret; \
+})
+#else
+#define vmls_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vmlsq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
+  float64x2_t __ret;
+  __ret = __p0 - __p1 * (float64x2_t) {__p2, __p2};
+  return __ret;
+}
+#else
+__ai float64x2_t vmlsq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = __rev0 - __rev1 * (float64x2_t) {__p2, __p2};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsl_high_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x2_t __s2 = __p2; \
+  uint64x2_t __ret; \
+  __ret = __s0 - vmull_u32(vget_high_u32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlsl_high_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x2_t __s2 = __p2; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = __rev0 - __noswap_vmull_u32(__noswap_vget_high_u32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsl_high_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x4_t __s2 = __p2; \
+  uint32x4_t __ret; \
+  __ret = __s0 - vmull_u16(vget_high_u16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlsl_high_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x4_t __s2 = __p2; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = __rev0 - __noswap_vmull_u16(__noswap_vget_high_u16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsl_high_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int64x2_t __ret; \
+  __ret = __s0 - vmull_s32(vget_high_s32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlsl_high_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __rev0 - __noswap_vmull_s32(__noswap_vget_high_s32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsl_high_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int32x4_t __ret; \
+  __ret = __s0 - vmull_s16(vget_high_s16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlsl_high_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __rev0 - __noswap_vmull_s16(__noswap_vget_high_s16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsl_high_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __s2 = __p2; \
+  uint64x2_t __ret; \
+  __ret = __s0 - vmull_u32(vget_high_u32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlsl_high_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __s2 = __p2; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = __rev0 - __noswap_vmull_u32(__noswap_vget_high_u32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsl_high_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __s2 = __p2; \
+  uint32x4_t __ret; \
+  __ret = __s0 - vmull_u16(vget_high_u16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlsl_high_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __s2 = __p2; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = __rev0 - __noswap_vmull_u16(__noswap_vget_high_u16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsl_high_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int64x2_t __ret; \
+  __ret = __s0 - vmull_s32(vget_high_s32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlsl_high_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __rev0 - __noswap_vmull_s32(__noswap_vget_high_s32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsl_high_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int32x4_t __ret; \
+  __ret = __s0 - vmull_s16(vget_high_s16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlsl_high_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __rev0 - __noswap_vmull_s16(__noswap_vget_high_s16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsl_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x4_t __s2 = __p2; \
+  uint64x2_t __ret; \
+  __ret = __s0 - vmull_u32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlsl_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x4_t __s2 = __p2; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = __rev0 - __noswap_vmull_u32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsl_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x8_t __s2 = __p2; \
+  uint32x4_t __ret; \
+  __ret = __s0 - vmull_u16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlsl_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x8_t __s2 = __p2; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = __rev0 - __noswap_vmull_u16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsl_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int64x2_t __ret; \
+  __ret = __s0 - vmull_s32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlsl_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __rev0 - __noswap_vmull_s32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsl_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int32x4_t __ret; \
+  __ret = __s0 - vmull_s16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlsl_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __rev0 - __noswap_vmull_s16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x1_t vmov_n_p64(poly64_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t) {__p0};
+  return __ret;
+}
+#else
+__ai poly64x1_t vmov_n_p64(poly64_t __p0) {
+  poly64x1_t __ret;
+  __ret = (poly64x1_t) {__p0};
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vmovq_n_p64(poly64_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t) {__p0, __p0};
+  return __ret;
+}
+#else
+__ai poly64x2_t vmovq_n_p64(poly64_t __p0) {
+  poly64x2_t __ret;
+  __ret = (poly64x2_t) {__p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vmovq_n_f64(float64_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) {__p0, __p0};
+  return __ret;
+}
+#else
+__ai float64x2_t vmovq_n_f64(float64_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) {__p0, __p0};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vmov_n_f64(float64_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) {__p0};
+  return __ret;
+}
+#else
+__ai float64x1_t vmov_n_f64(float64_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) {__p0};
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vmovl_high_u8(uint8x16_t __p0_116) {
+  uint16x8_t __ret_116;
+  uint8x8_t __a1_116 = vget_high_u8(__p0_116);
+  __ret_116 = (uint16x8_t)(vshll_n_u8(__a1_116, 0));
+  return __ret_116;
+}
+#else
+__ai uint16x8_t vmovl_high_u8(uint8x16_t __p0_117) {
+  uint8x16_t __rev0_117;  __rev0_117 = __builtin_shufflevector(__p0_117, __p0_117, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret_117;
+  uint8x8_t __a1_117 = __noswap_vget_high_u8(__rev0_117);
+  __ret_117 = (uint16x8_t)(__noswap_vshll_n_u8(__a1_117, 0));
+  __ret_117 = __builtin_shufflevector(__ret_117, __ret_117, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret_117;
+}
+__ai uint16x8_t __noswap_vmovl_high_u8(uint8x16_t __p0_118) {
+  uint16x8_t __ret_118;
+  uint8x8_t __a1_118 = __noswap_vget_high_u8(__p0_118);
+  __ret_118 = (uint16x8_t)(__noswap_vshll_n_u8(__a1_118, 0));
+  return __ret_118;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vmovl_high_u32(uint32x4_t __p0_119) {
+  uint64x2_t __ret_119;
+  uint32x2_t __a1_119 = vget_high_u32(__p0_119);
+  __ret_119 = (uint64x2_t)(vshll_n_u32(__a1_119, 0));
+  return __ret_119;
+}
+#else
+__ai uint64x2_t vmovl_high_u32(uint32x4_t __p0_120) {
+  uint32x4_t __rev0_120;  __rev0_120 = __builtin_shufflevector(__p0_120, __p0_120, 3, 2, 1, 0);
+  uint64x2_t __ret_120;
+  uint32x2_t __a1_120 = __noswap_vget_high_u32(__rev0_120);
+  __ret_120 = (uint64x2_t)(__noswap_vshll_n_u32(__a1_120, 0));
+  __ret_120 = __builtin_shufflevector(__ret_120, __ret_120, 1, 0);
+  return __ret_120;
+}
+__ai uint64x2_t __noswap_vmovl_high_u32(uint32x4_t __p0_121) {
+  uint64x2_t __ret_121;
+  uint32x2_t __a1_121 = __noswap_vget_high_u32(__p0_121);
+  __ret_121 = (uint64x2_t)(__noswap_vshll_n_u32(__a1_121, 0));
+  return __ret_121;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmovl_high_u16(uint16x8_t __p0_122) {
+  uint32x4_t __ret_122;
+  uint16x4_t __a1_122 = vget_high_u16(__p0_122);
+  __ret_122 = (uint32x4_t)(vshll_n_u16(__a1_122, 0));
+  return __ret_122;
+}
+#else
+__ai uint32x4_t vmovl_high_u16(uint16x8_t __p0_123) {
+  uint16x8_t __rev0_123;  __rev0_123 = __builtin_shufflevector(__p0_123, __p0_123, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint32x4_t __ret_123;
+  uint16x4_t __a1_123 = __noswap_vget_high_u16(__rev0_123);
+  __ret_123 = (uint32x4_t)(__noswap_vshll_n_u16(__a1_123, 0));
+  __ret_123 = __builtin_shufflevector(__ret_123, __ret_123, 3, 2, 1, 0);
+  return __ret_123;
+}
+__ai uint32x4_t __noswap_vmovl_high_u16(uint16x8_t __p0_124) {
+  uint32x4_t __ret_124;
+  uint16x4_t __a1_124 = __noswap_vget_high_u16(__p0_124);
+  __ret_124 = (uint32x4_t)(__noswap_vshll_n_u16(__a1_124, 0));
+  return __ret_124;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vmovl_high_s8(int8x16_t __p0_125) {
+  int16x8_t __ret_125;
+  int8x8_t __a1_125 = vget_high_s8(__p0_125);
+  __ret_125 = (int16x8_t)(vshll_n_s8(__a1_125, 0));
+  return __ret_125;
+}
+#else
+__ai int16x8_t vmovl_high_s8(int8x16_t __p0_126) {
+  int8x16_t __rev0_126;  __rev0_126 = __builtin_shufflevector(__p0_126, __p0_126, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret_126;
+  int8x8_t __a1_126 = __noswap_vget_high_s8(__rev0_126);
+  __ret_126 = (int16x8_t)(__noswap_vshll_n_s8(__a1_126, 0));
+  __ret_126 = __builtin_shufflevector(__ret_126, __ret_126, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret_126;
+}
+__ai int16x8_t __noswap_vmovl_high_s8(int8x16_t __p0_127) {
+  int16x8_t __ret_127;
+  int8x8_t __a1_127 = __noswap_vget_high_s8(__p0_127);
+  __ret_127 = (int16x8_t)(__noswap_vshll_n_s8(__a1_127, 0));
+  return __ret_127;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vmovl_high_s32(int32x4_t __p0_128) {
+  int64x2_t __ret_128;
+  int32x2_t __a1_128 = vget_high_s32(__p0_128);
+  __ret_128 = (int64x2_t)(vshll_n_s32(__a1_128, 0));
+  return __ret_128;
+}
+#else
+__ai int64x2_t vmovl_high_s32(int32x4_t __p0_129) {
+  int32x4_t __rev0_129;  __rev0_129 = __builtin_shufflevector(__p0_129, __p0_129, 3, 2, 1, 0);
+  int64x2_t __ret_129;
+  int32x2_t __a1_129 = __noswap_vget_high_s32(__rev0_129);
+  __ret_129 = (int64x2_t)(__noswap_vshll_n_s32(__a1_129, 0));
+  __ret_129 = __builtin_shufflevector(__ret_129, __ret_129, 1, 0);
+  return __ret_129;
+}
+__ai int64x2_t __noswap_vmovl_high_s32(int32x4_t __p0_130) {
+  int64x2_t __ret_130;
+  int32x2_t __a1_130 = __noswap_vget_high_s32(__p0_130);
+  __ret_130 = (int64x2_t)(__noswap_vshll_n_s32(__a1_130, 0));
+  return __ret_130;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmovl_high_s16(int16x8_t __p0_131) {
+  int32x4_t __ret_131;
+  int16x4_t __a1_131 = vget_high_s16(__p0_131);
+  __ret_131 = (int32x4_t)(vshll_n_s16(__a1_131, 0));
+  return __ret_131;
+}
+#else
+__ai int32x4_t vmovl_high_s16(int16x8_t __p0_132) {
+  int16x8_t __rev0_132;  __rev0_132 = __builtin_shufflevector(__p0_132, __p0_132, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret_132;
+  int16x4_t __a1_132 = __noswap_vget_high_s16(__rev0_132);
+  __ret_132 = (int32x4_t)(__noswap_vshll_n_s16(__a1_132, 0));
+  __ret_132 = __builtin_shufflevector(__ret_132, __ret_132, 3, 2, 1, 0);
+  return __ret_132;
+}
+__ai int32x4_t __noswap_vmovl_high_s16(int16x8_t __p0_133) {
+  int32x4_t __ret_133;
+  int16x4_t __a1_133 = __noswap_vget_high_s16(__p0_133);
+  __ret_133 = (int32x4_t)(__noswap_vshll_n_s16(__a1_133, 0));
+  return __ret_133;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vmovn_high_u32(uint16x4_t __p0, uint32x4_t __p1) {
+  uint16x8_t __ret;
+  __ret = vcombine_u16(__p0, vmovn_u32(__p1));
+  return __ret;
+}
+#else
+__ai uint16x8_t vmovn_high_u32(uint16x4_t __p0, uint32x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __noswap_vcombine_u16(__rev0, __noswap_vmovn_u32(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmovn_high_u64(uint32x2_t __p0, uint64x2_t __p1) {
+  uint32x4_t __ret;
+  __ret = vcombine_u32(__p0, vmovn_u64(__p1));
+  return __ret;
+}
+#else
+__ai uint32x4_t vmovn_high_u64(uint32x2_t __p0, uint64x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x4_t __ret;
+  __ret = __noswap_vcombine_u32(__rev0, __noswap_vmovn_u64(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vmovn_high_u16(uint8x8_t __p0, uint16x8_t __p1) {
+  uint8x16_t __ret;
+  __ret = vcombine_u8(__p0, vmovn_u16(__p1));
+  return __ret;
+}
+#else
+__ai uint8x16_t vmovn_high_u16(uint8x8_t __p0, uint16x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __noswap_vcombine_u8(__rev0, __noswap_vmovn_u16(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vmovn_high_s32(int16x4_t __p0, int32x4_t __p1) {
+  int16x8_t __ret;
+  __ret = vcombine_s16(__p0, vmovn_s32(__p1));
+  return __ret;
+}
+#else
+__ai int16x8_t vmovn_high_s32(int16x4_t __p0, int32x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __noswap_vcombine_s16(__rev0, __noswap_vmovn_s32(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmovn_high_s64(int32x2_t __p0, int64x2_t __p1) {
+  int32x4_t __ret;
+  __ret = vcombine_s32(__p0, vmovn_s64(__p1));
+  return __ret;
+}
+#else
+__ai int32x4_t vmovn_high_s64(int32x2_t __p0, int64x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vcombine_s32(__rev0, __noswap_vmovn_s64(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vmovn_high_s16(int8x8_t __p0, int16x8_t __p1) {
+  int8x16_t __ret;
+  __ret = vcombine_s8(__p0, vmovn_s16(__p1));
+  return __ret;
+}
+#else
+__ai int8x16_t vmovn_high_s16(int8x8_t __p0, int16x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __noswap_vcombine_s8(__rev0, __noswap_vmovn_s16(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vmulq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = __p0 * __p1;
+  return __ret;
+}
+#else
+__ai float64x2_t vmulq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = __rev0 * __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vmul_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = __p0 * __p1;
+  return __ret;
+}
+#else
+__ai float64x1_t vmul_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = __p0 * __p1;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmuld_lane_f64(__p0_134, __p1_134, __p2_134) __extension__ ({ \
+  float64_t __s0_134 = __p0_134; \
+  float64x1_t __s1_134 = __p1_134; \
+  float64_t __ret_134; \
+  __ret_134 = __s0_134 * vget_lane_f64(__s1_134, __p2_134); \
+  __ret_134; \
+})
+#else
+#define vmuld_lane_f64(__p0_135, __p1_135, __p2_135) __extension__ ({ \
+  float64_t __s0_135 = __p0_135; \
+  float64x1_t __s1_135 = __p1_135; \
+  float64_t __ret_135; \
+  __ret_135 = __s0_135 * __noswap_vget_lane_f64(__s1_135, __p2_135); \
+  __ret_135; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmuls_lane_f32(__p0_136, __p1_136, __p2_136) __extension__ ({ \
+  float32_t __s0_136 = __p0_136; \
+  float32x2_t __s1_136 = __p1_136; \
+  float32_t __ret_136; \
+  __ret_136 = __s0_136 * vget_lane_f32(__s1_136, __p2_136); \
+  __ret_136; \
+})
+#else
+#define vmuls_lane_f32(__p0_137, __p1_137, __p2_137) __extension__ ({ \
+  float32_t __s0_137 = __p0_137; \
+  float32x2_t __s1_137 = __p1_137; \
+  float32x2_t __rev1_137;  __rev1_137 = __builtin_shufflevector(__s1_137, __s1_137, 1, 0); \
+  float32_t __ret_137; \
+  __ret_137 = __s0_137 * __noswap_vget_lane_f32(__rev1_137, __p2_137); \
+  __ret_137; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmul_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64x1_t __s1 = __p1; \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vmul_lane_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 10); \
+  __ret; \
+})
+#else
+#define vmul_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64x1_t __s1 = __p1; \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vmul_lane_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 10); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulq_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x1_t __s1 = __p1; \
+  float64x2_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmulq_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x1_t __s1 = __p1; \
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float64x2_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__s1, __s1, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmuld_laneq_f64(__p0_138, __p1_138, __p2_138) __extension__ ({ \
+  float64_t __s0_138 = __p0_138; \
+  float64x2_t __s1_138 = __p1_138; \
+  float64_t __ret_138; \
+  __ret_138 = __s0_138 * vgetq_lane_f64(__s1_138, __p2_138); \
+  __ret_138; \
+})
+#else
+#define vmuld_laneq_f64(__p0_139, __p1_139, __p2_139) __extension__ ({ \
+  float64_t __s0_139 = __p0_139; \
+  float64x2_t __s1_139 = __p1_139; \
+  float64x2_t __rev1_139;  __rev1_139 = __builtin_shufflevector(__s1_139, __s1_139, 1, 0); \
+  float64_t __ret_139; \
+  __ret_139 = __s0_139 * __noswap_vgetq_lane_f64(__rev1_139, __p2_139); \
+  __ret_139; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmuls_laneq_f32(__p0_140, __p1_140, __p2_140) __extension__ ({ \
+  float32_t __s0_140 = __p0_140; \
+  float32x4_t __s1_140 = __p1_140; \
+  float32_t __ret_140; \
+  __ret_140 = __s0_140 * vgetq_lane_f32(__s1_140, __p2_140); \
+  __ret_140; \
+})
+#else
+#define vmuls_laneq_f32(__p0_141, __p1_141, __p2_141) __extension__ ({ \
+  float32_t __s0_141 = __p0_141; \
+  float32x4_t __s1_141 = __p1_141; \
+  float32x4_t __rev1_141;  __rev1_141 = __builtin_shufflevector(__s1_141, __s1_141, 3, 2, 1, 0); \
+  float32_t __ret_141; \
+  __ret_141 = __s0_141 * __noswap_vgetq_lane_f32(__rev1_141, __p2_141); \
+  __ret_141; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmul_laneq_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64x2_t __s1 = __p1; \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vmul_laneq_v((int8x8_t)__s0, (int8x16_t)__s1, __p2, 10); \
+  __ret; \
+})
+#else
+#define vmul_laneq_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1_t __s0 = __p0; \
+  float64x2_t __s1 = __p1; \
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vmul_laneq_v((int8x8_t)__s0, (int8x16_t)__rev1, __p2, 10); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulq_laneq_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmulq_laneq_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulq_laneq_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmulq_laneq_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulq_laneq_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x2_t __s1 = __p1; \
+  float64x2_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmulq_laneq_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x2_t __s1 = __p1; \
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float64x2_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulq_laneq_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x4_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmulq_laneq_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  float32x4_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulq_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmulq_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulq_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmulq_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmul_laneq_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x2_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmul_laneq_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint32x2_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmul_laneq_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x4_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmul_laneq_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x4_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmul_laneq_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x2_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmul_laneq_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  float32x2_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmul_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x2_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmul_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x2_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmul_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x4_t __ret; \
+  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
+  __ret; \
+})
+#else
+#define vmul_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vmul_n_f64(float64x1_t __p0, float64_t __p1) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vmul_n_f64((int8x8_t)__p0, __p1);
+  return __ret;
+}
+#else
+__ai float64x1_t vmul_n_f64(float64x1_t __p0, float64_t __p1) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vmul_n_f64((int8x8_t)__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vmulq_n_f64(float64x2_t __p0, float64_t __p1) {
+  float64x2_t __ret;
+  __ret = __p0 * (float64x2_t) {__p1, __p1};
+  return __ret;
+}
+#else
+__ai float64x2_t vmulq_n_f64(float64x2_t __p0, float64_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __ret;
+  __ret = __rev0 * (float64x2_t) {__p1, __p1};
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly128_t vmull_p64(poly64_t __p0, poly64_t __p1) {
+  poly128_t __ret;
+  __ret = (poly128_t) __builtin_neon_vmull_p64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai poly128_t vmull_p64(poly64_t __p0, poly64_t __p1) {
+  poly128_t __ret;
+  __ret = (poly128_t) __builtin_neon_vmull_p64(__p0, __p1);
+  return __ret;
+}
+__ai poly128_t __noswap_vmull_p64(poly64_t __p0, poly64_t __p1) {
+  poly128_t __ret;
+  __ret = (poly128_t) __builtin_neon_vmull_p64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vmull_high_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly16x8_t __ret;
+  __ret = vmull_p8(vget_high_p8(__p0), vget_high_p8(__p1));
+  return __ret;
+}
+#else
+__ai poly16x8_t vmull_high_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __ret;
+  __ret = __noswap_vmull_p8(__noswap_vget_high_p8(__rev0), __noswap_vget_high_p8(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vmull_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint16x8_t __ret;
+  __ret = vmull_u8(vget_high_u8(__p0), vget_high_u8(__p1));
+  return __ret;
+}
+#else
+__ai uint16x8_t vmull_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __noswap_vmull_u8(__noswap_vget_high_u8(__rev0), __noswap_vget_high_u8(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vmull_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint64x2_t __ret;
+  __ret = vmull_u32(vget_high_u32(__p0), vget_high_u32(__p1));
+  return __ret;
+}
+#else
+__ai uint64x2_t vmull_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint64x2_t __ret;
+  __ret = __noswap_vmull_u32(__noswap_vget_high_u32(__rev0), __noswap_vget_high_u32(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmull_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint32x4_t __ret;
+  __ret = vmull_u16(vget_high_u16(__p0), vget_high_u16(__p1));
+  return __ret;
+}
+#else
+__ai uint32x4_t vmull_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __noswap_vmull_u16(__noswap_vget_high_u16(__rev0), __noswap_vget_high_u16(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vmull_high_s8(int8x16_t __p0, int8x16_t __p1) {
+  int16x8_t __ret;
+  __ret = vmull_s8(vget_high_s8(__p0), vget_high_s8(__p1));
+  return __ret;
+}
+#else
+__ai int16x8_t vmull_high_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __noswap_vmull_s8(__noswap_vget_high_s8(__rev0), __noswap_vget_high_s8(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vmull_high_s32(int32x4_t __p0, int32x4_t __p1) {
+  int64x2_t __ret;
+  __ret = vmull_s32(vget_high_s32(__p0), vget_high_s32(__p1));
+  return __ret;
+}
+#else
+__ai int64x2_t vmull_high_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int64x2_t __ret;
+  __ret = __noswap_vmull_s32(__noswap_vget_high_s32(__rev0), __noswap_vget_high_s32(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmull_high_s16(int16x8_t __p0, int16x8_t __p1) {
+  int32x4_t __ret;
+  __ret = vmull_s16(vget_high_s16(__p0), vget_high_s16(__p1));
+  return __ret;
+}
+#else
+__ai int32x4_t vmull_high_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vmull_s16(__noswap_vget_high_s16(__rev0), __noswap_vget_high_s16(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly128_t vmull_high_p64(poly64x2_t __p0, poly64x2_t __p1) {
+  poly128_t __ret;
+  __ret = vmull_p64((poly64_t)(vget_high_p64(__p0)), (poly64_t)(vget_high_p64(__p1)));
+  return __ret;
+}
+#else
+__ai poly128_t vmull_high_p64(poly64x2_t __p0, poly64x2_t __p1) {
+  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  poly128_t __ret;
+  __ret = __noswap_vmull_p64((poly64_t)(__noswap_vget_high_p64(__rev0)), (poly64_t)(__noswap_vget_high_p64(__rev1)));
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmull_high_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint64x2_t __ret; \
+  __ret = vmull_u32(vget_high_u32(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmull_high_lane_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = __noswap_vmull_u32(__noswap_vget_high_u32(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmull_high_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint32x4_t __ret; \
+  __ret = vmull_u16(vget_high_u16(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmull_high_lane_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = __noswap_vmull_u16(__noswap_vget_high_u16(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmull_high_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int64x2_t __ret; \
+  __ret = vmull_s32(vget_high_s32(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmull_high_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __noswap_vmull_s32(__noswap_vget_high_s32(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmull_high_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = vmull_s16(vget_high_s16(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmull_high_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __noswap_vmull_s16(__noswap_vget_high_s16(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmull_high_laneq_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint64x2_t __ret; \
+  __ret = vmull_u32(vget_high_u32(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmull_high_laneq_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = __noswap_vmull_u32(__noswap_vget_high_u32(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmull_high_laneq_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint32x4_t __ret; \
+  __ret = vmull_u16(vget_high_u16(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmull_high_laneq_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x8_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = __noswap_vmull_u16(__noswap_vget_high_u16(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmull_high_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int64x2_t __ret; \
+  __ret = vmull_s32(vget_high_s32(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmull_high_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __noswap_vmull_s32(__noswap_vget_high_s32(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmull_high_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = vmull_s16(vget_high_s16(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmull_high_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __noswap_vmull_s16(__noswap_vget_high_s16(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vmull_high_n_u32(uint32x4_t __p0, uint32_t __p1) {
+  uint64x2_t __ret;
+  __ret = vmull_n_u32(vget_high_u32(__p0), __p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vmull_high_n_u32(uint32x4_t __p0, uint32_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint64x2_t __ret;
+  __ret = __noswap_vmull_n_u32(__noswap_vget_high_u32(__rev0), __p1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmull_high_n_u16(uint16x8_t __p0, uint16_t __p1) {
+  uint32x4_t __ret;
+  __ret = vmull_n_u16(vget_high_u16(__p0), __p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vmull_high_n_u16(uint16x8_t __p0, uint16_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __noswap_vmull_n_u16(__noswap_vget_high_u16(__rev0), __p1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vmull_high_n_s32(int32x4_t __p0, int32_t __p1) {
+  int64x2_t __ret;
+  __ret = vmull_n_s32(vget_high_s32(__p0), __p1);
+  return __ret;
+}
+#else
+__ai int64x2_t vmull_high_n_s32(int32x4_t __p0, int32_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int64x2_t __ret;
+  __ret = __noswap_vmull_n_s32(__noswap_vget_high_s32(__rev0), __p1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmull_high_n_s16(int16x8_t __p0, int16_t __p1) {
+  int32x4_t __ret;
+  __ret = vmull_n_s16(vget_high_s16(__p0), __p1);
+  return __ret;
+}
+#else
+__ai int32x4_t vmull_high_n_s16(int16x8_t __p0, int16_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vmull_n_s16(__noswap_vget_high_s16(__rev0), __p1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmull_laneq_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint64x2_t __ret; \
+  __ret = vmull_u32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmull_laneq_u32(__p0, __p1, __p2) __extension__ ({ \
+  uint32x2_t __s0 = __p0; \
+  uint32x4_t __s1 = __p1; \
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = __noswap_vmull_u32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmull_laneq_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint32x4_t __ret; \
+  __ret = vmull_u16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmull_laneq_u16(__p0, __p1, __p2) __extension__ ({ \
+  uint16x4_t __s0 = __p0; \
+  uint16x8_t __s1 = __p1; \
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = __noswap_vmull_u16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmull_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int64x2_t __ret; \
+  __ret = vmull_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmull_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __noswap_vmull_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmull_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = vmull_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmull_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __noswap_vmull_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vmulxq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vmulxq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vmulxq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vmulxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai float64x2_t __noswap_vmulxq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vmulxq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vmulxq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vmulxq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vmulxq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vmulxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai float32x4_t __noswap_vmulxq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vmulxq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vmulx_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vmulx_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vmulx_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vmulx_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vmulx_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vmulx_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vmulx_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vmulx_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai float32x2_t __noswap_vmulx_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vmulx_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64_t vmulxd_f64(float64_t __p0, float64_t __p1) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vmulxd_f64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai float64_t vmulxd_f64(float64_t __p0, float64_t __p1) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vmulxd_f64(__p0, __p1);
+  return __ret;
+}
+__ai float64_t __noswap_vmulxd_f64(float64_t __p0, float64_t __p1) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vmulxd_f64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vmulxs_f32(float32_t __p0, float32_t __p1) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vmulxs_f32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai float32_t vmulxs_f32(float32_t __p0, float32_t __p1) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vmulxs_f32(__p0, __p1);
+  return __ret;
+}
+__ai float32_t __noswap_vmulxs_f32(float32_t __p0, float32_t __p1) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vmulxs_f32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulxd_lane_f64(__p0_142, __p1_142, __p2_142) __extension__ ({ \
+  float64_t __s0_142 = __p0_142; \
+  float64x1_t __s1_142 = __p1_142; \
+  float64_t __ret_142; \
+  __ret_142 = vmulxd_f64(__s0_142, vget_lane_f64(__s1_142, __p2_142)); \
+  __ret_142; \
+})
+#else
+#define vmulxd_lane_f64(__p0_143, __p1_143, __p2_143) __extension__ ({ \
+  float64_t __s0_143 = __p0_143; \
+  float64x1_t __s1_143 = __p1_143; \
+  float64_t __ret_143; \
+  __ret_143 = __noswap_vmulxd_f64(__s0_143, __noswap_vget_lane_f64(__s1_143, __p2_143)); \
+  __ret_143; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulxs_lane_f32(__p0_144, __p1_144, __p2_144) __extension__ ({ \
+  float32_t __s0_144 = __p0_144; \
+  float32x2_t __s1_144 = __p1_144; \
+  float32_t __ret_144; \
+  __ret_144 = vmulxs_f32(__s0_144, vget_lane_f32(__s1_144, __p2_144)); \
+  __ret_144; \
+})
+#else
+#define vmulxs_lane_f32(__p0_145, __p1_145, __p2_145) __extension__ ({ \
+  float32_t __s0_145 = __p0_145; \
+  float32x2_t __s1_145 = __p1_145; \
+  float32x2_t __rev1_145;  __rev1_145 = __builtin_shufflevector(__s1_145, __s1_145, 1, 0); \
+  float32_t __ret_145; \
+  __ret_145 = __noswap_vmulxs_f32(__s0_145, __noswap_vget_lane_f32(__rev1_145, __p2_145)); \
+  __ret_145; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulxq_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x1_t __s1 = __p1; \
+  float64x2_t __ret; \
+  __ret = vmulxq_f64(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmulxq_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x1_t __s1 = __p1; \
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float64x2_t __ret; \
+  __ret = __noswap_vmulxq_f64(__rev0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulxq_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x4_t __ret; \
+  __ret = vmulxq_f32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmulxq_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float32x4_t __ret; \
+  __ret = __noswap_vmulxq_f32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulx_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x2_t __ret; \
+  __ret = vmulx_f32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmulx_lane_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x2_t __s1 = __p1; \
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float32x2_t __ret; \
+  __ret = __noswap_vmulx_f32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulxd_laneq_f64(__p0_146, __p1_146, __p2_146) __extension__ ({ \
+  float64_t __s0_146 = __p0_146; \
+  float64x2_t __s1_146 = __p1_146; \
+  float64_t __ret_146; \
+  __ret_146 = vmulxd_f64(__s0_146, vgetq_lane_f64(__s1_146, __p2_146)); \
+  __ret_146; \
+})
+#else
+#define vmulxd_laneq_f64(__p0_147, __p1_147, __p2_147) __extension__ ({ \
+  float64_t __s0_147 = __p0_147; \
+  float64x2_t __s1_147 = __p1_147; \
+  float64x2_t __rev1_147;  __rev1_147 = __builtin_shufflevector(__s1_147, __s1_147, 1, 0); \
+  float64_t __ret_147; \
+  __ret_147 = __noswap_vmulxd_f64(__s0_147, __noswap_vgetq_lane_f64(__rev1_147, __p2_147)); \
+  __ret_147; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulxs_laneq_f32(__p0_148, __p1_148, __p2_148) __extension__ ({ \
+  float32_t __s0_148 = __p0_148; \
+  float32x4_t __s1_148 = __p1_148; \
+  float32_t __ret_148; \
+  __ret_148 = vmulxs_f32(__s0_148, vgetq_lane_f32(__s1_148, __p2_148)); \
+  __ret_148; \
+})
+#else
+#define vmulxs_laneq_f32(__p0_149, __p1_149, __p2_149) __extension__ ({ \
+  float32_t __s0_149 = __p0_149; \
+  float32x4_t __s1_149 = __p1_149; \
+  float32x4_t __rev1_149;  __rev1_149 = __builtin_shufflevector(__s1_149, __s1_149, 3, 2, 1, 0); \
+  float32_t __ret_149; \
+  __ret_149 = __noswap_vmulxs_f32(__s0_149, __noswap_vgetq_lane_f32(__rev1_149, __p2_149)); \
+  __ret_149; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulxq_laneq_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x2_t __s1 = __p1; \
+  float64x2_t __ret; \
+  __ret = vmulxq_f64(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmulxq_laneq_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2_t __s0 = __p0; \
+  float64x2_t __s1 = __p1; \
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float64x2_t __ret; \
+  __ret = __noswap_vmulxq_f64(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulxq_laneq_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x4_t __ret; \
+  __ret = vmulxq_f32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmulxq_laneq_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x4_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  float32x4_t __ret; \
+  __ret = __noswap_vmulxq_f32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulx_laneq_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x2_t __ret; \
+  __ret = vmulx_f32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vmulx_laneq_f32(__p0, __p1, __p2) __extension__ ({ \
+  float32x2_t __s0 = __p0; \
+  float32x4_t __s1 = __p1; \
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  float32x2_t __ret; \
+  __ret = __noswap_vmulx_f32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vnegq_f64(float64x2_t __p0) {
+  float64x2_t __ret;
+  __ret = -__p0;
+  return __ret;
+}
+#else
+__ai float64x2_t vnegq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __ret;
+  __ret = -__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vnegq_s64(int64x2_t __p0) {
+  int64x2_t __ret;
+  __ret = -__p0;
+  return __ret;
+}
+#else
+__ai int64x2_t vnegq_s64(int64x2_t __p0) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __ret;
+  __ret = -__rev0;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vneg_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = -__p0;
+  return __ret;
+}
+#else
+__ai float64x1_t vneg_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = -__p0;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vneg_s64(int64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = -__p0;
+  return __ret;
+}
+#else
+__ai int64x1_t vneg_s64(int64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = -__p0;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vnegd_s64(int64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vnegd_s64(__p0);
+  return __ret;
+}
+#else
+__ai int64_t vnegd_s64(int64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vnegd_s64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vpaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vpaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vpaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vpaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vpaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vpaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vpaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vpaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vpaddq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vpaddq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vpaddq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vpaddq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vpaddq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vpaddq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vpaddq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vpaddq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vpaddq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vpaddq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vpaddq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vpaddq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vpaddd_u64(uint64x2_t __p0) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vpaddd_u64((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai uint64_t vpaddd_u64(uint64x2_t __p0) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vpaddd_u64((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64_t vpaddd_f64(float64x2_t __p0) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vpaddd_f64((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai float64_t vpaddd_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vpaddd_f64((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vpaddd_s64(int64x2_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vpaddd_s64((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai int64_t vpaddd_s64(int64x2_t __p0) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vpaddd_s64((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vpadds_f32(float32x2_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vpadds_f32((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai float32_t vpadds_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vpadds_f32((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vpmaxq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vpmaxq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vpmaxq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vpmaxq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vpmaxq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vpmaxq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vpmaxq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vpmaxq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vpmaxq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vpmaxq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vpmaxq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vpmaxq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vpmaxq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vpmaxq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vpmaxq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vpmaxq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64_t vpmaxqd_f64(float64x2_t __p0) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vpmaxqd_f64((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai float64_t vpmaxqd_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vpmaxqd_f64((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vpmaxs_f32(float32x2_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vpmaxs_f32((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai float32_t vpmaxs_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vpmaxs_f32((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vpmaxnmq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vpmaxnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vpmaxnmq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vpmaxnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vpmaxnmq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vpmaxnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vpmaxnmq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vpmaxnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vpmaxnm_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vpmaxnm_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vpmaxnm_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vpmaxnm_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64_t vpmaxnmqd_f64(float64x2_t __p0) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vpmaxnmqd_f64((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai float64_t vpmaxnmqd_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vpmaxnmqd_f64((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vpmaxnms_f32(float32x2_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vpmaxnms_f32((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai float32_t vpmaxnms_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vpmaxnms_f32((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vpminq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vpminq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vpminq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vpminq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vpminq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vpminq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vpminq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vpminq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vpminq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vpminq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vpminq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vpminq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vpminq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vpminq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vpminq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vpminq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64_t vpminqd_f64(float64x2_t __p0) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vpminqd_f64((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai float64_t vpminqd_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vpminqd_f64((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vpmins_f32(float32x2_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vpmins_f32((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai float32_t vpmins_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vpmins_f32((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vpminnmq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vpminnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vpminnmq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vpminnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vpminnmq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vpminnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vpminnmq_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vpminnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vpminnm_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vpminnm_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vpminnm_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vpminnm_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64_t vpminnmqd_f64(float64x2_t __p0) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vpminnmqd_f64((int8x16_t)__p0);
+  return __ret;
+}
+#else
+__ai float64_t vpminnmqd_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vpminnmqd_f64((int8x16_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vpminnms_f32(float32x2_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vpminnms_f32((int8x8_t)__p0);
+  return __ret;
+}
+#else
+__ai float32_t vpminnms_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vpminnms_f32((int8x8_t)__rev0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vqabsq_s64(int64x2_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqabsq_v((int8x16_t)__p0, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vqabsq_s64(int64x2_t __p0) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqabsq_v((int8x16_t)__rev0, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vqabs_s64(int64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vqabs_v((int8x8_t)__p0, 3);
+  return __ret;
+}
+#else
+__ai int64x1_t vqabs_s64(int64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vqabs_v((int8x8_t)__p0, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8_t vqabsb_s8(int8_t __p0) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vqabsb_s8(__p0);
+  return __ret;
+}
+#else
+__ai int8_t vqabsb_s8(int8_t __p0) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vqabsb_s8(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vqabss_s32(int32_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqabss_s32(__p0);
+  return __ret;
+}
+#else
+__ai int32_t vqabss_s32(int32_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqabss_s32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vqabsd_s64(int64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqabsd_s64(__p0);
+  return __ret;
+}
+#else
+__ai int64_t vqabsd_s64(int64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqabsd_s64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16_t vqabsh_s16(int16_t __p0) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqabsh_s16(__p0);
+  return __ret;
+}
+#else
+__ai int16_t vqabsh_s16(int16_t __p0) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqabsh_s16(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8_t vqaddb_u8(uint8_t __p0, uint8_t __p1) {
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vqaddb_u8(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint8_t vqaddb_u8(uint8_t __p0, uint8_t __p1) {
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vqaddb_u8(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vqadds_u32(uint32_t __p0, uint32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vqadds_u32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint32_t vqadds_u32(uint32_t __p0, uint32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vqadds_u32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vqaddd_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vqaddd_u64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vqaddd_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vqaddd_u64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16_t vqaddh_u16(uint16_t __p0, uint16_t __p1) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vqaddh_u16(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint16_t vqaddh_u16(uint16_t __p0, uint16_t __p1) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vqaddh_u16(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8_t vqaddb_s8(int8_t __p0, int8_t __p1) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vqaddb_s8(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int8_t vqaddb_s8(int8_t __p0, int8_t __p1) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vqaddb_s8(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vqadds_s32(int32_t __p0, int32_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqadds_s32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int32_t vqadds_s32(int32_t __p0, int32_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqadds_s32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vqaddd_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqaddd_s64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int64_t vqaddd_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqaddd_s64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16_t vqaddh_s16(int16_t __p0, int16_t __p1) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqaddh_s16(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int16_t vqaddh_s16(int16_t __p0, int16_t __p1) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqaddh_s16(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vqdmlals_s32(int64_t __p0, int32_t __p1, int32_t __p2) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqdmlals_s32(__p0, __p1, __p2);
+  return __ret;
+}
+#else
+__ai int64_t vqdmlals_s32(int64_t __p0, int32_t __p1, int32_t __p2) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqdmlals_s32(__p0, __p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vqdmlalh_s16(int32_t __p0, int16_t __p1, int16_t __p2) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqdmlalh_s16(__p0, __p1, __p2);
+  return __ret;
+}
+#else
+__ai int32_t vqdmlalh_s16(int32_t __p0, int16_t __p1, int16_t __p2) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqdmlalh_s16(__p0, __p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vqdmlal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int64x2_t __ret;
+  __ret = vqdmlal_s32(__p0, vget_high_s32(__p1), vget_high_s32(__p2));
+  return __ret;
+}
+#else
+__ai int64x2_t vqdmlal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int64x2_t __ret;
+  __ret = __noswap_vqdmlal_s32(__rev0, __noswap_vget_high_s32(__rev1), __noswap_vget_high_s32(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqdmlal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int32x4_t __ret;
+  __ret = vqdmlal_s16(__p0, vget_high_s16(__p1), vget_high_s16(__p2));
+  return __ret;
+}
+#else
+__ai int32x4_t vqdmlal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vqdmlal_s16(__rev0, __noswap_vget_high_s16(__rev1), __noswap_vget_high_s16(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlal_high_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int64x2_t __ret; \
+  __ret = vqdmlal_s32(__s0, vget_high_s32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vqdmlal_high_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __noswap_vqdmlal_s32(__rev0, __noswap_vget_high_s32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlal_high_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int32x4_t __ret; \
+  __ret = vqdmlal_s16(__s0, vget_high_s16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vqdmlal_high_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __noswap_vqdmlal_s16(__rev0, __noswap_vget_high_s16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlal_high_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int64x2_t __ret; \
+  __ret = vqdmlal_s32(__s0, vget_high_s32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vqdmlal_high_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __noswap_vqdmlal_s32(__rev0, __noswap_vget_high_s32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlal_high_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int32x4_t __ret; \
+  __ret = vqdmlal_s16(__s0, vget_high_s16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vqdmlal_high_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __noswap_vqdmlal_s16(__rev0, __noswap_vget_high_s16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vqdmlal_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
+  int64x2_t __ret;
+  __ret = vqdmlal_n_s32(__p0, vget_high_s32(__p1), __p2);
+  return __ret;
+}
+#else
+__ai int64x2_t vqdmlal_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int64x2_t __ret;
+  __ret = __noswap_vqdmlal_n_s32(__rev0, __noswap_vget_high_s32(__rev1), __p2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqdmlal_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
+  int32x4_t __ret;
+  __ret = vqdmlal_n_s16(__p0, vget_high_s16(__p1), __p2);
+  return __ret;
+}
+#else
+__ai int32x4_t vqdmlal_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vqdmlal_n_s16(__rev0, __noswap_vget_high_s16(__rev1), __p2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlals_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vqdmlals_lane_s32(__s0, __s1, (int8x8_t)__s2, __p3); \
+  __ret; \
+})
+#else
+#define vqdmlals_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vqdmlals_lane_s32(__s0, __s1, (int8x8_t)__rev2, __p3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlalh_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int16_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqdmlalh_lane_s16(__s0, __s1, (int8x8_t)__s2, __p3); \
+  __ret; \
+})
+#else
+#define vqdmlalh_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int16_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqdmlalh_lane_s16(__s0, __s1, (int8x8_t)__rev2, __p3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlals_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vqdmlals_laneq_s32(__s0, __s1, (int8x16_t)__s2, __p3); \
+  __ret; \
+})
+#else
+#define vqdmlals_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vqdmlals_laneq_s32(__s0, __s1, (int8x16_t)__rev2, __p3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlalh_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int16_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqdmlalh_laneq_s16(__s0, __s1, (int8x16_t)__s2, __p3); \
+  __ret; \
+})
+#else
+#define vqdmlalh_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int16_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqdmlalh_laneq_s16(__s0, __s1, (int8x16_t)__rev2, __p3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlal_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int64x2_t __ret; \
+  __ret = vqdmlal_s32(__s0, __s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vqdmlal_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __noswap_vqdmlal_s32(__rev0, __rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlal_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int32x4_t __ret; \
+  __ret = vqdmlal_s16(__s0, __s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vqdmlal_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __noswap_vqdmlal_s16(__rev0, __rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vqdmlsls_s32(int64_t __p0, int32_t __p1, int32_t __p2) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqdmlsls_s32(__p0, __p1, __p2);
+  return __ret;
+}
+#else
+__ai int64_t vqdmlsls_s32(int64_t __p0, int32_t __p1, int32_t __p2) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqdmlsls_s32(__p0, __p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vqdmlslh_s16(int32_t __p0, int16_t __p1, int16_t __p2) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqdmlslh_s16(__p0, __p1, __p2);
+  return __ret;
+}
+#else
+__ai int32_t vqdmlslh_s16(int32_t __p0, int16_t __p1, int16_t __p2) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqdmlslh_s16(__p0, __p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vqdmlsl_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int64x2_t __ret;
+  __ret = vqdmlsl_s32(__p0, vget_high_s32(__p1), vget_high_s32(__p2));
+  return __ret;
+}
+#else
+__ai int64x2_t vqdmlsl_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int64x2_t __ret;
+  __ret = __noswap_vqdmlsl_s32(__rev0, __noswap_vget_high_s32(__rev1), __noswap_vget_high_s32(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqdmlsl_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int32x4_t __ret;
+  __ret = vqdmlsl_s16(__p0, vget_high_s16(__p1), vget_high_s16(__p2));
+  return __ret;
+}
+#else
+__ai int32x4_t vqdmlsl_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vqdmlsl_s16(__rev0, __noswap_vget_high_s16(__rev1), __noswap_vget_high_s16(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlsl_high_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int64x2_t __ret; \
+  __ret = vqdmlsl_s32(__s0, vget_high_s32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vqdmlsl_high_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __noswap_vqdmlsl_s32(__rev0, __noswap_vget_high_s32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlsl_high_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int32x4_t __ret; \
+  __ret = vqdmlsl_s16(__s0, vget_high_s16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vqdmlsl_high_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __noswap_vqdmlsl_s16(__rev0, __noswap_vget_high_s16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlsl_high_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int64x2_t __ret; \
+  __ret = vqdmlsl_s32(__s0, vget_high_s32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vqdmlsl_high_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __noswap_vqdmlsl_s32(__rev0, __noswap_vget_high_s32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlsl_high_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int32x4_t __ret; \
+  __ret = vqdmlsl_s16(__s0, vget_high_s16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vqdmlsl_high_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __noswap_vqdmlsl_s16(__rev0, __noswap_vget_high_s16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vqdmlsl_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
+  int64x2_t __ret;
+  __ret = vqdmlsl_n_s32(__p0, vget_high_s32(__p1), __p2);
+  return __ret;
+}
+#else
+__ai int64x2_t vqdmlsl_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int64x2_t __ret;
+  __ret = __noswap_vqdmlsl_n_s32(__rev0, __noswap_vget_high_s32(__rev1), __p2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqdmlsl_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
+  int32x4_t __ret;
+  __ret = vqdmlsl_n_s16(__p0, vget_high_s16(__p1), __p2);
+  return __ret;
+}
+#else
+__ai int32x4_t vqdmlsl_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vqdmlsl_n_s16(__rev0, __noswap_vget_high_s16(__rev1), __p2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlsls_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vqdmlsls_lane_s32(__s0, __s1, (int8x8_t)__s2, __p3); \
+  __ret; \
+})
+#else
+#define vqdmlsls_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vqdmlsls_lane_s32(__s0, __s1, (int8x8_t)__rev2, __p3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlslh_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int16_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqdmlslh_lane_s16(__s0, __s1, (int8x8_t)__s2, __p3); \
+  __ret; \
+})
+#else
+#define vqdmlslh_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int16_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqdmlslh_lane_s16(__s0, __s1, (int8x8_t)__rev2, __p3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlsls_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vqdmlsls_laneq_s32(__s0, __s1, (int8x16_t)__s2, __p3); \
+  __ret; \
+})
+#else
+#define vqdmlsls_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vqdmlsls_laneq_s32(__s0, __s1, (int8x16_t)__rev2, __p3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlslh_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int16_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqdmlslh_laneq_s16(__s0, __s1, (int8x16_t)__s2, __p3); \
+  __ret; \
+})
+#else
+#define vqdmlslh_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int16_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqdmlslh_laneq_s16(__s0, __s1, (int8x16_t)__rev2, __p3); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlsl_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int64x2_t __ret; \
+  __ret = vqdmlsl_s32(__s0, __s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vqdmlsl_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x4_t __s2 = __p2; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __noswap_vqdmlsl_s32(__rev0, __rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmlsl_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int32x4_t __ret; \
+  __ret = vqdmlsl_s16(__s0, __s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vqdmlsl_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x8_t __s2 = __p2; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __noswap_vqdmlsl_s16(__rev0, __rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vqdmulhs_s32(int32_t __p0, int32_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqdmulhs_s32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int32_t vqdmulhs_s32(int32_t __p0, int32_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqdmulhs_s32(__p0, __p1);
+  return __ret;
+}
+__ai int32_t __noswap_vqdmulhs_s32(int32_t __p0, int32_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqdmulhs_s32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16_t vqdmulhh_s16(int16_t __p0, int16_t __p1) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqdmulhh_s16(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int16_t vqdmulhh_s16(int16_t __p0, int16_t __p1) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqdmulhh_s16(__p0, __p1);
+  return __ret;
+}
+__ai int16_t __noswap_vqdmulhh_s16(int16_t __p0, int16_t __p1) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqdmulhh_s16(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulhs_lane_s32(__p0_150, __p1_150, __p2_150) __extension__ ({ \
+  int32_t __s0_150 = __p0_150; \
+  int32x2_t __s1_150 = __p1_150; \
+  int32_t __ret_150; \
+  __ret_150 = vqdmulhs_s32(__s0_150, vget_lane_s32(__s1_150, __p2_150)); \
+  __ret_150; \
+})
+#else
+#define vqdmulhs_lane_s32(__p0_151, __p1_151, __p2_151) __extension__ ({ \
+  int32_t __s0_151 = __p0_151; \
+  int32x2_t __s1_151 = __p1_151; \
+  int32x2_t __rev1_151;  __rev1_151 = __builtin_shufflevector(__s1_151, __s1_151, 1, 0); \
+  int32_t __ret_151; \
+  __ret_151 = __noswap_vqdmulhs_s32(__s0_151, __noswap_vget_lane_s32(__rev1_151, __p2_151)); \
+  __ret_151; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulhh_lane_s16(__p0_152, __p1_152, __p2_152) __extension__ ({ \
+  int16_t __s0_152 = __p0_152; \
+  int16x4_t __s1_152 = __p1_152; \
+  int16_t __ret_152; \
+  __ret_152 = vqdmulhh_s16(__s0_152, vget_lane_s16(__s1_152, __p2_152)); \
+  __ret_152; \
+})
+#else
+#define vqdmulhh_lane_s16(__p0_153, __p1_153, __p2_153) __extension__ ({ \
+  int16_t __s0_153 = __p0_153; \
+  int16x4_t __s1_153 = __p1_153; \
+  int16x4_t __rev1_153;  __rev1_153 = __builtin_shufflevector(__s1_153, __s1_153, 3, 2, 1, 0); \
+  int16_t __ret_153; \
+  __ret_153 = __noswap_vqdmulhh_s16(__s0_153, __noswap_vget_lane_s16(__rev1_153, __p2_153)); \
+  __ret_153; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulhs_laneq_s32(__p0_154, __p1_154, __p2_154) __extension__ ({ \
+  int32_t __s0_154 = __p0_154; \
+  int32x4_t __s1_154 = __p1_154; \
+  int32_t __ret_154; \
+  __ret_154 = vqdmulhs_s32(__s0_154, vgetq_lane_s32(__s1_154, __p2_154)); \
+  __ret_154; \
+})
+#else
+#define vqdmulhs_laneq_s32(__p0_155, __p1_155, __p2_155) __extension__ ({ \
+  int32_t __s0_155 = __p0_155; \
+  int32x4_t __s1_155 = __p1_155; \
+  int32x4_t __rev1_155;  __rev1_155 = __builtin_shufflevector(__s1_155, __s1_155, 3, 2, 1, 0); \
+  int32_t __ret_155; \
+  __ret_155 = __noswap_vqdmulhs_s32(__s0_155, __noswap_vgetq_lane_s32(__rev1_155, __p2_155)); \
+  __ret_155; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulhh_laneq_s16(__p0_156, __p1_156, __p2_156) __extension__ ({ \
+  int16_t __s0_156 = __p0_156; \
+  int16x8_t __s1_156 = __p1_156; \
+  int16_t __ret_156; \
+  __ret_156 = vqdmulhh_s16(__s0_156, vgetq_lane_s16(__s1_156, __p2_156)); \
+  __ret_156; \
+})
+#else
+#define vqdmulhh_laneq_s16(__p0_157, __p1_157, __p2_157) __extension__ ({ \
+  int16_t __s0_157 = __p0_157; \
+  int16x8_t __s1_157 = __p1_157; \
+  int16x8_t __rev1_157;  __rev1_157 = __builtin_shufflevector(__s1_157, __s1_157, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16_t __ret_157; \
+  __ret_157 = __noswap_vqdmulhh_s16(__s0_157, __noswap_vgetq_lane_s16(__rev1_157, __p2_157)); \
+  __ret_157; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulhq_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = vqdmulhq_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqdmulhq_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __noswap_vqdmulhq_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulhq_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __ret; \
+  __ret = vqdmulhq_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqdmulhq_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = __noswap_vqdmulhq_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulh_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x2_t __ret; \
+  __ret = vqdmulh_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqdmulh_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x2_t __ret; \
+  __ret = __noswap_vqdmulh_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulh_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x4_t __ret; \
+  __ret = vqdmulh_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqdmulh_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = __noswap_vqdmulh_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vqdmulls_s32(int32_t __p0, int32_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqdmulls_s32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int64_t vqdmulls_s32(int32_t __p0, int32_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqdmulls_s32(__p0, __p1);
+  return __ret;
+}
+__ai int64_t __noswap_vqdmulls_s32(int32_t __p0, int32_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqdmulls_s32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vqdmullh_s16(int16_t __p0, int16_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqdmullh_s16(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int32_t vqdmullh_s16(int16_t __p0, int16_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqdmullh_s16(__p0, __p1);
+  return __ret;
+}
+__ai int32_t __noswap_vqdmullh_s16(int16_t __p0, int16_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqdmullh_s16(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vqdmull_high_s32(int32x4_t __p0, int32x4_t __p1) {
+  int64x2_t __ret;
+  __ret = vqdmull_s32(vget_high_s32(__p0), vget_high_s32(__p1));
+  return __ret;
+}
+#else
+__ai int64x2_t vqdmull_high_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int64x2_t __ret;
+  __ret = __noswap_vqdmull_s32(__noswap_vget_high_s32(__rev0), __noswap_vget_high_s32(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqdmull_high_s16(int16x8_t __p0, int16x8_t __p1) {
+  int32x4_t __ret;
+  __ret = vqdmull_s16(vget_high_s16(__p0), vget_high_s16(__p1));
+  return __ret;
+}
+#else
+__ai int32x4_t vqdmull_high_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vqdmull_s16(__noswap_vget_high_s16(__rev0), __noswap_vget_high_s16(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmull_high_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int64x2_t __ret; \
+  __ret = vqdmull_s32(vget_high_s32(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqdmull_high_lane_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __noswap_vqdmull_s32(__noswap_vget_high_s32(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmull_high_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = vqdmull_s16(vget_high_s16(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqdmull_high_lane_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __noswap_vqdmull_s16(__noswap_vget_high_s16(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmull_high_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int64x2_t __ret; \
+  __ret = vqdmull_s32(vget_high_s32(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqdmull_high_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __noswap_vqdmull_s32(__noswap_vget_high_s32(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmull_high_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = vqdmull_s16(vget_high_s16(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqdmull_high_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __noswap_vqdmull_s16(__noswap_vget_high_s16(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vqdmull_high_n_s32(int32x4_t __p0, int32_t __p1) {
+  int64x2_t __ret;
+  __ret = vqdmull_n_s32(vget_high_s32(__p0), __p1);
+  return __ret;
+}
+#else
+__ai int64x2_t vqdmull_high_n_s32(int32x4_t __p0, int32_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int64x2_t __ret;
+  __ret = __noswap_vqdmull_n_s32(__noswap_vget_high_s32(__rev0), __p1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqdmull_high_n_s16(int16x8_t __p0, int16_t __p1) {
+  int32x4_t __ret;
+  __ret = vqdmull_n_s16(vget_high_s16(__p0), __p1);
+  return __ret;
+}
+#else
+__ai int32x4_t vqdmull_high_n_s16(int16x8_t __p0, int16_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vqdmull_n_s16(__noswap_vget_high_s16(__rev0), __p1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulls_lane_s32(__p0_158, __p1_158, __p2_158) __extension__ ({ \
+  int32_t __s0_158 = __p0_158; \
+  int32x2_t __s1_158 = __p1_158; \
+  int64_t __ret_158; \
+  __ret_158 = vqdmulls_s32(__s0_158, vget_lane_s32(__s1_158, __p2_158)); \
+  __ret_158; \
+})
+#else
+#define vqdmulls_lane_s32(__p0_159, __p1_159, __p2_159) __extension__ ({ \
+  int32_t __s0_159 = __p0_159; \
+  int32x2_t __s1_159 = __p1_159; \
+  int32x2_t __rev1_159;  __rev1_159 = __builtin_shufflevector(__s1_159, __s1_159, 1, 0); \
+  int64_t __ret_159; \
+  __ret_159 = __noswap_vqdmulls_s32(__s0_159, __noswap_vget_lane_s32(__rev1_159, __p2_159)); \
+  __ret_159; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmullh_lane_s16(__p0_160, __p1_160, __p2_160) __extension__ ({ \
+  int16_t __s0_160 = __p0_160; \
+  int16x4_t __s1_160 = __p1_160; \
+  int32_t __ret_160; \
+  __ret_160 = vqdmullh_s16(__s0_160, vget_lane_s16(__s1_160, __p2_160)); \
+  __ret_160; \
+})
+#else
+#define vqdmullh_lane_s16(__p0_161, __p1_161, __p2_161) __extension__ ({ \
+  int16_t __s0_161 = __p0_161; \
+  int16x4_t __s1_161 = __p1_161; \
+  int16x4_t __rev1_161;  __rev1_161 = __builtin_shufflevector(__s1_161, __s1_161, 3, 2, 1, 0); \
+  int32_t __ret_161; \
+  __ret_161 = __noswap_vqdmullh_s16(__s0_161, __noswap_vget_lane_s16(__rev1_161, __p2_161)); \
+  __ret_161; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulls_laneq_s32(__p0_162, __p1_162, __p2_162) __extension__ ({ \
+  int32_t __s0_162 = __p0_162; \
+  int32x4_t __s1_162 = __p1_162; \
+  int64_t __ret_162; \
+  __ret_162 = vqdmulls_s32(__s0_162, vgetq_lane_s32(__s1_162, __p2_162)); \
+  __ret_162; \
+})
+#else
+#define vqdmulls_laneq_s32(__p0_163, __p1_163, __p2_163) __extension__ ({ \
+  int32_t __s0_163 = __p0_163; \
+  int32x4_t __s1_163 = __p1_163; \
+  int32x4_t __rev1_163;  __rev1_163 = __builtin_shufflevector(__s1_163, __s1_163, 3, 2, 1, 0); \
+  int64_t __ret_163; \
+  __ret_163 = __noswap_vqdmulls_s32(__s0_163, __noswap_vgetq_lane_s32(__rev1_163, __p2_163)); \
+  __ret_163; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmullh_laneq_s16(__p0_164, __p1_164, __p2_164) __extension__ ({ \
+  int16_t __s0_164 = __p0_164; \
+  int16x8_t __s1_164 = __p1_164; \
+  int32_t __ret_164; \
+  __ret_164 = vqdmullh_s16(__s0_164, vgetq_lane_s16(__s1_164, __p2_164)); \
+  __ret_164; \
+})
+#else
+#define vqdmullh_laneq_s16(__p0_165, __p1_165, __p2_165) __extension__ ({ \
+  int16_t __s0_165 = __p0_165; \
+  int16x8_t __s1_165 = __p1_165; \
+  int16x8_t __rev1_165;  __rev1_165 = __builtin_shufflevector(__s1_165, __s1_165, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int32_t __ret_165; \
+  __ret_165 = __noswap_vqdmullh_s16(__s0_165, __noswap_vgetq_lane_s16(__rev1_165, __p2_165)); \
+  __ret_165; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmull_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int64x2_t __ret; \
+  __ret = vqdmull_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqdmull_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __noswap_vqdmull_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmull_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = vqdmull_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqdmull_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __noswap_vqdmull_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16_t vqmovns_s32(int32_t __p0) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqmovns_s32(__p0);
+  return __ret;
+}
+#else
+__ai int16_t vqmovns_s32(int32_t __p0) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqmovns_s32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vqmovnd_s64(int64_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqmovnd_s64(__p0);
+  return __ret;
+}
+#else
+__ai int32_t vqmovnd_s64(int64_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqmovnd_s64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8_t vqmovnh_s16(int16_t __p0) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vqmovnh_s16(__p0);
+  return __ret;
+}
+#else
+__ai int8_t vqmovnh_s16(int16_t __p0) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vqmovnh_s16(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16_t vqmovns_u32(uint32_t __p0) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vqmovns_u32(__p0);
+  return __ret;
+}
+#else
+__ai uint16_t vqmovns_u32(uint32_t __p0) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vqmovns_u32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vqmovnd_u64(uint64_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vqmovnd_u64(__p0);
+  return __ret;
+}
+#else
+__ai uint32_t vqmovnd_u64(uint64_t __p0) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vqmovnd_u64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8_t vqmovnh_u16(uint16_t __p0) {
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vqmovnh_u16(__p0);
+  return __ret;
+}
+#else
+__ai uint8_t vqmovnh_u16(uint16_t __p0) {
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vqmovnh_u16(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vqmovn_high_u32(uint16x4_t __p0, uint32x4_t __p1) {
+  uint16x8_t __ret;
+  __ret = vcombine_u16(__p0, vqmovn_u32(__p1));
+  return __ret;
+}
+#else
+__ai uint16x8_t vqmovn_high_u32(uint16x4_t __p0, uint32x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __noswap_vcombine_u16(__rev0, __noswap_vqmovn_u32(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vqmovn_high_u64(uint32x2_t __p0, uint64x2_t __p1) {
+  uint32x4_t __ret;
+  __ret = vcombine_u32(__p0, vqmovn_u64(__p1));
+  return __ret;
+}
+#else
+__ai uint32x4_t vqmovn_high_u64(uint32x2_t __p0, uint64x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x4_t __ret;
+  __ret = __noswap_vcombine_u32(__rev0, __noswap_vqmovn_u64(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vqmovn_high_u16(uint8x8_t __p0, uint16x8_t __p1) {
+  uint8x16_t __ret;
+  __ret = vcombine_u8(__p0, vqmovn_u16(__p1));
+  return __ret;
+}
+#else
+__ai uint8x16_t vqmovn_high_u16(uint8x8_t __p0, uint16x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __noswap_vcombine_u8(__rev0, __noswap_vqmovn_u16(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vqmovn_high_s32(int16x4_t __p0, int32x4_t __p1) {
+  int16x8_t __ret;
+  __ret = vcombine_s16(__p0, vqmovn_s32(__p1));
+  return __ret;
+}
+#else
+__ai int16x8_t vqmovn_high_s32(int16x4_t __p0, int32x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __noswap_vcombine_s16(__rev0, __noswap_vqmovn_s32(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqmovn_high_s64(int32x2_t __p0, int64x2_t __p1) {
+  int32x4_t __ret;
+  __ret = vcombine_s32(__p0, vqmovn_s64(__p1));
+  return __ret;
+}
+#else
+__ai int32x4_t vqmovn_high_s64(int32x2_t __p0, int64x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vcombine_s32(__rev0, __noswap_vqmovn_s64(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vqmovn_high_s16(int8x8_t __p0, int16x8_t __p1) {
+  int8x16_t __ret;
+  __ret = vcombine_s8(__p0, vqmovn_s16(__p1));
+  return __ret;
+}
+#else
+__ai int8x16_t vqmovn_high_s16(int8x8_t __p0, int16x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __noswap_vcombine_s8(__rev0, __noswap_vqmovn_s16(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16_t vqmovuns_s32(int32_t __p0) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqmovuns_s32(__p0);
+  return __ret;
+}
+#else
+__ai int16_t vqmovuns_s32(int32_t __p0) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqmovuns_s32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vqmovund_s64(int64_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqmovund_s64(__p0);
+  return __ret;
+}
+#else
+__ai int32_t vqmovund_s64(int64_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqmovund_s64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8_t vqmovunh_s16(int16_t __p0) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vqmovunh_s16(__p0);
+  return __ret;
+}
+#else
+__ai int8_t vqmovunh_s16(int16_t __p0) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vqmovunh_s16(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vqmovun_high_s32(int16x4_t __p0, int32x4_t __p1) {
+  int16x8_t __ret;
+  __ret = vcombine_u16((uint16x4_t)(__p0), vqmovun_s32(__p1));
+  return __ret;
+}
+#else
+__ai int16x8_t vqmovun_high_s32(int16x4_t __p0, int32x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __noswap_vcombine_u16((uint16x4_t)(__rev0), __noswap_vqmovun_s32(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vqmovun_high_s64(int32x2_t __p0, int64x2_t __p1) {
+  int32x4_t __ret;
+  __ret = vcombine_u32((uint32x2_t)(__p0), vqmovun_s64(__p1));
+  return __ret;
+}
+#else
+__ai int32x4_t vqmovun_high_s64(int32x2_t __p0, int64x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vcombine_u32((uint32x2_t)(__rev0), __noswap_vqmovun_s64(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vqmovun_high_s16(int8x8_t __p0, int16x8_t __p1) {
+  int8x16_t __ret;
+  __ret = vcombine_u8((uint8x8_t)(__p0), vqmovun_s16(__p1));
+  return __ret;
+}
+#else
+__ai int8x16_t vqmovun_high_s16(int8x8_t __p0, int16x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __noswap_vcombine_u8((uint8x8_t)(__rev0), __noswap_vqmovun_s16(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vqnegq_s64(int64x2_t __p0) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqnegq_v((int8x16_t)__p0, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vqnegq_s64(int64x2_t __p0) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vqnegq_v((int8x16_t)__rev0, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vqneg_s64(int64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vqneg_v((int8x8_t)__p0, 3);
+  return __ret;
+}
+#else
+__ai int64x1_t vqneg_s64(int64x1_t __p0) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vqneg_v((int8x8_t)__p0, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8_t vqnegb_s8(int8_t __p0) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vqnegb_s8(__p0);
+  return __ret;
+}
+#else
+__ai int8_t vqnegb_s8(int8_t __p0) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vqnegb_s8(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vqnegs_s32(int32_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqnegs_s32(__p0);
+  return __ret;
+}
+#else
+__ai int32_t vqnegs_s32(int32_t __p0) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqnegs_s32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vqnegd_s64(int64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqnegd_s64(__p0);
+  return __ret;
+}
+#else
+__ai int64_t vqnegd_s64(int64_t __p0) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqnegd_s64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16_t vqnegh_s16(int16_t __p0) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqnegh_s16(__p0);
+  return __ret;
+}
+#else
+__ai int16_t vqnegh_s16(int16_t __p0) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqnegh_s16(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vqrdmulhs_s32(int32_t __p0, int32_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqrdmulhs_s32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int32_t vqrdmulhs_s32(int32_t __p0, int32_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqrdmulhs_s32(__p0, __p1);
+  return __ret;
+}
+__ai int32_t __noswap_vqrdmulhs_s32(int32_t __p0, int32_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqrdmulhs_s32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16_t vqrdmulhh_s16(int16_t __p0, int16_t __p1) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqrdmulhh_s16(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int16_t vqrdmulhh_s16(int16_t __p0, int16_t __p1) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqrdmulhh_s16(__p0, __p1);
+  return __ret;
+}
+__ai int16_t __noswap_vqrdmulhh_s16(int16_t __p0, int16_t __p1) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqrdmulhh_s16(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmulhs_lane_s32(__p0_166, __p1_166, __p2_166) __extension__ ({ \
+  int32_t __s0_166 = __p0_166; \
+  int32x2_t __s1_166 = __p1_166; \
+  int32_t __ret_166; \
+  __ret_166 = vqrdmulhs_s32(__s0_166, vget_lane_s32(__s1_166, __p2_166)); \
+  __ret_166; \
+})
+#else
+#define vqrdmulhs_lane_s32(__p0_167, __p1_167, __p2_167) __extension__ ({ \
+  int32_t __s0_167 = __p0_167; \
+  int32x2_t __s1_167 = __p1_167; \
+  int32x2_t __rev1_167;  __rev1_167 = __builtin_shufflevector(__s1_167, __s1_167, 1, 0); \
+  int32_t __ret_167; \
+  __ret_167 = __noswap_vqrdmulhs_s32(__s0_167, __noswap_vget_lane_s32(__rev1_167, __p2_167)); \
+  __ret_167; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmulhh_lane_s16(__p0_168, __p1_168, __p2_168) __extension__ ({ \
+  int16_t __s0_168 = __p0_168; \
+  int16x4_t __s1_168 = __p1_168; \
+  int16_t __ret_168; \
+  __ret_168 = vqrdmulhh_s16(__s0_168, vget_lane_s16(__s1_168, __p2_168)); \
+  __ret_168; \
+})
+#else
+#define vqrdmulhh_lane_s16(__p0_169, __p1_169, __p2_169) __extension__ ({ \
+  int16_t __s0_169 = __p0_169; \
+  int16x4_t __s1_169 = __p1_169; \
+  int16x4_t __rev1_169;  __rev1_169 = __builtin_shufflevector(__s1_169, __s1_169, 3, 2, 1, 0); \
+  int16_t __ret_169; \
+  __ret_169 = __noswap_vqrdmulhh_s16(__s0_169, __noswap_vget_lane_s16(__rev1_169, __p2_169)); \
+  __ret_169; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmulhs_laneq_s32(__p0_170, __p1_170, __p2_170) __extension__ ({ \
+  int32_t __s0_170 = __p0_170; \
+  int32x4_t __s1_170 = __p1_170; \
+  int32_t __ret_170; \
+  __ret_170 = vqrdmulhs_s32(__s0_170, vgetq_lane_s32(__s1_170, __p2_170)); \
+  __ret_170; \
+})
+#else
+#define vqrdmulhs_laneq_s32(__p0_171, __p1_171, __p2_171) __extension__ ({ \
+  int32_t __s0_171 = __p0_171; \
+  int32x4_t __s1_171 = __p1_171; \
+  int32x4_t __rev1_171;  __rev1_171 = __builtin_shufflevector(__s1_171, __s1_171, 3, 2, 1, 0); \
+  int32_t __ret_171; \
+  __ret_171 = __noswap_vqrdmulhs_s32(__s0_171, __noswap_vgetq_lane_s32(__rev1_171, __p2_171)); \
+  __ret_171; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmulhh_laneq_s16(__p0_172, __p1_172, __p2_172) __extension__ ({ \
+  int16_t __s0_172 = __p0_172; \
+  int16x8_t __s1_172 = __p1_172; \
+  int16_t __ret_172; \
+  __ret_172 = vqrdmulhh_s16(__s0_172, vgetq_lane_s16(__s1_172, __p2_172)); \
+  __ret_172; \
+})
+#else
+#define vqrdmulhh_laneq_s16(__p0_173, __p1_173, __p2_173) __extension__ ({ \
+  int16_t __s0_173 = __p0_173; \
+  int16x8_t __s1_173 = __p1_173; \
+  int16x8_t __rev1_173;  __rev1_173 = __builtin_shufflevector(__s1_173, __s1_173, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16_t __ret_173; \
+  __ret_173 = __noswap_vqrdmulhh_s16(__s0_173, __noswap_vgetq_lane_s16(__rev1_173, __p2_173)); \
+  __ret_173; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmulhq_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __ret; \
+  __ret = vqrdmulhq_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqrdmulhq_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __noswap_vqrdmulhq_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmulhq_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __ret; \
+  __ret = vqrdmulhq_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqrdmulhq_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x8_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret; \
+  __ret = __noswap_vqrdmulhq_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmulh_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x2_t __ret; \
+  __ret = vqrdmulh_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqrdmulh_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
+  int32x2_t __s0 = __p0; \
+  int32x4_t __s1 = __p1; \
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int32x2_t __ret; \
+  __ret = __noswap_vqrdmulh_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmulh_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x4_t __ret; \
+  __ret = vqrdmulh_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
+  __ret; \
+})
+#else
+#define vqrdmulh_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
+  int16x4_t __s0 = __p0; \
+  int16x8_t __s1 = __p1; \
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __ret; \
+  __ret = __noswap_vqrdmulh_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8_t vqrshlb_u8(uint8_t __p0, uint8_t __p1) {
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vqrshlb_u8(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint8_t vqrshlb_u8(uint8_t __p0, uint8_t __p1) {
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vqrshlb_u8(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vqrshls_u32(uint32_t __p0, uint32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vqrshls_u32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint32_t vqrshls_u32(uint32_t __p0, uint32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vqrshls_u32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vqrshld_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vqrshld_u64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vqrshld_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vqrshld_u64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16_t vqrshlh_u16(uint16_t __p0, uint16_t __p1) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vqrshlh_u16(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint16_t vqrshlh_u16(uint16_t __p0, uint16_t __p1) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vqrshlh_u16(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8_t vqrshlb_s8(int8_t __p0, int8_t __p1) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vqrshlb_s8(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int8_t vqrshlb_s8(int8_t __p0, int8_t __p1) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vqrshlb_s8(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vqrshls_s32(int32_t __p0, int32_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqrshls_s32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int32_t vqrshls_s32(int32_t __p0, int32_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqrshls_s32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vqrshld_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqrshld_s64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int64_t vqrshld_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqrshld_s64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16_t vqrshlh_s16(int16_t __p0, int16_t __p1) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqrshlh_s16(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int16_t vqrshlh_s16(int16_t __p0, int16_t __p1) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqrshlh_s16(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrn_high_n_u32(__p0_174, __p1_174, __p2_174) __extension__ ({ \
+  uint16x4_t __s0_174 = __p0_174; \
+  uint32x4_t __s1_174 = __p1_174; \
+  uint16x8_t __ret_174; \
+  __ret_174 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_174), (uint16x4_t)(vqrshrn_n_u32(__s1_174, __p2_174)))); \
+  __ret_174; \
+})
+#else
+#define vqrshrn_high_n_u32(__p0_175, __p1_175, __p2_175) __extension__ ({ \
+  uint16x4_t __s0_175 = __p0_175; \
+  uint32x4_t __s1_175 = __p1_175; \
+  uint16x4_t __rev0_175;  __rev0_175 = __builtin_shufflevector(__s0_175, __s0_175, 3, 2, 1, 0); \
+  uint32x4_t __rev1_175;  __rev1_175 = __builtin_shufflevector(__s1_175, __s1_175, 3, 2, 1, 0); \
+  uint16x8_t __ret_175; \
+  __ret_175 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_175), (uint16x4_t)(__noswap_vqrshrn_n_u32(__rev1_175, __p2_175)))); \
+  __ret_175 = __builtin_shufflevector(__ret_175, __ret_175, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_175; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrn_high_n_u64(__p0_176, __p1_176, __p2_176) __extension__ ({ \
+  uint32x2_t __s0_176 = __p0_176; \
+  uint64x2_t __s1_176 = __p1_176; \
+  uint32x4_t __ret_176; \
+  __ret_176 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_176), (uint32x2_t)(vqrshrn_n_u64(__s1_176, __p2_176)))); \
+  __ret_176; \
+})
+#else
+#define vqrshrn_high_n_u64(__p0_177, __p1_177, __p2_177) __extension__ ({ \
+  uint32x2_t __s0_177 = __p0_177; \
+  uint64x2_t __s1_177 = __p1_177; \
+  uint32x2_t __rev0_177;  __rev0_177 = __builtin_shufflevector(__s0_177, __s0_177, 1, 0); \
+  uint64x2_t __rev1_177;  __rev1_177 = __builtin_shufflevector(__s1_177, __s1_177, 1, 0); \
+  uint32x4_t __ret_177; \
+  __ret_177 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_177), (uint32x2_t)(__noswap_vqrshrn_n_u64(__rev1_177, __p2_177)))); \
+  __ret_177 = __builtin_shufflevector(__ret_177, __ret_177, 3, 2, 1, 0); \
+  __ret_177; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrn_high_n_u16(__p0_178, __p1_178, __p2_178) __extension__ ({ \
+  uint8x8_t __s0_178 = __p0_178; \
+  uint16x8_t __s1_178 = __p1_178; \
+  uint8x16_t __ret_178; \
+  __ret_178 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_178), (uint8x8_t)(vqrshrn_n_u16(__s1_178, __p2_178)))); \
+  __ret_178; \
+})
+#else
+#define vqrshrn_high_n_u16(__p0_179, __p1_179, __p2_179) __extension__ ({ \
+  uint8x8_t __s0_179 = __p0_179; \
+  uint16x8_t __s1_179 = __p1_179; \
+  uint8x8_t __rev0_179;  __rev0_179 = __builtin_shufflevector(__s0_179, __s0_179, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1_179;  __rev1_179 = __builtin_shufflevector(__s1_179, __s1_179, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret_179; \
+  __ret_179 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_179), (uint8x8_t)(__noswap_vqrshrn_n_u16(__rev1_179, __p2_179)))); \
+  __ret_179 = __builtin_shufflevector(__ret_179, __ret_179, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_179; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrn_high_n_s32(__p0_180, __p1_180, __p2_180) __extension__ ({ \
+  int16x4_t __s0_180 = __p0_180; \
+  int32x4_t __s1_180 = __p1_180; \
+  int16x8_t __ret_180; \
+  __ret_180 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_180), (int16x4_t)(vqrshrn_n_s32(__s1_180, __p2_180)))); \
+  __ret_180; \
+})
+#else
+#define vqrshrn_high_n_s32(__p0_181, __p1_181, __p2_181) __extension__ ({ \
+  int16x4_t __s0_181 = __p0_181; \
+  int32x4_t __s1_181 = __p1_181; \
+  int16x4_t __rev0_181;  __rev0_181 = __builtin_shufflevector(__s0_181, __s0_181, 3, 2, 1, 0); \
+  int32x4_t __rev1_181;  __rev1_181 = __builtin_shufflevector(__s1_181, __s1_181, 3, 2, 1, 0); \
+  int16x8_t __ret_181; \
+  __ret_181 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_181), (int16x4_t)(__noswap_vqrshrn_n_s32(__rev1_181, __p2_181)))); \
+  __ret_181 = __builtin_shufflevector(__ret_181, __ret_181, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_181; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrn_high_n_s64(__p0_182, __p1_182, __p2_182) __extension__ ({ \
+  int32x2_t __s0_182 = __p0_182; \
+  int64x2_t __s1_182 = __p1_182; \
+  int32x4_t __ret_182; \
+  __ret_182 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_182), (int32x2_t)(vqrshrn_n_s64(__s1_182, __p2_182)))); \
+  __ret_182; \
+})
+#else
+#define vqrshrn_high_n_s64(__p0_183, __p1_183, __p2_183) __extension__ ({ \
+  int32x2_t __s0_183 = __p0_183; \
+  int64x2_t __s1_183 = __p1_183; \
+  int32x2_t __rev0_183;  __rev0_183 = __builtin_shufflevector(__s0_183, __s0_183, 1, 0); \
+  int64x2_t __rev1_183;  __rev1_183 = __builtin_shufflevector(__s1_183, __s1_183, 1, 0); \
+  int32x4_t __ret_183; \
+  __ret_183 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_183), (int32x2_t)(__noswap_vqrshrn_n_s64(__rev1_183, __p2_183)))); \
+  __ret_183 = __builtin_shufflevector(__ret_183, __ret_183, 3, 2, 1, 0); \
+  __ret_183; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrn_high_n_s16(__p0_184, __p1_184, __p2_184) __extension__ ({ \
+  int8x8_t __s0_184 = __p0_184; \
+  int16x8_t __s1_184 = __p1_184; \
+  int8x16_t __ret_184; \
+  __ret_184 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_184), (int8x8_t)(vqrshrn_n_s16(__s1_184, __p2_184)))); \
+  __ret_184; \
+})
+#else
+#define vqrshrn_high_n_s16(__p0_185, __p1_185, __p2_185) __extension__ ({ \
+  int8x8_t __s0_185 = __p0_185; \
+  int16x8_t __s1_185 = __p1_185; \
+  int8x8_t __rev0_185;  __rev0_185 = __builtin_shufflevector(__s0_185, __s0_185, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1_185;  __rev1_185 = __builtin_shufflevector(__s1_185, __s1_185, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret_185; \
+  __ret_185 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_185), (int8x8_t)(__noswap_vqrshrn_n_s16(__rev1_185, __p2_185)))); \
+  __ret_185 = __builtin_shufflevector(__ret_185, __ret_185, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_185; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrns_n_u32(__p0, __p1) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vqrshrns_n_u32(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqrshrns_n_u32(__p0, __p1) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vqrshrns_n_u32(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrnd_n_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vqrshrnd_n_u64(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqrshrnd_n_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vqrshrnd_n_u64(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrnh_n_u16(__p0, __p1) __extension__ ({ \
+  uint16_t __s0 = __p0; \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vqrshrnh_n_u16(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqrshrnh_n_u16(__p0, __p1) __extension__ ({ \
+  uint16_t __s0 = __p0; \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vqrshrnh_n_u16(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrns_n_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vqrshrns_n_s32(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqrshrns_n_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vqrshrns_n_s32(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrnd_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqrshrnd_n_s64(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqrshrnd_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqrshrnd_n_s64(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrnh_n_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vqrshrnh_n_s16(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqrshrnh_n_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vqrshrnh_n_s16(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrun_high_n_s32(__p0_186, __p1_186, __p2_186) __extension__ ({ \
+  int16x4_t __s0_186 = __p0_186; \
+  int32x4_t __s1_186 = __p1_186; \
+  int16x8_t __ret_186; \
+  __ret_186 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_186), (int16x4_t)(vqrshrun_n_s32(__s1_186, __p2_186)))); \
+  __ret_186; \
+})
+#else
+#define vqrshrun_high_n_s32(__p0_187, __p1_187, __p2_187) __extension__ ({ \
+  int16x4_t __s0_187 = __p0_187; \
+  int32x4_t __s1_187 = __p1_187; \
+  int16x4_t __rev0_187;  __rev0_187 = __builtin_shufflevector(__s0_187, __s0_187, 3, 2, 1, 0); \
+  int32x4_t __rev1_187;  __rev1_187 = __builtin_shufflevector(__s1_187, __s1_187, 3, 2, 1, 0); \
+  int16x8_t __ret_187; \
+  __ret_187 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_187), (int16x4_t)(__noswap_vqrshrun_n_s32(__rev1_187, __p2_187)))); \
+  __ret_187 = __builtin_shufflevector(__ret_187, __ret_187, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_187; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrun_high_n_s64(__p0_188, __p1_188, __p2_188) __extension__ ({ \
+  int32x2_t __s0_188 = __p0_188; \
+  int64x2_t __s1_188 = __p1_188; \
+  int32x4_t __ret_188; \
+  __ret_188 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_188), (int32x2_t)(vqrshrun_n_s64(__s1_188, __p2_188)))); \
+  __ret_188; \
+})
+#else
+#define vqrshrun_high_n_s64(__p0_189, __p1_189, __p2_189) __extension__ ({ \
+  int32x2_t __s0_189 = __p0_189; \
+  int64x2_t __s1_189 = __p1_189; \
+  int32x2_t __rev0_189;  __rev0_189 = __builtin_shufflevector(__s0_189, __s0_189, 1, 0); \
+  int64x2_t __rev1_189;  __rev1_189 = __builtin_shufflevector(__s1_189, __s1_189, 1, 0); \
+  int32x4_t __ret_189; \
+  __ret_189 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_189), (int32x2_t)(__noswap_vqrshrun_n_s64(__rev1_189, __p2_189)))); \
+  __ret_189 = __builtin_shufflevector(__ret_189, __ret_189, 3, 2, 1, 0); \
+  __ret_189; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrun_high_n_s16(__p0_190, __p1_190, __p2_190) __extension__ ({ \
+  int8x8_t __s0_190 = __p0_190; \
+  int16x8_t __s1_190 = __p1_190; \
+  int8x16_t __ret_190; \
+  __ret_190 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_190), (int8x8_t)(vqrshrun_n_s16(__s1_190, __p2_190)))); \
+  __ret_190; \
+})
+#else
+#define vqrshrun_high_n_s16(__p0_191, __p1_191, __p2_191) __extension__ ({ \
+  int8x8_t __s0_191 = __p0_191; \
+  int16x8_t __s1_191 = __p1_191; \
+  int8x8_t __rev0_191;  __rev0_191 = __builtin_shufflevector(__s0_191, __s0_191, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1_191;  __rev1_191 = __builtin_shufflevector(__s1_191, __s1_191, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret_191; \
+  __ret_191 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_191), (int8x8_t)(__noswap_vqrshrun_n_s16(__rev1_191, __p2_191)))); \
+  __ret_191 = __builtin_shufflevector(__ret_191, __ret_191, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_191; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshruns_n_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vqrshruns_n_s32(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqrshruns_n_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vqrshruns_n_s32(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrund_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqrshrund_n_s64(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqrshrund_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqrshrund_n_s64(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrunh_n_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vqrshrunh_n_s16(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqrshrunh_n_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vqrshrunh_n_s16(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8_t vqshlb_u8(uint8_t __p0, uint8_t __p1) {
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vqshlb_u8(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint8_t vqshlb_u8(uint8_t __p0, uint8_t __p1) {
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vqshlb_u8(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vqshls_u32(uint32_t __p0, uint32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vqshls_u32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint32_t vqshls_u32(uint32_t __p0, uint32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vqshls_u32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vqshld_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vqshld_u64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vqshld_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vqshld_u64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16_t vqshlh_u16(uint16_t __p0, uint16_t __p1) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vqshlh_u16(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint16_t vqshlh_u16(uint16_t __p0, uint16_t __p1) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vqshlh_u16(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8_t vqshlb_s8(int8_t __p0, int8_t __p1) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vqshlb_s8(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int8_t vqshlb_s8(int8_t __p0, int8_t __p1) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vqshlb_s8(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vqshls_s32(int32_t __p0, int32_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqshls_s32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int32_t vqshls_s32(int32_t __p0, int32_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqshls_s32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vqshld_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqshld_s64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int64_t vqshld_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqshld_s64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16_t vqshlh_s16(int16_t __p0, int16_t __p1) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqshlh_s16(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int16_t vqshlh_s16(int16_t __p0, int16_t __p1) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqshlh_s16(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshlb_n_u8(__p0, __p1) __extension__ ({ \
+  uint8_t __s0 = __p0; \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vqshlb_n_u8(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshlb_n_u8(__p0, __p1) __extension__ ({ \
+  uint8_t __s0 = __p0; \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vqshlb_n_u8(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshls_n_u32(__p0, __p1) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vqshls_n_u32(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshls_n_u32(__p0, __p1) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vqshls_n_u32(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshld_n_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vqshld_n_u64(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshld_n_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vqshld_n_u64(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshlh_n_u16(__p0, __p1) __extension__ ({ \
+  uint16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vqshlh_n_u16(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshlh_n_u16(__p0, __p1) __extension__ ({ \
+  uint16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vqshlh_n_u16(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshlb_n_s8(__p0, __p1) __extension__ ({ \
+  int8_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vqshlb_n_s8(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshlb_n_s8(__p0, __p1) __extension__ ({ \
+  int8_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vqshlb_n_s8(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshls_n_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqshls_n_s32(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshls_n_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqshls_n_s32(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshld_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vqshld_n_s64(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshld_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vqshld_n_s64(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshlh_n_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vqshlh_n_s16(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshlh_n_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vqshlh_n_s16(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshlub_n_s8(__p0, __p1) __extension__ ({ \
+  int8_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vqshlub_n_s8(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshlub_n_s8(__p0, __p1) __extension__ ({ \
+  int8_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vqshlub_n_s8(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshlus_n_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqshlus_n_s32(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshlus_n_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqshlus_n_s32(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshlud_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vqshlud_n_s64(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshlud_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vqshlud_n_s64(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshluh_n_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vqshluh_n_s16(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshluh_n_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vqshluh_n_s16(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrn_high_n_u32(__p0_192, __p1_192, __p2_192) __extension__ ({ \
+  uint16x4_t __s0_192 = __p0_192; \
+  uint32x4_t __s1_192 = __p1_192; \
+  uint16x8_t __ret_192; \
+  __ret_192 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_192), (uint16x4_t)(vqshrn_n_u32(__s1_192, __p2_192)))); \
+  __ret_192; \
+})
+#else
+#define vqshrn_high_n_u32(__p0_193, __p1_193, __p2_193) __extension__ ({ \
+  uint16x4_t __s0_193 = __p0_193; \
+  uint32x4_t __s1_193 = __p1_193; \
+  uint16x4_t __rev0_193;  __rev0_193 = __builtin_shufflevector(__s0_193, __s0_193, 3, 2, 1, 0); \
+  uint32x4_t __rev1_193;  __rev1_193 = __builtin_shufflevector(__s1_193, __s1_193, 3, 2, 1, 0); \
+  uint16x8_t __ret_193; \
+  __ret_193 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_193), (uint16x4_t)(__noswap_vqshrn_n_u32(__rev1_193, __p2_193)))); \
+  __ret_193 = __builtin_shufflevector(__ret_193, __ret_193, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_193; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrn_high_n_u64(__p0_194, __p1_194, __p2_194) __extension__ ({ \
+  uint32x2_t __s0_194 = __p0_194; \
+  uint64x2_t __s1_194 = __p1_194; \
+  uint32x4_t __ret_194; \
+  __ret_194 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_194), (uint32x2_t)(vqshrn_n_u64(__s1_194, __p2_194)))); \
+  __ret_194; \
+})
+#else
+#define vqshrn_high_n_u64(__p0_195, __p1_195, __p2_195) __extension__ ({ \
+  uint32x2_t __s0_195 = __p0_195; \
+  uint64x2_t __s1_195 = __p1_195; \
+  uint32x2_t __rev0_195;  __rev0_195 = __builtin_shufflevector(__s0_195, __s0_195, 1, 0); \
+  uint64x2_t __rev1_195;  __rev1_195 = __builtin_shufflevector(__s1_195, __s1_195, 1, 0); \
+  uint32x4_t __ret_195; \
+  __ret_195 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_195), (uint32x2_t)(__noswap_vqshrn_n_u64(__rev1_195, __p2_195)))); \
+  __ret_195 = __builtin_shufflevector(__ret_195, __ret_195, 3, 2, 1, 0); \
+  __ret_195; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrn_high_n_u16(__p0_196, __p1_196, __p2_196) __extension__ ({ \
+  uint8x8_t __s0_196 = __p0_196; \
+  uint16x8_t __s1_196 = __p1_196; \
+  uint8x16_t __ret_196; \
+  __ret_196 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_196), (uint8x8_t)(vqshrn_n_u16(__s1_196, __p2_196)))); \
+  __ret_196; \
+})
+#else
+#define vqshrn_high_n_u16(__p0_197, __p1_197, __p2_197) __extension__ ({ \
+  uint8x8_t __s0_197 = __p0_197; \
+  uint16x8_t __s1_197 = __p1_197; \
+  uint8x8_t __rev0_197;  __rev0_197 = __builtin_shufflevector(__s0_197, __s0_197, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1_197;  __rev1_197 = __builtin_shufflevector(__s1_197, __s1_197, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret_197; \
+  __ret_197 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_197), (uint8x8_t)(__noswap_vqshrn_n_u16(__rev1_197, __p2_197)))); \
+  __ret_197 = __builtin_shufflevector(__ret_197, __ret_197, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_197; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrn_high_n_s32(__p0_198, __p1_198, __p2_198) __extension__ ({ \
+  int16x4_t __s0_198 = __p0_198; \
+  int32x4_t __s1_198 = __p1_198; \
+  int16x8_t __ret_198; \
+  __ret_198 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_198), (int16x4_t)(vqshrn_n_s32(__s1_198, __p2_198)))); \
+  __ret_198; \
+})
+#else
+#define vqshrn_high_n_s32(__p0_199, __p1_199, __p2_199) __extension__ ({ \
+  int16x4_t __s0_199 = __p0_199; \
+  int32x4_t __s1_199 = __p1_199; \
+  int16x4_t __rev0_199;  __rev0_199 = __builtin_shufflevector(__s0_199, __s0_199, 3, 2, 1, 0); \
+  int32x4_t __rev1_199;  __rev1_199 = __builtin_shufflevector(__s1_199, __s1_199, 3, 2, 1, 0); \
+  int16x8_t __ret_199; \
+  __ret_199 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_199), (int16x4_t)(__noswap_vqshrn_n_s32(__rev1_199, __p2_199)))); \
+  __ret_199 = __builtin_shufflevector(__ret_199, __ret_199, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_199; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrn_high_n_s64(__p0_200, __p1_200, __p2_200) __extension__ ({ \
+  int32x2_t __s0_200 = __p0_200; \
+  int64x2_t __s1_200 = __p1_200; \
+  int32x4_t __ret_200; \
+  __ret_200 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_200), (int32x2_t)(vqshrn_n_s64(__s1_200, __p2_200)))); \
+  __ret_200; \
+})
+#else
+#define vqshrn_high_n_s64(__p0_201, __p1_201, __p2_201) __extension__ ({ \
+  int32x2_t __s0_201 = __p0_201; \
+  int64x2_t __s1_201 = __p1_201; \
+  int32x2_t __rev0_201;  __rev0_201 = __builtin_shufflevector(__s0_201, __s0_201, 1, 0); \
+  int64x2_t __rev1_201;  __rev1_201 = __builtin_shufflevector(__s1_201, __s1_201, 1, 0); \
+  int32x4_t __ret_201; \
+  __ret_201 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_201), (int32x2_t)(__noswap_vqshrn_n_s64(__rev1_201, __p2_201)))); \
+  __ret_201 = __builtin_shufflevector(__ret_201, __ret_201, 3, 2, 1, 0); \
+  __ret_201; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrn_high_n_s16(__p0_202, __p1_202, __p2_202) __extension__ ({ \
+  int8x8_t __s0_202 = __p0_202; \
+  int16x8_t __s1_202 = __p1_202; \
+  int8x16_t __ret_202; \
+  __ret_202 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_202), (int8x8_t)(vqshrn_n_s16(__s1_202, __p2_202)))); \
+  __ret_202; \
+})
+#else
+#define vqshrn_high_n_s16(__p0_203, __p1_203, __p2_203) __extension__ ({ \
+  int8x8_t __s0_203 = __p0_203; \
+  int16x8_t __s1_203 = __p1_203; \
+  int8x8_t __rev0_203;  __rev0_203 = __builtin_shufflevector(__s0_203, __s0_203, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1_203;  __rev1_203 = __builtin_shufflevector(__s1_203, __s1_203, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret_203; \
+  __ret_203 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_203), (int8x8_t)(__noswap_vqshrn_n_s16(__rev1_203, __p2_203)))); \
+  __ret_203 = __builtin_shufflevector(__ret_203, __ret_203, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_203; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrns_n_u32(__p0, __p1) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vqshrns_n_u32(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshrns_n_u32(__p0, __p1) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vqshrns_n_u32(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrnd_n_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vqshrnd_n_u64(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshrnd_n_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vqshrnd_n_u64(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrnh_n_u16(__p0, __p1) __extension__ ({ \
+  uint16_t __s0 = __p0; \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vqshrnh_n_u16(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshrnh_n_u16(__p0, __p1) __extension__ ({ \
+  uint16_t __s0 = __p0; \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vqshrnh_n_u16(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrns_n_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vqshrns_n_s32(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshrns_n_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vqshrns_n_s32(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrnd_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqshrnd_n_s64(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshrnd_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqshrnd_n_s64(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrnh_n_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vqshrnh_n_s16(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshrnh_n_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vqshrnh_n_s16(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrun_high_n_s32(__p0_204, __p1_204, __p2_204) __extension__ ({ \
+  int16x4_t __s0_204 = __p0_204; \
+  int32x4_t __s1_204 = __p1_204; \
+  int16x8_t __ret_204; \
+  __ret_204 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_204), (int16x4_t)(vqshrun_n_s32(__s1_204, __p2_204)))); \
+  __ret_204; \
+})
+#else
+#define vqshrun_high_n_s32(__p0_205, __p1_205, __p2_205) __extension__ ({ \
+  int16x4_t __s0_205 = __p0_205; \
+  int32x4_t __s1_205 = __p1_205; \
+  int16x4_t __rev0_205;  __rev0_205 = __builtin_shufflevector(__s0_205, __s0_205, 3, 2, 1, 0); \
+  int32x4_t __rev1_205;  __rev1_205 = __builtin_shufflevector(__s1_205, __s1_205, 3, 2, 1, 0); \
+  int16x8_t __ret_205; \
+  __ret_205 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_205), (int16x4_t)(__noswap_vqshrun_n_s32(__rev1_205, __p2_205)))); \
+  __ret_205 = __builtin_shufflevector(__ret_205, __ret_205, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_205; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrun_high_n_s64(__p0_206, __p1_206, __p2_206) __extension__ ({ \
+  int32x2_t __s0_206 = __p0_206; \
+  int64x2_t __s1_206 = __p1_206; \
+  int32x4_t __ret_206; \
+  __ret_206 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_206), (int32x2_t)(vqshrun_n_s64(__s1_206, __p2_206)))); \
+  __ret_206; \
+})
+#else
+#define vqshrun_high_n_s64(__p0_207, __p1_207, __p2_207) __extension__ ({ \
+  int32x2_t __s0_207 = __p0_207; \
+  int64x2_t __s1_207 = __p1_207; \
+  int32x2_t __rev0_207;  __rev0_207 = __builtin_shufflevector(__s0_207, __s0_207, 1, 0); \
+  int64x2_t __rev1_207;  __rev1_207 = __builtin_shufflevector(__s1_207, __s1_207, 1, 0); \
+  int32x4_t __ret_207; \
+  __ret_207 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_207), (int32x2_t)(__noswap_vqshrun_n_s64(__rev1_207, __p2_207)))); \
+  __ret_207 = __builtin_shufflevector(__ret_207, __ret_207, 3, 2, 1, 0); \
+  __ret_207; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrun_high_n_s16(__p0_208, __p1_208, __p2_208) __extension__ ({ \
+  int8x8_t __s0_208 = __p0_208; \
+  int16x8_t __s1_208 = __p1_208; \
+  int8x16_t __ret_208; \
+  __ret_208 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_208), (int8x8_t)(vqshrun_n_s16(__s1_208, __p2_208)))); \
+  __ret_208; \
+})
+#else
+#define vqshrun_high_n_s16(__p0_209, __p1_209, __p2_209) __extension__ ({ \
+  int8x8_t __s0_209 = __p0_209; \
+  int16x8_t __s1_209 = __p1_209; \
+  int8x8_t __rev0_209;  __rev0_209 = __builtin_shufflevector(__s0_209, __s0_209, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1_209;  __rev1_209 = __builtin_shufflevector(__s1_209, __s1_209, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret_209; \
+  __ret_209 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_209), (int8x8_t)(__noswap_vqshrun_n_s16(__rev1_209, __p2_209)))); \
+  __ret_209 = __builtin_shufflevector(__ret_209, __ret_209, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_209; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshruns_n_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vqshruns_n_s32(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshruns_n_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vqshruns_n_s32(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrund_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqshrund_n_s64(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshrund_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqshrund_n_s64(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqshrunh_n_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vqshrunh_n_s16(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vqshrunh_n_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vqshrunh_n_s16(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8_t vqsubb_u8(uint8_t __p0, uint8_t __p1) {
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vqsubb_u8(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint8_t vqsubb_u8(uint8_t __p0, uint8_t __p1) {
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vqsubb_u8(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vqsubs_u32(uint32_t __p0, uint32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vqsubs_u32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint32_t vqsubs_u32(uint32_t __p0, uint32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vqsubs_u32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vqsubd_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vqsubd_u64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vqsubd_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vqsubd_u64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16_t vqsubh_u16(uint16_t __p0, uint16_t __p1) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vqsubh_u16(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint16_t vqsubh_u16(uint16_t __p0, uint16_t __p1) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vqsubh_u16(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8_t vqsubb_s8(int8_t __p0, int8_t __p1) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vqsubb_s8(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int8_t vqsubb_s8(int8_t __p0, int8_t __p1) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vqsubb_s8(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vqsubs_s32(int32_t __p0, int32_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqsubs_s32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int32_t vqsubs_s32(int32_t __p0, int32_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqsubs_s32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vqsubd_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqsubd_s64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int64_t vqsubd_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqsubd_s64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16_t vqsubh_s16(int16_t __p0, int16_t __p1) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqsubh_s16(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int16_t vqsubh_s16(int16_t __p0, int16_t __p1) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqsubh_s16(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vqtbl1_p8(poly8x16_t __p0, uint8x8_t __p1) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vqtbl1_v((int8x16_t)__p0, (int8x8_t)__p1, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vqtbl1_p8(poly8x16_t __p0, uint8x8_t __p1) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vqtbl1_v((int8x16_t)__rev0, (int8x8_t)__rev1, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vqtbl1q_p8(poly8x16_t __p0, uint8x16_t __p1) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vqtbl1q_v((int8x16_t)__p0, (int8x16_t)__p1, 36);
+  return __ret;
+}
+#else
+__ai poly8x16_t vqtbl1q_p8(poly8x16_t __p0, uint8x16_t __p1) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vqtbl1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, 36);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vqtbl1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqtbl1q_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vqtbl1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqtbl1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vqtbl1q_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqtbl1q_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vqtbl1q_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqtbl1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vqtbl1_u8(uint8x16_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqtbl1_v((int8x16_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vqtbl1_u8(uint8x16_t __p0, uint8x8_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqtbl1_v((int8x16_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vqtbl1_s8(int8x16_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqtbl1_v((int8x16_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vqtbl1_s8(int8x16_t __p0, int8x8_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqtbl1_v((int8x16_t)__rev0, (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vqtbl2_p8(poly8x16x2_t __p0, uint8x8_t __p1) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vqtbl2_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x8_t)__p1, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vqtbl2_p8(poly8x16x2_t __p0, uint8x8_t __p1) {
+  poly8x16x2_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vqtbl2_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x8_t)__rev1, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vqtbl2q_p8(poly8x16x2_t __p0, uint8x16_t __p1) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vqtbl2q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p1, 36);
+  return __ret;
+}
+#else
+__ai poly8x16_t vqtbl2q_p8(poly8x16x2_t __p0, uint8x16_t __p1) {
+  poly8x16x2_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vqtbl2q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev1, 36);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vqtbl2q_u8(uint8x16x2_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqtbl2q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vqtbl2q_u8(uint8x16x2_t __p0, uint8x16_t __p1) {
+  uint8x16x2_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqtbl2q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vqtbl2q_s8(int8x16x2_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqtbl2q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vqtbl2q_s8(int8x16x2_t __p0, int8x16_t __p1) {
+  int8x16x2_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqtbl2q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev1, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vqtbl2_u8(uint8x16x2_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqtbl2_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vqtbl2_u8(uint8x16x2_t __p0, uint8x8_t __p1) {
+  uint8x16x2_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqtbl2_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vqtbl2_s8(int8x16x2_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqtbl2_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vqtbl2_s8(int8x16x2_t __p0, int8x8_t __p1) {
+  int8x16x2_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqtbl2_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vqtbl3_p8(poly8x16x3_t __p0, uint8x8_t __p1) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vqtbl3_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x8_t)__p1, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vqtbl3_p8(poly8x16x3_t __p0, uint8x8_t __p1) {
+  poly8x16x3_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vqtbl3_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x8_t)__rev1, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vqtbl3q_p8(poly8x16x3_t __p0, uint8x16_t __p1) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vqtbl3q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p1, 36);
+  return __ret;
+}
+#else
+__ai poly8x16_t vqtbl3q_p8(poly8x16x3_t __p0, uint8x16_t __p1) {
+  poly8x16x3_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vqtbl3q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev1, 36);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vqtbl3q_u8(uint8x16x3_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqtbl3q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vqtbl3q_u8(uint8x16x3_t __p0, uint8x16_t __p1) {
+  uint8x16x3_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqtbl3q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vqtbl3q_s8(int8x16x3_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqtbl3q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vqtbl3q_s8(int8x16x3_t __p0, int8x16_t __p1) {
+  int8x16x3_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqtbl3q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev1, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vqtbl3_u8(uint8x16x3_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqtbl3_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vqtbl3_u8(uint8x16x3_t __p0, uint8x8_t __p1) {
+  uint8x16x3_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqtbl3_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vqtbl3_s8(int8x16x3_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqtbl3_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vqtbl3_s8(int8x16x3_t __p0, int8x8_t __p1) {
+  int8x16x3_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqtbl3_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vqtbl4_p8(poly8x16x4_t __p0, uint8x8_t __p1) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vqtbl4_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p0.val[3], (int8x8_t)__p1, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vqtbl4_p8(poly8x16x4_t __p0, uint8x8_t __p1) {
+  poly8x16x4_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vqtbl4_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev0.val[3], (int8x8_t)__rev1, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vqtbl4q_p8(poly8x16x4_t __p0, uint8x16_t __p1) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vqtbl4q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p0.val[3], (int8x16_t)__p1, 36);
+  return __ret;
+}
+#else
+__ai poly8x16_t vqtbl4q_p8(poly8x16x4_t __p0, uint8x16_t __p1) {
+  poly8x16x4_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vqtbl4q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev0.val[3], (int8x16_t)__rev1, 36);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vqtbl4q_u8(uint8x16x4_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqtbl4q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p0.val[3], (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vqtbl4q_u8(uint8x16x4_t __p0, uint8x16_t __p1) {
+  uint8x16x4_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqtbl4q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev0.val[3], (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vqtbl4q_s8(int8x16x4_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqtbl4q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p0.val[3], (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vqtbl4q_s8(int8x16x4_t __p0, int8x16_t __p1) {
+  int8x16x4_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqtbl4q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev0.val[3], (int8x16_t)__rev1, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vqtbl4_u8(uint8x16x4_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqtbl4_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p0.val[3], (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vqtbl4_u8(uint8x16x4_t __p0, uint8x8_t __p1) {
+  uint8x16x4_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqtbl4_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev0.val[3], (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vqtbl4_s8(int8x16x4_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqtbl4_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p0.val[3], (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vqtbl4_s8(int8x16x4_t __p0, int8x8_t __p1) {
+  int8x16x4_t __rev0;
+  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqtbl4_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev0.val[3], (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vqtbx1_p8(poly8x8_t __p0, poly8x16_t __p1, uint8x8_t __p2) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vqtbx1_v((int8x8_t)__p0, (int8x16_t)__p1, (int8x8_t)__p2, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vqtbx1_p8(poly8x8_t __p0, poly8x16_t __p1, uint8x8_t __p2) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vqtbx1_v((int8x8_t)__rev0, (int8x16_t)__rev1, (int8x8_t)__rev2, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vqtbx1q_p8(poly8x16_t __p0, poly8x16_t __p1, uint8x16_t __p2) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vqtbx1q_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 36);
+  return __ret;
+}
+#else
+__ai poly8x16_t vqtbx1q_p8(poly8x16_t __p0, poly8x16_t __p1, uint8x16_t __p2) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vqtbx1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 36);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vqtbx1q_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqtbx1q_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vqtbx1q_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqtbx1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vqtbx1q_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqtbx1q_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vqtbx1q_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqtbx1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vqtbx1_u8(uint8x8_t __p0, uint8x16_t __p1, uint8x8_t __p2) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqtbx1_v((int8x8_t)__p0, (int8x16_t)__p1, (int8x8_t)__p2, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vqtbx1_u8(uint8x8_t __p0, uint8x16_t __p1, uint8x8_t __p2) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqtbx1_v((int8x8_t)__rev0, (int8x16_t)__rev1, (int8x8_t)__rev2, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vqtbx1_s8(int8x8_t __p0, int8x16_t __p1, int8x8_t __p2) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqtbx1_v((int8x8_t)__p0, (int8x16_t)__p1, (int8x8_t)__p2, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vqtbx1_s8(int8x8_t __p0, int8x16_t __p1, int8x8_t __p2) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqtbx1_v((int8x8_t)__rev0, (int8x16_t)__rev1, (int8x8_t)__rev2, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vqtbx2_p8(poly8x8_t __p0, poly8x16x2_t __p1, uint8x8_t __p2) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vqtbx2_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x8_t)__p2, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vqtbx2_p8(poly8x8_t __p0, poly8x16x2_t __p1, uint8x8_t __p2) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16x2_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vqtbx2_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x8_t)__rev2, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vqtbx2q_p8(poly8x16_t __p0, poly8x16x2_t __p1, uint8x16_t __p2) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vqtbx2q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p2, 36);
+  return __ret;
+}
+#else
+__ai poly8x16_t vqtbx2q_p8(poly8x16_t __p0, poly8x16x2_t __p1, uint8x16_t __p2) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16x2_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vqtbx2q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev2, 36);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vqtbx2q_u8(uint8x16_t __p0, uint8x16x2_t __p1, uint8x16_t __p2) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqtbx2q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p2, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vqtbx2q_u8(uint8x16_t __p0, uint8x16x2_t __p1, uint8x16_t __p2) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16x2_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqtbx2q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev2, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vqtbx2q_s8(int8x16_t __p0, int8x16x2_t __p1, int8x16_t __p2) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqtbx2q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p2, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vqtbx2q_s8(int8x16_t __p0, int8x16x2_t __p1, int8x16_t __p2) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16x2_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqtbx2q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev2, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vqtbx2_u8(uint8x8_t __p0, uint8x16x2_t __p1, uint8x8_t __p2) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqtbx2_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x8_t)__p2, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vqtbx2_u8(uint8x8_t __p0, uint8x16x2_t __p1, uint8x8_t __p2) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16x2_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqtbx2_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x8_t)__rev2, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vqtbx2_s8(int8x8_t __p0, int8x16x2_t __p1, int8x8_t __p2) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqtbx2_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x8_t)__p2, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vqtbx2_s8(int8x8_t __p0, int8x16x2_t __p1, int8x8_t __p2) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16x2_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqtbx2_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x8_t)__rev2, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vqtbx3_p8(poly8x8_t __p0, poly8x16x3_t __p1, uint8x8_t __p2) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vqtbx3_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x8_t)__p2, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vqtbx3_p8(poly8x8_t __p0, poly8x16x3_t __p1, uint8x8_t __p2) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16x3_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vqtbx3_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x8_t)__rev2, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vqtbx3q_p8(poly8x16_t __p0, poly8x16x3_t __p1, uint8x16_t __p2) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vqtbx3q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p2, 36);
+  return __ret;
+}
+#else
+__ai poly8x16_t vqtbx3q_p8(poly8x16_t __p0, poly8x16x3_t __p1, uint8x16_t __p2) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16x3_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vqtbx3q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev2, 36);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vqtbx3q_u8(uint8x16_t __p0, uint8x16x3_t __p1, uint8x16_t __p2) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqtbx3q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p2, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vqtbx3q_u8(uint8x16_t __p0, uint8x16x3_t __p1, uint8x16_t __p2) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16x3_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqtbx3q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev2, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vqtbx3q_s8(int8x16_t __p0, int8x16x3_t __p1, int8x16_t __p2) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqtbx3q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p2, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vqtbx3q_s8(int8x16_t __p0, int8x16x3_t __p1, int8x16_t __p2) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16x3_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqtbx3q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev2, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vqtbx3_u8(uint8x8_t __p0, uint8x16x3_t __p1, uint8x8_t __p2) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqtbx3_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x8_t)__p2, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vqtbx3_u8(uint8x8_t __p0, uint8x16x3_t __p1, uint8x8_t __p2) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16x3_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqtbx3_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x8_t)__rev2, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vqtbx3_s8(int8x8_t __p0, int8x16x3_t __p1, int8x8_t __p2) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqtbx3_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x8_t)__p2, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vqtbx3_s8(int8x8_t __p0, int8x16x3_t __p1, int8x8_t __p2) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16x3_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqtbx3_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x8_t)__rev2, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vqtbx4_p8(poly8x8_t __p0, poly8x16x4_t __p1, uint8x8_t __p2) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vqtbx4_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p1.val[3], (int8x8_t)__p2, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vqtbx4_p8(poly8x8_t __p0, poly8x16x4_t __p1, uint8x8_t __p2) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16x4_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vqtbx4_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], (int8x8_t)__rev2, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vqtbx4q_p8(poly8x16_t __p0, poly8x16x4_t __p1, uint8x16_t __p2) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vqtbx4q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p1.val[3], (int8x16_t)__p2, 36);
+  return __ret;
+}
+#else
+__ai poly8x16_t vqtbx4q_p8(poly8x16_t __p0, poly8x16x4_t __p1, uint8x16_t __p2) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16x4_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vqtbx4q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], (int8x16_t)__rev2, 36);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vqtbx4q_u8(uint8x16_t __p0, uint8x16x4_t __p1, uint8x16_t __p2) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqtbx4q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p1.val[3], (int8x16_t)__p2, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vqtbx4q_u8(uint8x16_t __p0, uint8x16x4_t __p1, uint8x16_t __p2) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16x4_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vqtbx4q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], (int8x16_t)__rev2, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vqtbx4q_s8(int8x16_t __p0, int8x16x4_t __p1, int8x16_t __p2) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqtbx4q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p1.val[3], (int8x16_t)__p2, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vqtbx4q_s8(int8x16_t __p0, int8x16x4_t __p1, int8x16_t __p2) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16x4_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vqtbx4q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], (int8x16_t)__rev2, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vqtbx4_u8(uint8x8_t __p0, uint8x16x4_t __p1, uint8x8_t __p2) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqtbx4_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p1.val[3], (int8x8_t)__p2, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vqtbx4_u8(uint8x8_t __p0, uint8x16x4_t __p1, uint8x8_t __p2) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16x4_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vqtbx4_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], (int8x8_t)__rev2, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vqtbx4_s8(int8x8_t __p0, int8x16x4_t __p1, int8x8_t __p2) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqtbx4_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p1.val[3], (int8x8_t)__p2, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vqtbx4_s8(int8x8_t __p0, int8x16x4_t __p1, int8x8_t __p2) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16x4_t __rev1;
+  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vqtbx4_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], (int8x8_t)__rev2, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vraddhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint16x8_t __ret;
+  __ret = vcombine_u16(__p0, vraddhn_u32(__p1, __p2));
+  return __ret;
+}
+#else
+__ai uint16x8_t vraddhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __noswap_vcombine_u16(__rev0, __noswap_vraddhn_u32(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vraddhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
+  uint32x4_t __ret;
+  __ret = vcombine_u32(__p0, vraddhn_u64(__p1, __p2));
+  return __ret;
+}
+#else
+__ai uint32x4_t vraddhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __noswap_vcombine_u32(__rev0, __noswap_vraddhn_u64(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vraddhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint8x16_t __ret;
+  __ret = vcombine_u8(__p0, vraddhn_u16(__p1, __p2));
+  return __ret;
+}
+#else
+__ai uint8x16_t vraddhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __noswap_vcombine_u8(__rev0, __noswap_vraddhn_u16(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vraddhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int16x8_t __ret;
+  __ret = vcombine_s16(__p0, vraddhn_s32(__p1, __p2));
+  return __ret;
+}
+#else
+__ai int16x8_t vraddhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __noswap_vcombine_s16(__rev0, __noswap_vraddhn_s32(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vraddhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
+  int32x4_t __ret;
+  __ret = vcombine_s32(__p0, vraddhn_s64(__p1, __p2));
+  return __ret;
+}
+#else
+__ai int32x4_t vraddhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vcombine_s32(__rev0, __noswap_vraddhn_s64(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vraddhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int8x16_t __ret;
+  __ret = vcombine_s8(__p0, vraddhn_s16(__p1, __p2));
+  return __ret;
+}
+#else
+__ai int8x16_t vraddhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __noswap_vcombine_s8(__rev0, __noswap_vraddhn_s16(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vrbit_p8(poly8x8_t __p0) {
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vrbit_v((int8x8_t)__p0, 4);
+  return __ret;
+}
+#else
+__ai poly8x8_t vrbit_p8(poly8x8_t __p0) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = (poly8x8_t) __builtin_neon_vrbit_v((int8x8_t)__rev0, 4);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vrbitq_p8(poly8x16_t __p0) {
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vrbitq_v((int8x16_t)__p0, 36);
+  return __ret;
+}
+#else
+__ai poly8x16_t vrbitq_p8(poly8x16_t __p0) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = (poly8x16_t) __builtin_neon_vrbitq_v((int8x16_t)__rev0, 36);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vrbitq_u8(uint8x16_t __p0) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vrbitq_v((int8x16_t)__p0, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vrbitq_u8(uint8x16_t __p0) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vrbitq_v((int8x16_t)__rev0, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vrbitq_s8(int8x16_t __p0) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vrbitq_v((int8x16_t)__p0, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vrbitq_s8(int8x16_t __p0) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vrbitq_v((int8x16_t)__rev0, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vrbit_u8(uint8x8_t __p0) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vrbit_v((int8x8_t)__p0, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vrbit_u8(uint8x8_t __p0) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vrbit_v((int8x8_t)__rev0, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vrbit_s8(int8x8_t __p0) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vrbit_v((int8x8_t)__p0, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vrbit_s8(int8x8_t __p0) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vrbit_v((int8x8_t)__rev0, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vrecpeq_f64(float64x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrecpeq_v((int8x16_t)__p0, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vrecpeq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrecpeq_v((int8x16_t)__rev0, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vrecpe_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrecpe_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vrecpe_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrecpe_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64_t vrecped_f64(float64_t __p0) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vrecped_f64(__p0);
+  return __ret;
+}
+#else
+__ai float64_t vrecped_f64(float64_t __p0) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vrecped_f64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vrecpes_f32(float32_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vrecpes_f32(__p0);
+  return __ret;
+}
+#else
+__ai float32_t vrecpes_f32(float32_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vrecpes_f32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vrecpsq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrecpsq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vrecpsq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrecpsq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vrecps_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrecps_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vrecps_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrecps_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64_t vrecpsd_f64(float64_t __p0, float64_t __p1) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vrecpsd_f64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai float64_t vrecpsd_f64(float64_t __p0, float64_t __p1) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vrecpsd_f64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vrecpss_f32(float32_t __p0, float32_t __p1) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vrecpss_f32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai float32_t vrecpss_f32(float32_t __p0, float32_t __p1) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vrecpss_f32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64_t vrecpxd_f64(float64_t __p0) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vrecpxd_f64(__p0);
+  return __ret;
+}
+#else
+__ai float64_t vrecpxd_f64(float64_t __p0) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vrecpxd_f64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vrecpxs_f32(float32_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vrecpxs_f32(__p0);
+  return __ret;
+}
+#else
+__ai float32_t vrecpxs_f32(float32_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vrecpxs_f32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vrshld_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vrshld_u64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vrshld_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vrshld_u64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vrshld_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vrshld_s64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int64_t vrshld_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vrshld_s64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrd_n_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vrshrd_n_u64(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vrshrd_n_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vrshrd_n_u64(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrd_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vrshrd_n_s64(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vrshrd_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vrshrd_n_s64(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrn_high_n_u32(__p0_210, __p1_210, __p2_210) __extension__ ({ \
+  uint16x4_t __s0_210 = __p0_210; \
+  uint32x4_t __s1_210 = __p1_210; \
+  uint16x8_t __ret_210; \
+  __ret_210 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_210), (uint16x4_t)(vrshrn_n_u32(__s1_210, __p2_210)))); \
+  __ret_210; \
+})
+#else
+#define vrshrn_high_n_u32(__p0_211, __p1_211, __p2_211) __extension__ ({ \
+  uint16x4_t __s0_211 = __p0_211; \
+  uint32x4_t __s1_211 = __p1_211; \
+  uint16x4_t __rev0_211;  __rev0_211 = __builtin_shufflevector(__s0_211, __s0_211, 3, 2, 1, 0); \
+  uint32x4_t __rev1_211;  __rev1_211 = __builtin_shufflevector(__s1_211, __s1_211, 3, 2, 1, 0); \
+  uint16x8_t __ret_211; \
+  __ret_211 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_211), (uint16x4_t)(__noswap_vrshrn_n_u32(__rev1_211, __p2_211)))); \
+  __ret_211 = __builtin_shufflevector(__ret_211, __ret_211, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_211; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrn_high_n_u64(__p0_212, __p1_212, __p2_212) __extension__ ({ \
+  uint32x2_t __s0_212 = __p0_212; \
+  uint64x2_t __s1_212 = __p1_212; \
+  uint32x4_t __ret_212; \
+  __ret_212 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_212), (uint32x2_t)(vrshrn_n_u64(__s1_212, __p2_212)))); \
+  __ret_212; \
+})
+#else
+#define vrshrn_high_n_u64(__p0_213, __p1_213, __p2_213) __extension__ ({ \
+  uint32x2_t __s0_213 = __p0_213; \
+  uint64x2_t __s1_213 = __p1_213; \
+  uint32x2_t __rev0_213;  __rev0_213 = __builtin_shufflevector(__s0_213, __s0_213, 1, 0); \
+  uint64x2_t __rev1_213;  __rev1_213 = __builtin_shufflevector(__s1_213, __s1_213, 1, 0); \
+  uint32x4_t __ret_213; \
+  __ret_213 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_213), (uint32x2_t)(__noswap_vrshrn_n_u64(__rev1_213, __p2_213)))); \
+  __ret_213 = __builtin_shufflevector(__ret_213, __ret_213, 3, 2, 1, 0); \
+  __ret_213; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrn_high_n_u16(__p0_214, __p1_214, __p2_214) __extension__ ({ \
+  uint8x8_t __s0_214 = __p0_214; \
+  uint16x8_t __s1_214 = __p1_214; \
+  uint8x16_t __ret_214; \
+  __ret_214 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_214), (uint8x8_t)(vrshrn_n_u16(__s1_214, __p2_214)))); \
+  __ret_214; \
+})
+#else
+#define vrshrn_high_n_u16(__p0_215, __p1_215, __p2_215) __extension__ ({ \
+  uint8x8_t __s0_215 = __p0_215; \
+  uint16x8_t __s1_215 = __p1_215; \
+  uint8x8_t __rev0_215;  __rev0_215 = __builtin_shufflevector(__s0_215, __s0_215, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1_215;  __rev1_215 = __builtin_shufflevector(__s1_215, __s1_215, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret_215; \
+  __ret_215 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_215), (uint8x8_t)(__noswap_vrshrn_n_u16(__rev1_215, __p2_215)))); \
+  __ret_215 = __builtin_shufflevector(__ret_215, __ret_215, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_215; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrn_high_n_s32(__p0_216, __p1_216, __p2_216) __extension__ ({ \
+  int16x4_t __s0_216 = __p0_216; \
+  int32x4_t __s1_216 = __p1_216; \
+  int16x8_t __ret_216; \
+  __ret_216 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_216), (int16x4_t)(vrshrn_n_s32(__s1_216, __p2_216)))); \
+  __ret_216; \
+})
+#else
+#define vrshrn_high_n_s32(__p0_217, __p1_217, __p2_217) __extension__ ({ \
+  int16x4_t __s0_217 = __p0_217; \
+  int32x4_t __s1_217 = __p1_217; \
+  int16x4_t __rev0_217;  __rev0_217 = __builtin_shufflevector(__s0_217, __s0_217, 3, 2, 1, 0); \
+  int32x4_t __rev1_217;  __rev1_217 = __builtin_shufflevector(__s1_217, __s1_217, 3, 2, 1, 0); \
+  int16x8_t __ret_217; \
+  __ret_217 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_217), (int16x4_t)(__noswap_vrshrn_n_s32(__rev1_217, __p2_217)))); \
+  __ret_217 = __builtin_shufflevector(__ret_217, __ret_217, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_217; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrn_high_n_s64(__p0_218, __p1_218, __p2_218) __extension__ ({ \
+  int32x2_t __s0_218 = __p0_218; \
+  int64x2_t __s1_218 = __p1_218; \
+  int32x4_t __ret_218; \
+  __ret_218 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_218), (int32x2_t)(vrshrn_n_s64(__s1_218, __p2_218)))); \
+  __ret_218; \
+})
+#else
+#define vrshrn_high_n_s64(__p0_219, __p1_219, __p2_219) __extension__ ({ \
+  int32x2_t __s0_219 = __p0_219; \
+  int64x2_t __s1_219 = __p1_219; \
+  int32x2_t __rev0_219;  __rev0_219 = __builtin_shufflevector(__s0_219, __s0_219, 1, 0); \
+  int64x2_t __rev1_219;  __rev1_219 = __builtin_shufflevector(__s1_219, __s1_219, 1, 0); \
+  int32x4_t __ret_219; \
+  __ret_219 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_219), (int32x2_t)(__noswap_vrshrn_n_s64(__rev1_219, __p2_219)))); \
+  __ret_219 = __builtin_shufflevector(__ret_219, __ret_219, 3, 2, 1, 0); \
+  __ret_219; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrn_high_n_s16(__p0_220, __p1_220, __p2_220) __extension__ ({ \
+  int8x8_t __s0_220 = __p0_220; \
+  int16x8_t __s1_220 = __p1_220; \
+  int8x16_t __ret_220; \
+  __ret_220 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_220), (int8x8_t)(vrshrn_n_s16(__s1_220, __p2_220)))); \
+  __ret_220; \
+})
+#else
+#define vrshrn_high_n_s16(__p0_221, __p1_221, __p2_221) __extension__ ({ \
+  int8x8_t __s0_221 = __p0_221; \
+  int16x8_t __s1_221 = __p1_221; \
+  int8x8_t __rev0_221;  __rev0_221 = __builtin_shufflevector(__s0_221, __s0_221, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1_221;  __rev1_221 = __builtin_shufflevector(__s1_221, __s1_221, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret_221; \
+  __ret_221 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_221), (int8x8_t)(__noswap_vrshrn_n_s16(__rev1_221, __p2_221)))); \
+  __ret_221 = __builtin_shufflevector(__ret_221, __ret_221, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_221; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vrsqrteq_f64(float64x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrsqrteq_v((int8x16_t)__p0, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vrsqrteq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrsqrteq_v((int8x16_t)__rev0, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vrsqrte_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrsqrte_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vrsqrte_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrsqrte_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64_t vrsqrted_f64(float64_t __p0) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vrsqrted_f64(__p0);
+  return __ret;
+}
+#else
+__ai float64_t vrsqrted_f64(float64_t __p0) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vrsqrted_f64(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vrsqrtes_f32(float32_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vrsqrtes_f32(__p0);
+  return __ret;
+}
+#else
+__ai float32_t vrsqrtes_f32(float32_t __p0) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vrsqrtes_f32(__p0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vrsqrtsq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrsqrtsq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vrsqrtsq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vrsqrtsq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vrsqrts_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrsqrts_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vrsqrts_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vrsqrts_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64_t vrsqrtsd_f64(float64_t __p0, float64_t __p1) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vrsqrtsd_f64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai float64_t vrsqrtsd_f64(float64_t __p0, float64_t __p1) {
+  float64_t __ret;
+  __ret = (float64_t) __builtin_neon_vrsqrtsd_f64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32_t vrsqrtss_f32(float32_t __p0, float32_t __p1) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vrsqrtss_f32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai float32_t vrsqrtss_f32(float32_t __p0, float32_t __p1) {
+  float32_t __ret;
+  __ret = (float32_t) __builtin_neon_vrsqrtss_f32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrsrad_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64_t __s1 = __p1; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vrsrad_n_u64(__s0, __s1, __p2); \
+  __ret; \
+})
+#else
+#define vrsrad_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64_t __s1 = __p1; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vrsrad_n_u64(__s0, __s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrsrad_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __s1 = __p1; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vrsrad_n_s64(__s0, __s1, __p2); \
+  __ret; \
+})
+#else
+#define vrsrad_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __s1 = __p1; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vrsrad_n_s64(__s0, __s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vrsubhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint16x8_t __ret;
+  __ret = vcombine_u16(__p0, vrsubhn_u32(__p1, __p2));
+  return __ret;
+}
+#else
+__ai uint16x8_t vrsubhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __noswap_vcombine_u16(__rev0, __noswap_vrsubhn_u32(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vrsubhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
+  uint32x4_t __ret;
+  __ret = vcombine_u32(__p0, vrsubhn_u64(__p1, __p2));
+  return __ret;
+}
+#else
+__ai uint32x4_t vrsubhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __noswap_vcombine_u32(__rev0, __noswap_vrsubhn_u64(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vrsubhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint8x16_t __ret;
+  __ret = vcombine_u8(__p0, vrsubhn_u16(__p1, __p2));
+  return __ret;
+}
+#else
+__ai uint8x16_t vrsubhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __noswap_vcombine_u8(__rev0, __noswap_vrsubhn_u16(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vrsubhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int16x8_t __ret;
+  __ret = vcombine_s16(__p0, vrsubhn_s32(__p1, __p2));
+  return __ret;
+}
+#else
+__ai int16x8_t vrsubhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __noswap_vcombine_s16(__rev0, __noswap_vrsubhn_s32(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vrsubhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
+  int32x4_t __ret;
+  __ret = vcombine_s32(__p0, vrsubhn_s64(__p1, __p2));
+  return __ret;
+}
+#else
+__ai int32x4_t vrsubhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vcombine_s32(__rev0, __noswap_vrsubhn_s64(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vrsubhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int8x16_t __ret;
+  __ret = vcombine_s8(__p0, vrsubhn_s16(__p1, __p2));
+  return __ret;
+}
+#else
+__ai int8x16_t vrsubhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __noswap_vcombine_s8(__rev0, __noswap_vrsubhn_s16(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vset_lane_f16(__p0_222, __p1_222, __p2_222) __extension__ ({ \
+  float16_t __s0_222 = __p0_222; \
+  float16x4_t __s1_222 = __p1_222; \
+  float16x4_t __ret_222; \
+float16_t __reint_222 = __s0_222; \
+float16x4_t __reint1_222 = __s1_222; \
+int16x4_t __reint2_222 = vset_lane_s16(*(int16_t *) &__reint_222, *(int16x4_t *) &__reint1_222, __p2_222); \
+  __ret_222 = *(float16x4_t *) &__reint2_222; \
+  __ret_222; \
+})
+#else
+#define vset_lane_f16(__p0_223, __p1_223, __p2_223) __extension__ ({ \
+  float16_t __s0_223 = __p0_223; \
+  float16x4_t __s1_223 = __p1_223; \
+  float16x4_t __rev1_223;  __rev1_223 = __builtin_shufflevector(__s1_223, __s1_223, 3, 2, 1, 0); \
+  float16x4_t __ret_223; \
+float16_t __reint_223 = __s0_223; \
+float16x4_t __reint1_223 = __rev1_223; \
+int16x4_t __reint2_223 = __noswap_vset_lane_s16(*(int16_t *) &__reint_223, *(int16x4_t *) &__reint1_223, __p2_223); \
+  __ret_223 = *(float16x4_t *) &__reint2_223; \
+  __ret_223 = __builtin_shufflevector(__ret_223, __ret_223, 3, 2, 1, 0); \
+  __ret_223; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsetq_lane_f16(__p0_224, __p1_224, __p2_224) __extension__ ({ \
+  float16_t __s0_224 = __p0_224; \
+  float16x8_t __s1_224 = __p1_224; \
+  float16x8_t __ret_224; \
+float16_t __reint_224 = __s0_224; \
+float16x8_t __reint1_224 = __s1_224; \
+int16x8_t __reint2_224 = vsetq_lane_s16(*(int16_t *) &__reint_224, *(int16x8_t *) &__reint1_224, __p2_224); \
+  __ret_224 = *(float16x8_t *) &__reint2_224; \
+  __ret_224; \
+})
+#else
+#define vsetq_lane_f16(__p0_225, __p1_225, __p2_225) __extension__ ({ \
+  float16_t __s0_225 = __p0_225; \
+  float16x8_t __s1_225 = __p1_225; \
+  float16x8_t __rev1_225;  __rev1_225 = __builtin_shufflevector(__s1_225, __s1_225, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x8_t __ret_225; \
+float16_t __reint_225 = __s0_225; \
+float16x8_t __reint1_225 = __rev1_225; \
+int16x8_t __reint2_225 = __noswap_vsetq_lane_s16(*(int16_t *) &__reint_225, *(int16x8_t *) &__reint1_225, __p2_225); \
+  __ret_225 = *(float16x8_t *) &__reint2_225; \
+  __ret_225 = __builtin_shufflevector(__ret_225, __ret_225, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_225; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vset_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64_t __s0 = __p0; \
+  poly64x1_t __s1 = __p1; \
+  poly64x1_t __ret; \
+  __ret = (poly64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vset_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64_t __s0 = __p0; \
+  poly64x1_t __s1 = __p1; \
+  poly64x1_t __ret; \
+  __ret = (poly64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#define __noswap_vset_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64_t __s0 = __p0; \
+  poly64x1_t __s1 = __p1; \
+  poly64x1_t __ret; \
+  __ret = (poly64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsetq_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64_t __s0 = __p0; \
+  poly64x2_t __s1 = __p1; \
+  poly64x2_t __ret; \
+  __ret = (poly64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vsetq_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64_t __s0 = __p0; \
+  poly64x2_t __s1 = __p1; \
+  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  poly64x2_t __ret; \
+  __ret = (poly64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int8x16_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vsetq_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64_t __s0 = __p0; \
+  poly64x2_t __s1 = __p1; \
+  poly64x2_t __ret; \
+  __ret = (poly64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsetq_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64_t __s0 = __p0; \
+  float64x2_t __s1 = __p1; \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vsetq_lane_f64(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vsetq_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64_t __s0 = __p0; \
+  float64x2_t __s1 = __p1; \
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vsetq_lane_f64(__s0, (int8x16_t)__rev1, __p2); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#define __noswap_vsetq_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64_t __s0 = __p0; \
+  float64x2_t __s1 = __p1; \
+  float64x2_t __ret; \
+  __ret = (float64x2_t) __builtin_neon_vsetq_lane_f64(__s0, (int8x16_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vset_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64_t __s0 = __p0; \
+  float64x1_t __s1 = __p1; \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vset_lane_f64(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#else
+#define vset_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64_t __s0 = __p0; \
+  float64x1_t __s1 = __p1; \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vset_lane_f64(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#define __noswap_vset_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64_t __s0 = __p0; \
+  float64x1_t __s1 = __p1; \
+  float64x1_t __ret; \
+  __ret = (float64x1_t) __builtin_neon_vset_lane_f64(__s0, (int8x8_t)__s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vshld_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vshld_u64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vshld_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vshld_u64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vshld_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vshld_s64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int64_t vshld_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vshld_s64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshld_n_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vshld_n_u64(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vshld_n_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vshld_n_u64(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshld_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vshld_n_s64(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vshld_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vshld_n_s64(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshll_high_n_u8(__p0_226, __p1_226) __extension__ ({ \
+  uint8x16_t __s0_226 = __p0_226; \
+  uint16x8_t __ret_226; \
+  __ret_226 = (uint16x8_t)(vshll_n_u8(vget_high_u8(__s0_226), __p1_226)); \
+  __ret_226; \
+})
+#else
+#define vshll_high_n_u8(__p0_227, __p1_227) __extension__ ({ \
+  uint8x16_t __s0_227 = __p0_227; \
+  uint8x16_t __rev0_227;  __rev0_227 = __builtin_shufflevector(__s0_227, __s0_227, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret_227; \
+  __ret_227 = (uint16x8_t)(__noswap_vshll_n_u8(__noswap_vget_high_u8(__rev0_227), __p1_227)); \
+  __ret_227 = __builtin_shufflevector(__ret_227, __ret_227, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_227; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshll_high_n_u32(__p0_228, __p1_228) __extension__ ({ \
+  uint32x4_t __s0_228 = __p0_228; \
+  uint64x2_t __ret_228; \
+  __ret_228 = (uint64x2_t)(vshll_n_u32(vget_high_u32(__s0_228), __p1_228)); \
+  __ret_228; \
+})
+#else
+#define vshll_high_n_u32(__p0_229, __p1_229) __extension__ ({ \
+  uint32x4_t __s0_229 = __p0_229; \
+  uint32x4_t __rev0_229;  __rev0_229 = __builtin_shufflevector(__s0_229, __s0_229, 3, 2, 1, 0); \
+  uint64x2_t __ret_229; \
+  __ret_229 = (uint64x2_t)(__noswap_vshll_n_u32(__noswap_vget_high_u32(__rev0_229), __p1_229)); \
+  __ret_229 = __builtin_shufflevector(__ret_229, __ret_229, 1, 0); \
+  __ret_229; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshll_high_n_u16(__p0_230, __p1_230) __extension__ ({ \
+  uint16x8_t __s0_230 = __p0_230; \
+  uint32x4_t __ret_230; \
+  __ret_230 = (uint32x4_t)(vshll_n_u16(vget_high_u16(__s0_230), __p1_230)); \
+  __ret_230; \
+})
+#else
+#define vshll_high_n_u16(__p0_231, __p1_231) __extension__ ({ \
+  uint16x8_t __s0_231 = __p0_231; \
+  uint16x8_t __rev0_231;  __rev0_231 = __builtin_shufflevector(__s0_231, __s0_231, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint32x4_t __ret_231; \
+  __ret_231 = (uint32x4_t)(__noswap_vshll_n_u16(__noswap_vget_high_u16(__rev0_231), __p1_231)); \
+  __ret_231 = __builtin_shufflevector(__ret_231, __ret_231, 3, 2, 1, 0); \
+  __ret_231; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshll_high_n_s8(__p0_232, __p1_232) __extension__ ({ \
+  int8x16_t __s0_232 = __p0_232; \
+  int16x8_t __ret_232; \
+  __ret_232 = (int16x8_t)(vshll_n_s8(vget_high_s8(__s0_232), __p1_232)); \
+  __ret_232; \
+})
+#else
+#define vshll_high_n_s8(__p0_233, __p1_233) __extension__ ({ \
+  int8x16_t __s0_233 = __p0_233; \
+  int8x16_t __rev0_233;  __rev0_233 = __builtin_shufflevector(__s0_233, __s0_233, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret_233; \
+  __ret_233 = (int16x8_t)(__noswap_vshll_n_s8(__noswap_vget_high_s8(__rev0_233), __p1_233)); \
+  __ret_233 = __builtin_shufflevector(__ret_233, __ret_233, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_233; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshll_high_n_s32(__p0_234, __p1_234) __extension__ ({ \
+  int32x4_t __s0_234 = __p0_234; \
+  int64x2_t __ret_234; \
+  __ret_234 = (int64x2_t)(vshll_n_s32(vget_high_s32(__s0_234), __p1_234)); \
+  __ret_234; \
+})
+#else
+#define vshll_high_n_s32(__p0_235, __p1_235) __extension__ ({ \
+  int32x4_t __s0_235 = __p0_235; \
+  int32x4_t __rev0_235;  __rev0_235 = __builtin_shufflevector(__s0_235, __s0_235, 3, 2, 1, 0); \
+  int64x2_t __ret_235; \
+  __ret_235 = (int64x2_t)(__noswap_vshll_n_s32(__noswap_vget_high_s32(__rev0_235), __p1_235)); \
+  __ret_235 = __builtin_shufflevector(__ret_235, __ret_235, 1, 0); \
+  __ret_235; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshll_high_n_s16(__p0_236, __p1_236) __extension__ ({ \
+  int16x8_t __s0_236 = __p0_236; \
+  int32x4_t __ret_236; \
+  __ret_236 = (int32x4_t)(vshll_n_s16(vget_high_s16(__s0_236), __p1_236)); \
+  __ret_236; \
+})
+#else
+#define vshll_high_n_s16(__p0_237, __p1_237) __extension__ ({ \
+  int16x8_t __s0_237 = __p0_237; \
+  int16x8_t __rev0_237;  __rev0_237 = __builtin_shufflevector(__s0_237, __s0_237, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int32x4_t __ret_237; \
+  __ret_237 = (int32x4_t)(__noswap_vshll_n_s16(__noswap_vget_high_s16(__rev0_237), __p1_237)); \
+  __ret_237 = __builtin_shufflevector(__ret_237, __ret_237, 3, 2, 1, 0); \
+  __ret_237; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrd_n_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vshrd_n_u64(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vshrd_n_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vshrd_n_u64(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrd_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vshrd_n_s64(__s0, __p1); \
+  __ret; \
+})
+#else
+#define vshrd_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vshrd_n_s64(__s0, __p1); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrn_high_n_u32(__p0_238, __p1_238, __p2_238) __extension__ ({ \
+  uint16x4_t __s0_238 = __p0_238; \
+  uint32x4_t __s1_238 = __p1_238; \
+  uint16x8_t __ret_238; \
+  __ret_238 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_238), (uint16x4_t)(vshrn_n_u32(__s1_238, __p2_238)))); \
+  __ret_238; \
+})
+#else
+#define vshrn_high_n_u32(__p0_239, __p1_239, __p2_239) __extension__ ({ \
+  uint16x4_t __s0_239 = __p0_239; \
+  uint32x4_t __s1_239 = __p1_239; \
+  uint16x4_t __rev0_239;  __rev0_239 = __builtin_shufflevector(__s0_239, __s0_239, 3, 2, 1, 0); \
+  uint32x4_t __rev1_239;  __rev1_239 = __builtin_shufflevector(__s1_239, __s1_239, 3, 2, 1, 0); \
+  uint16x8_t __ret_239; \
+  __ret_239 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_239), (uint16x4_t)(__noswap_vshrn_n_u32(__rev1_239, __p2_239)))); \
+  __ret_239 = __builtin_shufflevector(__ret_239, __ret_239, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_239; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrn_high_n_u64(__p0_240, __p1_240, __p2_240) __extension__ ({ \
+  uint32x2_t __s0_240 = __p0_240; \
+  uint64x2_t __s1_240 = __p1_240; \
+  uint32x4_t __ret_240; \
+  __ret_240 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_240), (uint32x2_t)(vshrn_n_u64(__s1_240, __p2_240)))); \
+  __ret_240; \
+})
+#else
+#define vshrn_high_n_u64(__p0_241, __p1_241, __p2_241) __extension__ ({ \
+  uint32x2_t __s0_241 = __p0_241; \
+  uint64x2_t __s1_241 = __p1_241; \
+  uint32x2_t __rev0_241;  __rev0_241 = __builtin_shufflevector(__s0_241, __s0_241, 1, 0); \
+  uint64x2_t __rev1_241;  __rev1_241 = __builtin_shufflevector(__s1_241, __s1_241, 1, 0); \
+  uint32x4_t __ret_241; \
+  __ret_241 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_241), (uint32x2_t)(__noswap_vshrn_n_u64(__rev1_241, __p2_241)))); \
+  __ret_241 = __builtin_shufflevector(__ret_241, __ret_241, 3, 2, 1, 0); \
+  __ret_241; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrn_high_n_u16(__p0_242, __p1_242, __p2_242) __extension__ ({ \
+  uint8x8_t __s0_242 = __p0_242; \
+  uint16x8_t __s1_242 = __p1_242; \
+  uint8x16_t __ret_242; \
+  __ret_242 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_242), (uint8x8_t)(vshrn_n_u16(__s1_242, __p2_242)))); \
+  __ret_242; \
+})
+#else
+#define vshrn_high_n_u16(__p0_243, __p1_243, __p2_243) __extension__ ({ \
+  uint8x8_t __s0_243 = __p0_243; \
+  uint16x8_t __s1_243 = __p1_243; \
+  uint8x8_t __rev0_243;  __rev0_243 = __builtin_shufflevector(__s0_243, __s0_243, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1_243;  __rev1_243 = __builtin_shufflevector(__s1_243, __s1_243, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret_243; \
+  __ret_243 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_243), (uint8x8_t)(__noswap_vshrn_n_u16(__rev1_243, __p2_243)))); \
+  __ret_243 = __builtin_shufflevector(__ret_243, __ret_243, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_243; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrn_high_n_s32(__p0_244, __p1_244, __p2_244) __extension__ ({ \
+  int16x4_t __s0_244 = __p0_244; \
+  int32x4_t __s1_244 = __p1_244; \
+  int16x8_t __ret_244; \
+  __ret_244 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_244), (int16x4_t)(vshrn_n_s32(__s1_244, __p2_244)))); \
+  __ret_244; \
+})
+#else
+#define vshrn_high_n_s32(__p0_245, __p1_245, __p2_245) __extension__ ({ \
+  int16x4_t __s0_245 = __p0_245; \
+  int32x4_t __s1_245 = __p1_245; \
+  int16x4_t __rev0_245;  __rev0_245 = __builtin_shufflevector(__s0_245, __s0_245, 3, 2, 1, 0); \
+  int32x4_t __rev1_245;  __rev1_245 = __builtin_shufflevector(__s1_245, __s1_245, 3, 2, 1, 0); \
+  int16x8_t __ret_245; \
+  __ret_245 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_245), (int16x4_t)(__noswap_vshrn_n_s32(__rev1_245, __p2_245)))); \
+  __ret_245 = __builtin_shufflevector(__ret_245, __ret_245, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_245; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrn_high_n_s64(__p0_246, __p1_246, __p2_246) __extension__ ({ \
+  int32x2_t __s0_246 = __p0_246; \
+  int64x2_t __s1_246 = __p1_246; \
+  int32x4_t __ret_246; \
+  __ret_246 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_246), (int32x2_t)(vshrn_n_s64(__s1_246, __p2_246)))); \
+  __ret_246; \
+})
+#else
+#define vshrn_high_n_s64(__p0_247, __p1_247, __p2_247) __extension__ ({ \
+  int32x2_t __s0_247 = __p0_247; \
+  int64x2_t __s1_247 = __p1_247; \
+  int32x2_t __rev0_247;  __rev0_247 = __builtin_shufflevector(__s0_247, __s0_247, 1, 0); \
+  int64x2_t __rev1_247;  __rev1_247 = __builtin_shufflevector(__s1_247, __s1_247, 1, 0); \
+  int32x4_t __ret_247; \
+  __ret_247 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_247), (int32x2_t)(__noswap_vshrn_n_s64(__rev1_247, __p2_247)))); \
+  __ret_247 = __builtin_shufflevector(__ret_247, __ret_247, 3, 2, 1, 0); \
+  __ret_247; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshrn_high_n_s16(__p0_248, __p1_248, __p2_248) __extension__ ({ \
+  int8x8_t __s0_248 = __p0_248; \
+  int16x8_t __s1_248 = __p1_248; \
+  int8x16_t __ret_248; \
+  __ret_248 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_248), (int8x8_t)(vshrn_n_s16(__s1_248, __p2_248)))); \
+  __ret_248; \
+})
+#else
+#define vshrn_high_n_s16(__p0_249, __p1_249, __p2_249) __extension__ ({ \
+  int8x8_t __s0_249 = __p0_249; \
+  int16x8_t __s1_249 = __p1_249; \
+  int8x8_t __rev0_249;  __rev0_249 = __builtin_shufflevector(__s0_249, __s0_249, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1_249;  __rev1_249 = __builtin_shufflevector(__s1_249, __s1_249, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret_249; \
+  __ret_249 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_249), (int8x8_t)(__noswap_vshrn_n_s16(__rev1_249, __p2_249)))); \
+  __ret_249 = __builtin_shufflevector(__ret_249, __ret_249, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_249; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vslid_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64_t __s1 = __p1; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vslid_n_u64(__s0, __s1, __p2); \
+  __ret; \
+})
+#else
+#define vslid_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64_t __s1 = __p1; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vslid_n_u64(__s0, __s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vslid_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __s1 = __p1; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vslid_n_s64(__s0, __s1, __p2); \
+  __ret; \
+})
+#else
+#define vslid_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __s1 = __p1; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vslid_n_s64(__s0, __s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsli_n_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1_t __s0 = __p0; \
+  poly64x1_t __s1 = __p1; \
+  poly64x1_t __ret; \
+  __ret = (poly64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
+  __ret; \
+})
+#else
+#define vsli_n_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1_t __s0 = __p0; \
+  poly64x1_t __s1 = __p1; \
+  poly64x1_t __ret; \
+  __ret = (poly64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsliq_n_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2_t __s0 = __p0; \
+  poly64x2_t __s1 = __p1; \
+  poly64x2_t __ret; \
+  __ret = (poly64x2_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 38); \
+  __ret; \
+})
+#else
+#define vsliq_n_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2_t __s0 = __p0; \
+  poly64x2_t __s1 = __p1; \
+  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  poly64x2_t __ret; \
+  __ret = (poly64x2_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 38); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8_t vsqaddb_u8(uint8_t __p0, uint8_t __p1) {
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vsqaddb_u8(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint8_t vsqaddb_u8(uint8_t __p0, uint8_t __p1) {
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vsqaddb_u8(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32_t vsqadds_u32(uint32_t __p0, uint32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vsqadds_u32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint32_t vsqadds_u32(uint32_t __p0, uint32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vsqadds_u32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vsqaddd_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vsqaddd_u64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vsqaddd_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vsqaddd_u64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16_t vsqaddh_u16(uint16_t __p0, uint16_t __p1) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vsqaddh_u16(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint16_t vsqaddh_u16(uint16_t __p0, uint16_t __p1) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vsqaddh_u16(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vsqaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vsqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
+  return __ret;
+}
+#else
+__ai uint8x16_t vsqaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = (uint8x16_t) __builtin_neon_vsqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vsqaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vsqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
+  return __ret;
+}
+#else
+__ai uint32x4_t vsqaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t) __builtin_neon_vsqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vsqaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vsqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vsqaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vsqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vsqaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vsqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
+  return __ret;
+}
+#else
+__ai uint16x8_t vsqaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t) __builtin_neon_vsqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vsqadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vsqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
+  return __ret;
+}
+#else
+__ai uint8x8_t vsqadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = (uint8x8_t) __builtin_neon_vsqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vsqadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vsqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
+  return __ret;
+}
+#else
+__ai uint32x2_t vsqadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = (uint32x2_t) __builtin_neon_vsqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vsqadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vsqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vsqadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vsqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vsqadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vsqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
+  return __ret;
+}
+#else
+__ai uint16x4_t vsqadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = (uint16x4_t) __builtin_neon_vsqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vsqrtq_f64(float64x2_t __p0) {
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vsqrtq_v((int8x16_t)__p0, 42);
+  return __ret;
+}
+#else
+__ai float64x2_t vsqrtq_f64(float64x2_t __p0) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __ret;
+  __ret = (float64x2_t) __builtin_neon_vsqrtq_v((int8x16_t)__rev0, 42);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vsqrtq_f32(float32x4_t __p0) {
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vsqrtq_v((int8x16_t)__p0, 41);
+  return __ret;
+}
+#else
+__ai float32x4_t vsqrtq_f32(float32x4_t __p0) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = (float32x4_t) __builtin_neon_vsqrtq_v((int8x16_t)__rev0, 41);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vsqrt_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vsqrt_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#else
+__ai float64x1_t vsqrt_f64(float64x1_t __p0) {
+  float64x1_t __ret;
+  __ret = (float64x1_t) __builtin_neon_vsqrt_v((int8x8_t)__p0, 10);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vsqrt_f32(float32x2_t __p0) {
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vsqrt_v((int8x8_t)__p0, 9);
+  return __ret;
+}
+#else
+__ai float32x2_t vsqrt_f32(float32x2_t __p0) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __ret;
+  __ret = (float32x2_t) __builtin_neon_vsqrt_v((int8x8_t)__rev0, 9);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsrad_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64_t __s1 = __p1; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vsrad_n_u64(__s0, __s1, __p2); \
+  __ret; \
+})
+#else
+#define vsrad_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64_t __s1 = __p1; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vsrad_n_u64(__s0, __s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsrad_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __s1 = __p1; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vsrad_n_s64(__s0, __s1, __p2); \
+  __ret; \
+})
+#else
+#define vsrad_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __s1 = __p1; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vsrad_n_s64(__s0, __s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsrid_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64_t __s1 = __p1; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vsrid_n_u64(__s0, __s1, __p2); \
+  __ret; \
+})
+#else
+#define vsrid_n_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64_t __s1 = __p1; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vsrid_n_u64(__s0, __s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsrid_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __s1 = __p1; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vsrid_n_s64(__s0, __s1, __p2); \
+  __ret; \
+})
+#else
+#define vsrid_n_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __s1 = __p1; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vsrid_n_s64(__s0, __s1, __p2); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsri_n_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1_t __s0 = __p0; \
+  poly64x1_t __s1 = __p1; \
+  poly64x1_t __ret; \
+  __ret = (poly64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
+  __ret; \
+})
+#else
+#define vsri_n_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1_t __s0 = __p0; \
+  poly64x1_t __s1 = __p1; \
+  poly64x1_t __ret; \
+  __ret = (poly64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vsriq_n_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2_t __s0 = __p0; \
+  poly64x2_t __s1 = __p1; \
+  poly64x2_t __ret; \
+  __ret = (poly64x2_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 38); \
+  __ret; \
+})
+#else
+#define vsriq_n_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2_t __s0 = __p0; \
+  poly64x2_t __s1 = __p1; \
+  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  poly64x2_t __ret; \
+  __ret = (poly64x2_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 38); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_p64(__p0, __p1) __extension__ ({ \
+  poly64x1_t __s1 = __p1; \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 6); \
+})
+#else
+#define vst1_p64(__p0, __p1) __extension__ ({ \
+  poly64x1_t __s1 = __p1; \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 6); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_p64(__p0, __p1) __extension__ ({ \
+  poly64x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 38); \
+})
+#else
+#define vst1q_p64(__p0, __p1) __extension__ ({ \
+  poly64x2_t __s1 = __p1; \
+  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 38); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_f64(__p0, __p1) __extension__ ({ \
+  float64x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 42); \
+})
+#else
+#define vst1q_f64(__p0, __p1) __extension__ ({ \
+  float64x2_t __s1 = __p1; \
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 42); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_f64(__p0, __p1) __extension__ ({ \
+  float64x1_t __s1 = __p1; \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 10); \
+})
+#else
+#define vst1_f64(__p0, __p1) __extension__ ({ \
+  float64x1_t __s1 = __p1; \
+  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 10); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1_t __s1 = __p1; \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 6); \
+})
+#else
+#define vst1_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1_t __s1 = __p1; \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 6); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 38); \
+})
+#else
+#define vst1q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2_t __s1 = __p1; \
+  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 38); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 42); \
+})
+#else
+#define vst1q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2_t __s1 = __p1; \
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 42); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1_t __s1 = __p1; \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 10); \
+})
+#else
+#define vst1_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1_t __s1 = __p1; \
+  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 10); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_p8_x2(__p0, __p1) __extension__ ({ \
+  poly8x8x2_t __s1 = __p1; \
+  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 4); \
+})
+#else
+#define vst1_p8_x2(__p0, __p1) __extension__ ({ \
+  poly8x8x2_t __s1 = __p1; \
+  poly8x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 4); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_p64_x2(__p0, __p1) __extension__ ({ \
+  poly64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 6); \
+})
+#else
+#define vst1_p64_x2(__p0, __p1) __extension__ ({ \
+  poly64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 6); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_p16_x2(__p0, __p1) __extension__ ({ \
+  poly16x4x2_t __s1 = __p1; \
+  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 5); \
+})
+#else
+#define vst1_p16_x2(__p0, __p1) __extension__ ({ \
+  poly16x4x2_t __s1 = __p1; \
+  poly16x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 5); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_p8_x2(__p0, __p1) __extension__ ({ \
+  poly8x16x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 36); \
+})
+#else
+#define vst1q_p8_x2(__p0, __p1) __extension__ ({ \
+  poly8x16x2_t __s1 = __p1; \
+  poly8x16x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 36); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_p64_x2(__p0, __p1) __extension__ ({ \
+  poly64x2x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 38); \
+})
+#else
+#define vst1q_p64_x2(__p0, __p1) __extension__ ({ \
+  poly64x2x2_t __s1 = __p1; \
+  poly64x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 38); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_p16_x2(__p0, __p1) __extension__ ({ \
+  poly16x8x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 37); \
+})
+#else
+#define vst1q_p16_x2(__p0, __p1) __extension__ ({ \
+  poly16x8x2_t __s1 = __p1; \
+  poly16x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 37); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_u8_x2(__p0, __p1) __extension__ ({ \
+  uint8x16x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 48); \
+})
+#else
+#define vst1q_u8_x2(__p0, __p1) __extension__ ({ \
+  uint8x16x2_t __s1 = __p1; \
+  uint8x16x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 48); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_u32_x2(__p0, __p1) __extension__ ({ \
+  uint32x4x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 50); \
+})
+#else
+#define vst1q_u32_x2(__p0, __p1) __extension__ ({ \
+  uint32x4x2_t __s1 = __p1; \
+  uint32x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 50); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_u64_x2(__p0, __p1) __extension__ ({ \
+  uint64x2x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 51); \
+})
+#else
+#define vst1q_u64_x2(__p0, __p1) __extension__ ({ \
+  uint64x2x2_t __s1 = __p1; \
+  uint64x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 51); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_u16_x2(__p0, __p1) __extension__ ({ \
+  uint16x8x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 49); \
+})
+#else
+#define vst1q_u16_x2(__p0, __p1) __extension__ ({ \
+  uint16x8x2_t __s1 = __p1; \
+  uint16x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 49); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_s8_x2(__p0, __p1) __extension__ ({ \
+  int8x16x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 32); \
+})
+#else
+#define vst1q_s8_x2(__p0, __p1) __extension__ ({ \
+  int8x16x2_t __s1 = __p1; \
+  int8x16x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 32); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_f64_x2(__p0, __p1) __extension__ ({ \
+  float64x2x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_x2_v(__p0, __s1.val[0], __s1.val[1], 42); \
+})
+#else
+#define vst1q_f64_x2(__p0, __p1) __extension__ ({ \
+  float64x2x2_t __s1 = __p1; \
+  float64x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst1q_x2_v(__p0, __rev1.val[0], __rev1.val[1], 42); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_f32_x2(__p0, __p1) __extension__ ({ \
+  float32x4x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_x2_v(__p0, __s1.val[0], __s1.val[1], 41); \
+})
+#else
+#define vst1q_f32_x2(__p0, __p1) __extension__ ({ \
+  float32x4x2_t __s1 = __p1; \
+  float32x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x2_v(__p0, __rev1.val[0], __rev1.val[1], 41); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_f16_x2(__p0, __p1) __extension__ ({ \
+  float16x8x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_x2_v(__p0, __s1.val[0], __s1.val[1], 40); \
+})
+#else
+#define vst1q_f16_x2(__p0, __p1) __extension__ ({ \
+  float16x8x2_t __s1 = __p1; \
+  float16x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x2_v(__p0, __rev1.val[0], __rev1.val[1], 40); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_s32_x2(__p0, __p1) __extension__ ({ \
+  int32x4x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_x2_v(__p0, __s1.val[0], __s1.val[1], 34); \
+})
+#else
+#define vst1q_s32_x2(__p0, __p1) __extension__ ({ \
+  int32x4x2_t __s1 = __p1; \
+  int32x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x2_v(__p0, __rev1.val[0], __rev1.val[1], 34); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_s64_x2(__p0, __p1) __extension__ ({ \
+  int64x2x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_x2_v(__p0, __s1.val[0], __s1.val[1], 35); \
+})
+#else
+#define vst1q_s64_x2(__p0, __p1) __extension__ ({ \
+  int64x2x2_t __s1 = __p1; \
+  int64x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst1q_x2_v(__p0, __rev1.val[0], __rev1.val[1], 35); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_s16_x2(__p0, __p1) __extension__ ({ \
+  int16x8x2_t __s1 = __p1; \
+  __builtin_neon_vst1q_x2_v(__p0, __s1.val[0], __s1.val[1], 33); \
+})
+#else
+#define vst1q_s16_x2(__p0, __p1) __extension__ ({ \
+  int16x8x2_t __s1 = __p1; \
+  int16x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x2_v(__p0, __rev1.val[0], __rev1.val[1], 33); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_u8_x2(__p0, __p1) __extension__ ({ \
+  uint8x8x2_t __s1 = __p1; \
+  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 16); \
+})
+#else
+#define vst1_u8_x2(__p0, __p1) __extension__ ({ \
+  uint8x8x2_t __s1 = __p1; \
+  uint8x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 16); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_u32_x2(__p0, __p1) __extension__ ({ \
+  uint32x2x2_t __s1 = __p1; \
+  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 18); \
+})
+#else
+#define vst1_u32_x2(__p0, __p1) __extension__ ({ \
+  uint32x2x2_t __s1 = __p1; \
+  uint32x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 18); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_u64_x2(__p0, __p1) __extension__ ({ \
+  uint64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 19); \
+})
+#else
+#define vst1_u64_x2(__p0, __p1) __extension__ ({ \
+  uint64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 19); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_u16_x2(__p0, __p1) __extension__ ({ \
+  uint16x4x2_t __s1 = __p1; \
+  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 17); \
+})
+#else
+#define vst1_u16_x2(__p0, __p1) __extension__ ({ \
+  uint16x4x2_t __s1 = __p1; \
+  uint16x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 17); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_s8_x2(__p0, __p1) __extension__ ({ \
+  int8x8x2_t __s1 = __p1; \
+  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 0); \
+})
+#else
+#define vst1_s8_x2(__p0, __p1) __extension__ ({ \
+  int8x8x2_t __s1 = __p1; \
+  int8x8x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 0); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_f64_x2(__p0, __p1) __extension__ ({ \
+  float64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 10); \
+})
+#else
+#define vst1_f64_x2(__p0, __p1) __extension__ ({ \
+  float64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 10); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_f32_x2(__p0, __p1) __extension__ ({ \
+  float32x2x2_t __s1 = __p1; \
+  __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 9); \
+})
+#else
+#define vst1_f32_x2(__p0, __p1) __extension__ ({ \
+  float32x2x2_t __s1 = __p1; \
+  float32x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst1_x2_v(__p0, __rev1.val[0], __rev1.val[1], 9); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_f16_x2(__p0, __p1) __extension__ ({ \
+  float16x4x2_t __s1 = __p1; \
+  __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 8); \
+})
+#else
+#define vst1_f16_x2(__p0, __p1) __extension__ ({ \
+  float16x4x2_t __s1 = __p1; \
+  float16x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst1_x2_v(__p0, __rev1.val[0], __rev1.val[1], 8); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_s32_x2(__p0, __p1) __extension__ ({ \
+  int32x2x2_t __s1 = __p1; \
+  __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 2); \
+})
+#else
+#define vst1_s32_x2(__p0, __p1) __extension__ ({ \
+  int32x2x2_t __s1 = __p1; \
+  int32x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst1_x2_v(__p0, __rev1.val[0], __rev1.val[1], 2); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_s64_x2(__p0, __p1) __extension__ ({ \
+  int64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 3); \
+})
+#else
+#define vst1_s64_x2(__p0, __p1) __extension__ ({ \
+  int64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 3); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_s16_x2(__p0, __p1) __extension__ ({ \
+  int16x4x2_t __s1 = __p1; \
+  __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 1); \
+})
+#else
+#define vst1_s16_x2(__p0, __p1) __extension__ ({ \
+  int16x4x2_t __s1 = __p1; \
+  int16x4x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __builtin_neon_vst1_x2_v(__p0, __rev1.val[0], __rev1.val[1], 1); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_p8_x3(__p0, __p1) __extension__ ({ \
+  poly8x8x3_t __s1 = __p1; \
+  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 4); \
+})
+#else
+#define vst1_p8_x3(__p0, __p1) __extension__ ({ \
+  poly8x8x3_t __s1 = __p1; \
+  poly8x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 4); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_p64_x3(__p0, __p1) __extension__ ({ \
+  poly64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 6); \
+})
+#else
+#define vst1_p64_x3(__p0, __p1) __extension__ ({ \
+  poly64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 6); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_p16_x3(__p0, __p1) __extension__ ({ \
+  poly16x4x3_t __s1 = __p1; \
+  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 5); \
+})
+#else
+#define vst1_p16_x3(__p0, __p1) __extension__ ({ \
+  poly16x4x3_t __s1 = __p1; \
+  poly16x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 5); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_p8_x3(__p0, __p1) __extension__ ({ \
+  poly8x16x3_t __s1 = __p1; \
+  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 36); \
+})
+#else
+#define vst1q_p8_x3(__p0, __p1) __extension__ ({ \
+  poly8x16x3_t __s1 = __p1; \
+  poly8x16x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 36); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_p64_x3(__p0, __p1) __extension__ ({ \
+  poly64x2x3_t __s1 = __p1; \
+  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 38); \
+})
+#else
+#define vst1q_p64_x3(__p0, __p1) __extension__ ({ \
+  poly64x2x3_t __s1 = __p1; \
+  poly64x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 38); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_p16_x3(__p0, __p1) __extension__ ({ \
+  poly16x8x3_t __s1 = __p1; \
+  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 37); \
+})
+#else
+#define vst1q_p16_x3(__p0, __p1) __extension__ ({ \
+  poly16x8x3_t __s1 = __p1; \
+  poly16x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 37); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_u8_x3(__p0, __p1) __extension__ ({ \
+  uint8x16x3_t __s1 = __p1; \
+  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 48); \
+})
+#else
+#define vst1q_u8_x3(__p0, __p1) __extension__ ({ \
+  uint8x16x3_t __s1 = __p1; \
+  uint8x16x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 48); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_u32_x3(__p0, __p1) __extension__ ({ \
+  uint32x4x3_t __s1 = __p1; \
+  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 50); \
+})
+#else
+#define vst1q_u32_x3(__p0, __p1) __extension__ ({ \
+  uint32x4x3_t __s1 = __p1; \
+  uint32x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 50); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_u64_x3(__p0, __p1) __extension__ ({ \
+  uint64x2x3_t __s1 = __p1; \
+  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 51); \
+})
+#else
+#define vst1q_u64_x3(__p0, __p1) __extension__ ({ \
+  uint64x2x3_t __s1 = __p1; \
+  uint64x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 51); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_u16_x3(__p0, __p1) __extension__ ({ \
+  uint16x8x3_t __s1 = __p1; \
+  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 49); \
+})
+#else
+#define vst1q_u16_x3(__p0, __p1) __extension__ ({ \
+  uint16x8x3_t __s1 = __p1; \
+  uint16x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 49); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_s8_x3(__p0, __p1) __extension__ ({ \
+  int8x16x3_t __s1 = __p1; \
+  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 32); \
+})
+#else
+#define vst1q_s8_x3(__p0, __p1) __extension__ ({ \
+  int8x16x3_t __s1 = __p1; \
+  int8x16x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 32); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_f64_x3(__p0, __p1) __extension__ ({ \
+  float64x2x3_t __s1 = __p1; \
+  __builtin_neon_vst1q_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 42); \
+})
+#else
+#define vst1q_f64_x3(__p0, __p1) __extension__ ({ \
+  float64x2x3_t __s1 = __p1; \
+  float64x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst1q_x3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 42); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_f32_x3(__p0, __p1) __extension__ ({ \
+  float32x4x3_t __s1 = __p1; \
+  __builtin_neon_vst1q_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 41); \
+})
+#else
+#define vst1q_f32_x3(__p0, __p1) __extension__ ({ \
+  float32x4x3_t __s1 = __p1; \
+  float32x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 41); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_f16_x3(__p0, __p1) __extension__ ({ \
+  float16x8x3_t __s1 = __p1; \
+  __builtin_neon_vst1q_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 40); \
+})
+#else
+#define vst1q_f16_x3(__p0, __p1) __extension__ ({ \
+  float16x8x3_t __s1 = __p1; \
+  float16x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 40); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_s32_x3(__p0, __p1) __extension__ ({ \
+  int32x4x3_t __s1 = __p1; \
+  __builtin_neon_vst1q_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 34); \
+})
+#else
+#define vst1q_s32_x3(__p0, __p1) __extension__ ({ \
+  int32x4x3_t __s1 = __p1; \
+  int32x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 34); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_s64_x3(__p0, __p1) __extension__ ({ \
+  int64x2x3_t __s1 = __p1; \
+  __builtin_neon_vst1q_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 35); \
+})
+#else
+#define vst1q_s64_x3(__p0, __p1) __extension__ ({ \
+  int64x2x3_t __s1 = __p1; \
+  int64x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst1q_x3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 35); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_s16_x3(__p0, __p1) __extension__ ({ \
+  int16x8x3_t __s1 = __p1; \
+  __builtin_neon_vst1q_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 33); \
+})
+#else
+#define vst1q_s16_x3(__p0, __p1) __extension__ ({ \
+  int16x8x3_t __s1 = __p1; \
+  int16x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 33); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_u8_x3(__p0, __p1) __extension__ ({ \
+  uint8x8x3_t __s1 = __p1; \
+  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 16); \
+})
+#else
+#define vst1_u8_x3(__p0, __p1) __extension__ ({ \
+  uint8x8x3_t __s1 = __p1; \
+  uint8x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 16); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_u32_x3(__p0, __p1) __extension__ ({ \
+  uint32x2x3_t __s1 = __p1; \
+  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 18); \
+})
+#else
+#define vst1_u32_x3(__p0, __p1) __extension__ ({ \
+  uint32x2x3_t __s1 = __p1; \
+  uint32x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 18); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_u64_x3(__p0, __p1) __extension__ ({ \
+  uint64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 19); \
+})
+#else
+#define vst1_u64_x3(__p0, __p1) __extension__ ({ \
+  uint64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 19); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_u16_x3(__p0, __p1) __extension__ ({ \
+  uint16x4x3_t __s1 = __p1; \
+  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 17); \
+})
+#else
+#define vst1_u16_x3(__p0, __p1) __extension__ ({ \
+  uint16x4x3_t __s1 = __p1; \
+  uint16x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 17); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_s8_x3(__p0, __p1) __extension__ ({ \
+  int8x8x3_t __s1 = __p1; \
+  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 0); \
+})
+#else
+#define vst1_s8_x3(__p0, __p1) __extension__ ({ \
+  int8x8x3_t __s1 = __p1; \
+  int8x8x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 0); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_f64_x3(__p0, __p1) __extension__ ({ \
+  float64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 10); \
+})
+#else
+#define vst1_f64_x3(__p0, __p1) __extension__ ({ \
+  float64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 10); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_f32_x3(__p0, __p1) __extension__ ({ \
+  float32x2x3_t __s1 = __p1; \
+  __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 9); \
+})
+#else
+#define vst1_f32_x3(__p0, __p1) __extension__ ({ \
+  float32x2x3_t __s1 = __p1; \
+  float32x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst1_x3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 9); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_f16_x3(__p0, __p1) __extension__ ({ \
+  float16x4x3_t __s1 = __p1; \
+  __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 8); \
+})
+#else
+#define vst1_f16_x3(__p0, __p1) __extension__ ({ \
+  float16x4x3_t __s1 = __p1; \
+  float16x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst1_x3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 8); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_s32_x3(__p0, __p1) __extension__ ({ \
+  int32x2x3_t __s1 = __p1; \
+  __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 2); \
+})
+#else
+#define vst1_s32_x3(__p0, __p1) __extension__ ({ \
+  int32x2x3_t __s1 = __p1; \
+  int32x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst1_x3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 2); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_s64_x3(__p0, __p1) __extension__ ({ \
+  int64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 3); \
+})
+#else
+#define vst1_s64_x3(__p0, __p1) __extension__ ({ \
+  int64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 3); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_s16_x3(__p0, __p1) __extension__ ({ \
+  int16x4x3_t __s1 = __p1; \
+  __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 1); \
+})
+#else
+#define vst1_s16_x3(__p0, __p1) __extension__ ({ \
+  int16x4x3_t __s1 = __p1; \
+  int16x4x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __builtin_neon_vst1_x3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 1); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_p8_x4(__p0, __p1) __extension__ ({ \
+  poly8x8x4_t __s1 = __p1; \
+  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 4); \
+})
+#else
+#define vst1_p8_x4(__p0, __p1) __extension__ ({ \
+  poly8x8x4_t __s1 = __p1; \
+  poly8x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 4); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_p64_x4(__p0, __p1) __extension__ ({ \
+  poly64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 6); \
+})
+#else
+#define vst1_p64_x4(__p0, __p1) __extension__ ({ \
+  poly64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 6); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_p16_x4(__p0, __p1) __extension__ ({ \
+  poly16x4x4_t __s1 = __p1; \
+  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 5); \
+})
+#else
+#define vst1_p16_x4(__p0, __p1) __extension__ ({ \
+  poly16x4x4_t __s1 = __p1; \
+  poly16x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 5); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_p8_x4(__p0, __p1) __extension__ ({ \
+  poly8x16x4_t __s1 = __p1; \
+  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 36); \
+})
+#else
+#define vst1q_p8_x4(__p0, __p1) __extension__ ({ \
+  poly8x16x4_t __s1 = __p1; \
+  poly8x16x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 36); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_p64_x4(__p0, __p1) __extension__ ({ \
+  poly64x2x4_t __s1 = __p1; \
+  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 38); \
+})
+#else
+#define vst1q_p64_x4(__p0, __p1) __extension__ ({ \
+  poly64x2x4_t __s1 = __p1; \
+  poly64x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 38); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_p16_x4(__p0, __p1) __extension__ ({ \
+  poly16x8x4_t __s1 = __p1; \
+  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 37); \
+})
+#else
+#define vst1q_p16_x4(__p0, __p1) __extension__ ({ \
+  poly16x8x4_t __s1 = __p1; \
+  poly16x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 37); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_u8_x4(__p0, __p1) __extension__ ({ \
+  uint8x16x4_t __s1 = __p1; \
+  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 48); \
+})
+#else
+#define vst1q_u8_x4(__p0, __p1) __extension__ ({ \
+  uint8x16x4_t __s1 = __p1; \
+  uint8x16x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 48); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_u32_x4(__p0, __p1) __extension__ ({ \
+  uint32x4x4_t __s1 = __p1; \
+  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 50); \
+})
+#else
+#define vst1q_u32_x4(__p0, __p1) __extension__ ({ \
+  uint32x4x4_t __s1 = __p1; \
+  uint32x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 50); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_u64_x4(__p0, __p1) __extension__ ({ \
+  uint64x2x4_t __s1 = __p1; \
+  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 51); \
+})
+#else
+#define vst1q_u64_x4(__p0, __p1) __extension__ ({ \
+  uint64x2x4_t __s1 = __p1; \
+  uint64x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 51); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_u16_x4(__p0, __p1) __extension__ ({ \
+  uint16x8x4_t __s1 = __p1; \
+  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 49); \
+})
+#else
+#define vst1q_u16_x4(__p0, __p1) __extension__ ({ \
+  uint16x8x4_t __s1 = __p1; \
+  uint16x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 49); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_s8_x4(__p0, __p1) __extension__ ({ \
+  int8x16x4_t __s1 = __p1; \
+  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 32); \
+})
+#else
+#define vst1q_s8_x4(__p0, __p1) __extension__ ({ \
+  int8x16x4_t __s1 = __p1; \
+  int8x16x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 32); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_f64_x4(__p0, __p1) __extension__ ({ \
+  float64x2x4_t __s1 = __p1; \
+  __builtin_neon_vst1q_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 42); \
+})
+#else
+#define vst1q_f64_x4(__p0, __p1) __extension__ ({ \
+  float64x2x4_t __s1 = __p1; \
+  float64x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst1q_x4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 42); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_f32_x4(__p0, __p1) __extension__ ({ \
+  float32x4x4_t __s1 = __p1; \
+  __builtin_neon_vst1q_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 41); \
+})
+#else
+#define vst1q_f32_x4(__p0, __p1) __extension__ ({ \
+  float32x4x4_t __s1 = __p1; \
+  float32x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 41); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_f16_x4(__p0, __p1) __extension__ ({ \
+  float16x8x4_t __s1 = __p1; \
+  __builtin_neon_vst1q_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 40); \
+})
+#else
+#define vst1q_f16_x4(__p0, __p1) __extension__ ({ \
+  float16x8x4_t __s1 = __p1; \
+  float16x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 40); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_s32_x4(__p0, __p1) __extension__ ({ \
+  int32x4x4_t __s1 = __p1; \
+  __builtin_neon_vst1q_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 34); \
+})
+#else
+#define vst1q_s32_x4(__p0, __p1) __extension__ ({ \
+  int32x4x4_t __s1 = __p1; \
+  int32x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 34); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_s64_x4(__p0, __p1) __extension__ ({ \
+  int64x2x4_t __s1 = __p1; \
+  __builtin_neon_vst1q_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 35); \
+})
+#else
+#define vst1q_s64_x4(__p0, __p1) __extension__ ({ \
+  int64x2x4_t __s1 = __p1; \
+  int64x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst1q_x4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 35); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1q_s16_x4(__p0, __p1) __extension__ ({ \
+  int16x8x4_t __s1 = __p1; \
+  __builtin_neon_vst1q_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 33); \
+})
+#else
+#define vst1q_s16_x4(__p0, __p1) __extension__ ({ \
+  int16x8x4_t __s1 = __p1; \
+  int16x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1q_x4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 33); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_u8_x4(__p0, __p1) __extension__ ({ \
+  uint8x8x4_t __s1 = __p1; \
+  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 16); \
+})
+#else
+#define vst1_u8_x4(__p0, __p1) __extension__ ({ \
+  uint8x8x4_t __s1 = __p1; \
+  uint8x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 16); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_u32_x4(__p0, __p1) __extension__ ({ \
+  uint32x2x4_t __s1 = __p1; \
+  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 18); \
+})
+#else
+#define vst1_u32_x4(__p0, __p1) __extension__ ({ \
+  uint32x2x4_t __s1 = __p1; \
+  uint32x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 18); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_u64_x4(__p0, __p1) __extension__ ({ \
+  uint64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 19); \
+})
+#else
+#define vst1_u64_x4(__p0, __p1) __extension__ ({ \
+  uint64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 19); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_u16_x4(__p0, __p1) __extension__ ({ \
+  uint16x4x4_t __s1 = __p1; \
+  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 17); \
+})
+#else
+#define vst1_u16_x4(__p0, __p1) __extension__ ({ \
+  uint16x4x4_t __s1 = __p1; \
+  uint16x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 17); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_s8_x4(__p0, __p1) __extension__ ({ \
+  int8x8x4_t __s1 = __p1; \
+  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 0); \
+})
+#else
+#define vst1_s8_x4(__p0, __p1) __extension__ ({ \
+  int8x8x4_t __s1 = __p1; \
+  int8x8x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 0); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_f64_x4(__p0, __p1) __extension__ ({ \
+  float64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 10); \
+})
+#else
+#define vst1_f64_x4(__p0, __p1) __extension__ ({ \
+  float64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 10); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_f32_x4(__p0, __p1) __extension__ ({ \
+  float32x2x4_t __s1 = __p1; \
+  __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 9); \
+})
+#else
+#define vst1_f32_x4(__p0, __p1) __extension__ ({ \
+  float32x2x4_t __s1 = __p1; \
+  float32x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst1_x4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 9); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_f16_x4(__p0, __p1) __extension__ ({ \
+  float16x4x4_t __s1 = __p1; \
+  __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 8); \
+})
+#else
+#define vst1_f16_x4(__p0, __p1) __extension__ ({ \
+  float16x4x4_t __s1 = __p1; \
+  float16x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst1_x4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 8); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_s32_x4(__p0, __p1) __extension__ ({ \
+  int32x2x4_t __s1 = __p1; \
+  __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 2); \
+})
+#else
+#define vst1_s32_x4(__p0, __p1) __extension__ ({ \
+  int32x2x4_t __s1 = __p1; \
+  int32x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst1_x4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 2); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_s64_x4(__p0, __p1) __extension__ ({ \
+  int64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 3); \
+})
+#else
+#define vst1_s64_x4(__p0, __p1) __extension__ ({ \
+  int64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 3); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst1_s16_x4(__p0, __p1) __extension__ ({ \
+  int16x4x4_t __s1 = __p1; \
+  __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 1); \
+})
+#else
+#define vst1_s16_x4(__p0, __p1) __extension__ ({ \
+  int16x4x4_t __s1 = __p1; \
+  int16x4x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
+  __builtin_neon_vst1_x4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 1); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_p64(__p0, __p1) __extension__ ({ \
+  poly64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 6); \
+})
+#else
+#define vst2_p64(__p0, __p1) __extension__ ({ \
+  poly64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 6); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_p64(__p0, __p1) __extension__ ({ \
+  poly64x2x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 38); \
+})
+#else
+#define vst2q_p64(__p0, __p1) __extension__ ({ \
+  poly64x2x2_t __s1 = __p1; \
+  poly64x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 38); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_u64(__p0, __p1) __extension__ ({ \
+  uint64x2x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 51); \
+})
+#else
+#define vst2q_u64(__p0, __p1) __extension__ ({ \
+  uint64x2x2_t __s1 = __p1; \
+  uint64x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 51); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_f64(__p0, __p1) __extension__ ({ \
+  float64x2x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_v(__p0, __s1.val[0], __s1.val[1], 42); \
+})
+#else
+#define vst2q_f64(__p0, __p1) __extension__ ({ \
+  float64x2x2_t __s1 = __p1; \
+  float64x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst2q_v(__p0, __rev1.val[0], __rev1.val[1], 42); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_s64(__p0, __p1) __extension__ ({ \
+  int64x2x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_v(__p0, __s1.val[0], __s1.val[1], 35); \
+})
+#else
+#define vst2q_s64(__p0, __p1) __extension__ ({ \
+  int64x2x2_t __s1 = __p1; \
+  int64x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst2q_v(__p0, __rev1.val[0], __rev1.val[1], 35); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_f64(__p0, __p1) __extension__ ({ \
+  float64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 10); \
+})
+#else
+#define vst2_f64(__p0, __p1) __extension__ ({ \
+  float64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 10); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 6); \
+})
+#else
+#define vst2_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 6); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 36); \
+})
+#else
+#define vst2q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16x2_t __s1 = __p1; \
+  poly8x16x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 36); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 38); \
+})
+#else
+#define vst2q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2x2_t __s1 = __p1; \
+  poly64x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 38); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 48); \
+})
+#else
+#define vst2q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16x2_t __s1 = __p1; \
+  uint8x16x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 48); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 51); \
+})
+#else
+#define vst2q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2x2_t __s1 = __p1; \
+  uint64x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 51); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 32); \
+})
+#else
+#define vst2q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16x2_t __s1 = __p1; \
+  int8x16x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 32); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 42); \
+})
+#else
+#define vst2q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2x2_t __s1 = __p1; \
+  float64x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst2q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __p2, 42); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2x2_t __s1 = __p1; \
+  __builtin_neon_vst2q_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 35); \
+})
+#else
+#define vst2q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2x2_t __s1 = __p1; \
+  int64x2x2_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __builtin_neon_vst2q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __p2, 35); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 19); \
+})
+#else
+#define vst2_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 19); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 10); \
+})
+#else
+#define vst2_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 10); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst2_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 3); \
+})
+#else
+#define vst2_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1x2_t __s1 = __p1; \
+  __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 3); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_p64(__p0, __p1) __extension__ ({ \
+  poly64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 6); \
+})
+#else
+#define vst3_p64(__p0, __p1) __extension__ ({ \
+  poly64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 6); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_p64(__p0, __p1) __extension__ ({ \
+  poly64x2x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 38); \
+})
+#else
+#define vst3q_p64(__p0, __p1) __extension__ ({ \
+  poly64x2x3_t __s1 = __p1; \
+  poly64x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 38); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_u64(__p0, __p1) __extension__ ({ \
+  uint64x2x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 51); \
+})
+#else
+#define vst3q_u64(__p0, __p1) __extension__ ({ \
+  uint64x2x3_t __s1 = __p1; \
+  uint64x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 51); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_f64(__p0, __p1) __extension__ ({ \
+  float64x2x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 42); \
+})
+#else
+#define vst3q_f64(__p0, __p1) __extension__ ({ \
+  float64x2x3_t __s1 = __p1; \
+  float64x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst3q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 42); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_s64(__p0, __p1) __extension__ ({ \
+  int64x2x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 35); \
+})
+#else
+#define vst3q_s64(__p0, __p1) __extension__ ({ \
+  int64x2x3_t __s1 = __p1; \
+  int64x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst3q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 35); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_f64(__p0, __p1) __extension__ ({ \
+  float64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 10); \
+})
+#else
+#define vst3_f64(__p0, __p1) __extension__ ({ \
+  float64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 10); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 6); \
+})
+#else
+#define vst3_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 6); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 36); \
+})
+#else
+#define vst3q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16x3_t __s1 = __p1; \
+  poly8x16x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 36); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 38); \
+})
+#else
+#define vst3q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2x3_t __s1 = __p1; \
+  poly64x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 38); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 48); \
+})
+#else
+#define vst3q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16x3_t __s1 = __p1; \
+  uint8x16x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 48); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 51); \
+})
+#else
+#define vst3q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2x3_t __s1 = __p1; \
+  uint64x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 51); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 32); \
+})
+#else
+#define vst3q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16x3_t __s1 = __p1; \
+  int8x16x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 32); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 42); \
+})
+#else
+#define vst3q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2x3_t __s1 = __p1; \
+  float64x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst3q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 42); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2x3_t __s1 = __p1; \
+  __builtin_neon_vst3q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 35); \
+})
+#else
+#define vst3q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2x3_t __s1 = __p1; \
+  int64x2x3_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __builtin_neon_vst3q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 35); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 19); \
+})
+#else
+#define vst3_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 19); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 10); \
+})
+#else
+#define vst3_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 10); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst3_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 3); \
+})
+#else
+#define vst3_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1x3_t __s1 = __p1; \
+  __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 3); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_p64(__p0, __p1) __extension__ ({ \
+  poly64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 6); \
+})
+#else
+#define vst4_p64(__p0, __p1) __extension__ ({ \
+  poly64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 6); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_p64(__p0, __p1) __extension__ ({ \
+  poly64x2x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 38); \
+})
+#else
+#define vst4q_p64(__p0, __p1) __extension__ ({ \
+  poly64x2x4_t __s1 = __p1; \
+  poly64x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 38); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_u64(__p0, __p1) __extension__ ({ \
+  uint64x2x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 51); \
+})
+#else
+#define vst4q_u64(__p0, __p1) __extension__ ({ \
+  uint64x2x4_t __s1 = __p1; \
+  uint64x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 51); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_f64(__p0, __p1) __extension__ ({ \
+  float64x2x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 42); \
+})
+#else
+#define vst4q_f64(__p0, __p1) __extension__ ({ \
+  float64x2x4_t __s1 = __p1; \
+  float64x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst4q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 42); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_s64(__p0, __p1) __extension__ ({ \
+  int64x2x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 35); \
+})
+#else
+#define vst4q_s64(__p0, __p1) __extension__ ({ \
+  int64x2x4_t __s1 = __p1; \
+  int64x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst4q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 35); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_f64(__p0, __p1) __extension__ ({ \
+  float64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 10); \
+})
+#else
+#define vst4_f64(__p0, __p1) __extension__ ({ \
+  float64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 10); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 6); \
+})
+#else
+#define vst4_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 6); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 36); \
+})
+#else
+#define vst4q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
+  poly8x16x4_t __s1 = __p1; \
+  poly8x16x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 36); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 38); \
+})
+#else
+#define vst4q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
+  poly64x2x4_t __s1 = __p1; \
+  poly64x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 38); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 48); \
+})
+#else
+#define vst4q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
+  uint8x16x4_t __s1 = __p1; \
+  uint8x16x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 48); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 51); \
+})
+#else
+#define vst4q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x2x4_t __s1 = __p1; \
+  uint64x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 51); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 32); \
+})
+#else
+#define vst4q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
+  int8x16x4_t __s1 = __p1; \
+  int8x16x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 32); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 42); \
+})
+#else
+#define vst4q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x2x4_t __s1 = __p1; \
+  float64x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst4q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 42); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2x4_t __s1 = __p1; \
+  __builtin_neon_vst4q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 35); \
+})
+#else
+#define vst4q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x2x4_t __s1 = __p1; \
+  int64x2x4_t __rev1; \
+  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
+  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
+  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
+  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
+  __builtin_neon_vst4q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 35); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 19); \
+})
+#else
+#define vst4_lane_u64(__p0, __p1, __p2) __extension__ ({ \
+  uint64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 19); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 10); \
+})
+#else
+#define vst4_lane_f64(__p0, __p1, __p2) __extension__ ({ \
+  float64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 10); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vst4_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 3); \
+})
+#else
+#define vst4_lane_s64(__p0, __p1, __p2) __extension__ ({ \
+  int64x1x4_t __s1 = __p1; \
+  __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 3); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vstrq_p128(__p0, __p1) __extension__ ({ \
+  poly128_t __s1 = __p1; \
+  __builtin_neon_vstrq_p128(__p0, __s1); \
+})
+#else
+#define vstrq_p128(__p0, __p1) __extension__ ({ \
+  poly128_t __s1 = __p1; \
+  __builtin_neon_vstrq_p128(__p0, __s1); \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vsubd_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vsubd_u64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vsubd_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vsubd_u64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vsubd_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vsubd_s64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int64_t vsubd_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vsubd_s64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vsubq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#else
+__ai float64x2_t vsubq_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = __rev0 - __rev1;
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x1_t vsub_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#else
+__ai float64x1_t vsub_f64(float64x1_t __p0, float64x1_t __p1) {
+  float64x1_t __ret;
+  __ret = __p0 - __p1;
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vsubhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint16x8_t __ret;
+  __ret = vcombine_u16(__p0, vsubhn_u32(__p1, __p2));
+  return __ret;
+}
+#else
+__ai uint16x8_t vsubhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __noswap_vcombine_u16(__rev0, __noswap_vsubhn_u32(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vsubhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
+  uint32x4_t __ret;
+  __ret = vcombine_u32(__p0, vsubhn_u64(__p1, __p2));
+  return __ret;
+}
+#else
+__ai uint32x4_t vsubhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __noswap_vcombine_u32(__rev0, __noswap_vsubhn_u64(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vsubhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint8x16_t __ret;
+  __ret = vcombine_u8(__p0, vsubhn_u16(__p1, __p2));
+  return __ret;
+}
+#else
+__ai uint8x16_t vsubhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __noswap_vcombine_u8(__rev0, __noswap_vsubhn_u16(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vsubhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int16x8_t __ret;
+  __ret = vcombine_s16(__p0, vsubhn_s32(__p1, __p2));
+  return __ret;
+}
+#else
+__ai int16x8_t vsubhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __noswap_vcombine_s16(__rev0, __noswap_vsubhn_s32(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vsubhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
+  int32x4_t __ret;
+  __ret = vcombine_s32(__p0, vsubhn_s64(__p1, __p2));
+  return __ret;
+}
+#else
+__ai int32x4_t vsubhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vcombine_s32(__rev0, __noswap_vsubhn_s64(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vsubhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int8x16_t __ret;
+  __ret = vcombine_s8(__p0, vsubhn_s16(__p1, __p2));
+  return __ret;
+}
+#else
+__ai int8x16_t vsubhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __noswap_vcombine_s8(__rev0, __noswap_vsubhn_s16(__rev1, __rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vsubl_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint16x8_t __ret;
+  __ret = vmovl_high_u8(__p0) - vmovl_high_u8(__p1);
+  return __ret;
+}
+#else
+__ai uint16x8_t vsubl_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __noswap_vmovl_high_u8(__rev0) - __noswap_vmovl_high_u8(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vsubl_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint64x2_t __ret;
+  __ret = vmovl_high_u32(__p0) - vmovl_high_u32(__p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vsubl_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint64x2_t __ret;
+  __ret = __noswap_vmovl_high_u32(__rev0) - __noswap_vmovl_high_u32(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vsubl_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint32x4_t __ret;
+  __ret = vmovl_high_u16(__p0) - vmovl_high_u16(__p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vsubl_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __noswap_vmovl_high_u16(__rev0) - __noswap_vmovl_high_u16(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vsubl_high_s8(int8x16_t __p0, int8x16_t __p1) {
+  int16x8_t __ret;
+  __ret = vmovl_high_s8(__p0) - vmovl_high_s8(__p1);
+  return __ret;
+}
+#else
+__ai int16x8_t vsubl_high_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __noswap_vmovl_high_s8(__rev0) - __noswap_vmovl_high_s8(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vsubl_high_s32(int32x4_t __p0, int32x4_t __p1) {
+  int64x2_t __ret;
+  __ret = vmovl_high_s32(__p0) - vmovl_high_s32(__p1);
+  return __ret;
+}
+#else
+__ai int64x2_t vsubl_high_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int64x2_t __ret;
+  __ret = __noswap_vmovl_high_s32(__rev0) - __noswap_vmovl_high_s32(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vsubl_high_s16(int16x8_t __p0, int16x8_t __p1) {
+  int32x4_t __ret;
+  __ret = vmovl_high_s16(__p0) - vmovl_high_s16(__p1);
+  return __ret;
+}
+#else
+__ai int32x4_t vsubl_high_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vmovl_high_s16(__rev0) - __noswap_vmovl_high_s16(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vsubw_high_u8(uint16x8_t __p0, uint8x16_t __p1) {
+  uint16x8_t __ret;
+  __ret = __p0 - vmovl_high_u8(__p1);
+  return __ret;
+}
+#else
+__ai uint16x8_t vsubw_high_u8(uint16x8_t __p0, uint8x16_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 - __noswap_vmovl_high_u8(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vsubw_high_u32(uint64x2_t __p0, uint32x4_t __p1) {
+  uint64x2_t __ret;
+  __ret = __p0 - vmovl_high_u32(__p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vsubw_high_u32(uint64x2_t __p0, uint32x4_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint64x2_t __ret;
+  __ret = __rev0 - __noswap_vmovl_high_u32(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vsubw_high_u16(uint32x4_t __p0, uint16x8_t __p1) {
+  uint32x4_t __ret;
+  __ret = __p0 - vmovl_high_u16(__p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vsubw_high_u16(uint32x4_t __p0, uint16x8_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 - __noswap_vmovl_high_u16(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vsubw_high_s8(int16x8_t __p0, int8x16_t __p1) {
+  int16x8_t __ret;
+  __ret = __p0 - vmovl_high_s8(__p1);
+  return __ret;
+}
+#else
+__ai int16x8_t vsubw_high_s8(int16x8_t __p0, int8x16_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 - __noswap_vmovl_high_s8(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vsubw_high_s32(int64x2_t __p0, int32x4_t __p1) {
+  int64x2_t __ret;
+  __ret = __p0 - vmovl_high_s32(__p1);
+  return __ret;
+}
+#else
+__ai int64x2_t vsubw_high_s32(int64x2_t __p0, int32x4_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int64x2_t __ret;
+  __ret = __rev0 - __noswap_vmovl_high_s32(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vsubw_high_s16(int32x4_t __p0, int16x8_t __p1) {
+  int32x4_t __ret;
+  __ret = __p0 - vmovl_high_s16(__p1);
+  return __ret;
+}
+#else
+__ai int32x4_t vsubw_high_s16(int32x4_t __p0, int16x8_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 - __noswap_vmovl_high_s16(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vtrn1_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 2, 10, 4, 12, 6, 14);
+  return __ret;
+}
+#else
+__ai poly8x8_t vtrn1_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 2, 10, 4, 12, 6, 14);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vtrn1_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 2, 6);
+  return __ret;
+}
+#else
+__ai poly16x4_t vtrn1_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  poly16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 2, 6);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vtrn1q_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30);
+  return __ret;
+}
+#else
+__ai poly8x16_t vtrn1q_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vtrn1q_p64(poly64x2_t __p0, poly64x2_t __p1) {
+  poly64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai poly64x2_t vtrn1q_p64(poly64x2_t __p0, poly64x2_t __p1) {
+  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  poly64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vtrn1q_p16(poly16x8_t __p0, poly16x8_t __p1) {
+  poly16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 2, 10, 4, 12, 6, 14);
+  return __ret;
+}
+#else
+__ai poly16x8_t vtrn1q_p16(poly16x8_t __p0, poly16x8_t __p1) {
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 2, 10, 4, 12, 6, 14);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vtrn1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30);
+  return __ret;
+}
+#else
+__ai uint8x16_t vtrn1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vtrn1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 2, 6);
+  return __ret;
+}
+#else
+__ai uint32x4_t vtrn1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 2, 6);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vtrn1q_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai uint64x2_t vtrn1q_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vtrn1q_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 2, 10, 4, 12, 6, 14);
+  return __ret;
+}
+#else
+__ai uint16x8_t vtrn1q_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 2, 10, 4, 12, 6, 14);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vtrn1q_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30);
+  return __ret;
+}
+#else
+__ai int8x16_t vtrn1q_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vtrn1q_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai float64x2_t vtrn1q_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vtrn1q_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 2, 6);
+  return __ret;
+}
+#else
+__ai float32x4_t vtrn1q_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 2, 6);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vtrn1q_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 2, 6);
+  return __ret;
+}
+#else
+__ai int32x4_t vtrn1q_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 2, 6);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vtrn1q_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai int64x2_t vtrn1q_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vtrn1q_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 2, 10, 4, 12, 6, 14);
+  return __ret;
+}
+#else
+__ai int16x8_t vtrn1q_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 2, 10, 4, 12, 6, 14);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vtrn1_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 2, 10, 4, 12, 6, 14);
+  return __ret;
+}
+#else
+__ai uint8x8_t vtrn1_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 2, 10, 4, 12, 6, 14);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vtrn1_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai uint32x2_t vtrn1_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vtrn1_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 2, 6);
+  return __ret;
+}
+#else
+__ai uint16x4_t vtrn1_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 2, 6);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vtrn1_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 2, 10, 4, 12, 6, 14);
+  return __ret;
+}
+#else
+__ai int8x8_t vtrn1_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 2, 10, 4, 12, 6, 14);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vtrn1_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai float32x2_t vtrn1_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vtrn1_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vtrn1_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vtrn1_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 2, 6);
+  return __ret;
+}
+#else
+__ai int16x4_t vtrn1_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 2, 6);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vtrn2_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 9, 3, 11, 5, 13, 7, 15);
+  return __ret;
+}
+#else
+__ai poly8x8_t vtrn2_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 9, 3, 11, 5, 13, 7, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vtrn2_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 5, 3, 7);
+  return __ret;
+}
+#else
+__ai poly16x4_t vtrn2_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  poly16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 5, 3, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vtrn2q_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31);
+  return __ret;
+}
+#else
+__ai poly8x16_t vtrn2q_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vtrn2q_p64(poly64x2_t __p0, poly64x2_t __p1) {
+  poly64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai poly64x2_t vtrn2q_p64(poly64x2_t __p0, poly64x2_t __p1) {
+  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  poly64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vtrn2q_p16(poly16x8_t __p0, poly16x8_t __p1) {
+  poly16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 9, 3, 11, 5, 13, 7, 15);
+  return __ret;
+}
+#else
+__ai poly16x8_t vtrn2q_p16(poly16x8_t __p0, poly16x8_t __p1) {
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 9, 3, 11, 5, 13, 7, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vtrn2q_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31);
+  return __ret;
+}
+#else
+__ai uint8x16_t vtrn2q_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vtrn2q_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 5, 3, 7);
+  return __ret;
+}
+#else
+__ai uint32x4_t vtrn2q_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 5, 3, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vtrn2q_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai uint64x2_t vtrn2q_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vtrn2q_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 9, 3, 11, 5, 13, 7, 15);
+  return __ret;
+}
+#else
+__ai uint16x8_t vtrn2q_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 9, 3, 11, 5, 13, 7, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vtrn2q_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31);
+  return __ret;
+}
+#else
+__ai int8x16_t vtrn2q_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vtrn2q_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai float64x2_t vtrn2q_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vtrn2q_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 5, 3, 7);
+  return __ret;
+}
+#else
+__ai float32x4_t vtrn2q_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 5, 3, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vtrn2q_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 5, 3, 7);
+  return __ret;
+}
+#else
+__ai int32x4_t vtrn2q_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 5, 3, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vtrn2q_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai int64x2_t vtrn2q_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vtrn2q_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 9, 3, 11, 5, 13, 7, 15);
+  return __ret;
+}
+#else
+__ai int16x8_t vtrn2q_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 9, 3, 11, 5, 13, 7, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vtrn2_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 9, 3, 11, 5, 13, 7, 15);
+  return __ret;
+}
+#else
+__ai uint8x8_t vtrn2_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 9, 3, 11, 5, 13, 7, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vtrn2_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai uint32x2_t vtrn2_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vtrn2_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 5, 3, 7);
+  return __ret;
+}
+#else
+__ai uint16x4_t vtrn2_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 5, 3, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vtrn2_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 9, 3, 11, 5, 13, 7, 15);
+  return __ret;
+}
+#else
+__ai int8x8_t vtrn2_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 9, 3, 11, 5, 13, 7, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vtrn2_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai float32x2_t vtrn2_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vtrn2_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai int32x2_t vtrn2_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vtrn2_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 5, 3, 7);
+  return __ret;
+}
+#else
+__ai int16x4_t vtrn2_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 5, 3, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vtst_p64(poly64x1_t __p0, poly64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vtst_p64(poly64x1_t __p0, poly64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vtstq_p64(poly64x2_t __p0, poly64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vtstq_p64(poly64x2_t __p0, poly64x2_t __p1) {
+  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vtstq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vtstq_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vtstq_s64(int64x2_t __p0, int64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
+  return __ret;
+}
+#else
+__ai uint64x2_t vtstq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vtst_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vtst_u64(uint64x1_t __p0, uint64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x1_t vtst_s64(int64x1_t __p0, int64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#else
+__ai uint64x1_t vtst_s64(int64x1_t __p0, int64x1_t __p1) {
+  uint64x1_t __ret;
+  __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64_t vtstd_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vtstd_u64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai uint64_t vtstd_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vtstd_u64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vtstd_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vtstd_s64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int64_t vtstd_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vtstd_s64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8_t vuqaddb_s8(int8_t __p0, int8_t __p1) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vuqaddb_s8(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int8_t vuqaddb_s8(int8_t __p0, int8_t __p1) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vuqaddb_s8(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32_t vuqadds_s32(int32_t __p0, int32_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vuqadds_s32(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int32_t vuqadds_s32(int32_t __p0, int32_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vuqadds_s32(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64_t vuqaddd_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vuqaddd_s64(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int64_t vuqaddd_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vuqaddd_s64(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16_t vuqaddh_s16(int16_t __p0, int16_t __p1) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vuqaddh_s16(__p0, __p1);
+  return __ret;
+}
+#else
+__ai int16_t vuqaddh_s16(int16_t __p0, int16_t __p1) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vuqaddh_s16(__p0, __p1);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vuqaddq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vuqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
+  return __ret;
+}
+#else
+__ai int8x16_t vuqaddq_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = (int8x16_t) __builtin_neon_vuqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vuqaddq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vuqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
+  return __ret;
+}
+#else
+__ai int32x4_t vuqaddq_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t) __builtin_neon_vuqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vuqaddq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vuqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
+  return __ret;
+}
+#else
+__ai int64x2_t vuqaddq_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t) __builtin_neon_vuqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vuqaddq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vuqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
+  return __ret;
+}
+#else
+__ai int16x8_t vuqaddq_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t) __builtin_neon_vuqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vuqadd_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vuqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
+  return __ret;
+}
+#else
+__ai int8x8_t vuqadd_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = (int8x8_t) __builtin_neon_vuqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vuqadd_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vuqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vuqadd_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = (int32x2_t) __builtin_neon_vuqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x1_t vuqadd_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vuqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
+  return __ret;
+}
+#else
+__ai int64x1_t vuqadd_s64(int64x1_t __p0, int64x1_t __p1) {
+  int64x1_t __ret;
+  __ret = (int64x1_t) __builtin_neon_vuqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vuqadd_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vuqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
+  return __ret;
+}
+#else
+__ai int16x4_t vuqadd_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = (int16x4_t) __builtin_neon_vuqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vuzp1_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14);
+  return __ret;
+}
+#else
+__ai poly8x8_t vuzp1_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vuzp1_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6);
+  return __ret;
+}
+#else
+__ai poly16x4_t vuzp1_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  poly16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vuzp1q_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30);
+  return __ret;
+}
+#else
+__ai poly8x16_t vuzp1q_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vuzp1q_p64(poly64x2_t __p0, poly64x2_t __p1) {
+  poly64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai poly64x2_t vuzp1q_p64(poly64x2_t __p0, poly64x2_t __p1) {
+  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  poly64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vuzp1q_p16(poly16x8_t __p0, poly16x8_t __p1) {
+  poly16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14);
+  return __ret;
+}
+#else
+__ai poly16x8_t vuzp1q_p16(poly16x8_t __p0, poly16x8_t __p1) {
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vuzp1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30);
+  return __ret;
+}
+#else
+__ai uint8x16_t vuzp1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vuzp1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6);
+  return __ret;
+}
+#else
+__ai uint32x4_t vuzp1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vuzp1q_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai uint64x2_t vuzp1q_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vuzp1q_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14);
+  return __ret;
+}
+#else
+__ai uint16x8_t vuzp1q_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vuzp1q_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30);
+  return __ret;
+}
+#else
+__ai int8x16_t vuzp1q_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vuzp1q_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai float64x2_t vuzp1q_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vuzp1q_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6);
+  return __ret;
+}
+#else
+__ai float32x4_t vuzp1q_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vuzp1q_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6);
+  return __ret;
+}
+#else
+__ai int32x4_t vuzp1q_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vuzp1q_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai int64x2_t vuzp1q_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vuzp1q_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14);
+  return __ret;
+}
+#else
+__ai int16x8_t vuzp1q_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vuzp1_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14);
+  return __ret;
+}
+#else
+__ai uint8x8_t vuzp1_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vuzp1_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai uint32x2_t vuzp1_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vuzp1_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6);
+  return __ret;
+}
+#else
+__ai uint16x4_t vuzp1_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vuzp1_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14);
+  return __ret;
+}
+#else
+__ai int8x8_t vuzp1_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vuzp1_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai float32x2_t vuzp1_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vuzp1_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vuzp1_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vuzp1_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6);
+  return __ret;
+}
+#else
+__ai int16x4_t vuzp1_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vuzp2_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15);
+  return __ret;
+}
+#else
+__ai poly8x8_t vuzp2_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vuzp2_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7);
+  return __ret;
+}
+#else
+__ai poly16x4_t vuzp2_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  poly16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vuzp2q_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31);
+  return __ret;
+}
+#else
+__ai poly8x16_t vuzp2q_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vuzp2q_p64(poly64x2_t __p0, poly64x2_t __p1) {
+  poly64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai poly64x2_t vuzp2q_p64(poly64x2_t __p0, poly64x2_t __p1) {
+  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  poly64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vuzp2q_p16(poly16x8_t __p0, poly16x8_t __p1) {
+  poly16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15);
+  return __ret;
+}
+#else
+__ai poly16x8_t vuzp2q_p16(poly16x8_t __p0, poly16x8_t __p1) {
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vuzp2q_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31);
+  return __ret;
+}
+#else
+__ai uint8x16_t vuzp2q_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vuzp2q_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7);
+  return __ret;
+}
+#else
+__ai uint32x4_t vuzp2q_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vuzp2q_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai uint64x2_t vuzp2q_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vuzp2q_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15);
+  return __ret;
+}
+#else
+__ai uint16x8_t vuzp2q_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vuzp2q_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31);
+  return __ret;
+}
+#else
+__ai int8x16_t vuzp2q_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vuzp2q_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai float64x2_t vuzp2q_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vuzp2q_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7);
+  return __ret;
+}
+#else
+__ai float32x4_t vuzp2q_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vuzp2q_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7);
+  return __ret;
+}
+#else
+__ai int32x4_t vuzp2q_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vuzp2q_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai int64x2_t vuzp2q_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vuzp2q_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15);
+  return __ret;
+}
+#else
+__ai int16x8_t vuzp2q_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vuzp2_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15);
+  return __ret;
+}
+#else
+__ai uint8x8_t vuzp2_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vuzp2_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai uint32x2_t vuzp2_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vuzp2_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7);
+  return __ret;
+}
+#else
+__ai uint16x4_t vuzp2_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vuzp2_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15);
+  return __ret;
+}
+#else
+__ai int8x8_t vuzp2_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vuzp2_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai float32x2_t vuzp2_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vuzp2_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai int32x2_t vuzp2_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vuzp2_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7);
+  return __ret;
+}
+#else
+__ai int16x4_t vuzp2_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vzip1_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 1, 9, 2, 10, 3, 11);
+  return __ret;
+}
+#else
+__ai poly8x8_t vzip1_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 1, 9, 2, 10, 3, 11);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vzip1_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 1, 5);
+  return __ret;
+}
+#else
+__ai poly16x4_t vzip1_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  poly16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 1, 5);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vzip1q_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23);
+  return __ret;
+}
+#else
+__ai poly8x16_t vzip1q_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vzip1q_p64(poly64x2_t __p0, poly64x2_t __p1) {
+  poly64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai poly64x2_t vzip1q_p64(poly64x2_t __p0, poly64x2_t __p1) {
+  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  poly64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vzip1q_p16(poly16x8_t __p0, poly16x8_t __p1) {
+  poly16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 1, 9, 2, 10, 3, 11);
+  return __ret;
+}
+#else
+__ai poly16x8_t vzip1q_p16(poly16x8_t __p0, poly16x8_t __p1) {
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 1, 9, 2, 10, 3, 11);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vzip1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23);
+  return __ret;
+}
+#else
+__ai uint8x16_t vzip1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vzip1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 1, 5);
+  return __ret;
+}
+#else
+__ai uint32x4_t vzip1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 1, 5);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vzip1q_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai uint64x2_t vzip1q_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vzip1q_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 1, 9, 2, 10, 3, 11);
+  return __ret;
+}
+#else
+__ai uint16x8_t vzip1q_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 1, 9, 2, 10, 3, 11);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vzip1q_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23);
+  return __ret;
+}
+#else
+__ai int8x16_t vzip1q_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vzip1q_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai float64x2_t vzip1q_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vzip1q_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 1, 5);
+  return __ret;
+}
+#else
+__ai float32x4_t vzip1q_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 1, 5);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vzip1q_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 1, 5);
+  return __ret;
+}
+#else
+__ai int32x4_t vzip1q_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 1, 5);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vzip1q_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai int64x2_t vzip1q_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vzip1q_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 1, 9, 2, 10, 3, 11);
+  return __ret;
+}
+#else
+__ai int16x8_t vzip1q_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 1, 9, 2, 10, 3, 11);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vzip1_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 1, 9, 2, 10, 3, 11);
+  return __ret;
+}
+#else
+__ai uint8x8_t vzip1_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 1, 9, 2, 10, 3, 11);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vzip1_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai uint32x2_t vzip1_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vzip1_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 1, 5);
+  return __ret;
+}
+#else
+__ai uint16x4_t vzip1_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 1, 5);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vzip1_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 1, 9, 2, 10, 3, 11);
+  return __ret;
+}
+#else
+__ai int8x8_t vzip1_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 1, 9, 2, 10, 3, 11);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vzip1_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai float32x2_t vzip1_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vzip1_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
+  return __ret;
+}
+#else
+__ai int32x2_t vzip1_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vzip1_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 1, 5);
+  return __ret;
+}
+#else
+__ai int16x4_t vzip1_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 1, 5);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x8_t vzip2_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 4, 12, 5, 13, 6, 14, 7, 15);
+  return __ret;
+}
+#else
+__ai poly8x8_t vzip2_p8(poly8x8_t __p0, poly8x8_t __p1) {
+  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 4, 12, 5, 13, 6, 14, 7, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x4_t vzip2_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 2, 6, 3, 7);
+  return __ret;
+}
+#else
+__ai poly16x4_t vzip2_p16(poly16x4_t __p0, poly16x4_t __p1) {
+  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  poly16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 2, 6, 3, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly8x16_t vzip2q_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31);
+  return __ret;
+}
+#else
+__ai poly8x16_t vzip2q_p8(poly8x16_t __p0, poly8x16_t __p1) {
+  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly64x2_t vzip2q_p64(poly64x2_t __p0, poly64x2_t __p1) {
+  poly64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai poly64x2_t vzip2q_p64(poly64x2_t __p0, poly64x2_t __p1) {
+  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  poly64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai poly16x8_t vzip2q_p16(poly16x8_t __p0, poly16x8_t __p1) {
+  poly16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 4, 12, 5, 13, 6, 14, 7, 15);
+  return __ret;
+}
+#else
+__ai poly16x8_t vzip2q_p16(poly16x8_t __p0, poly16x8_t __p1) {
+  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  poly16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 4, 12, 5, 13, 6, 14, 7, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vzip2q_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31);
+  return __ret;
+}
+#else
+__ai uint8x16_t vzip2q_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vzip2q_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 2, 6, 3, 7);
+  return __ret;
+}
+#else
+__ai uint32x4_t vzip2q_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 2, 6, 3, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vzip2q_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai uint64x2_t vzip2q_u64(uint64x2_t __p0, uint64x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vzip2q_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 4, 12, 5, 13, 6, 14, 7, 15);
+  return __ret;
+}
+#else
+__ai uint16x8_t vzip2q_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 4, 12, 5, 13, 6, 14, 7, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vzip2q_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31);
+  return __ret;
+}
+#else
+__ai int8x16_t vzip2q_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float64x2_t vzip2q_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai float64x2_t vzip2q_f64(float64x2_t __p0, float64x2_t __p1) {
+  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x4_t vzip2q_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 2, 6, 3, 7);
+  return __ret;
+}
+#else
+__ai float32x4_t vzip2q_f32(float32x4_t __p0, float32x4_t __p1) {
+  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  float32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 2, 6, 3, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vzip2q_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 2, 6, 3, 7);
+  return __ret;
+}
+#else
+__ai int32x4_t vzip2q_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 2, 6, 3, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vzip2q_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai int64x2_t vzip2q_s64(int64x2_t __p0, int64x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vzip2q_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 4, 12, 5, 13, 6, 14, 7, 15);
+  return __ret;
+}
+#else
+__ai int16x8_t vzip2q_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 4, 12, 5, 13, 6, 14, 7, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vzip2_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 4, 12, 5, 13, 6, 14, 7, 15);
+  return __ret;
+}
+#else
+__ai uint8x8_t vzip2_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 4, 12, 5, 13, 6, 14, 7, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vzip2_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai uint32x2_t vzip2_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vzip2_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 2, 6, 3, 7);
+  return __ret;
+}
+#else
+__ai uint16x4_t vzip2_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 2, 6, 3, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vzip2_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 4, 12, 5, 13, 6, 14, 7, 15);
+  return __ret;
+}
+#else
+__ai int8x8_t vzip2_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 4, 12, 5, 13, 6, 14, 7, 15);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai float32x2_t vzip2_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai float32x2_t vzip2_f32(float32x2_t __p0, float32x2_t __p1) {
+  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  float32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vzip2_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
+  return __ret;
+}
+#else
+__ai int32x2_t vzip2_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vzip2_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__p0, __p1, 2, 6, 3, 7);
+  return __ret;
+}
+#else
+__ai int16x4_t vzip2_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __builtin_shufflevector(__rev0, __rev1, 2, 6, 3, 7);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#endif
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x16_t vabaq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
+  uint8x16_t __ret;
+  __ret = __p0 + vabdq_u8(__p1, __p2);
+  return __ret;
+}
+#else
+__ai uint8x16_t vabaq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __ret;
+  __ret = __rev0 + __noswap_vabdq_u8(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vabaq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint32x4_t __ret;
+  __ret = __p0 + vabdq_u32(__p1, __p2);
+  return __ret;
+}
+#else
+__ai uint32x4_t vabaq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 + __noswap_vabdq_u32(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vabaq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint16x8_t __ret;
+  __ret = __p0 + vabdq_u16(__p1, __p2);
+  return __ret;
+}
+#else
+__ai uint16x8_t vabaq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 + __noswap_vabdq_u16(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x16_t vabaq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
+  int8x16_t __ret;
+  __ret = __p0 + vabdq_s8(__p1, __p2);
+  return __ret;
+}
+#else
+__ai int8x16_t vabaq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __ret;
+  __ret = __rev0 + __noswap_vabdq_s8(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vabaq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int32x4_t __ret;
+  __ret = __p0 + vabdq_s32(__p1, __p2);
+  return __ret;
+}
+#else
+__ai int32x4_t vabaq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 + __noswap_vabdq_s32(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vabaq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int16x8_t __ret;
+  __ret = __p0 + vabdq_s16(__p1, __p2);
+  return __ret;
+}
+#else
+__ai int16x8_t vabaq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 + __noswap_vabdq_s16(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint8x8_t vaba_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
+  uint8x8_t __ret;
+  __ret = __p0 + vabd_u8(__p1, __p2);
+  return __ret;
+}
+#else
+__ai uint8x8_t vaba_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __ret;
+  __ret = __rev0 + __noswap_vabd_u8(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x2_t vaba_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
+  uint32x2_t __ret;
+  __ret = __p0 + vabd_u32(__p1, __p2);
+  return __ret;
+}
+#else
+__ai uint32x2_t vaba_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  uint32x2_t __ret;
+  __ret = __rev0 + __noswap_vabd_u32(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x4_t vaba_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
+  uint16x4_t __ret;
+  __ret = __p0 + vabd_u16(__p1, __p2);
+  return __ret;
+}
+#else
+__ai uint16x4_t vaba_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint16x4_t __ret;
+  __ret = __rev0 + __noswap_vabd_u16(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int8x8_t vaba_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
+  int8x8_t __ret;
+  __ret = __p0 + vabd_s8(__p1, __p2);
+  return __ret;
+}
+#else
+__ai int8x8_t vaba_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __ret;
+  __ret = __rev0 + __noswap_vabd_s8(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x2_t vaba_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int32x2_t __ret;
+  __ret = __p0 + vabd_s32(__p1, __p2);
+  return __ret;
+}
+#else
+__ai int32x2_t vaba_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  int32x2_t __ret;
+  __ret = __rev0 + __noswap_vabd_s32(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x4_t vaba_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int16x4_t __ret;
+  __ret = __p0 + vabd_s16(__p1, __p2);
+  return __ret;
+}
+#else
+__ai int16x4_t vaba_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int16x4_t __ret;
+  __ret = __rev0 + __noswap_vabd_s16(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vabdl_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(vmovl_u8((uint8x8_t)(vabd_u8(__p0, __p1))));
+  return __ret;
+}
+#else
+__ai uint16x8_t vabdl_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__noswap_vmovl_u8((uint8x8_t)(__noswap_vabd_u8(__rev0, __rev1))));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint16x8_t __noswap_vabdl_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = (uint16x8_t)(__noswap_vmovl_u8((uint8x8_t)(__noswap_vabd_u8(__p0, __p1))));
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vabdl_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(vmovl_u32((uint32x2_t)(vabd_u32(__p0, __p1))));
+  return __ret;
+}
+#else
+__ai uint64x2_t vabdl_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__noswap_vmovl_u32((uint32x2_t)(__noswap_vabd_u32(__rev0, __rev1))));
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai uint64x2_t __noswap_vabdl_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = (uint64x2_t)(__noswap_vmovl_u32((uint32x2_t)(__noswap_vabd_u32(__p0, __p1))));
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vabdl_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(vmovl_u16((uint16x4_t)(vabd_u16(__p0, __p1))));
+  return __ret;
+}
+#else
+__ai uint32x4_t vabdl_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__noswap_vmovl_u16((uint16x4_t)(__noswap_vabd_u16(__rev0, __rev1))));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint32x4_t __noswap_vabdl_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = (uint32x4_t)(__noswap_vmovl_u16((uint16x4_t)(__noswap_vabd_u16(__p0, __p1))));
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vabdl_s8(int8x8_t __p0, int8x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(vmovl_u8((uint8x8_t)(vabd_s8(__p0, __p1))));
+  return __ret;
+}
+#else
+__ai int16x8_t vabdl_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__noswap_vmovl_u8((uint8x8_t)(__noswap_vabd_s8(__rev0, __rev1))));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int16x8_t __noswap_vabdl_s8(int8x8_t __p0, int8x8_t __p1) {
+  int16x8_t __ret;
+  __ret = (int16x8_t)(__noswap_vmovl_u8((uint8x8_t)(__noswap_vabd_s8(__p0, __p1))));
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vabdl_s32(int32x2_t __p0, int32x2_t __p1) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(vmovl_u32((uint32x2_t)(vabd_s32(__p0, __p1))));
+  return __ret;
+}
+#else
+__ai int64x2_t vabdl_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__noswap_vmovl_u32((uint32x2_t)(__noswap_vabd_s32(__rev0, __rev1))));
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int64x2_t __noswap_vabdl_s32(int32x2_t __p0, int32x2_t __p1) {
+  int64x2_t __ret;
+  __ret = (int64x2_t)(__noswap_vmovl_u32((uint32x2_t)(__noswap_vabd_s32(__p0, __p1))));
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vabdl_s16(int16x4_t __p0, int16x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(vmovl_u16((uint16x4_t)(vabd_s16(__p0, __p1))));
+  return __ret;
+}
+#else
+__ai int32x4_t vabdl_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__noswap_vmovl_u16((uint16x4_t)(__noswap_vabd_s16(__rev0, __rev1))));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int32x4_t __noswap_vabdl_s16(int16x4_t __p0, int16x4_t __p1) {
+  int32x4_t __ret;
+  __ret = (int32x4_t)(__noswap_vmovl_u16((uint16x4_t)(__noswap_vabd_s16(__p0, __p1))));
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vaddl_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = vmovl_u8(__p0) + vmovl_u8(__p1);
+  return __ret;
+}
+#else
+__ai uint16x8_t vaddl_u8(uint8x8_t __p0, uint8x8_t __p1) {
+  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __noswap_vmovl_u8(__rev0) + __noswap_vmovl_u8(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vaddl_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = vmovl_u32(__p0) + vmovl_u32(__p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vaddl_u32(uint32x2_t __p0, uint32x2_t __p1) {
+  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = __noswap_vmovl_u32(__rev0) + __noswap_vmovl_u32(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vaddl_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = vmovl_u16(__p0) + vmovl_u16(__p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vaddl_u16(uint16x4_t __p0, uint16x4_t __p1) {
+  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __noswap_vmovl_u16(__rev0) + __noswap_vmovl_u16(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vaddl_s8(int8x8_t __p0, int8x8_t __p1) {
+  int16x8_t __ret;
+  __ret = vmovl_s8(__p0) + vmovl_s8(__p1);
+  return __ret;
+}
+#else
+__ai int16x8_t vaddl_s8(int8x8_t __p0, int8x8_t __p1) {
+  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __noswap_vmovl_s8(__rev0) + __noswap_vmovl_s8(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vaddl_s32(int32x2_t __p0, int32x2_t __p1) {
+  int64x2_t __ret;
+  __ret = vmovl_s32(__p0) + vmovl_s32(__p1);
+  return __ret;
+}
+#else
+__ai int64x2_t vaddl_s32(int32x2_t __p0, int32x2_t __p1) {
+  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = __noswap_vmovl_s32(__rev0) + __noswap_vmovl_s32(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vaddl_s16(int16x4_t __p0, int16x4_t __p1) {
+  int32x4_t __ret;
+  __ret = vmovl_s16(__p0) + vmovl_s16(__p1);
+  return __ret;
+}
+#else
+__ai int32x4_t vaddl_s16(int16x4_t __p0, int16x4_t __p1) {
+  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vmovl_s16(__rev0) + __noswap_vmovl_s16(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vaddw_u8(uint16x8_t __p0, uint8x8_t __p1) {
+  uint16x8_t __ret;
+  __ret = __p0 + vmovl_u8(__p1);
+  return __ret;
+}
+#else
+__ai uint16x8_t vaddw_u8(uint16x8_t __p0, uint8x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 + __noswap_vmovl_u8(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vaddw_u32(uint64x2_t __p0, uint32x2_t __p1) {
+  uint64x2_t __ret;
+  __ret = __p0 + vmovl_u32(__p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vaddw_u32(uint64x2_t __p0, uint32x2_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = __rev0 + __noswap_vmovl_u32(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vaddw_u16(uint32x4_t __p0, uint16x4_t __p1) {
+  uint32x4_t __ret;
+  __ret = __p0 + vmovl_u16(__p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vaddw_u16(uint32x4_t __p0, uint16x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 + __noswap_vmovl_u16(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vaddw_s8(int16x8_t __p0, int8x8_t __p1) {
+  int16x8_t __ret;
+  __ret = __p0 + vmovl_s8(__p1);
+  return __ret;
+}
+#else
+__ai int16x8_t vaddw_s8(int16x8_t __p0, int8x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 + __noswap_vmovl_s8(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vaddw_s32(int64x2_t __p0, int32x2_t __p1) {
+  int64x2_t __ret;
+  __ret = __p0 + vmovl_s32(__p1);
+  return __ret;
+}
+#else
+__ai int64x2_t vaddw_s32(int64x2_t __p0, int32x2_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = __rev0 + __noswap_vmovl_s32(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vaddw_s16(int32x4_t __p0, int16x4_t __p1) {
+  int32x4_t __ret;
+  __ret = __p0 + vmovl_s16(__p1);
+  return __ret;
+}
+#else
+__ai int32x4_t vaddw_s16(int32x4_t __p0, int16x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 + __noswap_vmovl_s16(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vmlal_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
+  uint16x8_t __ret;
+  __ret = __p0 + vmull_u8(__p1, __p2);
+  return __ret;
+}
+#else
+__ai uint16x8_t vmlal_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 + __noswap_vmull_u8(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint16x8_t __noswap_vmlal_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
+  uint16x8_t __ret;
+  __ret = __p0 + __noswap_vmull_u8(__p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vmlal_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
+  uint64x2_t __ret;
+  __ret = __p0 + vmull_u32(__p1, __p2);
+  return __ret;
+}
+#else
+__ai uint64x2_t vmlal_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  uint64x2_t __ret;
+  __ret = __rev0 + __noswap_vmull_u32(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai uint64x2_t __noswap_vmlal_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
+  uint64x2_t __ret;
+  __ret = __p0 + __noswap_vmull_u32(__p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmlal_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
+  uint32x4_t __ret;
+  __ret = __p0 + vmull_u16(__p1, __p2);
+  return __ret;
+}
+#else
+__ai uint32x4_t vmlal_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 + __noswap_vmull_u16(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint32x4_t __noswap_vmlal_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
+  uint32x4_t __ret;
+  __ret = __p0 + __noswap_vmull_u16(__p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vmlal_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
+  int16x8_t __ret;
+  __ret = __p0 + vmull_s8(__p1, __p2);
+  return __ret;
+}
+#else
+__ai int16x8_t vmlal_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 + __noswap_vmull_s8(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int16x8_t __noswap_vmlal_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
+  int16x8_t __ret;
+  __ret = __p0 + __noswap_vmull_s8(__p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vmlal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int64x2_t __ret;
+  __ret = __p0 + vmull_s32(__p1, __p2);
+  return __ret;
+}
+#else
+__ai int64x2_t vmlal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  int64x2_t __ret;
+  __ret = __rev0 + __noswap_vmull_s32(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int64x2_t __noswap_vmlal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int64x2_t __ret;
+  __ret = __p0 + __noswap_vmull_s32(__p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int32x4_t __ret;
+  __ret = __p0 + vmull_s16(__p1, __p2);
+  return __ret;
+}
+#else
+__ai int32x4_t vmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 + __noswap_vmull_s16(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int32x4_t __noswap_vmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int32x4_t __ret;
+  __ret = __p0 + __noswap_vmull_s16(__p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlal_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __s2 = __p2; \
+  uint64x2_t __ret; \
+  __ret = __s0 + vmull_u32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlal_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __s2 = __p2; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = __rev0 + __noswap_vmull_u32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlal_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __s2 = __p2; \
+  uint32x4_t __ret; \
+  __ret = __s0 + vmull_u16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlal_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __s2 = __p2; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = __rev0 + __noswap_vmull_u16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlal_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int64x2_t __ret; \
+  __ret = __s0 + vmull_s32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlal_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __rev0 + __noswap_vmull_s32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlal_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int32x4_t __ret; \
+  __ret = __s0 + vmull_s16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlal_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __rev0 + __noswap_vmull_s16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vmlal_n_u32(uint64x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
+  uint64x2_t __ret;
+  __ret = __p0 + vmull_u32(__p1, (uint32x2_t) {__p2, __p2});
+  return __ret;
+}
+#else
+__ai uint64x2_t vmlal_n_u32(uint64x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = __rev0 + __noswap_vmull_u32(__rev1, (uint32x2_t) {__p2, __p2});
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai uint64x2_t __noswap_vmlal_n_u32(uint64x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
+  uint64x2_t __ret;
+  __ret = __p0 + __noswap_vmull_u32(__p1, (uint32x2_t) {__p2, __p2});
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmlal_n_u16(uint32x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
+  uint32x4_t __ret;
+  __ret = __p0 + vmull_u16(__p1, (uint16x4_t) {__p2, __p2, __p2, __p2});
+  return __ret;
+}
+#else
+__ai uint32x4_t vmlal_n_u16(uint32x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 + __noswap_vmull_u16(__rev1, (uint16x4_t) {__p2, __p2, __p2, __p2});
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint32x4_t __noswap_vmlal_n_u16(uint32x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
+  uint32x4_t __ret;
+  __ret = __p0 + __noswap_vmull_u16(__p1, (uint16x4_t) {__p2, __p2, __p2, __p2});
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vmlal_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
+  int64x2_t __ret;
+  __ret = __p0 + vmull_s32(__p1, (int32x2_t) {__p2, __p2});
+  return __ret;
+}
+#else
+__ai int64x2_t vmlal_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = __rev0 + __noswap_vmull_s32(__rev1, (int32x2_t) {__p2, __p2});
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int64x2_t __noswap_vmlal_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
+  int64x2_t __ret;
+  __ret = __p0 + __noswap_vmull_s32(__p1, (int32x2_t) {__p2, __p2});
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmlal_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
+  int32x4_t __ret;
+  __ret = __p0 + vmull_s16(__p1, (int16x4_t) {__p2, __p2, __p2, __p2});
+  return __ret;
+}
+#else
+__ai int32x4_t vmlal_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 + __noswap_vmull_s16(__rev1, (int16x4_t) {__p2, __p2, __p2, __p2});
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int32x4_t __noswap_vmlal_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
+  int32x4_t __ret;
+  __ret = __p0 + __noswap_vmull_s16(__p1, (int16x4_t) {__p2, __p2, __p2, __p2});
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vmlsl_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
+  uint16x8_t __ret;
+  __ret = __p0 - vmull_u8(__p1, __p2);
+  return __ret;
+}
+#else
+__ai uint16x8_t vmlsl_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 - __noswap_vmull_u8(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint16x8_t __noswap_vmlsl_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
+  uint16x8_t __ret;
+  __ret = __p0 - __noswap_vmull_u8(__p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vmlsl_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
+  uint64x2_t __ret;
+  __ret = __p0 - vmull_u32(__p1, __p2);
+  return __ret;
+}
+#else
+__ai uint64x2_t vmlsl_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  uint64x2_t __ret;
+  __ret = __rev0 - __noswap_vmull_u32(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai uint64x2_t __noswap_vmlsl_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
+  uint64x2_t __ret;
+  __ret = __p0 - __noswap_vmull_u32(__p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmlsl_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
+  uint32x4_t __ret;
+  __ret = __p0 - vmull_u16(__p1, __p2);
+  return __ret;
+}
+#else
+__ai uint32x4_t vmlsl_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 - __noswap_vmull_u16(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint32x4_t __noswap_vmlsl_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
+  uint32x4_t __ret;
+  __ret = __p0 - __noswap_vmull_u16(__p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vmlsl_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
+  int16x8_t __ret;
+  __ret = __p0 - vmull_s8(__p1, __p2);
+  return __ret;
+}
+#else
+__ai int16x8_t vmlsl_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 - __noswap_vmull_s8(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int16x8_t __noswap_vmlsl_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
+  int16x8_t __ret;
+  __ret = __p0 - __noswap_vmull_s8(__p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vmlsl_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int64x2_t __ret;
+  __ret = __p0 - vmull_s32(__p1, __p2);
+  return __ret;
+}
+#else
+__ai int64x2_t vmlsl_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  int64x2_t __ret;
+  __ret = __rev0 - __noswap_vmull_s32(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int64x2_t __noswap_vmlsl_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int64x2_t __ret;
+  __ret = __p0 - __noswap_vmull_s32(__p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int32x4_t __ret;
+  __ret = __p0 - vmull_s16(__p1, __p2);
+  return __ret;
+}
+#else
+__ai int32x4_t vmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 - __noswap_vmull_s16(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int32x4_t __noswap_vmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int32x4_t __ret;
+  __ret = __p0 - __noswap_vmull_s16(__p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsl_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __s2 = __p2; \
+  uint64x2_t __ret; \
+  __ret = __s0 - vmull_u32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlsl_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint64x2_t __s0 = __p0; \
+  uint32x2_t __s1 = __p1; \
+  uint32x2_t __s2 = __p2; \
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  uint64x2_t __ret; \
+  __ret = __rev0 - __noswap_vmull_u32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsl_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __s2 = __p2; \
+  uint32x4_t __ret; \
+  __ret = __s0 - vmull_u16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlsl_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  uint32x4_t __s0 = __p0; \
+  uint16x4_t __s1 = __p1; \
+  uint16x4_t __s2 = __p2; \
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  uint32x4_t __ret; \
+  __ret = __rev0 - __noswap_vmull_u16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsl_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int64x2_t __ret; \
+  __ret = __s0 - vmull_s32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlsl_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int64x2_t __s0 = __p0; \
+  int32x2_t __s1 = __p1; \
+  int32x2_t __s2 = __p2; \
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
+  int64x2_t __ret; \
+  __ret = __rev0 - __noswap_vmull_s32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmlsl_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int32x4_t __ret; \
+  __ret = __s0 - vmull_s16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
+  __ret; \
+})
+#else
+#define vmlsl_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
+  int32x4_t __s0 = __p0; \
+  int16x4_t __s1 = __p1; \
+  int16x4_t __s2 = __p2; \
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
+  int32x4_t __ret; \
+  __ret = __rev0 - __noswap_vmull_s16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
+  __ret; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vmlsl_n_u32(uint64x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
+  uint64x2_t __ret;
+  __ret = __p0 - vmull_u32(__p1, (uint32x2_t) {__p2, __p2});
+  return __ret;
+}
+#else
+__ai uint64x2_t vmlsl_n_u32(uint64x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint64x2_t __ret;
+  __ret = __rev0 - __noswap_vmull_u32(__rev1, (uint32x2_t) {__p2, __p2});
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai uint64x2_t __noswap_vmlsl_n_u32(uint64x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
+  uint64x2_t __ret;
+  __ret = __p0 - __noswap_vmull_u32(__p1, (uint32x2_t) {__p2, __p2});
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmlsl_n_u16(uint32x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
+  uint32x4_t __ret;
+  __ret = __p0 - vmull_u16(__p1, (uint16x4_t) {__p2, __p2, __p2, __p2});
+  return __ret;
+}
+#else
+__ai uint32x4_t vmlsl_n_u16(uint32x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 - __noswap_vmull_u16(__rev1, (uint16x4_t) {__p2, __p2, __p2, __p2});
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint32x4_t __noswap_vmlsl_n_u16(uint32x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
+  uint32x4_t __ret;
+  __ret = __p0 - __noswap_vmull_u16(__p1, (uint16x4_t) {__p2, __p2, __p2, __p2});
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vmlsl_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
+  int64x2_t __ret;
+  __ret = __p0 - vmull_s32(__p1, (int32x2_t) {__p2, __p2});
+  return __ret;
+}
+#else
+__ai int64x2_t vmlsl_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int64x2_t __ret;
+  __ret = __rev0 - __noswap_vmull_s32(__rev1, (int32x2_t) {__p2, __p2});
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int64x2_t __noswap_vmlsl_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
+  int64x2_t __ret;
+  __ret = __p0 - __noswap_vmull_s32(__p1, (int32x2_t) {__p2, __p2});
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
+  int32x4_t __ret;
+  __ret = __p0 - vmull_s16(__p1, (int16x4_t) {__p2, __p2, __p2, __p2});
+  return __ret;
+}
+#else
+__ai int32x4_t vmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 - __noswap_vmull_s16(__rev1, (int16x4_t) {__p2, __p2, __p2, __p2});
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int32x4_t __noswap_vmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
+  int32x4_t __ret;
+  __ret = __p0 - __noswap_vmull_s16(__p1, (int16x4_t) {__p2, __p2, __p2, __p2});
+  return __ret;
+}
+#endif
+
+#if defined(__aarch64__)
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vabdl_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint16x8_t __ret;
+  __ret = vabdl_u8(vget_high_u8(__p0), vget_high_u8(__p1));
+  return __ret;
+}
+#else
+__ai uint16x8_t vabdl_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __noswap_vabdl_u8(__noswap_vget_high_u8(__rev0), __noswap_vget_high_u8(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vabdl_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint64x2_t __ret;
+  __ret = vabdl_u32(vget_high_u32(__p0), vget_high_u32(__p1));
+  return __ret;
+}
+#else
+__ai uint64x2_t vabdl_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint64x2_t __ret;
+  __ret = __noswap_vabdl_u32(__noswap_vget_high_u32(__rev0), __noswap_vget_high_u32(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vabdl_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint32x4_t __ret;
+  __ret = vabdl_u16(vget_high_u16(__p0), vget_high_u16(__p1));
+  return __ret;
+}
+#else
+__ai uint32x4_t vabdl_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __noswap_vabdl_u16(__noswap_vget_high_u16(__rev0), __noswap_vget_high_u16(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vabdl_high_s8(int8x16_t __p0, int8x16_t __p1) {
+  int16x8_t __ret;
+  __ret = vabdl_s8(vget_high_s8(__p0), vget_high_s8(__p1));
+  return __ret;
+}
+#else
+__ai int16x8_t vabdl_high_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __noswap_vabdl_s8(__noswap_vget_high_s8(__rev0), __noswap_vget_high_s8(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vabdl_high_s32(int32x4_t __p0, int32x4_t __p1) {
+  int64x2_t __ret;
+  __ret = vabdl_s32(vget_high_s32(__p0), vget_high_s32(__p1));
+  return __ret;
+}
+#else
+__ai int64x2_t vabdl_high_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int64x2_t __ret;
+  __ret = __noswap_vabdl_s32(__noswap_vget_high_s32(__rev0), __noswap_vget_high_s32(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vabdl_high_s16(int16x8_t __p0, int16x8_t __p1) {
+  int32x4_t __ret;
+  __ret = vabdl_s16(vget_high_s16(__p0), vget_high_s16(__p1));
+  return __ret;
+}
+#else
+__ai int32x4_t vabdl_high_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vabdl_s16(__noswap_vget_high_s16(__rev0), __noswap_vget_high_s16(__rev1));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vaddl_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint16x8_t __ret;
+  __ret = vmovl_high_u8(__p0) + vmovl_high_u8(__p1);
+  return __ret;
+}
+#else
+__ai uint16x8_t vaddl_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
+  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __noswap_vmovl_high_u8(__rev0) + __noswap_vmovl_high_u8(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vaddl_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint64x2_t __ret;
+  __ret = vmovl_high_u32(__p0) + vmovl_high_u32(__p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vaddl_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint64x2_t __ret;
+  __ret = __noswap_vmovl_high_u32(__rev0) + __noswap_vmovl_high_u32(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vaddl_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint32x4_t __ret;
+  __ret = vmovl_high_u16(__p0) + vmovl_high_u16(__p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vaddl_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __noswap_vmovl_high_u16(__rev0) + __noswap_vmovl_high_u16(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vaddl_high_s8(int8x16_t __p0, int8x16_t __p1) {
+  int16x8_t __ret;
+  __ret = vmovl_high_s8(__p0) + vmovl_high_s8(__p1);
+  return __ret;
+}
+#else
+__ai int16x8_t vaddl_high_s8(int8x16_t __p0, int8x16_t __p1) {
+  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __noswap_vmovl_high_s8(__rev0) + __noswap_vmovl_high_s8(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vaddl_high_s32(int32x4_t __p0, int32x4_t __p1) {
+  int64x2_t __ret;
+  __ret = vmovl_high_s32(__p0) + vmovl_high_s32(__p1);
+  return __ret;
+}
+#else
+__ai int64x2_t vaddl_high_s32(int32x4_t __p0, int32x4_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int64x2_t __ret;
+  __ret = __noswap_vmovl_high_s32(__rev0) + __noswap_vmovl_high_s32(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vaddl_high_s16(int16x8_t __p0, int16x8_t __p1) {
+  int32x4_t __ret;
+  __ret = vmovl_high_s16(__p0) + vmovl_high_s16(__p1);
+  return __ret;
+}
+#else
+__ai int32x4_t vaddl_high_s16(int16x8_t __p0, int16x8_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vmovl_high_s16(__rev0) + __noswap_vmovl_high_s16(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vaddw_high_u8(uint16x8_t __p0, uint8x16_t __p1) {
+  uint16x8_t __ret;
+  __ret = __p0 + vmovl_high_u8(__p1);
+  return __ret;
+}
+#else
+__ai uint16x8_t vaddw_high_u8(uint16x8_t __p0, uint8x16_t __p1) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 + __noswap_vmovl_high_u8(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vaddw_high_u32(uint64x2_t __p0, uint32x4_t __p1) {
+  uint64x2_t __ret;
+  __ret = __p0 + vmovl_high_u32(__p1);
+  return __ret;
+}
+#else
+__ai uint64x2_t vaddw_high_u32(uint64x2_t __p0, uint32x4_t __p1) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint64x2_t __ret;
+  __ret = __rev0 + __noswap_vmovl_high_u32(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vaddw_high_u16(uint32x4_t __p0, uint16x8_t __p1) {
+  uint32x4_t __ret;
+  __ret = __p0 + vmovl_high_u16(__p1);
+  return __ret;
+}
+#else
+__ai uint32x4_t vaddw_high_u16(uint32x4_t __p0, uint16x8_t __p1) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 + __noswap_vmovl_high_u16(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vaddw_high_s8(int16x8_t __p0, int8x16_t __p1) {
+  int16x8_t __ret;
+  __ret = __p0 + vmovl_high_s8(__p1);
+  return __ret;
+}
+#else
+__ai int16x8_t vaddw_high_s8(int16x8_t __p0, int8x16_t __p1) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 + __noswap_vmovl_high_s8(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vaddw_high_s32(int64x2_t __p0, int32x4_t __p1) {
+  int64x2_t __ret;
+  __ret = __p0 + vmovl_high_s32(__p1);
+  return __ret;
+}
+#else
+__ai int64x2_t vaddw_high_s32(int64x2_t __p0, int32x4_t __p1) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int64x2_t __ret;
+  __ret = __rev0 + __noswap_vmovl_high_s32(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vaddw_high_s16(int32x4_t __p0, int16x8_t __p1) {
+  int32x4_t __ret;
+  __ret = __p0 + vmovl_high_s16(__p1);
+  return __ret;
+}
+#else
+__ai int32x4_t vaddw_high_s16(int32x4_t __p0, int16x8_t __p1) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 + __noswap_vmovl_high_s16(__rev1);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_lane_p64(__p0_250, __p1_250, __p2_250, __p3_250) __extension__ ({ \
+  poly64x2_t __s0_250 = __p0_250; \
+  poly64x1_t __s2_250 = __p2_250; \
+  poly64x2_t __ret_250; \
+  __ret_250 = vsetq_lane_p64(vget_lane_p64(__s2_250, __p3_250), __s0_250, __p1_250); \
+  __ret_250; \
+})
+#else
+#define vcopyq_lane_p64(__p0_251, __p1_251, __p2_251, __p3_251) __extension__ ({ \
+  poly64x2_t __s0_251 = __p0_251; \
+  poly64x1_t __s2_251 = __p2_251; \
+  poly64x2_t __rev0_251;  __rev0_251 = __builtin_shufflevector(__s0_251, __s0_251, 1, 0); \
+  poly64x2_t __ret_251; \
+  __ret_251 = __noswap_vsetq_lane_p64(__noswap_vget_lane_p64(__s2_251, __p3_251), __rev0_251, __p1_251); \
+  __ret_251 = __builtin_shufflevector(__ret_251, __ret_251, 1, 0); \
+  __ret_251; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_lane_f64(__p0_252, __p1_252, __p2_252, __p3_252) __extension__ ({ \
+  float64x2_t __s0_252 = __p0_252; \
+  float64x1_t __s2_252 = __p2_252; \
+  float64x2_t __ret_252; \
+  __ret_252 = vsetq_lane_f64(vget_lane_f64(__s2_252, __p3_252), __s0_252, __p1_252); \
+  __ret_252; \
+})
+#else
+#define vcopyq_lane_f64(__p0_253, __p1_253, __p2_253, __p3_253) __extension__ ({ \
+  float64x2_t __s0_253 = __p0_253; \
+  float64x1_t __s2_253 = __p2_253; \
+  float64x2_t __rev0_253;  __rev0_253 = __builtin_shufflevector(__s0_253, __s0_253, 1, 0); \
+  float64x2_t __ret_253; \
+  __ret_253 = __noswap_vsetq_lane_f64(__noswap_vget_lane_f64(__s2_253, __p3_253), __rev0_253, __p1_253); \
+  __ret_253 = __builtin_shufflevector(__ret_253, __ret_253, 1, 0); \
+  __ret_253; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_p64(__p0_254, __p1_254, __p2_254, __p3_254) __extension__ ({ \
+  poly64x1_t __s0_254 = __p0_254; \
+  poly64x1_t __s2_254 = __p2_254; \
+  poly64x1_t __ret_254; \
+  __ret_254 = vset_lane_p64(vget_lane_p64(__s2_254, __p3_254), __s0_254, __p1_254); \
+  __ret_254; \
+})
+#else
+#define vcopy_lane_p64(__p0_255, __p1_255, __p2_255, __p3_255) __extension__ ({ \
+  poly64x1_t __s0_255 = __p0_255; \
+  poly64x1_t __s2_255 = __p2_255; \
+  poly64x1_t __ret_255; \
+  __ret_255 = __noswap_vset_lane_p64(__noswap_vget_lane_p64(__s2_255, __p3_255), __s0_255, __p1_255); \
+  __ret_255; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_f64(__p0_256, __p1_256, __p2_256, __p3_256) __extension__ ({ \
+  float64x1_t __s0_256 = __p0_256; \
+  float64x1_t __s2_256 = __p2_256; \
+  float64x1_t __ret_256; \
+  __ret_256 = vset_lane_f64(vget_lane_f64(__s2_256, __p3_256), __s0_256, __p1_256); \
+  __ret_256; \
+})
+#else
+#define vcopy_lane_f64(__p0_257, __p1_257, __p2_257, __p3_257) __extension__ ({ \
+  float64x1_t __s0_257 = __p0_257; \
+  float64x1_t __s2_257 = __p2_257; \
+  float64x1_t __ret_257; \
+  __ret_257 = __noswap_vset_lane_f64(__noswap_vget_lane_f64(__s2_257, __p3_257), __s0_257, __p1_257); \
+  __ret_257; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_laneq_p64(__p0_258, __p1_258, __p2_258, __p3_258) __extension__ ({ \
+  poly64x2_t __s0_258 = __p0_258; \
+  poly64x2_t __s2_258 = __p2_258; \
+  poly64x2_t __ret_258; \
+  __ret_258 = vsetq_lane_p64(vgetq_lane_p64(__s2_258, __p3_258), __s0_258, __p1_258); \
+  __ret_258; \
+})
+#else
+#define vcopyq_laneq_p64(__p0_259, __p1_259, __p2_259, __p3_259) __extension__ ({ \
+  poly64x2_t __s0_259 = __p0_259; \
+  poly64x2_t __s2_259 = __p2_259; \
+  poly64x2_t __rev0_259;  __rev0_259 = __builtin_shufflevector(__s0_259, __s0_259, 1, 0); \
+  poly64x2_t __rev2_259;  __rev2_259 = __builtin_shufflevector(__s2_259, __s2_259, 1, 0); \
+  poly64x2_t __ret_259; \
+  __ret_259 = __noswap_vsetq_lane_p64(__noswap_vgetq_lane_p64(__rev2_259, __p3_259), __rev0_259, __p1_259); \
+  __ret_259 = __builtin_shufflevector(__ret_259, __ret_259, 1, 0); \
+  __ret_259; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_laneq_f64(__p0_260, __p1_260, __p2_260, __p3_260) __extension__ ({ \
+  float64x2_t __s0_260 = __p0_260; \
+  float64x2_t __s2_260 = __p2_260; \
+  float64x2_t __ret_260; \
+  __ret_260 = vsetq_lane_f64(vgetq_lane_f64(__s2_260, __p3_260), __s0_260, __p1_260); \
+  __ret_260; \
+})
+#else
+#define vcopyq_laneq_f64(__p0_261, __p1_261, __p2_261, __p3_261) __extension__ ({ \
+  float64x2_t __s0_261 = __p0_261; \
+  float64x2_t __s2_261 = __p2_261; \
+  float64x2_t __rev0_261;  __rev0_261 = __builtin_shufflevector(__s0_261, __s0_261, 1, 0); \
+  float64x2_t __rev2_261;  __rev2_261 = __builtin_shufflevector(__s2_261, __s2_261, 1, 0); \
+  float64x2_t __ret_261; \
+  __ret_261 = __noswap_vsetq_lane_f64(__noswap_vgetq_lane_f64(__rev2_261, __p3_261), __rev0_261, __p1_261); \
+  __ret_261 = __builtin_shufflevector(__ret_261, __ret_261, 1, 0); \
+  __ret_261; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_laneq_p64(__p0_262, __p1_262, __p2_262, __p3_262) __extension__ ({ \
+  poly64x1_t __s0_262 = __p0_262; \
+  poly64x2_t __s2_262 = __p2_262; \
+  poly64x1_t __ret_262; \
+  __ret_262 = vset_lane_p64(vgetq_lane_p64(__s2_262, __p3_262), __s0_262, __p1_262); \
+  __ret_262; \
+})
+#else
+#define vcopy_laneq_p64(__p0_263, __p1_263, __p2_263, __p3_263) __extension__ ({ \
+  poly64x1_t __s0_263 = __p0_263; \
+  poly64x2_t __s2_263 = __p2_263; \
+  poly64x2_t __rev2_263;  __rev2_263 = __builtin_shufflevector(__s2_263, __s2_263, 1, 0); \
+  poly64x1_t __ret_263; \
+  __ret_263 = __noswap_vset_lane_p64(__noswap_vgetq_lane_p64(__rev2_263, __p3_263), __s0_263, __p1_263); \
+  __ret_263; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_laneq_f64(__p0_264, __p1_264, __p2_264, __p3_264) __extension__ ({ \
+  float64x1_t __s0_264 = __p0_264; \
+  float64x2_t __s2_264 = __p2_264; \
+  float64x1_t __ret_264; \
+  __ret_264 = vset_lane_f64(vgetq_lane_f64(__s2_264, __p3_264), __s0_264, __p1_264); \
+  __ret_264; \
+})
+#else
+#define vcopy_laneq_f64(__p0_265, __p1_265, __p2_265, __p3_265) __extension__ ({ \
+  float64x1_t __s0_265 = __p0_265; \
+  float64x2_t __s2_265 = __p2_265; \
+  float64x2_t __rev2_265;  __rev2_265 = __builtin_shufflevector(__s2_265, __s2_265, 1, 0); \
+  float64x1_t __ret_265; \
+  __ret_265 = __noswap_vset_lane_f64(__noswap_vgetq_lane_f64(__rev2_265, __p3_265), __s0_265, __p1_265); \
+  __ret_265; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vmlal_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
+  uint16x8_t __ret;
+  __ret = vmlal_u8(__p0, vget_high_u8(__p1), vget_high_u8(__p2));
+  return __ret;
+}
+#else
+__ai uint16x8_t vmlal_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __noswap_vmlal_u8(__rev0, __noswap_vget_high_u8(__rev1), __noswap_vget_high_u8(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vmlal_high_u32(uint64x2_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint64x2_t __ret;
+  __ret = vmlal_u32(__p0, vget_high_u32(__p1), vget_high_u32(__p2));
+  return __ret;
+}
+#else
+__ai uint64x2_t vmlal_high_u32(uint64x2_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint64x2_t __ret;
+  __ret = __noswap_vmlal_u32(__rev0, __noswap_vget_high_u32(__rev1), __noswap_vget_high_u32(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmlal_high_u16(uint32x4_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint32x4_t __ret;
+  __ret = vmlal_u16(__p0, vget_high_u16(__p1), vget_high_u16(__p2));
+  return __ret;
+}
+#else
+__ai uint32x4_t vmlal_high_u16(uint32x4_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __noswap_vmlal_u16(__rev0, __noswap_vget_high_u16(__rev1), __noswap_vget_high_u16(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vmlal_high_s8(int16x8_t __p0, int8x16_t __p1, int8x16_t __p2) {
+  int16x8_t __ret;
+  __ret = vmlal_s8(__p0, vget_high_s8(__p1), vget_high_s8(__p2));
+  return __ret;
+}
+#else
+__ai int16x8_t vmlal_high_s8(int16x8_t __p0, int8x16_t __p1, int8x16_t __p2) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __noswap_vmlal_s8(__rev0, __noswap_vget_high_s8(__rev1), __noswap_vget_high_s8(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vmlal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int64x2_t __ret;
+  __ret = vmlal_s32(__p0, vget_high_s32(__p1), vget_high_s32(__p2));
+  return __ret;
+}
+#else
+__ai int64x2_t vmlal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int64x2_t __ret;
+  __ret = __noswap_vmlal_s32(__rev0, __noswap_vget_high_s32(__rev1), __noswap_vget_high_s32(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmlal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int32x4_t __ret;
+  __ret = vmlal_s16(__p0, vget_high_s16(__p1), vget_high_s16(__p2));
+  return __ret;
+}
+#else
+__ai int32x4_t vmlal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vmlal_s16(__rev0, __noswap_vget_high_s16(__rev1), __noswap_vget_high_s16(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vmlal_high_n_u32(uint64x2_t __p0, uint32x4_t __p1, uint32_t __p2) {
+  uint64x2_t __ret;
+  __ret = vmlal_n_u32(__p0, vget_high_u32(__p1), __p2);
+  return __ret;
+}
+#else
+__ai uint64x2_t vmlal_high_n_u32(uint64x2_t __p0, uint32x4_t __p1, uint32_t __p2) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint64x2_t __ret;
+  __ret = __noswap_vmlal_n_u32(__rev0, __noswap_vget_high_u32(__rev1), __p2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmlal_high_n_u16(uint32x4_t __p0, uint16x8_t __p1, uint16_t __p2) {
+  uint32x4_t __ret;
+  __ret = vmlal_n_u16(__p0, vget_high_u16(__p1), __p2);
+  return __ret;
+}
+#else
+__ai uint32x4_t vmlal_high_n_u16(uint32x4_t __p0, uint16x8_t __p1, uint16_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __noswap_vmlal_n_u16(__rev0, __noswap_vget_high_u16(__rev1), __p2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vmlal_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
+  int64x2_t __ret;
+  __ret = vmlal_n_s32(__p0, vget_high_s32(__p1), __p2);
+  return __ret;
+}
+#else
+__ai int64x2_t vmlal_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int64x2_t __ret;
+  __ret = __noswap_vmlal_n_s32(__rev0, __noswap_vget_high_s32(__rev1), __p2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmlal_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
+  int32x4_t __ret;
+  __ret = vmlal_n_s16(__p0, vget_high_s16(__p1), __p2);
+  return __ret;
+}
+#else
+__ai int32x4_t vmlal_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vmlal_n_s16(__rev0, __noswap_vget_high_s16(__rev1), __p2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vmlsl_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
+  uint16x8_t __ret;
+  __ret = vmlsl_u8(__p0, vget_high_u8(__p1), vget_high_u8(__p2));
+  return __ret;
+}
+#else
+__ai uint16x8_t vmlsl_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __noswap_vmlsl_u8(__rev0, __noswap_vget_high_u8(__rev1), __noswap_vget_high_u8(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vmlsl_high_u32(uint64x2_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint64x2_t __ret;
+  __ret = vmlsl_u32(__p0, vget_high_u32(__p1), vget_high_u32(__p2));
+  return __ret;
+}
+#else
+__ai uint64x2_t vmlsl_high_u32(uint64x2_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint64x2_t __ret;
+  __ret = __noswap_vmlsl_u32(__rev0, __noswap_vget_high_u32(__rev1), __noswap_vget_high_u32(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmlsl_high_u16(uint32x4_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint32x4_t __ret;
+  __ret = vmlsl_u16(__p0, vget_high_u16(__p1), vget_high_u16(__p2));
+  return __ret;
+}
+#else
+__ai uint32x4_t vmlsl_high_u16(uint32x4_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __noswap_vmlsl_u16(__rev0, __noswap_vget_high_u16(__rev1), __noswap_vget_high_u16(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vmlsl_high_s8(int16x8_t __p0, int8x16_t __p1, int8x16_t __p2) {
+  int16x8_t __ret;
+  __ret = vmlsl_s8(__p0, vget_high_s8(__p1), vget_high_s8(__p2));
+  return __ret;
+}
+#else
+__ai int16x8_t vmlsl_high_s8(int16x8_t __p0, int8x16_t __p1, int8x16_t __p2) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __noswap_vmlsl_s8(__rev0, __noswap_vget_high_s8(__rev1), __noswap_vget_high_s8(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vmlsl_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int64x2_t __ret;
+  __ret = vmlsl_s32(__p0, vget_high_s32(__p1), vget_high_s32(__p2));
+  return __ret;
+}
+#else
+__ai int64x2_t vmlsl_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int64x2_t __ret;
+  __ret = __noswap_vmlsl_s32(__rev0, __noswap_vget_high_s32(__rev1), __noswap_vget_high_s32(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmlsl_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int32x4_t __ret;
+  __ret = vmlsl_s16(__p0, vget_high_s16(__p1), vget_high_s16(__p2));
+  return __ret;
+}
+#else
+__ai int32x4_t vmlsl_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vmlsl_s16(__rev0, __noswap_vget_high_s16(__rev1), __noswap_vget_high_s16(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vmlsl_high_n_u32(uint64x2_t __p0, uint32x4_t __p1, uint32_t __p2) {
+  uint64x2_t __ret;
+  __ret = vmlsl_n_u32(__p0, vget_high_u32(__p1), __p2);
+  return __ret;
+}
+#else
+__ai uint64x2_t vmlsl_high_n_u32(uint64x2_t __p0, uint32x4_t __p1, uint32_t __p2) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint64x2_t __ret;
+  __ret = __noswap_vmlsl_n_u32(__rev0, __noswap_vget_high_u32(__rev1), __p2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmlsl_high_n_u16(uint32x4_t __p0, uint16x8_t __p1, uint16_t __p2) {
+  uint32x4_t __ret;
+  __ret = vmlsl_n_u16(__p0, vget_high_u16(__p1), __p2);
+  return __ret;
+}
+#else
+__ai uint32x4_t vmlsl_high_n_u16(uint32x4_t __p0, uint16x8_t __p1, uint16_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __noswap_vmlsl_n_u16(__rev0, __noswap_vget_high_u16(__rev1), __p2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vmlsl_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
+  int64x2_t __ret;
+  __ret = vmlsl_n_s32(__p0, vget_high_s32(__p1), __p2);
+  return __ret;
+}
+#else
+__ai int64x2_t vmlsl_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int64x2_t __ret;
+  __ret = __noswap_vmlsl_n_s32(__rev0, __noswap_vget_high_s32(__rev1), __p2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmlsl_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
+  int32x4_t __ret;
+  __ret = vmlsl_n_s16(__p0, vget_high_s16(__p1), __p2);
+  return __ret;
+}
+#else
+__ai int32x4_t vmlsl_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vmlsl_n_s16(__rev0, __noswap_vget_high_s16(__rev1), __p2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulx_lane_f64(__p0_266, __p1_266, __p2_266) __extension__ ({ \
+  float64x1_t __s0_266 = __p0_266; \
+  float64x1_t __s1_266 = __p1_266; \
+  float64x1_t __ret_266; \
+  float64_t __x_266 = vget_lane_f64(__s0_266, 0); \
+  float64_t __y_266 = vget_lane_f64(__s1_266, __p2_266); \
+  float64_t __z_266 = vmulxd_f64(__x_266, __y_266); \
+  __ret_266 = vset_lane_f64(__z_266, __s0_266, __p2_266); \
+  __ret_266; \
+})
+#else
+#define vmulx_lane_f64(__p0_267, __p1_267, __p2_267) __extension__ ({ \
+  float64x1_t __s0_267 = __p0_267; \
+  float64x1_t __s1_267 = __p1_267; \
+  float64x1_t __ret_267; \
+  float64_t __x_267 = __noswap_vget_lane_f64(__s0_267, 0); \
+  float64_t __y_267 = __noswap_vget_lane_f64(__s1_267, __p2_267); \
+  float64_t __z_267 = __noswap_vmulxd_f64(__x_267, __y_267); \
+  __ret_267 = __noswap_vset_lane_f64(__z_267, __s0_267, __p2_267); \
+  __ret_267; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vmulx_laneq_f64(__p0_268, __p1_268, __p2_268) __extension__ ({ \
+  float64x1_t __s0_268 = __p0_268; \
+  float64x2_t __s1_268 = __p1_268; \
+  float64x1_t __ret_268; \
+  float64_t __x_268 = vget_lane_f64(__s0_268, 0); \
+  float64_t __y_268 = vgetq_lane_f64(__s1_268, __p2_268); \
+  float64_t __z_268 = vmulxd_f64(__x_268, __y_268); \
+  __ret_268 = vset_lane_f64(__z_268, __s0_268, 0); \
+  __ret_268; \
+})
+#else
+#define vmulx_laneq_f64(__p0_269, __p1_269, __p2_269) __extension__ ({ \
+  float64x1_t __s0_269 = __p0_269; \
+  float64x2_t __s1_269 = __p1_269; \
+  float64x2_t __rev1_269;  __rev1_269 = __builtin_shufflevector(__s1_269, __s1_269, 1, 0); \
+  float64x1_t __ret_269; \
+  float64_t __x_269 = __noswap_vget_lane_f64(__s0_269, 0); \
+  float64_t __y_269 = __noswap_vgetq_lane_f64(__rev1_269, __p2_269); \
+  float64_t __z_269 = __noswap_vmulxd_f64(__x_269, __y_269); \
+  __ret_269 = __noswap_vset_lane_f64(__z_269, __s0_269, 0); \
+  __ret_269; \
+})
+#endif
+
+#endif
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vabal_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
+  uint16x8_t __ret;
+  __ret = __p0 + vabdl_u8(__p1, __p2);
+  return __ret;
+}
+#else
+__ai uint16x8_t vabal_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __rev0 + __noswap_vabdl_u8(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint16x8_t __noswap_vabal_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
+  uint16x8_t __ret;
+  __ret = __p0 + __noswap_vabdl_u8(__p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vabal_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
+  uint64x2_t __ret;
+  __ret = __p0 + vabdl_u32(__p1, __p2);
+  return __ret;
+}
+#else
+__ai uint64x2_t vabal_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  uint64x2_t __ret;
+  __ret = __rev0 + __noswap_vabdl_u32(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai uint64x2_t __noswap_vabal_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
+  uint64x2_t __ret;
+  __ret = __p0 + __noswap_vabdl_u32(__p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vabal_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
+  uint32x4_t __ret;
+  __ret = __p0 + vabdl_u16(__p1, __p2);
+  return __ret;
+}
+#else
+__ai uint32x4_t vabal_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __rev0 + __noswap_vabdl_u16(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai uint32x4_t __noswap_vabal_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
+  uint32x4_t __ret;
+  __ret = __p0 + __noswap_vabdl_u16(__p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vabal_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
+  int16x8_t __ret;
+  __ret = __p0 + vabdl_s8(__p1, __p2);
+  return __ret;
+}
+#else
+__ai int16x8_t vabal_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __rev0 + __noswap_vabdl_s8(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int16x8_t __noswap_vabal_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
+  int16x8_t __ret;
+  __ret = __p0 + __noswap_vabdl_s8(__p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vabal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int64x2_t __ret;
+  __ret = __p0 + vabdl_s32(__p1, __p2);
+  return __ret;
+}
+#else
+__ai int64x2_t vabal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
+  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
+  int64x2_t __ret;
+  __ret = __rev0 + __noswap_vabdl_s32(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+__ai int64x2_t __noswap_vabal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
+  int64x2_t __ret;
+  __ret = __p0 + __noswap_vabdl_s32(__p1, __p2);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vabal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int32x4_t __ret;
+  __ret = __p0 + vabdl_s16(__p1, __p2);
+  return __ret;
+}
+#else
+__ai int32x4_t vabal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __rev0 + __noswap_vabdl_s16(__rev1, __rev2);
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+__ai int32x4_t __noswap_vabal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
+  int32x4_t __ret;
+  __ret = __p0 + __noswap_vabdl_s16(__p1, __p2);
+  return __ret;
+}
+#endif
+
+#if defined(__aarch64__)
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vabal_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
+  uint16x8_t __ret;
+  __ret = vabal_u8(__p0, vget_high_u8(__p1), vget_high_u8(__p2));
+  return __ret;
+}
+#else
+__ai uint16x8_t vabal_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
+  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret;
+  __ret = __noswap_vabal_u8(__rev0, __noswap_vget_high_u8(__rev1), __noswap_vget_high_u8(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint64x2_t vabal_high_u32(uint64x2_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint64x2_t __ret;
+  __ret = vabal_u32(__p0, vget_high_u32(__p1), vget_high_u32(__p2));
+  return __ret;
+}
+#else
+__ai uint64x2_t vabal_high_u32(uint64x2_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
+  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  uint64x2_t __ret;
+  __ret = __noswap_vabal_u32(__rev0, __noswap_vget_high_u32(__rev1), __noswap_vget_high_u32(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vabal_high_u16(uint32x4_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint32x4_t __ret;
+  __ret = vabal_u16(__p0, vget_high_u16(__p1), vget_high_u16(__p2));
+  return __ret;
+}
+#else
+__ai uint32x4_t vabal_high_u16(uint32x4_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
+  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint32x4_t __ret;
+  __ret = __noswap_vabal_u16(__rev0, __noswap_vget_high_u16(__rev1), __noswap_vget_high_u16(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vabal_high_s8(int16x8_t __p0, int8x16_t __p1, int8x16_t __p2) {
+  int16x8_t __ret;
+  __ret = vabal_s8(__p0, vget_high_s8(__p1), vget_high_s8(__p2));
+  return __ret;
+}
+#else
+__ai int16x8_t vabal_high_s8(int16x8_t __p0, int8x16_t __p1, int8x16_t __p2) {
+  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret;
+  __ret = __noswap_vabal_s8(__rev0, __noswap_vget_high_s8(__rev1), __noswap_vget_high_s8(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vabal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int64x2_t __ret;
+  __ret = vabal_s32(__p0, vget_high_s32(__p1), vget_high_s32(__p2));
+  return __ret;
+}
+#else
+__ai int64x2_t vabal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
+  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
+  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
+  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
+  int64x2_t __ret;
+  __ret = __noswap_vabal_s32(__rev0, __noswap_vget_high_s32(__rev1), __noswap_vget_high_s32(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
+  return __ret;
+}
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vabal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int32x4_t __ret;
+  __ret = vabal_s16(__p0, vget_high_s16(__p1), vget_high_s16(__p2));
+  return __ret;
+}
+#else
+__ai int32x4_t vabal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
+  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
+  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret;
+  __ret = __noswap_vabal_s16(__rev0, __noswap_vget_high_s16(__rev1), __noswap_vget_high_s16(__rev2));
+  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
+  return __ret;
+}
+#endif
+
+#endif
+
+#undef __ai
+
+#endif /* __ARM_NEON_H */
diff --git a/lib/clang/3.6/include/avx2intrin.h b/lib/clang/3.6/include/avx2intrin.h
new file mode 100644
index 0000000..394fdfe
--- /dev/null
+++ b/lib/clang/3.6/include/avx2intrin.h
@@ -0,0 +1,1234 @@
+/*===---- avx2intrin.h - AVX2 intrinsics -----------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __IMMINTRIN_H
+#error "Never use <avx2intrin.h> directly; include <immintrin.h> instead."
+#endif
+
+#ifndef __AVX2INTRIN_H
+#define __AVX2INTRIN_H
+
+/* SSE4 Multiple Packed Sums of Absolute Difference.  */
+#define _mm256_mpsadbw_epu8(X, Y, M) __builtin_ia32_mpsadbw256((X), (Y), (M))
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_abs_epi8(__m256i __a)
+{
+    return (__m256i)__builtin_ia32_pabsb256((__v32qi)__a);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_abs_epi16(__m256i __a)
+{
+    return (__m256i)__builtin_ia32_pabsw256((__v16hi)__a);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_abs_epi32(__m256i __a)
+{
+    return (__m256i)__builtin_ia32_pabsd256((__v8si)__a);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_packs_epi16(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_packsswb256((__v16hi)__a, (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_packs_epi32(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_packssdw256((__v8si)__a, (__v8si)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_packus_epi16(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_packuswb256((__v16hi)__a, (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_packus_epi32(__m256i __V1, __m256i __V2)
+{
+  return (__m256i) __builtin_ia32_packusdw256((__v8si)__V1, (__v8si)__V2);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_add_epi8(__m256i __a, __m256i __b)
+{
+  return (__m256i)((__v32qi)__a + (__v32qi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_add_epi16(__m256i __a, __m256i __b)
+{
+  return (__m256i)((__v16hi)__a + (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_add_epi32(__m256i __a, __m256i __b)
+{
+  return (__m256i)((__v8si)__a + (__v8si)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_add_epi64(__m256i __a, __m256i __b)
+{
+  return __a + __b;
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_adds_epi8(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_paddsb256((__v32qi)__a, (__v32qi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_adds_epi16(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_paddsw256((__v16hi)__a, (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_adds_epu8(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_paddusb256((__v32qi)__a, (__v32qi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_adds_epu16(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_paddusw256((__v16hi)__a, (__v16hi)__b);
+}
+
+#define _mm256_alignr_epi8(a, b, n) __extension__ ({ \
+  __m256i __a = (a); \
+  __m256i __b = (b); \
+  (__m256i)__builtin_ia32_palignr256((__v32qi)__a, (__v32qi)__b, (n)); })
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_and_si256(__m256i __a, __m256i __b)
+{
+  return __a & __b;
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_andnot_si256(__m256i __a, __m256i __b)
+{
+  return ~__a & __b;
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_avg_epu8(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_pavgb256((__v32qi)__a, (__v32qi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_avg_epu16(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_pavgw256((__v16hi)__a, (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_blendv_epi8(__m256i __V1, __m256i __V2, __m256i __M)
+{
+  return (__m256i)__builtin_ia32_pblendvb256((__v32qi)__V1, (__v32qi)__V2,
+                                              (__v32qi)__M);
+}
+
+#define _mm256_blend_epi16(V1, V2, M) __extension__ ({ \
+  __m256i __V1 = (V1); \
+  __m256i __V2 = (V2); \
+  (__m256d)__builtin_shufflevector((__v16hi)__V1, (__v16hi)__V2, \
+                                   (((M) & 0x01) ? 16 : 0), \
+                                   (((M) & 0x02) ? 17 : 1), \
+                                   (((M) & 0x04) ? 18 : 2), \
+                                   (((M) & 0x08) ? 19 : 3), \
+                                   (((M) & 0x10) ? 20 : 4), \
+                                   (((M) & 0x20) ? 21 : 5), \
+                                   (((M) & 0x40) ? 22 : 6), \
+                                   (((M) & 0x80) ? 23 : 7), \
+                                   (((M) & 0x01) ? 24 : 8), \
+                                   (((M) & 0x02) ? 25 : 9), \
+                                   (((M) & 0x04) ? 26 : 10), \
+                                   (((M) & 0x08) ? 27 : 11), \
+                                   (((M) & 0x10) ? 28 : 12), \
+                                   (((M) & 0x20) ? 29 : 13), \
+                                   (((M) & 0x40) ? 30 : 14), \
+                                   (((M) & 0x80) ? 31 : 15)); })
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cmpeq_epi8(__m256i __a, __m256i __b)
+{
+  return (__m256i)((__v32qi)__a == (__v32qi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cmpeq_epi16(__m256i __a, __m256i __b)
+{
+  return (__m256i)((__v16hi)__a == (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cmpeq_epi32(__m256i __a, __m256i __b)
+{
+  return (__m256i)((__v8si)__a == (__v8si)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cmpeq_epi64(__m256i __a, __m256i __b)
+{
+  return (__m256i)(__a == __b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cmpgt_epi8(__m256i __a, __m256i __b)
+{
+  return (__m256i)((__v32qi)__a > (__v32qi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cmpgt_epi16(__m256i __a, __m256i __b)
+{
+  return (__m256i)((__v16hi)__a > (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cmpgt_epi32(__m256i __a, __m256i __b)
+{
+  return (__m256i)((__v8si)__a > (__v8si)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cmpgt_epi64(__m256i __a, __m256i __b)
+{
+  return (__m256i)(__a > __b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_hadd_epi16(__m256i __a, __m256i __b)
+{
+    return (__m256i)__builtin_ia32_phaddw256((__v16hi)__a, (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_hadd_epi32(__m256i __a, __m256i __b)
+{
+    return (__m256i)__builtin_ia32_phaddd256((__v8si)__a, (__v8si)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_hadds_epi16(__m256i __a, __m256i __b)
+{
+    return (__m256i)__builtin_ia32_phaddsw256((__v16hi)__a, (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_hsub_epi16(__m256i __a, __m256i __b)
+{
+    return (__m256i)__builtin_ia32_phsubw256((__v16hi)__a, (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_hsub_epi32(__m256i __a, __m256i __b)
+{
+    return (__m256i)__builtin_ia32_phsubd256((__v8si)__a, (__v8si)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_hsubs_epi16(__m256i __a, __m256i __b)
+{
+    return (__m256i)__builtin_ia32_phsubsw256((__v16hi)__a, (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_maddubs_epi16(__m256i __a, __m256i __b)
+{
+    return (__m256i)__builtin_ia32_pmaddubsw256((__v32qi)__a, (__v32qi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_madd_epi16(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_pmaddwd256((__v16hi)__a, (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_max_epi8(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_pmaxsb256((__v32qi)__a, (__v32qi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_max_epi16(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_pmaxsw256((__v16hi)__a, (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_max_epi32(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_pmaxsd256((__v8si)__a, (__v8si)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_max_epu8(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_pmaxub256((__v32qi)__a, (__v32qi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_max_epu16(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_pmaxuw256((__v16hi)__a, (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_max_epu32(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_pmaxud256((__v8si)__a, (__v8si)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_min_epi8(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_pminsb256((__v32qi)__a, (__v32qi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_min_epi16(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_pminsw256((__v16hi)__a, (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_min_epi32(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_pminsd256((__v8si)__a, (__v8si)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_min_epu8(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_pminub256((__v32qi)__a, (__v32qi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_min_epu16(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_pminuw256 ((__v16hi)__a, (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_min_epu32(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_pminud256((__v8si)__a, (__v8si)__b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm256_movemask_epi8(__m256i __a)
+{
+  return __builtin_ia32_pmovmskb256((__v32qi)__a);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepi8_epi16(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovsxbw256((__v16qi)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepi8_epi32(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovsxbd256((__v16qi)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepi8_epi64(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovsxbq256((__v16qi)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepi16_epi32(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovsxwd256((__v8hi)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepi16_epi64(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovsxwq256((__v8hi)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepi32_epi64(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovsxdq256((__v4si)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepu8_epi16(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovzxbw256((__v16qi)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepu8_epi32(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovzxbd256((__v16qi)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepu8_epi64(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovzxbq256((__v16qi)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepu16_epi32(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovzxwd256((__v8hi)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepu16_epi64(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovzxwq256((__v8hi)__V);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepu32_epi64(__m128i __V)
+{
+  return (__m256i)__builtin_ia32_pmovzxdq256((__v4si)__V);
+}
+
+static __inline__  __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_mul_epi32(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_pmuldq256((__v8si)__a, (__v8si)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_mulhrs_epi16(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_pmulhrsw256((__v16hi)__a, (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_mulhi_epu16(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_pmulhuw256((__v16hi)__a, (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_mulhi_epi16(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_pmulhw256((__v16hi)__a, (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_mullo_epi16(__m256i __a, __m256i __b)
+{
+  return (__m256i)((__v16hi)__a * (__v16hi)__b);
+}
+
+static __inline__  __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_mullo_epi32 (__m256i __a, __m256i __b)
+{
+  return (__m256i)((__v8si)__a * (__v8si)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_mul_epu32(__m256i __a, __m256i __b)
+{
+  return __builtin_ia32_pmuludq256((__v8si)__a, (__v8si)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_or_si256(__m256i __a, __m256i __b)
+{
+  return __a | __b;
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sad_epu8(__m256i __a, __m256i __b)
+{
+  return __builtin_ia32_psadbw256((__v32qi)__a, (__v32qi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_shuffle_epi8(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_pshufb256((__v32qi)__a, (__v32qi)__b);
+}
+
+#define _mm256_shuffle_epi32(a, imm) __extension__ ({ \
+  __m256i __a = (a); \
+  (__m256i)__builtin_shufflevector((__v8si)__a, (__v8si)_mm256_set1_epi32(0), \
+                                   (imm) & 0x3, ((imm) & 0xc) >> 2, \
+                                   ((imm) & 0x30) >> 4, ((imm) & 0xc0) >> 6, \
+                                   4 + (((imm) & 0x03) >> 0), \
+                                   4 + (((imm) & 0x0c) >> 2), \
+                                   4 + (((imm) & 0x30) >> 4), \
+                                   4 + (((imm) & 0xc0) >> 6)); })
+
+#define _mm256_shufflehi_epi16(a, imm) __extension__ ({ \
+  __m256i __a = (a); \
+  (__m256i)__builtin_shufflevector((__v16hi)__a, (__v16hi)_mm256_set1_epi16(0), \
+                                   0, 1, 2, 3, \
+                                   4 + (((imm) & 0x03) >> 0), \
+                                   4 + (((imm) & 0x0c) >> 2), \
+                                   4 + (((imm) & 0x30) >> 4), \
+                                   4 + (((imm) & 0xc0) >> 6), \
+                                   8, 9, 10, 11, \
+                                   12 + (((imm) & 0x03) >> 0), \
+                                   12 + (((imm) & 0x0c) >> 2), \
+                                   12 + (((imm) & 0x30) >> 4), \
+                                   12 + (((imm) & 0xc0) >> 6)); })
+
+#define _mm256_shufflelo_epi16(a, imm) __extension__ ({ \
+  __m256i __a = (a); \
+  (__m256i)__builtin_shufflevector((__v16hi)__a, (__v16hi)_mm256_set1_epi16(0), \
+                                   (imm) & 0x3,((imm) & 0xc) >> 2, \
+                                   ((imm) & 0x30) >> 4, ((imm) & 0xc0) >> 6, \
+                                   4, 5, 6, 7, \
+                                   8 + (((imm) & 0x03) >> 0), \
+                                   8 + (((imm) & 0x0c) >> 2), \
+                                   8 + (((imm) & 0x30) >> 4), \
+                                   8 + (((imm) & 0xc0) >> 6), \
+                                   12, 13, 14, 15); })
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sign_epi8(__m256i __a, __m256i __b)
+{
+    return (__m256i)__builtin_ia32_psignb256((__v32qi)__a, (__v32qi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sign_epi16(__m256i __a, __m256i __b)
+{
+    return (__m256i)__builtin_ia32_psignw256((__v16hi)__a, (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sign_epi32(__m256i __a, __m256i __b)
+{
+    return (__m256i)__builtin_ia32_psignd256((__v8si)__a, (__v8si)__b);
+}
+
+#define _mm256_slli_si256(a, count) __extension__ ({ \
+  __m256i __a = (a); \
+  (__m256i)__builtin_ia32_pslldqi256(__a, (count)*8); })
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_slli_epi16(__m256i __a, int __count)
+{
+  return (__m256i)__builtin_ia32_psllwi256((__v16hi)__a, __count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sll_epi16(__m256i __a, __m128i __count)
+{
+  return (__m256i)__builtin_ia32_psllw256((__v16hi)__a, (__v8hi)__count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_slli_epi32(__m256i __a, int __count)
+{
+  return (__m256i)__builtin_ia32_pslldi256((__v8si)__a, __count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sll_epi32(__m256i __a, __m128i __count)
+{
+  return (__m256i)__builtin_ia32_pslld256((__v8si)__a, (__v4si)__count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_slli_epi64(__m256i __a, int __count)
+{
+  return __builtin_ia32_psllqi256(__a, __count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sll_epi64(__m256i __a, __m128i __count)
+{
+  return __builtin_ia32_psllq256(__a, __count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srai_epi16(__m256i __a, int __count)
+{
+  return (__m256i)__builtin_ia32_psrawi256((__v16hi)__a, __count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sra_epi16(__m256i __a, __m128i __count)
+{
+  return (__m256i)__builtin_ia32_psraw256((__v16hi)__a, (__v8hi)__count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srai_epi32(__m256i __a, int __count)
+{
+  return (__m256i)__builtin_ia32_psradi256((__v8si)__a, __count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sra_epi32(__m256i __a, __m128i __count)
+{
+  return (__m256i)__builtin_ia32_psrad256((__v8si)__a, (__v4si)__count);
+}
+
+#define _mm256_srli_si256(a, count) __extension__ ({ \
+  __m256i __a = (a); \
+  (__m256i)__builtin_ia32_psrldqi256(__a, (count)*8); })
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srli_epi16(__m256i __a, int __count)
+{
+  return (__m256i)__builtin_ia32_psrlwi256((__v16hi)__a, __count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srl_epi16(__m256i __a, __m128i __count)
+{
+  return (__m256i)__builtin_ia32_psrlw256((__v16hi)__a, (__v8hi)__count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srli_epi32(__m256i __a, int __count)
+{
+  return (__m256i)__builtin_ia32_psrldi256((__v8si)__a, __count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srl_epi32(__m256i __a, __m128i __count)
+{
+  return (__m256i)__builtin_ia32_psrld256((__v8si)__a, (__v4si)__count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srli_epi64(__m256i __a, int __count)
+{
+  return __builtin_ia32_psrlqi256(__a, __count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srl_epi64(__m256i __a, __m128i __count)
+{
+  return __builtin_ia32_psrlq256(__a, __count);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sub_epi8(__m256i __a, __m256i __b)
+{
+  return (__m256i)((__v32qi)__a - (__v32qi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sub_epi16(__m256i __a, __m256i __b)
+{
+  return (__m256i)((__v16hi)__a - (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sub_epi32(__m256i __a, __m256i __b)
+{
+  return (__m256i)((__v8si)__a - (__v8si)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sub_epi64(__m256i __a, __m256i __b)
+{
+  return __a - __b;
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_subs_epi8(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_psubsb256((__v32qi)__a, (__v32qi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_subs_epi16(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_psubsw256((__v16hi)__a, (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_subs_epu8(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_psubusb256((__v32qi)__a, (__v32qi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_subs_epu16(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_psubusw256((__v16hi)__a, (__v16hi)__b);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_unpackhi_epi8(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_shufflevector((__v32qi)__a, (__v32qi)__b, 8, 32+8, 9, 32+9, 10, 32+10, 11, 32+11, 12, 32+12, 13, 32+13, 14, 32+14, 15, 32+15, 24, 32+24, 25, 32+25, 26, 32+26, 27, 32+27, 28, 32+28, 29, 32+29, 30, 32+30, 31, 32+31);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_unpackhi_epi16(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_shufflevector((__v16hi)__a, (__v16hi)__b, 4, 16+4, 5, 16+5, 6, 16+6, 7, 16+7, 12, 16+12, 13, 16+13, 14, 16+14, 15, 16+15);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_unpackhi_epi32(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_shufflevector((__v8si)__a, (__v8si)__b, 2, 8+2, 3, 8+3, 6, 8+6, 7, 8+7);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_unpackhi_epi64(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_shufflevector(__a, __b, 1, 4+1, 3, 4+3);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_unpacklo_epi8(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_shufflevector((__v32qi)__a, (__v32qi)__b, 0, 32+0, 1, 32+1, 2, 32+2, 3, 32+3, 4, 32+4, 5, 32+5, 6, 32+6, 7, 32+7, 16, 32+16, 17, 32+17, 18, 32+18, 19, 32+19, 20, 32+20, 21, 32+21, 22, 32+22, 23, 32+23);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_unpacklo_epi16(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_shufflevector((__v16hi)__a, (__v16hi)__b, 0, 16+0, 1, 16+1, 2, 16+2, 3, 16+3, 8, 16+8, 9, 16+9, 10, 16+10, 11, 16+11);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_unpacklo_epi32(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_shufflevector((__v8si)__a, (__v8si)__b, 0, 8+0, 1, 8+1, 4, 8+4, 5, 8+5);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_unpacklo_epi64(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_shufflevector(__a, __b, 0, 4+0, 2, 4+2);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_xor_si256(__m256i __a, __m256i __b)
+{
+  return __a ^ __b;
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_stream_load_si256(__m256i *__V)
+{
+  return (__m256i)__builtin_ia32_movntdqa256((__v4di *)__V);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_broadcastss_ps(__m128 __X)
+{
+  return (__m128)__builtin_ia32_vbroadcastss_ps((__v4sf)__X);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcastss_ps(__m128 __X)
+{
+  return (__m256)__builtin_ia32_vbroadcastss_ps256((__v4sf)__X);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcastsd_pd(__m128d __X)
+{
+  return (__m256d)__builtin_ia32_vbroadcastsd_pd256((__v2df)__X);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcastsi128_si256(__m128i __X)
+{
+  return (__m256i)__builtin_ia32_vbroadcastsi256(__X);
+}
+
+#define _mm_blend_epi32(V1, V2, M) __extension__ ({ \
+  __m128i __V1 = (V1); \
+  __m128i __V2 = (V2); \
+  (__m128i)__builtin_shufflevector((__v4si)__V1, (__v4si)__V2, \
+                                   (((M) & 0x01) ? 4 : 0), \
+                                   (((M) & 0x02) ? 5 : 1), \
+                                   (((M) & 0x04) ? 6 : 2), \
+                                   (((M) & 0x08) ? 7 : 3)); })
+
+#define _mm256_blend_epi32(V1, V2, M) __extension__ ({ \
+  __m256i __V1 = (V1); \
+  __m256i __V2 = (V2); \
+  (__m256i)__builtin_shufflevector((__v8si)__V1, (__v8si)__V2, \
+                                   (((M) & 0x01) ?  8 : 0), \
+                                   (((M) & 0x02) ?  9 : 1), \
+                                   (((M) & 0x04) ? 10 : 2), \
+                                   (((M) & 0x08) ? 11 : 3), \
+                                   (((M) & 0x10) ? 12 : 4), \
+                                   (((M) & 0x20) ? 13 : 5), \
+                                   (((M) & 0x40) ? 14 : 6), \
+                                   (((M) & 0x80) ? 15 : 7)); })
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcastb_epi8(__m128i __X)
+{
+  return (__m256i)__builtin_ia32_pbroadcastb256((__v16qi)__X);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcastw_epi16(__m128i __X)
+{
+  return (__m256i)__builtin_ia32_pbroadcastw256((__v8hi)__X);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcastd_epi32(__m128i __X)
+{
+  return (__m256i)__builtin_ia32_pbroadcastd256((__v4si)__X);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcastq_epi64(__m128i __X)
+{
+  return (__m256i)__builtin_ia32_pbroadcastq256(__X);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_broadcastb_epi8(__m128i __X)
+{
+  return (__m128i)__builtin_ia32_pbroadcastb128((__v16qi)__X);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_broadcastw_epi16(__m128i __X)
+{
+  return (__m128i)__builtin_ia32_pbroadcastw128((__v8hi)__X);
+}
+
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_broadcastd_epi32(__m128i __X)
+{
+  return (__m128i)__builtin_ia32_pbroadcastd128((__v4si)__X);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_broadcastq_epi64(__m128i __X)
+{
+  return (__m128i)__builtin_ia32_pbroadcastq128(__X);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_permutevar8x32_epi32(__m256i __a, __m256i __b)
+{
+  return (__m256i)__builtin_ia32_permvarsi256((__v8si)__a, (__v8si)__b);
+}
+
+#define _mm256_permute4x64_pd(V, M) __extension__ ({ \
+  __m256d __V = (V); \
+  (__m256d)__builtin_shufflevector((__v4df)__V, (__v4df) _mm256_setzero_pd(), \
+                                   (M) & 0x3, ((M) & 0xc) >> 2, \
+                                   ((M) & 0x30) >> 4, ((M) & 0xc0) >> 6); })
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_permutevar8x32_ps(__m256 __a, __m256 __b)
+{
+  return (__m256)__builtin_ia32_permvarsf256((__v8sf)__a, (__v8sf)__b);
+}
+
+#define _mm256_permute4x64_epi64(V, M) __extension__ ({ \
+  __m256i __V = (V); \
+  (__m256i)__builtin_shufflevector((__v4di)__V, (__v4di) _mm256_setzero_si256(), \
+                                   (M) & 0x3, ((M) & 0xc) >> 2, \
+                                   ((M) & 0x30) >> 4, ((M) & 0xc0) >> 6); })
+
+#define _mm256_permute2x128_si256(V1, V2, M) __extension__ ({ \
+  __m256i __V1 = (V1); \
+  __m256i __V2 = (V2); \
+  (__m256i)__builtin_ia32_permti256(__V1, __V2, (M)); })
+
+#define _mm256_extracti128_si256(A, O) __extension__ ({ \
+  __m256i __A = (A); \
+  (__m128i)__builtin_ia32_extract128i256(__A, (O)); })
+
+#define _mm256_inserti128_si256(V1, V2, O) __extension__ ({ \
+  __m256i __V1 = (V1); \
+  __m128i __V2 = (V2); \
+  (__m256i)__builtin_ia32_insert128i256(__V1, __V2, (O)); })
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_maskload_epi32(int const *__X, __m256i __M)
+{
+  return (__m256i)__builtin_ia32_maskloadd256((const __v8si *)__X, (__v8si)__M);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_maskload_epi64(long long const *__X, __m256i __M)
+{
+  return (__m256i)__builtin_ia32_maskloadq256((const __v4di *)__X, __M);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_maskload_epi32(int const *__X, __m128i __M)
+{
+  return (__m128i)__builtin_ia32_maskloadd((const __v4si *)__X, (__v4si)__M);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_maskload_epi64(long long const *__X, __m128i __M)
+{
+  return (__m128i)__builtin_ia32_maskloadq((const __v2di *)__X, (__v2di)__M);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm256_maskstore_epi32(int *__X, __m256i __M, __m256i __Y)
+{
+  __builtin_ia32_maskstored256((__v8si *)__X, (__v8si)__M, (__v8si)__Y);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm256_maskstore_epi64(long long *__X, __m256i __M, __m256i __Y)
+{
+  __builtin_ia32_maskstoreq256((__v4di *)__X, __M, __Y);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_maskstore_epi32(int *__X, __m128i __M, __m128i __Y)
+{
+  __builtin_ia32_maskstored((__v4si *)__X, (__v4si)__M, (__v4si)__Y);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_maskstore_epi64(long long *__X, __m128i __M, __m128i __Y)
+{
+  __builtin_ia32_maskstoreq(( __v2di *)__X, __M, __Y);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sllv_epi32(__m256i __X, __m256i __Y)
+{
+  return (__m256i)__builtin_ia32_psllv8si((__v8si)__X, (__v8si)__Y);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sllv_epi32(__m128i __X, __m128i __Y)
+{
+  return (__m128i)__builtin_ia32_psllv4si((__v4si)__X, (__v4si)__Y);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_sllv_epi64(__m256i __X, __m256i __Y)
+{
+  return (__m256i)__builtin_ia32_psllv4di(__X, __Y);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sllv_epi64(__m128i __X, __m128i __Y)
+{
+  return (__m128i)__builtin_ia32_psllv2di(__X, __Y);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srav_epi32(__m256i __X, __m256i __Y)
+{
+  return (__m256i)__builtin_ia32_psrav8si((__v8si)__X, (__v8si)__Y);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srav_epi32(__m128i __X, __m128i __Y)
+{
+  return (__m128i)__builtin_ia32_psrav4si((__v4si)__X, (__v4si)__Y);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srlv_epi32(__m256i __X, __m256i __Y)
+{
+  return (__m256i)__builtin_ia32_psrlv8si((__v8si)__X, (__v8si)__Y);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srlv_epi32(__m128i __X, __m128i __Y)
+{
+  return (__m128i)__builtin_ia32_psrlv4si((__v4si)__X, (__v4si)__Y);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_srlv_epi64(__m256i __X, __m256i __Y)
+{
+  return (__m256i)__builtin_ia32_psrlv4di(__X, __Y);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srlv_epi64(__m128i __X, __m128i __Y)
+{
+  return (__m128i)__builtin_ia32_psrlv2di(__X, __Y);
+}
+
+#define _mm_mask_i32gather_pd(a, m, i, mask, s) __extension__ ({ \
+  __m128d __a = (a); \
+  double const *__m = (m); \
+  __m128i __i = (i); \
+  __m128d __mask = (mask); \
+  (__m128d)__builtin_ia32_gatherd_pd((__v2df)__a, (const __v2df *)__m, \
+             (__v4si)__i, (__v2df)__mask, (s)); })
+
+#define _mm256_mask_i32gather_pd(a, m, i, mask, s) __extension__ ({ \
+  __m256d __a = (a); \
+  double const *__m = (m); \
+  __m128i __i = (i); \
+  __m256d __mask = (mask); \
+  (__m256d)__builtin_ia32_gatherd_pd256((__v4df)__a, (const __v4df *)__m, \
+             (__v4si)__i, (__v4df)__mask, (s)); })
+
+#define _mm_mask_i64gather_pd(a, m, i, mask, s) __extension__ ({ \
+  __m128d __a = (a); \
+  double const *__m = (m); \
+  __m128i __i = (i); \
+  __m128d __mask = (mask); \
+  (__m128d)__builtin_ia32_gatherq_pd((__v2df)__a, (const __v2df *)__m, \
+             (__v2di)__i, (__v2df)__mask, (s)); })
+
+#define _mm256_mask_i64gather_pd(a, m, i, mask, s) __extension__ ({ \
+  __m256d __a = (a); \
+  double const *__m = (m); \
+  __m256i __i = (i); \
+  __m256d __mask = (mask); \
+  (__m256d)__builtin_ia32_gatherq_pd256((__v4df)__a, (const __v4df *)__m, \
+             (__v4di)__i, (__v4df)__mask, (s)); })
+
+#define _mm_mask_i32gather_ps(a, m, i, mask, s) __extension__ ({ \
+  __m128 __a = (a); \
+  float const *__m = (m); \
+  __m128i __i = (i); \
+  __m128 __mask = (mask); \
+  (__m128)__builtin_ia32_gatherd_ps((__v4sf)__a, (const __v4sf *)__m, \
+            (__v4si)__i, (__v4sf)__mask, (s)); })
+
+#define _mm256_mask_i32gather_ps(a, m, i, mask, s) __extension__ ({ \
+  __m256 __a = (a); \
+  float const *__m = (m); \
+  __m256i __i = (i); \
+  __m256 __mask = (mask); \
+  (__m256)__builtin_ia32_gatherd_ps256((__v8sf)__a, (const __v8sf *)__m, \
+            (__v8si)__i, (__v8sf)__mask, (s)); })
+
+#define _mm_mask_i64gather_ps(a, m, i, mask, s) __extension__ ({ \
+  __m128 __a = (a); \
+  float const *__m = (m); \
+  __m128i __i = (i); \
+  __m128 __mask = (mask); \
+  (__m128)__builtin_ia32_gatherq_ps((__v4sf)__a, (const __v4sf *)__m, \
+            (__v2di)__i, (__v4sf)__mask, (s)); })
+
+#define _mm256_mask_i64gather_ps(a, m, i, mask, s) __extension__ ({ \
+  __m128 __a = (a); \
+  float const *__m = (m); \
+  __m256i __i = (i); \
+  __m128 __mask = (mask); \
+  (__m128)__builtin_ia32_gatherq_ps256((__v4sf)__a, (const __v4sf *)__m, \
+            (__v4di)__i, (__v4sf)__mask, (s)); })
+
+#define _mm_mask_i32gather_epi32(a, m, i, mask, s) __extension__ ({ \
+  __m128i __a = (a); \
+  int const *__m = (m); \
+  __m128i __i = (i); \
+  __m128i __mask = (mask); \
+  (__m128i)__builtin_ia32_gatherd_d((__v4si)__a, (const __v4si *)__m, \
+            (__v4si)__i, (__v4si)__mask, (s)); })
+
+#define _mm256_mask_i32gather_epi32(a, m, i, mask, s) __extension__ ({ \
+  __m256i __a = (a); \
+  int const *__m = (m); \
+  __m256i __i = (i); \
+  __m256i __mask = (mask); \
+  (__m256i)__builtin_ia32_gatherd_d256((__v8si)__a, (const __v8si *)__m, \
+            (__v8si)__i, (__v8si)__mask, (s)); })
+
+#define _mm_mask_i64gather_epi32(a, m, i, mask, s) __extension__ ({ \
+  __m128i __a = (a); \
+  int const *__m = (m); \
+  __m128i __i = (i); \
+  __m128i __mask = (mask); \
+  (__m128i)__builtin_ia32_gatherq_d((__v4si)__a, (const __v4si *)__m, \
+            (__v2di)__i, (__v4si)__mask, (s)); })
+
+#define _mm256_mask_i64gather_epi32(a, m, i, mask, s) __extension__ ({ \
+  __m128i __a = (a); \
+  int const *__m = (m); \
+  __m256i __i = (i); \
+  __m128i __mask = (mask); \
+  (__m128i)__builtin_ia32_gatherq_d256((__v4si)__a, (const __v4si *)__m, \
+            (__v4di)__i, (__v4si)__mask, (s)); })
+
+#define _mm_mask_i32gather_epi64(a, m, i, mask, s) __extension__ ({ \
+  __m128i __a = (a); \
+  long long const *__m = (m); \
+  __m128i __i = (i); \
+  __m128i __mask = (mask); \
+  (__m128i)__builtin_ia32_gatherd_q((__v2di)__a, (const __v2di *)__m, \
+             (__v4si)__i, (__v2di)__mask, (s)); })
+
+#define _mm256_mask_i32gather_epi64(a, m, i, mask, s) __extension__ ({ \
+  __m256i __a = (a); \
+  long long const *__m = (m); \
+  __m128i __i = (i); \
+  __m256i __mask = (mask); \
+  (__m256i)__builtin_ia32_gatherd_q256((__v4di)__a, (const __v4di *)__m, \
+             (__v4si)__i, (__v4di)__mask, (s)); })
+
+#define _mm_mask_i64gather_epi64(a, m, i, mask, s) __extension__ ({ \
+  __m128i __a = (a); \
+  long long const *__m = (m); \
+  __m128i __i = (i); \
+  __m128i __mask = (mask); \
+  (__m128i)__builtin_ia32_gatherq_q((__v2di)__a, (const __v2di *)__m, \
+             (__v2di)__i, (__v2di)__mask, (s)); })
+
+#define _mm256_mask_i64gather_epi64(a, m, i, mask, s) __extension__ ({ \
+  __m256i __a = (a); \
+  long long const *__m = (m); \
+  __m256i __i = (i); \
+  __m256i __mask = (mask); \
+  (__m256i)__builtin_ia32_gatherq_q256((__v4di)__a, (const __v4di *)__m, \
+             (__v4di)__i, (__v4di)__mask, (s)); })
+
+#define _mm_i32gather_pd(m, i, s) __extension__ ({ \
+  double const *__m = (m); \
+  __m128i __i = (i); \
+  (__m128d)__builtin_ia32_gatherd_pd((__v2df)_mm_setzero_pd(), \
+             (const __v2df *)__m, (__v4si)__i, \
+             (__v2df)_mm_set1_pd((double)(long long int)-1), (s)); })
+
+#define _mm256_i32gather_pd(m, i, s) __extension__ ({ \
+  double const *__m = (m); \
+  __m128i __i = (i); \
+  (__m256d)__builtin_ia32_gatherd_pd256((__v4df)_mm256_setzero_pd(), \
+             (const __v4df *)__m, (__v4si)__i, \
+             (__v4df)_mm256_set1_pd((double)(long long int)-1), (s)); })
+
+#define _mm_i64gather_pd(m, i, s) __extension__ ({ \
+  double const *__m = (m); \
+  __m128i __i = (i); \
+  (__m128d)__builtin_ia32_gatherq_pd((__v2df)_mm_setzero_pd(), \
+             (const __v2df *)__m, (__v2di)__i, \
+             (__v2df)_mm_set1_pd((double)(long long int)-1), (s)); })
+
+#define _mm256_i64gather_pd(m, i, s) __extension__ ({ \
+  double const *__m = (m); \
+  __m256i __i = (i); \
+  (__m256d)__builtin_ia32_gatherq_pd256((__v4df)_mm256_setzero_pd(), \
+             (const __v4df *)__m, (__v4di)__i, \
+             (__v4df)_mm256_set1_pd((double)(long long int)-1), (s)); })
+
+#define _mm_i32gather_ps(m, i, s) __extension__ ({ \
+  float const *__m = (m); \
+  __m128i __i = (i); \
+  (__m128)__builtin_ia32_gatherd_ps((__v4sf)_mm_setzero_ps(), \
+             (const __v4sf *)__m, (__v4si)__i, \
+             (__v4sf)_mm_set1_ps((float)(int)-1), (s)); })
+
+#define _mm256_i32gather_ps(m, i, s) __extension__ ({ \
+  float const *__m = (m); \
+  __m256i __i = (i); \
+  (__m256)__builtin_ia32_gatherd_ps256((__v8sf)_mm256_setzero_ps(), \
+             (const __v8sf *)__m, (__v8si)__i, \
+             (__v8sf)_mm256_set1_ps((float)(int)-1), (s)); })
+
+#define _mm_i64gather_ps(m, i, s) __extension__ ({ \
+  float const *__m = (m); \
+  __m128i __i = (i); \
+  (__m128)__builtin_ia32_gatherq_ps((__v4sf)_mm_setzero_ps(), \
+             (const __v4sf *)__m, (__v2di)__i, \
+             (__v4sf)_mm_set1_ps((float)(int)-1), (s)); })
+
+#define _mm256_i64gather_ps(m, i, s) __extension__ ({ \
+  float const *__m = (m); \
+  __m256i __i = (i); \
+  (__m128)__builtin_ia32_gatherq_ps256((__v4sf)_mm_setzero_ps(), \
+             (const __v4sf *)__m, (__v4di)__i, \
+             (__v4sf)_mm_set1_ps((float)(int)-1), (s)); })
+
+#define _mm_i32gather_epi32(m, i, s) __extension__ ({ \
+  int const *__m = (m); \
+  __m128i __i = (i); \
+  (__m128i)__builtin_ia32_gatherd_d((__v4si)_mm_setzero_si128(), \
+            (const __v4si *)__m, (__v4si)__i, \
+            (__v4si)_mm_set1_epi32(-1), (s)); })
+
+#define _mm256_i32gather_epi32(m, i, s) __extension__ ({ \
+  int const *__m = (m); \
+  __m256i __i = (i); \
+  (__m256i)__builtin_ia32_gatherd_d256((__v8si)_mm256_setzero_si256(), \
+            (const __v8si *)__m, (__v8si)__i, \
+            (__v8si)_mm256_set1_epi32(-1), (s)); })
+
+#define _mm_i64gather_epi32(m, i, s) __extension__ ({ \
+  int const *__m = (m); \
+  __m128i __i = (i); \
+  (__m128i)__builtin_ia32_gatherq_d((__v4si)_mm_setzero_si128(), \
+            (const __v4si *)__m, (__v2di)__i, \
+            (__v4si)_mm_set1_epi32(-1), (s)); })
+
+#define _mm256_i64gather_epi32(m, i, s) __extension__ ({ \
+  int const *__m = (m); \
+  __m256i __i = (i); \
+  (__m128i)__builtin_ia32_gatherq_d256((__v4si)_mm_setzero_si128(), \
+            (const __v4si *)__m, (__v4di)__i, \
+            (__v4si)_mm_set1_epi32(-1), (s)); })
+
+#define _mm_i32gather_epi64(m, i, s) __extension__ ({ \
+  long long const *__m = (m); \
+  __m128i __i = (i); \
+  (__m128i)__builtin_ia32_gatherd_q((__v2di)_mm_setzero_si128(), \
+             (const __v2di *)__m, (__v4si)__i, \
+             (__v2di)_mm_set1_epi64x(-1), (s)); })
+
+#define _mm256_i32gather_epi64(m, i, s) __extension__ ({ \
+  long long const *__m = (m); \
+  __m128i __i = (i); \
+  (__m256i)__builtin_ia32_gatherd_q256((__v4di)_mm256_setzero_si256(), \
+             (const __v4di *)__m, (__v4si)__i, \
+             (__v4di)_mm256_set1_epi64x(-1), (s)); })
+
+#define _mm_i64gather_epi64(m, i, s) __extension__ ({ \
+  long long const *__m = (m); \
+  __m128i __i = (i); \
+  (__m128i)__builtin_ia32_gatherq_q((__v2di)_mm_setzero_si128(), \
+             (const __v2di *)__m, (__v2di)__i, \
+             (__v2di)_mm_set1_epi64x(-1), (s)); })
+
+#define _mm256_i64gather_epi64(m, i, s) __extension__ ({ \
+  long long const *__m = (m); \
+  __m256i __i = (i); \
+  (__m256i)__builtin_ia32_gatherq_q256((__v4di)_mm256_setzero_si256(), \
+             (const __v4di *)__m, (__v4di)__i, \
+             (__v4di)_mm256_set1_epi64x(-1), (s)); })
+
+#endif /* __AVX2INTRIN_H */
diff --git a/lib/clang/3.6/include/avx512bwintrin.h b/lib/clang/3.6/include/avx512bwintrin.h
new file mode 100644
index 0000000..bc4d4ac
--- /dev/null
+++ b/lib/clang/3.6/include/avx512bwintrin.h
@@ -0,0 +1,60 @@
+/*===------------- avx512bwintrin.h - AVX512BW intrinsics ------------------===
+ *
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __AVX512BWINTRIN_H
+#define __AVX512BWINTRIN_H
+
+typedef unsigned int __mmask32;
+typedef unsigned long long __mmask64;
+typedef char __v64qi __attribute__ ((vector_size (64)));
+typedef short __v32hi __attribute__ ((__vector_size__ (64)));
+
+
+/* Integer compare */
+
+static __inline__ __mmask64 __attribute__((__always_inline__, __nodebug__))
+_mm512_cmpeq_epi8_mask(__m512i __a, __m512i __b) {
+  return (__mmask64)__builtin_ia32_pcmpeqb512_mask((__v64qi)__a, (__v64qi)__b,
+                                                   (__mmask64)-1);
+}
+
+static __inline__ __mmask64 __attribute__((__always_inline__, __nodebug__))
+_mm512_mask_cmpeq_epi8_mask(__mmask64 __u, __m512i __a, __m512i __b) {
+  return (__mmask64)__builtin_ia32_pcmpeqb512_mask((__v64qi)__a, (__v64qi)__b,
+                                                   __u);
+}
+
+static __inline__ __mmask32 __attribute__((__always_inline__, __nodebug__))
+_mm512_cmpeq_epi16_mask(__m512i __a, __m512i __b) {
+  return (__mmask32)__builtin_ia32_pcmpeqw512_mask((__v32hi)__a, (__v32hi)__b,
+                                                   (__mmask32)-1);
+}
+
+static __inline__ __mmask32 __attribute__((__always_inline__, __nodebug__))
+_mm512_mask_cmpeq_epi16_mask(__mmask32 __u, __m512i __a, __m512i __b) {
+  return (__mmask32)__builtin_ia32_pcmpeqw512_mask((__v32hi)__a, (__v32hi)__b,
+                                                   __u);
+}
+
+#endif
diff --git a/lib/clang/3.6/include/avx512erintrin.h b/lib/clang/3.6/include/avx512erintrin.h
new file mode 100644
index 0000000..1a5ea15
--- /dev/null
+++ b/lib/clang/3.6/include/avx512erintrin.h
@@ -0,0 +1,112 @@
+/*===---- avx512fintrin.h - AVX2 intrinsics -----------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+#ifndef __IMMINTRIN_H
+#error "Never use <avx512erintrin.h> directly; include <immintrin.h> instead."
+#endif
+
+#ifndef __AVX512ERINTRIN_H
+#define __AVX512ERINTRIN_H
+
+
+// rsqrt28
+static  __inline__ __m512d __attribute__((__always_inline__, __nodebug__))
+_mm512_rsqrt28_round_pd (__m512d __A, int __R)
+{
+  return (__m512d)__builtin_ia32_rsqrt28pd_mask ((__v8df)__A,
+                                                 (__v8df)_mm512_setzero_pd(),
+                                                 (__mmask8)-1,
+                                                 __R);
+}
+static  __inline__ __m512 __attribute__((__always_inline__, __nodebug__))
+_mm512_rsqrt28_round_ps(__m512 __A, int __R)
+{
+  return (__m512)__builtin_ia32_rsqrt28ps_mask ((__v16sf)__A,
+                                                (__v16sf)_mm512_setzero_ps(),
+                                                (__mmask16)-1,
+                                                __R);
+}
+
+static  __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_rsqrt28_round_ss(__m128 __A, __m128 __B, int __R)
+{
+  return (__m128) __builtin_ia32_rsqrt28ss_mask ((__v4sf) __A,
+             (__v4sf) __B,
+             (__v4sf)
+             _mm_setzero_ps (),
+             (__mmask8) -1,
+             __R);
+}
+
+static  __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_rsqrt28_round_sd (__m128d __A, __m128d __B, int __R)
+{
+  return (__m128d) __builtin_ia32_rsqrt28sd_mask ((__v2df) __A,
+              (__v2df) __B,
+              (__v2df)
+              _mm_setzero_pd (),
+              (__mmask8) -1,
+             __R);
+}
+
+
+// rcp28
+static  __inline__ __m512d __attribute__((__always_inline__, __nodebug__))
+_mm512_rcp28_round_pd (__m512d __A, int __R)
+{
+  return (__m512d)__builtin_ia32_rcp28pd_mask ((__v8df)__A,
+                                               (__v8df)_mm512_setzero_pd(),
+                                               (__mmask8)-1,
+                                               __R);
+}
+
+static  __inline__ __m512 __attribute__((__always_inline__, __nodebug__))
+_mm512_rcp28_round_ps (__m512 __A, int __R)
+{
+  return (__m512)__builtin_ia32_rcp28ps_mask ((__v16sf)__A,
+                                              (__v16sf)_mm512_setzero_ps (),
+                                              (__mmask16)-1,
+                                              __R);
+}
+
+static  __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_rcp28_round_ss (__m128 __A, __m128 __B, int __R)
+{
+  return (__m128) __builtin_ia32_rcp28ss_mask ((__v4sf) __A,
+             (__v4sf) __B,
+             (__v4sf)
+             _mm_setzero_ps (),
+             (__mmask8) -1,
+             __R);
+}
+static  __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_rcp28_round_sd (__m128d __A, __m128d __B, int __R)
+{
+  return (__m128d) __builtin_ia32_rcp28sd_mask ((__v2df) __A,
+              (__v2df) __B,
+              (__v2df)
+              _mm_setzero_pd (),
+              (__mmask8) -1,
+             __R);
+}
+
+#endif // __AVX512ERINTRIN_H
diff --git a/lib/clang/3.6/include/avx512fintrin.h b/lib/clang/3.6/include/avx512fintrin.h
new file mode 100644
index 0000000..9591dcf
--- /dev/null
+++ b/lib/clang/3.6/include/avx512fintrin.h
@@ -0,0 +1,1003 @@
+/*===---- avx512fintrin.h - AVX2 intrinsics --------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+#ifndef __IMMINTRIN_H
+#error "Never use <avx512fintrin.h> directly; include <immintrin.h> instead."
+#endif
+
+#ifndef __AVX512FINTRIN_H
+#define __AVX512FINTRIN_H
+
+typedef double __v8df __attribute__((__vector_size__(64)));
+typedef float __v16sf __attribute__((__vector_size__(64)));
+typedef long long __v8di __attribute__((__vector_size__(64)));
+typedef int __v16si __attribute__((__vector_size__(64)));
+
+typedef float __m512 __attribute__((__vector_size__(64)));
+typedef double __m512d __attribute__((__vector_size__(64)));
+typedef long long __m512i __attribute__((__vector_size__(64)));
+
+typedef unsigned char __mmask8;
+typedef unsigned short __mmask16;
+
+/* Rounding mode macros.  */
+#define _MM_FROUND_TO_NEAREST_INT   0x00
+#define _MM_FROUND_TO_NEG_INF       0x01
+#define _MM_FROUND_TO_POS_INF       0x02
+#define _MM_FROUND_TO_ZERO          0x03
+#define _MM_FROUND_CUR_DIRECTION    0x04
+
+/* Create vectors with repeated elements */
+
+static  __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_setzero_si512(void)
+{
+  return (__m512i)(__v8di){ 0, 0, 0, 0, 0, 0, 0, 0 };
+}
+
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_maskz_set1_epi32(__mmask16 __M, int __A)
+{
+  return (__m512i) __builtin_ia32_pbroadcastd512_gpr_mask (__A,
+                 (__v16si)
+                 _mm512_setzero_si512 (),
+                 __M);
+}
+
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_maskz_set1_epi64(__mmask8 __M, long long __A)
+{
+#ifdef __x86_64__
+  return (__m512i) __builtin_ia32_pbroadcastq512_gpr_mask (__A,
+                 (__v8di)
+                 _mm512_setzero_si512 (),
+                 __M);
+#else
+  return (__m512i) __builtin_ia32_pbroadcastq512_mem_mask (__A,
+                 (__v8di)
+                 _mm512_setzero_si512 (),
+                 __M);
+#endif
+}
+
+static __inline __m512 __attribute__ ((__always_inline__, __nodebug__))
+_mm512_setzero_ps(void)
+{
+  return (__m512){ 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,
+                   0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0 };
+}
+static  __inline __m512d __attribute__ ((__always_inline__, __nodebug__))
+_mm512_setzero_pd(void)
+{
+  return (__m512d){ 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0 };
+}
+
+static __inline __m512 __attribute__((__always_inline__, __nodebug__))
+_mm512_set1_ps(float __w)
+{
+  return (__m512){ __w, __w, __w, __w, __w, __w, __w, __w,
+                   __w, __w, __w, __w, __w, __w, __w, __w  };
+}
+
+static __inline __m512d __attribute__((__always_inline__, __nodebug__))
+_mm512_set1_pd(double __w)
+{
+  return (__m512d){ __w, __w, __w, __w, __w, __w, __w, __w };
+}
+
+static __inline __m512i __attribute__((__always_inline__, __nodebug__))
+_mm512_set1_epi32(int __s)
+{
+  return (__m512i)(__v16si){ __s, __s, __s, __s, __s, __s, __s, __s,
+                             __s, __s, __s, __s, __s, __s, __s, __s };
+}
+
+static __inline __m512i __attribute__((__always_inline__, __nodebug__))
+_mm512_set1_epi64(long long __d)
+{
+  return (__m512i)(__v8di){ __d, __d, __d, __d, __d, __d, __d, __d };
+}
+
+static __inline__ __m512 __attribute__((__always_inline__, __nodebug__))
+_mm512_broadcastss_ps(__m128 __X)
+{
+  float __f = __X[0];
+  return (__v16sf){ __f, __f, __f, __f,
+                    __f, __f, __f, __f,
+                    __f, __f, __f, __f,
+                    __f, __f, __f, __f };
+}
+
+static __inline__ __m512d __attribute__((__always_inline__, __nodebug__))
+_mm512_broadcastsd_pd(__m128d __X)
+{
+  double __d = __X[0];
+  return (__v8df){ __d, __d, __d, __d,
+                   __d, __d, __d, __d };
+}
+
+/* Cast between vector types */
+
+static __inline __m512d __attribute__((__always_inline__, __nodebug__))
+_mm512_castpd256_pd512(__m256d __a)
+{
+  return __builtin_shufflevector(__a, __a, 0, 1, 2, 3, -1, -1, -1, -1);
+}
+
+static __inline __m512 __attribute__((__always_inline__, __nodebug__))
+_mm512_castps256_ps512(__m256 __a)
+{
+  return __builtin_shufflevector(__a, __a, 0,  1,  2,  3,  4,  5,  6,  7,
+                                          -1, -1, -1, -1, -1, -1, -1, -1);
+}
+
+static __inline __m128d __attribute__((__always_inline__, __nodebug__))
+_mm512_castpd512_pd128(__m512d __a)
+{
+  return __builtin_shufflevector(__a, __a, 0, 1);
+}
+
+static __inline __m128 __attribute__((__always_inline__, __nodebug__))
+_mm512_castps512_ps128(__m512 __a)
+{
+  return __builtin_shufflevector(__a, __a, 0, 1, 2, 3);
+}
+
+/* Arithmetic */
+
+static __inline __m512d __attribute__((__always_inline__, __nodebug__))
+_mm512_add_pd(__m512d __a, __m512d __b)
+{
+  return __a + __b;
+}
+
+static __inline __m512 __attribute__((__always_inline__, __nodebug__))
+_mm512_add_ps(__m512 __a, __m512 __b)
+{
+  return __a + __b;
+}
+
+static __inline __m512d __attribute__((__always_inline__, __nodebug__))
+_mm512_mul_pd(__m512d __a, __m512d __b)
+{
+  return __a * __b;
+}
+
+static __inline __m512 __attribute__((__always_inline__, __nodebug__))
+_mm512_mul_ps(__m512 __a, __m512 __b)
+{
+  return __a * __b;
+}
+
+static __inline __m512d __attribute__((__always_inline__, __nodebug__))
+_mm512_sub_pd(__m512d __a, __m512d __b)
+{
+  return __a - __b;
+}
+
+static __inline __m512 __attribute__((__always_inline__, __nodebug__))
+_mm512_sub_ps(__m512 __a, __m512 __b)
+{
+  return __a - __b;
+}
+
+static  __inline__ __m512d __attribute__((__always_inline__, __nodebug__))
+_mm512_max_pd(__m512d __A, __m512d __B)
+{
+  return (__m512d) __builtin_ia32_maxpd512_mask ((__v8df) __A,
+             (__v8df) __B,
+             (__v8df)
+             _mm512_setzero_pd (),
+             (__mmask8) -1,
+             _MM_FROUND_CUR_DIRECTION);
+}
+
+static  __inline__ __m512 __attribute__((__always_inline__, __nodebug__))
+_mm512_max_ps(__m512 __A, __m512 __B)
+{
+  return (__m512) __builtin_ia32_maxps512_mask ((__v16sf) __A,
+            (__v16sf) __B,
+            (__v16sf)
+            _mm512_setzero_ps (),
+            (__mmask16) -1,
+            _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline __m512i
+__attribute__ ((__always_inline__, __nodebug__))
+_mm512_max_epi32(__m512i __A, __m512i __B)
+{
+  return (__m512i) __builtin_ia32_pmaxsd512_mask ((__v16si) __A,
+              (__v16si) __B,
+              (__v16si)
+              _mm512_setzero_si512 (),
+              (__mmask16) -1);
+}
+
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_max_epu32(__m512i __A, __m512i __B)
+{
+  return (__m512i) __builtin_ia32_pmaxud512_mask ((__v16si) __A,
+              (__v16si) __B,
+              (__v16si)
+              _mm512_setzero_si512 (),
+              (__mmask16) -1);
+}
+
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_max_epi64(__m512i __A, __m512i __B)
+{
+  return (__m512i) __builtin_ia32_pmaxsq512_mask ((__v8di) __A,
+              (__v8di) __B,
+              (__v8di)
+              _mm512_setzero_si512 (),
+              (__mmask8) -1);
+}
+
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_max_epu64(__m512i __A, __m512i __B)
+{
+  return (__m512i) __builtin_ia32_pmaxuq512_mask ((__v8di) __A,
+              (__v8di) __B,
+              (__v8di)
+              _mm512_setzero_si512 (),
+              (__mmask8) -1);
+}
+
+static  __inline__ __m512d __attribute__((__always_inline__, __nodebug__))
+_mm512_min_pd(__m512d __A, __m512d __B)
+{
+  return (__m512d) __builtin_ia32_minpd512_mask ((__v8df) __A,
+             (__v8df) __B,
+             (__v8df)
+             _mm512_setzero_pd (),
+             (__mmask8) -1,
+             _MM_FROUND_CUR_DIRECTION);
+}
+
+static  __inline__ __m512 __attribute__((__always_inline__, __nodebug__))
+_mm512_min_ps(__m512 __A, __m512 __B)
+{
+  return (__m512) __builtin_ia32_minps512_mask ((__v16sf) __A,
+            (__v16sf) __B,
+            (__v16sf)
+            _mm512_setzero_ps (),
+            (__mmask16) -1,
+            _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline __m512i
+__attribute__ ((__always_inline__, __nodebug__))
+_mm512_min_epi32(__m512i __A, __m512i __B)
+{
+  return (__m512i) __builtin_ia32_pminsd512_mask ((__v16si) __A,
+              (__v16si) __B,
+              (__v16si)
+              _mm512_setzero_si512 (),
+              (__mmask16) -1);
+}
+
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_min_epu32(__m512i __A, __m512i __B)
+{
+  return (__m512i) __builtin_ia32_pminud512_mask ((__v16si) __A,
+              (__v16si) __B,
+              (__v16si)
+              _mm512_setzero_si512 (),
+              (__mmask16) -1);
+}
+
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_min_epi64(__m512i __A, __m512i __B)
+{
+  return (__m512i) __builtin_ia32_pminsq512_mask ((__v8di) __A,
+              (__v8di) __B,
+              (__v8di)
+              _mm512_setzero_si512 (),
+              (__mmask8) -1);
+}
+
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_min_epu64(__m512i __A, __m512i __B)
+{
+  return (__m512i) __builtin_ia32_pminuq512_mask ((__v8di) __A,
+              (__v8di) __B,
+              (__v8di)
+              _mm512_setzero_si512 (),
+              (__mmask8) -1);
+}
+
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_mul_epi32(__m512i __X, __m512i __Y)
+{
+  return (__m512i) __builtin_ia32_pmuldq512_mask ((__v16si) __X,
+              (__v16si) __Y,
+              (__v8di)
+              _mm512_setzero_si512 (),
+              (__mmask8) -1);
+}
+
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_mul_epu32(__m512i __X, __m512i __Y)
+{
+  return (__m512i) __builtin_ia32_pmuludq512_mask ((__v16si) __X,
+               (__v16si) __Y,
+               (__v8di)
+               _mm512_setzero_si512 (),
+               (__mmask8) -1);
+}
+
+static  __inline__ __m512d __attribute__((__always_inline__, __nodebug__))
+_mm512_sqrt_pd(__m512d a)
+{
+  return (__m512d)__builtin_ia32_sqrtpd512_mask((__v8df)a,
+                                                (__v8df) _mm512_setzero_pd (),
+                                                (__mmask8) -1,
+                                                _MM_FROUND_CUR_DIRECTION);
+}
+
+static  __inline__ __m512 __attribute__((__always_inline__, __nodebug__))
+_mm512_sqrt_ps(__m512 a)
+{
+  return (__m512)__builtin_ia32_sqrtps512_mask((__v16sf)a,
+                                               (__v16sf) _mm512_setzero_ps (),
+                                               (__mmask16) -1,
+                                               _MM_FROUND_CUR_DIRECTION);
+}
+
+static  __inline__ __m512d __attribute__((__always_inline__, __nodebug__))
+_mm512_rsqrt14_pd(__m512d __A)
+{
+  return (__m512d) __builtin_ia32_rsqrt14pd512_mask ((__v8df) __A,
+                 (__v8df)
+                 _mm512_setzero_pd (),
+                 (__mmask8) -1);}
+
+static  __inline__ __m512 __attribute__((__always_inline__, __nodebug__))
+_mm512_rsqrt14_ps(__m512 __A)
+{
+  return (__m512) __builtin_ia32_rsqrt14ps512_mask ((__v16sf) __A,
+                (__v16sf)
+                _mm512_setzero_ps (),
+                (__mmask16) -1);
+}
+
+static  __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_rsqrt14_ss(__m128 __A, __m128 __B)
+{
+  return (__m128) __builtin_ia32_rsqrt14ss_mask ((__v4sf) __A,
+             (__v4sf) __B,
+             (__v4sf)
+             _mm_setzero_ps (),
+             (__mmask8) -1);
+}
+
+static  __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_rsqrt14_sd(__m128d __A, __m128d __B)
+{
+  return (__m128d) __builtin_ia32_rsqrt14sd_mask ((__v2df) __A,
+              (__v2df) __B,
+              (__v2df)
+              _mm_setzero_pd (),
+              (__mmask8) -1);
+}
+
+static  __inline__ __m512d __attribute__((__always_inline__, __nodebug__))
+_mm512_rcp14_pd(__m512d __A)
+{
+  return (__m512d) __builtin_ia32_rcp14pd512_mask ((__v8df) __A,
+               (__v8df)
+               _mm512_setzero_pd (),
+               (__mmask8) -1);
+}
+
+static  __inline__ __m512 __attribute__((__always_inline__, __nodebug__))
+_mm512_rcp14_ps(__m512 __A)
+{
+  return (__m512) __builtin_ia32_rcp14ps512_mask ((__v16sf) __A,
+              (__v16sf)
+              _mm512_setzero_ps (),
+              (__mmask16) -1);
+}
+static  __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_rcp14_ss(__m128 __A, __m128 __B)
+{
+  return (__m128) __builtin_ia32_rcp14ss_mask ((__v4sf) __A,
+                 (__v4sf) __B,
+                 (__v4sf)
+                 _mm_setzero_ps (),
+                 (__mmask8) -1);
+}
+
+static  __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_rcp14_sd(__m128d __A, __m128d __B)
+{
+  return (__m128d) __builtin_ia32_rcp14sd_mask ((__v2df) __A,
+            (__v2df) __B,
+            (__v2df)
+            _mm_setzero_pd (),
+            (__mmask8) -1);
+}
+
+static __inline __m512 __attribute__ ((__always_inline__, __nodebug__))
+_mm512_floor_ps(__m512 __A)
+{
+  return (__m512) __builtin_ia32_rndscaleps_mask ((__v16sf) __A,
+                                                  _MM_FROUND_FLOOR,
+                                                  (__v16sf) __A, -1,
+                                                  _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline __m512d __attribute__ ((__always_inline__, __nodebug__))
+_mm512_floor_pd(__m512d __A)
+{
+  return (__m512d) __builtin_ia32_rndscalepd_mask ((__v8df) __A,
+                                                   _MM_FROUND_FLOOR,
+                                                   (__v8df) __A, -1,
+                                                   _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline __m512 __attribute__ ((__always_inline__, __nodebug__))
+_mm512_ceil_ps(__m512 __A)
+{
+  return (__m512) __builtin_ia32_rndscaleps_mask ((__v16sf) __A,
+                                                  _MM_FROUND_CEIL,
+                                                  (__v16sf) __A, -1,
+                                                  _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline __m512d __attribute__ ((__always_inline__, __nodebug__))
+_mm512_ceil_pd(__m512d __A)
+{
+  return (__m512d) __builtin_ia32_rndscalepd_mask ((__v8df) __A,
+                                                   _MM_FROUND_CEIL,
+                                                   (__v8df) __A, -1,
+                                                   _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline __m512i __attribute__ (( __always_inline__, __nodebug__))
+_mm512_abs_epi64(__m512i __A)
+{
+  return (__m512i) __builtin_ia32_pabsq512_mask ((__v8di) __A,
+             (__v8di)
+             _mm512_setzero_si512 (),
+             (__mmask8) -1);
+}
+
+static __inline __m512i __attribute__ (( __always_inline__, __nodebug__))
+_mm512_abs_epi32(__m512i __A)
+{
+  return (__m512i) __builtin_ia32_pabsd512_mask ((__v16si) __A,
+             (__v16si)
+             _mm512_setzero_si512 (),
+             (__mmask16) -1);
+}
+
+static __inline __m512 __attribute__ ((__always_inline__, __nodebug__))
+_mm512_roundscale_ps(__m512 __A, const int __imm)
+{
+  return (__m512) __builtin_ia32_rndscaleps_mask ((__v16sf) __A, __imm,
+              (__v16sf) __A, -1,
+              _MM_FROUND_CUR_DIRECTION);
+}
+static __inline __m512d __attribute__ ((__always_inline__, __nodebug__))
+_mm512_roundscale_pd(__m512d __A, const int __imm)
+{
+  return (__m512d) __builtin_ia32_rndscalepd_mask ((__v8df) __A, __imm,
+               (__v8df) __A, -1,
+               _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m512d __attribute__((__always_inline__, __nodebug__))
+_mm512_fmadd_pd(__m512d __A, __m512d __B, __m512d __C)
+{
+  return (__m512d)
+    __builtin_ia32_vfmaddpd512_mask(__A,
+                                    __B,
+                                    __C,
+                                    (__mmask8) -1,
+                                    _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m512d __attribute__((__always_inline__, __nodebug__))
+_mm512_fmsub_pd(__m512d __A, __m512d __B, __m512d __C)
+{
+  return (__m512d)
+    __builtin_ia32_vfmsubpd512_mask(__A,
+                                    __B,
+                                    __C,
+                                    (__mmask8) -1,
+                                    _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m512d __attribute__((__always_inline__, __nodebug__))
+_mm512_fnmadd_pd(__m512d __A, __m512d __B, __m512d __C)
+{
+  return (__m512d)
+    __builtin_ia32_vfnmaddpd512_mask(__A,
+                                     __B,
+                                     __C,
+                                     (__mmask8) -1,
+                                     _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m512 __attribute__((__always_inline__, __nodebug__))
+_mm512_fmadd_ps(__m512 __A, __m512 __B, __m512 __C)
+{
+  return (__m512)
+    __builtin_ia32_vfmaddps512_mask(__A,
+                                    __B,
+                                    __C,
+                                    (__mmask16) -1,
+                                    _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m512 __attribute__((__always_inline__, __nodebug__))
+_mm512_fmsub_ps(__m512 __A, __m512 __B, __m512 __C)
+{
+  return (__m512)
+    __builtin_ia32_vfmsubps512_mask(__A,
+                                    __B,
+                                    __C,
+                                    (__mmask16) -1,
+                                    _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m512 __attribute__((__always_inline__, __nodebug__))
+_mm512_fnmadd_ps(__m512 __A, __m512 __B, __m512 __C)
+{
+  return (__m512)
+    __builtin_ia32_vfnmaddps512_mask(__A,
+                                     __B,
+                                     __C,
+                                     (__mmask16) -1,
+                                     _MM_FROUND_CUR_DIRECTION);
+}
+
+/* Vector permutations */
+
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_permutex2var_epi32(__m512i __A, __m512i __I, __m512i __B)
+{
+  return (__m512i) __builtin_ia32_vpermt2vard512_mask ((__v16si) __I
+                                                       /* idx */ ,
+                                                       (__v16si) __A,
+                                                       (__v16si) __B,
+                                                       (__mmask16) -1);
+}
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_permutex2var_epi64(__m512i __A, __m512i __I, __m512i __B)
+{
+  return (__m512i) __builtin_ia32_vpermt2varq512_mask ((__v8di) __I
+                                                       /* idx */ ,
+                                                       (__v8di) __A,
+                                                       (__v8di) __B,
+                                                       (__mmask8) -1);
+}
+
+static __inline __m512d __attribute__ ((__always_inline__, __nodebug__))
+_mm512_permutex2var_pd(__m512d __A, __m512i __I, __m512d __B)
+{
+  return (__m512d) __builtin_ia32_vpermt2varpd512_mask ((__v8di) __I
+                                                        /* idx */ ,
+                                                        (__v8df) __A,
+                                                        (__v8df) __B,
+                                                        (__mmask8) -1);
+}
+static __inline __m512 __attribute__ ((__always_inline__, __nodebug__))
+_mm512_permutex2var_ps(__m512 __A, __m512i __I, __m512 __B)
+{
+  return (__m512) __builtin_ia32_vpermt2varps512_mask ((__v16si) __I
+                                                       /* idx */ ,
+                                                       (__v16sf) __A,
+                                                       (__v16sf) __B,
+                                                       (__mmask16) -1);
+}
+
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_valign_epi64(__m512i __A, __m512i __B, const int __I)
+{
+  return (__m512i) __builtin_ia32_alignq512_mask((__v8di)__A,
+                                                 (__v8di)__B,
+                                                 __I,
+                                                 (__v8di)_mm512_setzero_si512(),
+                                                 (__mmask8) -1);
+}
+
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_valign_epi32(__m512i __A, __m512i __B, const int __I)
+{
+  return (__m512i)__builtin_ia32_alignd512_mask((__v16si)__A,
+                                                (__v16si)__B,
+                                                __I,
+                                                (__v16si)_mm512_setzero_si512(),
+                                                (__mmask16) -1);
+}
+
+/* Vector Blend */
+
+static __inline __m512d __attribute__ ((__always_inline__, __nodebug__))
+_mm512_mask_blend_pd(__mmask8 __U, __m512d __A, __m512d __W)
+{
+  return (__m512d) __builtin_ia32_blendmpd_512_mask ((__v8df) __A,
+                 (__v8df) __W,
+                 (__mmask8) __U);
+}
+
+static __inline __m512 __attribute__ ((__always_inline__, __nodebug__))
+_mm512_mask_blend_ps(__mmask16 __U, __m512 __A, __m512 __W)
+{
+  return (__m512) __builtin_ia32_blendmps_512_mask ((__v16sf) __A,
+                (__v16sf) __W,
+                (__mmask16) __U);
+}
+
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_mask_blend_epi64(__mmask8 __U, __m512i __A, __m512i __W)
+{
+  return (__m512i) __builtin_ia32_blendmq_512_mask ((__v8di) __A,
+                (__v8di) __W,
+                (__mmask8) __U);
+}
+
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_mask_blend_epi32(__mmask16 __U, __m512i __A, __m512i __W)
+{
+  return (__m512i) __builtin_ia32_blendmd_512_mask ((__v16si) __A,
+                (__v16si) __W,
+                (__mmask16) __U);
+}
+
+/* Compare */
+
+static __inline __mmask16 __attribute__ ((__always_inline__, __nodebug__))
+_mm512_cmp_ps_mask(__m512 a, __m512 b, const int p)
+{
+  return (__mmask16) __builtin_ia32_cmpps512_mask ((__v16sf) a,
+               (__v16sf) b, p, (__mmask16) -1,
+               _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline __mmask8 __attribute__ ((__always_inline__, __nodebug__)) 
+_mm512_cmp_pd_mask(__m512d __X, __m512d __Y, const int __P)
+{
+  return (__mmask8) __builtin_ia32_cmppd512_mask ((__v8df) __X,
+              (__v8df) __Y, __P,
+              (__mmask8) -1,
+              _MM_FROUND_CUR_DIRECTION);
+}
+
+/* Conversion */
+
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_cvttps_epu32(__m512 __A)
+{
+  return (__m512i) __builtin_ia32_cvttps2udq512_mask ((__v16sf) __A,
+                  (__v16si)
+                  _mm512_setzero_si512 (),
+                  (__mmask16) -1,
+                  _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline __m512 __attribute__ (( __always_inline__, __nodebug__))
+_mm512_cvt_roundepi32_ps(__m512i __A, const int __R)
+{
+  return (__m512) __builtin_ia32_cvtdq2ps512_mask ((__v16si) __A,
+               (__v16sf)
+               _mm512_setzero_ps (),
+               (__mmask16) -1,
+               __R);
+}
+
+static __inline __m512 __attribute__ (( __always_inline__, __nodebug__))
+_mm512_cvt_roundepu32_ps(__m512i __A, const int __R)
+{
+  return (__m512) __builtin_ia32_cvtudq2ps512_mask ((__v16si) __A,
+               (__v16sf)
+               _mm512_setzero_ps (),
+               (__mmask16) -1,
+               __R);
+}
+
+static __inline __m512d __attribute__ (( __always_inline__, __nodebug__))
+_mm512_cvtepi32_pd(__m256i __A)
+{
+  return (__m512d) __builtin_ia32_cvtdq2pd512_mask ((__v8si) __A,
+                (__v8df)
+                _mm512_setzero_pd (),
+                (__mmask8) -1);
+}
+
+static __inline __m512d __attribute__ (( __always_inline__, __nodebug__))
+_mm512_cvtepu32_pd(__m256i __A)
+{
+  return (__m512d) __builtin_ia32_cvtudq2pd512_mask ((__v8si) __A,
+                (__v8df)
+                _mm512_setzero_pd (),
+                (__mmask8) -1);
+}
+static __inline __m256 __attribute__ (( __always_inline__, __nodebug__))
+_mm512_cvt_roundpd_ps(__m512d __A, const int __R)
+{
+  return (__m256) __builtin_ia32_cvtpd2ps512_mask ((__v8df) __A,
+               (__v8sf)
+               _mm256_setzero_ps (),
+               (__mmask8) -1,
+               __R);
+}
+
+static  __inline __m256i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_cvtps_ph(__m512 __A, const int __I)
+{
+  return (__m256i) __builtin_ia32_vcvtps2ph512_mask ((__v16sf) __A,
+                 __I,
+                 (__v16hi)
+                 _mm256_setzero_si256 (),
+                 -1);
+}
+
+static  __inline __m512 __attribute__ ((__always_inline__, __nodebug__))
+_mm512_cvtph_ps(__m256i __A)
+{
+  return (__m512) __builtin_ia32_vcvtph2ps512_mask ((__v16hi) __A,
+                (__v16sf)
+                _mm512_setzero_ps (),
+                (__mmask16) -1,
+                _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline __m512i __attribute__((__always_inline__, __nodebug__))
+_mm512_cvttps_epi32(__m512 a)
+{
+  return (__m512i)
+    __builtin_ia32_cvttps2dq512_mask((__v16sf) a,
+                                     (__v16si) _mm512_setzero_si512 (),
+                                     (__mmask16) -1, _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm512_cvttpd_epi32(__m512d a)
+{
+  return (__m256i)__builtin_ia32_cvttpd2dq512_mask((__v8df) a,
+                                                   (__v8si)_mm256_setzero_si256(),
+                                                   (__mmask8) -1,
+                                                    _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline __m256i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_cvtt_roundpd_epi32(__m512d __A, const int __R)
+{
+  return (__m256i) __builtin_ia32_cvttpd2dq512_mask ((__v8df) __A,
+                 (__v8si)
+                 _mm256_setzero_si256 (),
+                 (__mmask8) -1,
+                 __R);
+}
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_cvtt_roundps_epi32(__m512 __A, const int __R)
+{
+  return (__m512i) __builtin_ia32_cvttps2dq512_mask ((__v16sf) __A,
+                 (__v16si)
+                 _mm512_setzero_si512 (),
+                 (__mmask16) -1,
+                 __R);
+}
+
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_cvt_roundps_epi32(__m512 __A, const int __R)
+{
+  return (__m512i) __builtin_ia32_cvtps2dq512_mask ((__v16sf) __A,
+                (__v16si)
+                _mm512_setzero_si512 (),
+                (__mmask16) -1,
+                __R);
+}
+static __inline __m256i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_cvt_roundpd_epi32(__m512d __A, const int __R)
+{
+  return (__m256i) __builtin_ia32_cvtpd2dq512_mask ((__v8df) __A,
+                (__v8si)
+                _mm256_setzero_si256 (),
+                (__mmask8) -1,
+                __R);
+}
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_cvt_roundps_epu32(__m512 __A, const int __R)
+{
+  return (__m512i) __builtin_ia32_cvtps2udq512_mask ((__v16sf) __A,
+                (__v16si)
+                _mm512_setzero_si512 (),
+                (__mmask16) -1,
+                __R);
+}
+static __inline __m256i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_cvt_roundpd_epu32(__m512d __A, const int __R)
+{
+  return (__m256i) __builtin_ia32_cvtpd2udq512_mask ((__v8df) __A,
+                (__v8si)
+                _mm256_setzero_si256 (),
+                (__mmask8) -1,
+                __R);
+}
+
+/* Bit Test */
+
+static __inline __mmask16 __attribute__ ((__always_inline__, __nodebug__))
+_mm512_test_epi32_mask(__m512i __A, __m512i __B)
+{
+  return (__mmask16) __builtin_ia32_ptestmd512 ((__v16si) __A,
+            (__v16si) __B,
+            (__mmask16) -1);
+}
+
+static __inline __mmask8 __attribute__ ((__always_inline__, __nodebug__))
+_mm512_test_epi64_mask(__m512i __A, __m512i __B)
+{
+  return (__mmask8) __builtin_ia32_ptestmq512 ((__v8di) __A,
+                 (__v8di) __B,
+                 (__mmask8) -1);
+}
+
+/* SIMD load ops */
+
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_maskz_loadu_epi32(__mmask16 __U, void const *__P)
+{
+  return (__m512i) __builtin_ia32_loaddqusi512_mask ((const __v16si *)__P,
+                                                     (__v16si)
+                                                     _mm512_setzero_si512 (),
+                                                     (__mmask16) __U);
+}
+
+static __inline __m512i __attribute__ ((__always_inline__, __nodebug__))
+_mm512_maskz_loadu_epi64(__mmask8 __U, void const *__P)
+{
+  return (__m512i) __builtin_ia32_loaddqudi512_mask ((const __v8di *)__P,
+                                                     (__v8di)
+                                                     _mm512_setzero_si512 (),
+                                                     (__mmask8) __U);
+}
+
+static __inline __m512 __attribute__ ((__always_inline__, __nodebug__))
+_mm512_maskz_loadu_ps(__mmask16 __U, void const *__P)
+{
+  return (__m512) __builtin_ia32_loadups512_mask ((const __v16sf *)__P,
+                                                  (__v16sf)
+                                                  _mm512_setzero_ps (),
+                                                  (__mmask16) __U);
+}
+
+static __inline __m512d __attribute__ ((__always_inline__, __nodebug__))
+_mm512_maskz_loadu_pd(__mmask8 __U, void const *__P)
+{
+  return (__m512d) __builtin_ia32_loadupd512_mask ((const __v8df *)__P,
+                                                   (__v8df)
+                                                   _mm512_setzero_pd (),
+                                                   (__mmask8) __U);
+}
+
+static __inline __m512d __attribute__((__always_inline__, __nodebug__))
+_mm512_loadu_pd(double const *__p)
+{
+  struct __loadu_pd {
+    __m512d __v;
+  } __attribute__((packed, may_alias));
+  return ((struct __loadu_pd*)__p)->__v;
+}
+
+static __inline __m512 __attribute__((__always_inline__, __nodebug__))
+_mm512_loadu_ps(float const *__p)
+{
+  struct __loadu_ps {
+    __m512 __v;
+  } __attribute__((packed, may_alias));
+  return ((struct __loadu_ps*)__p)->__v;
+}
+
+/* SIMD store ops */
+
+static __inline void __attribute__ ((__always_inline__, __nodebug__))
+_mm512_mask_storeu_epi64(void *__P, __mmask8 __U, __m512i __A)
+{
+  __builtin_ia32_storedqudi512_mask ((__v8di *)__P, (__v8di) __A,
+                                     (__mmask8) __U);
+}
+
+static __inline void __attribute__ ((__always_inline__, __nodebug__))
+_mm512_mask_storeu_epi32(void *__P, __mmask16 __U, __m512i __A)
+{
+  __builtin_ia32_storedqusi512_mask ((__v16si *)__P, (__v16si) __A,
+                                     (__mmask16) __U);
+}
+
+static __inline void __attribute__ ((__always_inline__, __nodebug__))
+_mm512_mask_storeu_pd(void *__P, __mmask8 __U, __m512d __A)
+{
+  __builtin_ia32_storeupd512_mask ((__v8df *)__P, (__v8df) __A, (__mmask8) __U);
+}
+
+static __inline void __attribute__ ((__always_inline__, __nodebug__))
+_mm512_storeu_pd(void *__P, __m512d __A)
+{
+  __builtin_ia32_storeupd512_mask((__v8df *)__P, (__v8df)__A, (__mmask8)-1);
+}
+
+static __inline void __attribute__ ((__always_inline__, __nodebug__))
+_mm512_mask_storeu_ps(void *__P, __mmask16 __U, __m512 __A)
+{
+  __builtin_ia32_storeups512_mask ((__v16sf *)__P, (__v16sf) __A,
+                                   (__mmask16) __U);
+}
+
+static __inline void __attribute__ ((__always_inline__, __nodebug__))
+_mm512_storeu_ps(void *__P, __m512 __A)
+{
+  __builtin_ia32_storeups512_mask((__v16sf *)__P, (__v16sf)__A, (__mmask16)-1);
+}
+
+static __inline void __attribute__ ((__always_inline__, __nodebug__))
+_mm512_store_ps(void *__P, __m512 __A)
+{
+  *(__m512*)__P = __A;
+}
+
+static __inline void __attribute__ ((__always_inline__, __nodebug__))
+_mm512_store_pd(void *__P, __m512d __A)
+{
+  *(__m512d*)__P = __A;
+}
+
+/* Mask ops */
+
+static __inline __mmask16 __attribute__ ((__always_inline__, __nodebug__))
+_mm512_knot(__mmask16 __M)
+{
+  return __builtin_ia32_knothi(__M);
+}
+
+/* Integer compare */
+
+static __inline__ __mmask16 __attribute__((__always_inline__, __nodebug__))
+_mm512_cmpeq_epi32_mask(__m512i __a, __m512i __b) {
+  return (__mmask16)__builtin_ia32_pcmpeqd512_mask((__v16si)__a, (__v16si)__b,
+                                                   (__mmask16)-1);
+}
+
+static __inline__ __mmask16 __attribute__((__always_inline__, __nodebug__))
+_mm512_mask_cmpeq_epi32_mask(__mmask16 __u, __m512i __a, __m512i __b) {
+  return (__mmask16)__builtin_ia32_pcmpeqd512_mask((__v16si)__a, (__v16si)__b,
+                                                   __u);
+}
+
+static __inline__ __mmask8 __attribute__((__always_inline__, __nodebug__))
+_mm512_mask_cmpeq_epi64_mask(__mmask8 __u, __m512i __a, __m512i __b) {
+  return (__mmask8)__builtin_ia32_pcmpeqq512_mask((__v8di)__a, (__v8di)__b,
+                                                  __u);
+}
+
+static __inline__ __mmask8 __attribute__((__always_inline__, __nodebug__))
+_mm512_cmpeq_epi64_mask(__m512i __a, __m512i __b) {
+  return (__mmask8)__builtin_ia32_pcmpeqq512_mask((__v8di)__a, (__v8di)__b,
+                                                  (__mmask8)-1);
+}
+
+#endif // __AVX512FINTRIN_H
diff --git a/lib/clang/3.6/include/avx512vlbwintrin.h b/lib/clang/3.6/include/avx512vlbwintrin.h
new file mode 100644
index 0000000..11333f8
--- /dev/null
+++ b/lib/clang/3.6/include/avx512vlbwintrin.h
@@ -0,0 +1,83 @@
+/*===---- avx512vlbwintrin.h - AVX512VL and AVX512BW intrinsics ----------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __IMMINTRIN_H
+#error "Never use <avx512vlbwintrin.h> directly; include <immintrin.h> instead."
+#endif
+
+#ifndef __AVX512VLBWINTRIN_H
+#define __AVX512VLBWINTRIN_H
+
+/* Integer compare */
+
+static __inline__ __mmask16 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_epi8_mask(__m128i __a, __m128i __b) {
+  return (__mmask16)__builtin_ia32_pcmpeqb128_mask((__v16qi)__a, (__v16qi)__b,
+                                                   (__mmask16)-1);
+}
+
+static __inline__ __mmask16 __attribute__((__always_inline__, __nodebug__))
+_mm_mask_cmpeq_epi8_mask(__mmask16 __u, __m128i __a, __m128i __b) {
+  return (__mmask16)__builtin_ia32_pcmpeqb128_mask((__v16qi)__a, (__v16qi)__b,
+                                                   __u);
+}
+
+
+static __inline__ __mmask32 __attribute__((__always_inline__, __nodebug__))
+_mm256_cmpeq_epi8_mask(__m256i __a, __m256i __b) {
+  return (__mmask32)__builtin_ia32_pcmpeqb256_mask((__v32qi)__a, (__v32qi)__b,
+                                                   (__mmask32)-1);
+}
+
+static __inline__ __mmask32 __attribute__((__always_inline__, __nodebug__))
+_mm256_mask_cmpeq_epi8_mask(__mmask32 __u, __m256i __a, __m256i __b) {
+  return (__mmask32)__builtin_ia32_pcmpeqb256_mask((__v32qi)__a, (__v32qi)__b,
+                                                   __u);
+}
+
+static __inline__ __mmask8 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_epi16_mask(__m128i __a, __m128i __b) {
+  return (__mmask8)__builtin_ia32_pcmpeqw128_mask((__v8hi)__a, (__v8hi)__b,
+                                                  (__mmask8)-1);
+}
+
+static __inline__ __mmask8 __attribute__((__always_inline__, __nodebug__))
+_mm_mask_cmpeq_epi16_mask(__mmask8 __u, __m128i __a, __m128i __b) {
+  return (__mmask8)__builtin_ia32_pcmpeqw128_mask((__v8hi)__a, (__v8hi)__b,
+                                                  __u);
+}
+
+
+static __inline__ __mmask16 __attribute__((__always_inline__, __nodebug__))
+_mm256_cmpeq_epi16_mask(__m256i __a, __m256i __b) {
+  return (__mmask16)__builtin_ia32_pcmpeqw256_mask((__v16hi)__a, (__v16hi)__b,
+                                                   (__mmask16)-1);
+}
+
+static __inline__ __mmask16 __attribute__((__always_inline__, __nodebug__))
+_mm256_mask_cmpeq_epi16_mask(__mmask16 __u, __m256i __a, __m256i __b) {
+  return (__mmask16)__builtin_ia32_pcmpeqw256_mask((__v16hi)__a, (__v16hi)__b,
+                                                   __u);
+}
+
+#endif /* __AVX512VLBWINTRIN_H */
diff --git a/lib/clang/3.6/include/avx512vlintrin.h b/lib/clang/3.6/include/avx512vlintrin.h
new file mode 100644
index 0000000..8a374b1
--- /dev/null
+++ b/lib/clang/3.6/include/avx512vlintrin.h
@@ -0,0 +1,83 @@
+/*===---- avx512vlintrin.h - AVX512VL intrinsics ---------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __IMMINTRIN_H
+#error "Never use <avx512vlintrin.h> directly; include <immintrin.h> instead."
+#endif
+
+#ifndef __AVX512VLINTRIN_H
+#define __AVX512VLINTRIN_H
+
+/* Integer compare */
+
+static __inline__ __mmask8 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_epi32_mask(__m128i __a, __m128i __b) {
+  return (__mmask8)__builtin_ia32_pcmpeqd128_mask((__v4si)__a, (__v4si)__b,
+                                                  (__mmask8)-1);
+}
+
+static __inline__ __mmask8 __attribute__((__always_inline__, __nodebug__))
+_mm_mask_cmpeq_epi32_mask(__mmask8 __u, __m128i __a, __m128i __b) {
+  return (__mmask8)__builtin_ia32_pcmpeqd128_mask((__v4si)__a, (__v4si)__b,
+                                                  __u);
+}
+
+
+static __inline__ __mmask8 __attribute__((__always_inline__, __nodebug__))
+_mm256_cmpeq_epi32_mask(__m256i __a, __m256i __b) {
+  return (__mmask8)__builtin_ia32_pcmpeqd256_mask((__v8si)__a, (__v8si)__b,
+                                                  (__mmask8)-1);
+}
+
+static __inline__ __mmask8 __attribute__((__always_inline__, __nodebug__))
+_mm256_mask_cmpeq_epi32_mask(__mmask8 __u, __m256i __a, __m256i __b) {
+  return (__mmask8)__builtin_ia32_pcmpeqd256_mask((__v8si)__a, (__v8si)__b,
+                                                  __u);
+}
+
+static __inline__ __mmask8 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_epi64_mask(__m128i __a, __m128i __b) {
+  return (__mmask8)__builtin_ia32_pcmpeqq128_mask((__v2di)__a, (__v2di)__b,
+                                                  (__mmask8)-1);
+}
+
+static __inline__ __mmask8 __attribute__((__always_inline__, __nodebug__))
+_mm_mask_cmpeq_epi64_mask(__mmask8 __u, __m128i __a, __m128i __b) {
+  return (__mmask8)__builtin_ia32_pcmpeqq128_mask((__v2di)__a, (__v2di)__b,
+                                                  __u);
+}
+
+
+static __inline__ __mmask8 __attribute__((__always_inline__, __nodebug__))
+_mm256_cmpeq_epi64_mask(__m256i __a, __m256i __b) {
+  return (__mmask8)__builtin_ia32_pcmpeqq256_mask((__v4di)__a, (__v4di)__b,
+                                                  (__mmask8)-1);
+}
+
+static __inline__ __mmask8 __attribute__((__always_inline__, __nodebug__))
+_mm256_mask_cmpeq_epi64_mask(__mmask8 __u, __m256i __a, __m256i __b) {
+  return (__mmask8)__builtin_ia32_pcmpeqq256_mask((__v4di)__a, (__v4di)__b,
+                                                  __u);
+}
+
+#endif /* __AVX512VLINTRIN_H */
diff --git a/lib/clang/3.6/include/avxintrin.h b/lib/clang/3.6/include/avxintrin.h
new file mode 100644
index 0000000..4e1044a
--- /dev/null
+++ b/lib/clang/3.6/include/avxintrin.h
@@ -0,0 +1,1239 @@
+/*===---- avxintrin.h - AVX intrinsics -------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __IMMINTRIN_H
+#error "Never use <avxintrin.h> directly; include <immintrin.h> instead."
+#endif
+
+#ifndef __AVXINTRIN_H
+#define __AVXINTRIN_H
+
+typedef double __v4df __attribute__ ((__vector_size__ (32)));
+typedef float __v8sf __attribute__ ((__vector_size__ (32)));
+typedef long long __v4di __attribute__ ((__vector_size__ (32)));
+typedef int __v8si __attribute__ ((__vector_size__ (32)));
+typedef short __v16hi __attribute__ ((__vector_size__ (32)));
+typedef char __v32qi __attribute__ ((__vector_size__ (32)));
+
+typedef float __m256 __attribute__ ((__vector_size__ (32)));
+typedef double __m256d __attribute__((__vector_size__(32)));
+typedef long long __m256i __attribute__((__vector_size__(32)));
+
+/* Arithmetic */
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_add_pd(__m256d __a, __m256d __b)
+{
+  return __a+__b;
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_add_ps(__m256 __a, __m256 __b)
+{
+  return __a+__b;
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_sub_pd(__m256d __a, __m256d __b)
+{
+  return __a-__b;
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_sub_ps(__m256 __a, __m256 __b)
+{
+  return __a-__b;
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_addsub_pd(__m256d __a, __m256d __b)
+{
+  return (__m256d)__builtin_ia32_addsubpd256((__v4df)__a, (__v4df)__b);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_addsub_ps(__m256 __a, __m256 __b)
+{
+  return (__m256)__builtin_ia32_addsubps256((__v8sf)__a, (__v8sf)__b);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_div_pd(__m256d __a, __m256d __b)
+{
+  return __a / __b;
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_div_ps(__m256 __a, __m256 __b)
+{
+  return __a / __b;
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_max_pd(__m256d __a, __m256d __b)
+{
+  return (__m256d)__builtin_ia32_maxpd256((__v4df)__a, (__v4df)__b);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_max_ps(__m256 __a, __m256 __b)
+{
+  return (__m256)__builtin_ia32_maxps256((__v8sf)__a, (__v8sf)__b);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_min_pd(__m256d __a, __m256d __b)
+{
+  return (__m256d)__builtin_ia32_minpd256((__v4df)__a, (__v4df)__b);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_min_ps(__m256 __a, __m256 __b)
+{
+  return (__m256)__builtin_ia32_minps256((__v8sf)__a, (__v8sf)__b);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_mul_pd(__m256d __a, __m256d __b)
+{
+  return __a * __b;
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_mul_ps(__m256 __a, __m256 __b)
+{
+  return __a * __b;
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_sqrt_pd(__m256d __a)
+{
+  return (__m256d)__builtin_ia32_sqrtpd256((__v4df)__a);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_sqrt_ps(__m256 __a)
+{
+  return (__m256)__builtin_ia32_sqrtps256((__v8sf)__a);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_rsqrt_ps(__m256 __a)
+{
+  return (__m256)__builtin_ia32_rsqrtps256((__v8sf)__a);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_rcp_ps(__m256 __a)
+{
+  return (__m256)__builtin_ia32_rcpps256((__v8sf)__a);
+}
+
+#define _mm256_round_pd(V, M) __extension__ ({ \
+    __m256d __V = (V); \
+    (__m256d)__builtin_ia32_roundpd256((__v4df)__V, (M)); })
+
+#define _mm256_round_ps(V, M) __extension__ ({ \
+  __m256 __V = (V); \
+  (__m256)__builtin_ia32_roundps256((__v8sf)__V, (M)); })
+
+#define _mm256_ceil_pd(V)  _mm256_round_pd((V), _MM_FROUND_CEIL)
+#define _mm256_floor_pd(V) _mm256_round_pd((V), _MM_FROUND_FLOOR)
+#define _mm256_ceil_ps(V)  _mm256_round_ps((V), _MM_FROUND_CEIL)
+#define _mm256_floor_ps(V) _mm256_round_ps((V), _MM_FROUND_FLOOR)
+
+/* Logical */
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_and_pd(__m256d __a, __m256d __b)
+{
+  return (__m256d)((__v4di)__a & (__v4di)__b);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_and_ps(__m256 __a, __m256 __b)
+{
+  return (__m256)((__v8si)__a & (__v8si)__b);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_andnot_pd(__m256d __a, __m256d __b)
+{
+  return (__m256d)(~(__v4di)__a & (__v4di)__b);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_andnot_ps(__m256 __a, __m256 __b)
+{
+  return (__m256)(~(__v8si)__a & (__v8si)__b);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_or_pd(__m256d __a, __m256d __b)
+{
+  return (__m256d)((__v4di)__a | (__v4di)__b);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_or_ps(__m256 __a, __m256 __b)
+{
+  return (__m256)((__v8si)__a | (__v8si)__b);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_xor_pd(__m256d __a, __m256d __b)
+{
+  return (__m256d)((__v4di)__a ^ (__v4di)__b);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_xor_ps(__m256 __a, __m256 __b)
+{
+  return (__m256)((__v8si)__a ^ (__v8si)__b);
+}
+
+/* Horizontal arithmetic */
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_hadd_pd(__m256d __a, __m256d __b)
+{
+  return (__m256d)__builtin_ia32_haddpd256((__v4df)__a, (__v4df)__b);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_hadd_ps(__m256 __a, __m256 __b)
+{
+  return (__m256)__builtin_ia32_haddps256((__v8sf)__a, (__v8sf)__b);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_hsub_pd(__m256d __a, __m256d __b)
+{
+  return (__m256d)__builtin_ia32_hsubpd256((__v4df)__a, (__v4df)__b);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_hsub_ps(__m256 __a, __m256 __b)
+{
+  return (__m256)__builtin_ia32_hsubps256((__v8sf)__a, (__v8sf)__b);
+}
+
+/* Vector permutations */
+static __inline __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_permutevar_pd(__m128d __a, __m128i __c)
+{
+  return (__m128d)__builtin_ia32_vpermilvarpd((__v2df)__a, (__v2di)__c);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_permutevar_pd(__m256d __a, __m256i __c)
+{
+  return (__m256d)__builtin_ia32_vpermilvarpd256((__v4df)__a, (__v4di)__c);
+}
+
+static __inline __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_permutevar_ps(__m128 __a, __m128i __c)
+{
+  return (__m128)__builtin_ia32_vpermilvarps((__v4sf)__a, (__v4si)__c);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_permutevar_ps(__m256 __a, __m256i __c)
+{
+  return (__m256)__builtin_ia32_vpermilvarps256((__v8sf)__a,
+						  (__v8si)__c);
+}
+
+#define _mm_permute_pd(A, C) __extension__ ({ \
+  __m128d __A = (A); \
+  (__m128d)__builtin_shufflevector((__v2df)__A, (__v2df) _mm_setzero_pd(), \
+                                   (C) & 0x1, ((C) & 0x2) >> 1); })
+
+#define _mm256_permute_pd(A, C) __extension__ ({ \
+  __m256d __A = (A); \
+  (__m256d)__builtin_shufflevector((__v4df)__A, (__v4df) _mm256_setzero_pd(), \
+                                   (C) & 0x1, ((C) & 0x2) >> 1, \
+                                   2 + (((C) & 0x4) >> 2), \
+                                   2 + (((C) & 0x8) >> 3)); })
+
+#define _mm_permute_ps(A, C) __extension__ ({ \
+  __m128 __A = (A); \
+  (__m128)__builtin_shufflevector((__v4sf)__A, (__v4sf) _mm_setzero_ps(), \
+                                   (C) & 0x3, ((C) & 0xc) >> 2, \
+                                   ((C) & 0x30) >> 4, ((C) & 0xc0) >> 6); })
+
+#define _mm256_permute_ps(A, C) __extension__ ({ \
+  __m256 __A = (A); \
+  (__m256)__builtin_shufflevector((__v8sf)__A, (__v8sf) _mm256_setzero_ps(), \
+                                  (C) & 0x3, ((C) & 0xc) >> 2, \
+                                  ((C) & 0x30) >> 4, ((C) & 0xc0) >> 6, \
+                                  4 + (((C) & 0x03) >> 0), \
+                                  4 + (((C) & 0x0c) >> 2), \
+                                  4 + (((C) & 0x30) >> 4), \
+                                  4 + (((C) & 0xc0) >> 6)); })
+
+#define _mm256_permute2f128_pd(V1, V2, M) __extension__ ({ \
+  __m256d __V1 = (V1); \
+  __m256d __V2 = (V2); \
+  (__m256d)__builtin_ia32_vperm2f128_pd256((__v4df)__V1, (__v4df)__V2, (M)); })
+
+#define _mm256_permute2f128_ps(V1, V2, M) __extension__ ({ \
+  __m256 __V1 = (V1); \
+  __m256 __V2 = (V2); \
+  (__m256)__builtin_ia32_vperm2f128_ps256((__v8sf)__V1, (__v8sf)__V2, (M)); })
+
+#define _mm256_permute2f128_si256(V1, V2, M) __extension__ ({ \
+  __m256i __V1 = (V1); \
+  __m256i __V2 = (V2); \
+  (__m256i)__builtin_ia32_vperm2f128_si256((__v8si)__V1, (__v8si)__V2, (M)); })
+
+/* Vector Blend */
+#define _mm256_blend_pd(V1, V2, M) __extension__ ({ \
+  __m256d __V1 = (V1); \
+  __m256d __V2 = (V2); \
+  (__m256d)__builtin_shufflevector((__v4df)__V1, (__v4df)__V2, \
+                                   (((M) & 0x01) ? 4 : 0), \
+                                   (((M) & 0x02) ? 5 : 1), \
+                                   (((M) & 0x04) ? 6 : 2), \
+                                   (((M) & 0x08) ? 7 : 3)); })
+
+#define _mm256_blend_ps(V1, V2, M) __extension__ ({ \
+  __m256 __V1 = (V1); \
+  __m256 __V2 = (V2); \
+  (__m256)__builtin_shufflevector((__v8sf)__V1, (__v8sf)__V2, \
+                                  (((M) & 0x01) ?  8 : 0), \
+                                  (((M) & 0x02) ?  9 : 1), \
+                                  (((M) & 0x04) ? 10 : 2), \
+                                  (((M) & 0x08) ? 11 : 3), \
+                                  (((M) & 0x10) ? 12 : 4), \
+                                  (((M) & 0x20) ? 13 : 5), \
+                                  (((M) & 0x40) ? 14 : 6), \
+                                  (((M) & 0x80) ? 15 : 7)); })
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_blendv_pd(__m256d __a, __m256d __b, __m256d __c)
+{
+  return (__m256d)__builtin_ia32_blendvpd256(
+    (__v4df)__a, (__v4df)__b, (__v4df)__c);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_blendv_ps(__m256 __a, __m256 __b, __m256 __c)
+{
+  return (__m256)__builtin_ia32_blendvps256(
+    (__v8sf)__a, (__v8sf)__b, (__v8sf)__c);
+}
+
+/* Vector Dot Product */
+#define _mm256_dp_ps(V1, V2, M) __extension__ ({ \
+  __m256 __V1 = (V1); \
+  __m256 __V2 = (V2); \
+  (__m256)__builtin_ia32_dpps256((__v8sf)__V1, (__v8sf)__V2, (M)); })
+
+/* Vector shuffle */
+#define _mm256_shuffle_ps(a, b, mask) __extension__ ({ \
+        __m256 __a = (a); \
+        __m256 __b = (b); \
+        (__m256)__builtin_shufflevector((__v8sf)__a, (__v8sf)__b, \
+        (mask) & 0x3,                ((mask) & 0xc) >> 2, \
+        (((mask) & 0x30) >> 4) + 8,  (((mask) & 0xc0) >> 6) + 8, \
+        ((mask) & 0x3) + 4,          (((mask) & 0xc) >> 2) + 4, \
+        (((mask) & 0x30) >> 4) + 12, (((mask) & 0xc0) >> 6) + 12); })
+
+#define _mm256_shuffle_pd(a, b, mask) __extension__ ({ \
+        __m256d __a = (a); \
+        __m256d __b = (b); \
+        (__m256d)__builtin_shufflevector((__v4df)__a, (__v4df)__b, \
+        (mask) & 0x1, \
+        (((mask) & 0x2) >> 1) + 4, \
+        (((mask) & 0x4) >> 2) + 2, \
+        (((mask) & 0x8) >> 3) + 6); })
+
+/* Compare */
+#define _CMP_EQ_OQ    0x00 /* Equal (ordered, non-signaling)  */
+#define _CMP_LT_OS    0x01 /* Less-than (ordered, signaling)  */
+#define _CMP_LE_OS    0x02 /* Less-than-or-equal (ordered, signaling)  */
+#define _CMP_UNORD_Q  0x03 /* Unordered (non-signaling)  */
+#define _CMP_NEQ_UQ   0x04 /* Not-equal (unordered, non-signaling)  */
+#define _CMP_NLT_US   0x05 /* Not-less-than (unordered, signaling)  */
+#define _CMP_NLE_US   0x06 /* Not-less-than-or-equal (unordered, signaling)  */
+#define _CMP_ORD_Q    0x07 /* Ordered (nonsignaling)   */
+#define _CMP_EQ_UQ    0x08 /* Equal (unordered, non-signaling)  */
+#define _CMP_NGE_US   0x09 /* Not-greater-than-or-equal (unord, signaling)  */
+#define _CMP_NGT_US   0x0a /* Not-greater-than (unordered, signaling)  */
+#define _CMP_FALSE_OQ 0x0b /* False (ordered, non-signaling)  */
+#define _CMP_NEQ_OQ   0x0c /* Not-equal (ordered, non-signaling)  */
+#define _CMP_GE_OS    0x0d /* Greater-than-or-equal (ordered, signaling)  */
+#define _CMP_GT_OS    0x0e /* Greater-than (ordered, signaling)  */
+#define _CMP_TRUE_UQ  0x0f /* True (unordered, non-signaling)  */
+#define _CMP_EQ_OS    0x10 /* Equal (ordered, signaling)  */
+#define _CMP_LT_OQ    0x11 /* Less-than (ordered, non-signaling)  */
+#define _CMP_LE_OQ    0x12 /* Less-than-or-equal (ordered, non-signaling)  */
+#define _CMP_UNORD_S  0x13 /* Unordered (signaling)  */
+#define _CMP_NEQ_US   0x14 /* Not-equal (unordered, signaling)  */
+#define _CMP_NLT_UQ   0x15 /* Not-less-than (unordered, non-signaling)  */
+#define _CMP_NLE_UQ   0x16 /* Not-less-than-or-equal (unord, non-signaling)  */
+#define _CMP_ORD_S    0x17 /* Ordered (signaling)  */
+#define _CMP_EQ_US    0x18 /* Equal (unordered, signaling)  */
+#define _CMP_NGE_UQ   0x19 /* Not-greater-than-or-equal (unord, non-sign)  */
+#define _CMP_NGT_UQ   0x1a /* Not-greater-than (unordered, non-signaling)  */
+#define _CMP_FALSE_OS 0x1b /* False (ordered, signaling)  */
+#define _CMP_NEQ_OS   0x1c /* Not-equal (ordered, signaling)  */
+#define _CMP_GE_OQ    0x1d /* Greater-than-or-equal (ordered, non-signaling)  */
+#define _CMP_GT_OQ    0x1e /* Greater-than (ordered, non-signaling)  */
+#define _CMP_TRUE_US  0x1f /* True (unordered, signaling)  */
+
+#define _mm_cmp_pd(a, b, c) __extension__ ({ \
+  __m128d __a = (a); \
+  __m128d __b = (b); \
+  (__m128d)__builtin_ia32_cmppd((__v2df)__a, (__v2df)__b, (c)); })
+
+#define _mm_cmp_ps(a, b, c) __extension__ ({ \
+  __m128 __a = (a); \
+  __m128 __b = (b); \
+  (__m128)__builtin_ia32_cmpps((__v4sf)__a, (__v4sf)__b, (c)); })
+
+#define _mm256_cmp_pd(a, b, c) __extension__ ({ \
+  __m256d __a = (a); \
+  __m256d __b = (b); \
+  (__m256d)__builtin_ia32_cmppd256((__v4df)__a, (__v4df)__b, (c)); })
+
+#define _mm256_cmp_ps(a, b, c) __extension__ ({ \
+  __m256 __a = (a); \
+  __m256 __b = (b); \
+  (__m256)__builtin_ia32_cmpps256((__v8sf)__a, (__v8sf)__b, (c)); })
+
+#define _mm_cmp_sd(a, b, c) __extension__ ({ \
+  __m128d __a = (a); \
+  __m128d __b = (b); \
+  (__m128d)__builtin_ia32_cmpsd((__v2df)__a, (__v2df)__b, (c)); })
+
+#define _mm_cmp_ss(a, b, c) __extension__ ({ \
+  __m128 __a = (a); \
+  __m128 __b = (b); \
+  (__m128)__builtin_ia32_cmpss((__v4sf)__a, (__v4sf)__b, (c)); })
+
+/* Vector extract */
+#define _mm256_extractf128_pd(A, O) __extension__ ({ \
+  __m256d __A = (A); \
+  (__m128d)__builtin_ia32_vextractf128_pd256((__v4df)__A, (O)); })
+
+#define _mm256_extractf128_ps(A, O) __extension__ ({ \
+  __m256 __A = (A); \
+  (__m128)__builtin_ia32_vextractf128_ps256((__v8sf)__A, (O)); })
+
+#define _mm256_extractf128_si256(A, O) __extension__ ({ \
+  __m256i __A = (A); \
+  (__m128i)__builtin_ia32_vextractf128_si256((__v8si)__A, (O)); })
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_extract_epi32(__m256i __a, int const __imm)
+{
+  __v8si __b = (__v8si)__a;
+  return __b[__imm & 7];
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_extract_epi16(__m256i __a, int const __imm)
+{
+  __v16hi __b = (__v16hi)__a;
+  return __b[__imm & 15];
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_extract_epi8(__m256i __a, int const __imm)
+{
+  __v32qi __b = (__v32qi)__a;
+  return __b[__imm & 31];
+}
+
+#ifdef __x86_64__
+static __inline long long  __attribute__((__always_inline__, __nodebug__))
+_mm256_extract_epi64(__m256i __a, const int __imm)
+{
+  __v4di __b = (__v4di)__a;
+  return __b[__imm & 3];
+}
+#endif
+
+/* Vector insert */
+#define _mm256_insertf128_pd(V1, V2, O) __extension__ ({ \
+  __m256d __V1 = (V1); \
+  __m128d __V2 = (V2); \
+  (__m256d)__builtin_ia32_vinsertf128_pd256((__v4df)__V1, (__v2df)__V2, (O)); })
+
+#define _mm256_insertf128_ps(V1, V2, O) __extension__ ({ \
+  __m256 __V1 = (V1); \
+  __m128 __V2 = (V2); \
+  (__m256)__builtin_ia32_vinsertf128_ps256((__v8sf)__V1, (__v4sf)__V2, (O)); })
+
+#define _mm256_insertf128_si256(V1, V2, O) __extension__ ({ \
+  __m256i __V1 = (V1); \
+  __m128i __V2 = (V2); \
+  (__m256i)__builtin_ia32_vinsertf128_si256((__v8si)__V1, (__v4si)__V2, (O)); })
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_insert_epi32(__m256i __a, int __b, int const __imm)
+{
+  __v8si __c = (__v8si)__a;
+  __c[__imm & 7] = __b;
+  return (__m256i)__c;
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_insert_epi16(__m256i __a, int __b, int const __imm)
+{
+  __v16hi __c = (__v16hi)__a;
+  __c[__imm & 15] = __b;
+  return (__m256i)__c;
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_insert_epi8(__m256i __a, int __b, int const __imm)
+{
+  __v32qi __c = (__v32qi)__a;
+  __c[__imm & 31] = __b;
+  return (__m256i)__c;
+}
+
+#ifdef __x86_64__
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_insert_epi64(__m256i __a, int __b, int const __imm)
+{
+  __v4di __c = (__v4di)__a;
+  __c[__imm & 3] = __b;
+  return (__m256i)__c;
+}
+#endif
+
+/* Conversion */
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepi32_pd(__m128i __a)
+{
+  return (__m256d)__builtin_ia32_cvtdq2pd256((__v4si) __a);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtepi32_ps(__m256i __a)
+{
+  return (__m256)__builtin_ia32_cvtdq2ps256((__v8si) __a);
+}
+
+static __inline __m128 __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtpd_ps(__m256d __a)
+{
+  return (__m128)__builtin_ia32_cvtpd2ps256((__v4df) __a);
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtps_epi32(__m256 __a)
+{
+  return (__m256i)__builtin_ia32_cvtps2dq256((__v8sf) __a);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtps_pd(__m128 __a)
+{
+  return (__m256d)__builtin_ia32_cvtps2pd256((__v4sf) __a);
+}
+
+static __inline __m128i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvttpd_epi32(__m256d __a)
+{
+  return (__m128i)__builtin_ia32_cvttpd2dq256((__v4df) __a);
+}
+
+static __inline __m128i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtpd_epi32(__m256d __a)
+{
+  return (__m128i)__builtin_ia32_cvtpd2dq256((__v4df) __a);
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cvttps_epi32(__m256 __a)
+{
+  return (__m256i)__builtin_ia32_cvttps2dq256((__v8sf) __a);
+}
+
+/* Vector replicate */
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_movehdup_ps(__m256 __a)
+{
+  return __builtin_shufflevector(__a, __a, 1, 1, 3, 3, 5, 5, 7, 7);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_moveldup_ps(__m256 __a)
+{
+  return __builtin_shufflevector(__a, __a, 0, 0, 2, 2, 4, 4, 6, 6);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_movedup_pd(__m256d __a)
+{
+  return __builtin_shufflevector(__a, __a, 0, 0, 2, 2);
+}
+
+/* Unpack and Interleave */
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_unpackhi_pd(__m256d __a, __m256d __b)
+{
+  return __builtin_shufflevector(__a, __b, 1, 5, 1+2, 5+2);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_unpacklo_pd(__m256d __a, __m256d __b)
+{
+  return __builtin_shufflevector(__a, __b, 0, 4, 0+2, 4+2);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_unpackhi_ps(__m256 __a, __m256 __b)
+{
+  return __builtin_shufflevector(__a, __b, 2, 10, 2+1, 10+1, 6, 14, 6+1, 14+1);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_unpacklo_ps(__m256 __a, __m256 __b)
+{
+  return __builtin_shufflevector(__a, __b, 0, 8, 0+1, 8+1, 4, 12, 4+1, 12+1);
+}
+
+/* Bit Test */
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm_testz_pd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_vtestzpd((__v2df)__a, (__v2df)__b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm_testc_pd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_vtestcpd((__v2df)__a, (__v2df)__b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm_testnzc_pd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_vtestnzcpd((__v2df)__a, (__v2df)__b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm_testz_ps(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_vtestzps((__v4sf)__a, (__v4sf)__b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm_testc_ps(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_vtestcps((__v4sf)__a, (__v4sf)__b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm_testnzc_ps(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_vtestnzcps((__v4sf)__a, (__v4sf)__b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_testz_pd(__m256d __a, __m256d __b)
+{
+  return __builtin_ia32_vtestzpd256((__v4df)__a, (__v4df)__b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_testc_pd(__m256d __a, __m256d __b)
+{
+  return __builtin_ia32_vtestcpd256((__v4df)__a, (__v4df)__b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_testnzc_pd(__m256d __a, __m256d __b)
+{
+  return __builtin_ia32_vtestnzcpd256((__v4df)__a, (__v4df)__b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_testz_ps(__m256 __a, __m256 __b)
+{
+  return __builtin_ia32_vtestzps256((__v8sf)__a, (__v8sf)__b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_testc_ps(__m256 __a, __m256 __b)
+{
+  return __builtin_ia32_vtestcps256((__v8sf)__a, (__v8sf)__b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_testnzc_ps(__m256 __a, __m256 __b)
+{
+  return __builtin_ia32_vtestnzcps256((__v8sf)__a, (__v8sf)__b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_testz_si256(__m256i __a, __m256i __b)
+{
+  return __builtin_ia32_ptestz256((__v4di)__a, (__v4di)__b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_testc_si256(__m256i __a, __m256i __b)
+{
+  return __builtin_ia32_ptestc256((__v4di)__a, (__v4di)__b);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_testnzc_si256(__m256i __a, __m256i __b)
+{
+  return __builtin_ia32_ptestnzc256((__v4di)__a, (__v4di)__b);
+}
+
+/* Vector extract sign mask */
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_movemask_pd(__m256d __a)
+{
+  return __builtin_ia32_movmskpd256((__v4df)__a);
+}
+
+static __inline int __attribute__((__always_inline__, __nodebug__))
+_mm256_movemask_ps(__m256 __a)
+{
+  return __builtin_ia32_movmskps256((__v8sf)__a);
+}
+
+/* Vector __zero */
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_zeroall(void)
+{
+  __builtin_ia32_vzeroall();
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_zeroupper(void)
+{
+  __builtin_ia32_vzeroupper();
+}
+
+/* Vector load with broadcast */
+static __inline __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_broadcast_ss(float const *__a)
+{
+  float __f = *__a;
+  return (__m128)(__v4sf){ __f, __f, __f, __f };
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcast_sd(double const *__a)
+{
+  double __d = *__a;
+  return (__m256d)(__v4df){ __d, __d, __d, __d };
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcast_ss(float const *__a)
+{
+  float __f = *__a;
+  return (__m256)(__v8sf){ __f, __f, __f, __f, __f, __f, __f, __f };
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcast_pd(__m128d const *__a)
+{
+  return (__m256d)__builtin_ia32_vbroadcastf128_pd256(__a);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_broadcast_ps(__m128 const *__a)
+{
+  return (__m256)__builtin_ia32_vbroadcastf128_ps256(__a);
+}
+
+/* SIMD load ops */
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_load_pd(double const *__p)
+{
+  return *(__m256d *)__p;
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_load_ps(float const *__p)
+{
+  return *(__m256 *)__p;
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_loadu_pd(double const *__p)
+{
+  struct __loadu_pd {
+    __m256d __v;
+  } __attribute__((packed, may_alias));
+  return ((struct __loadu_pd*)__p)->__v;
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_loadu_ps(float const *__p)
+{
+  struct __loadu_ps {
+    __m256 __v;
+  } __attribute__((packed, may_alias));
+  return ((struct __loadu_ps*)__p)->__v;
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_load_si256(__m256i const *__p)
+{
+  return *__p;
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_loadu_si256(__m256i const *__p)
+{
+  struct __loadu_si256 {
+    __m256i __v;
+  } __attribute__((packed, may_alias));
+  return ((struct __loadu_si256*)__p)->__v;
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_lddqu_si256(__m256i const *__p)
+{
+  return (__m256i)__builtin_ia32_lddqu256((char const *)__p);
+}
+
+/* SIMD store ops */
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_store_pd(double *__p, __m256d __a)
+{
+  *(__m256d *)__p = __a;
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_store_ps(float *__p, __m256 __a)
+{
+  *(__m256 *)__p = __a;
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_storeu_pd(double *__p, __m256d __a)
+{
+  __builtin_ia32_storeupd256(__p, (__v4df)__a);
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_storeu_ps(float *__p, __m256 __a)
+{
+  __builtin_ia32_storeups256(__p, (__v8sf)__a);
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_store_si256(__m256i *__p, __m256i __a)
+{
+  *__p = __a;
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_storeu_si256(__m256i *__p, __m256i __a)
+{
+  __builtin_ia32_storedqu256((char *)__p, (__v32qi)__a);
+}
+
+/* Conditional load ops */
+static __inline __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_maskload_pd(double const *__p, __m128d __m)
+{
+  return (__m128d)__builtin_ia32_maskloadpd((const __v2df *)__p, (__v2df)__m);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_maskload_pd(double const *__p, __m256d __m)
+{
+  return (__m256d)__builtin_ia32_maskloadpd256((const __v4df *)__p,
+                                               (__v4df)__m);
+}
+
+static __inline __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_maskload_ps(float const *__p, __m128 __m)
+{
+  return (__m128)__builtin_ia32_maskloadps((const __v4sf *)__p, (__v4sf)__m);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_maskload_ps(float const *__p, __m256 __m)
+{
+  return (__m256)__builtin_ia32_maskloadps256((const __v8sf *)__p, (__v8sf)__m);
+}
+
+/* Conditional store ops */
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_maskstore_ps(float *__p, __m256 __m, __m256 __a)
+{
+  __builtin_ia32_maskstoreps256((__v8sf *)__p, (__v8sf)__m, (__v8sf)__a);
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm_maskstore_pd(double *__p, __m128d __m, __m128d __a)
+{
+  __builtin_ia32_maskstorepd((__v2df *)__p, (__v2df)__m, (__v2df)__a);
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_maskstore_pd(double *__p, __m256d __m, __m256d __a)
+{
+  __builtin_ia32_maskstorepd256((__v4df *)__p, (__v4df)__m, (__v4df)__a);
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm_maskstore_ps(float *__p, __m128 __m, __m128 __a)
+{
+  __builtin_ia32_maskstoreps((__v4sf *)__p, (__v4sf)__m, (__v4sf)__a);
+}
+
+/* Cacheability support ops */
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_stream_si256(__m256i *__a, __m256i __b)
+{
+  __builtin_ia32_movntdq256((__v4di *)__a, (__v4di)__b);
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_stream_pd(double *__a, __m256d __b)
+{
+  __builtin_ia32_movntpd256(__a, (__v4df)__b);
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_stream_ps(float *__p, __m256 __a)
+{
+  __builtin_ia32_movntps256(__p, (__v8sf)__a);
+}
+
+/* Create vectors */
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_set_pd(double __a, double __b, double __c, double __d)
+{
+  return (__m256d){ __d, __c, __b, __a };
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_set_ps(float __a, float __b, float __c, float __d,
+	            float __e, float __f, float __g, float __h)
+{
+  return (__m256){ __h, __g, __f, __e, __d, __c, __b, __a };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_set_epi32(int __i0, int __i1, int __i2, int __i3,
+		             int __i4, int __i5, int __i6, int __i7)
+{
+  return (__m256i)(__v8si){ __i7, __i6, __i5, __i4, __i3, __i2, __i1, __i0 };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_set_epi16(short __w15, short __w14, short __w13, short __w12,
+		             short __w11, short __w10, short __w09, short __w08,
+		             short __w07, short __w06, short __w05, short __w04,
+		             short __w03, short __w02, short __w01, short __w00)
+{
+  return (__m256i)(__v16hi){ __w00, __w01, __w02, __w03, __w04, __w05, __w06,
+    __w07, __w08, __w09, __w10, __w11, __w12, __w13, __w14, __w15 };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_set_epi8(char __b31, char __b30, char __b29, char __b28,
+		            char __b27, char __b26, char __b25, char __b24,
+		            char __b23, char __b22, char __b21, char __b20,
+		            char __b19, char __b18, char __b17, char __b16,
+		            char __b15, char __b14, char __b13, char __b12,
+		            char __b11, char __b10, char __b09, char __b08,
+		            char __b07, char __b06, char __b05, char __b04,
+		            char __b03, char __b02, char __b01, char __b00)
+{
+  return (__m256i)(__v32qi){
+    __b00, __b01, __b02, __b03, __b04, __b05, __b06, __b07,
+    __b08, __b09, __b10, __b11, __b12, __b13, __b14, __b15,
+    __b16, __b17, __b18, __b19, __b20, __b21, __b22, __b23,
+    __b24, __b25, __b26, __b27, __b28, __b29, __b30, __b31
+  };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_set_epi64x(long long __a, long long __b, long long __c, long long __d)
+{
+  return (__m256i)(__v4di){ __d, __c, __b, __a };
+}
+
+/* Create vectors with elements in reverse order */
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_setr_pd(double __a, double __b, double __c, double __d)
+{
+  return (__m256d){ __a, __b, __c, __d };
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_setr_ps(float __a, float __b, float __c, float __d,
+		           float __e, float __f, float __g, float __h)
+{
+  return (__m256){ __a, __b, __c, __d, __e, __f, __g, __h };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_setr_epi32(int __i0, int __i1, int __i2, int __i3,
+		              int __i4, int __i5, int __i6, int __i7)
+{
+  return (__m256i)(__v8si){ __i0, __i1, __i2, __i3, __i4, __i5, __i6, __i7 };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_setr_epi16(short __w15, short __w14, short __w13, short __w12,
+		   short __w11, short __w10, short __w09, short __w08,
+		   short __w07, short __w06, short __w05, short __w04,
+		   short __w03, short __w02, short __w01, short __w00)
+{
+  return (__m256i)(__v16hi){ __w15, __w14, __w13, __w12, __w11, __w10, __w09,
+    __w08, __w07, __w06, __w05, __w04, __w03, __w02, __w01, __w00 };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_setr_epi8(char __b31, char __b30, char __b29, char __b28,
+		             char __b27, char __b26, char __b25, char __b24,
+		             char __b23, char __b22, char __b21, char __b20,
+		             char __b19, char __b18, char __b17, char __b16,
+		             char __b15, char __b14, char __b13, char __b12,
+		             char __b11, char __b10, char __b09, char __b08,
+		             char __b07, char __b06, char __b05, char __b04,
+		             char __b03, char __b02, char __b01, char __b00)
+{
+  return (__m256i)(__v32qi){
+    __b31, __b30, __b29, __b28, __b27, __b26, __b25, __b24,
+		__b23, __b22, __b21, __b20, __b19, __b18, __b17, __b16,
+		__b15, __b14, __b13, __b12, __b11, __b10, __b09, __b08,
+		__b07, __b06, __b05, __b04, __b03, __b02, __b01, __b00 };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_setr_epi64x(long long __a, long long __b, long long __c, long long __d)
+{
+  return (__m256i)(__v4di){ __a, __b, __c, __d };
+}
+
+/* Create vectors with repeated elements */
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_set1_pd(double __w)
+{
+  return (__m256d){ __w, __w, __w, __w };
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_set1_ps(float __w)
+{
+  return (__m256){ __w, __w, __w, __w, __w, __w, __w, __w };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_set1_epi32(int __i)
+{
+  return (__m256i)(__v8si){ __i, __i, __i, __i, __i, __i, __i, __i };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_set1_epi16(short __w)
+{
+  return (__m256i)(__v16hi){ __w, __w, __w, __w, __w, __w, __w, __w, __w, __w,
+    __w, __w, __w, __w, __w, __w };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_set1_epi8(char __b)
+{
+  return (__m256i)(__v32qi){ __b, __b, __b, __b, __b, __b, __b, __b, __b, __b,
+    __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b,
+    __b, __b, __b, __b, __b, __b, __b };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_set1_epi64x(long long __q)
+{
+  return (__m256i)(__v4di){ __q, __q, __q, __q };
+}
+
+/* Create __zeroed vectors */
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_setzero_pd(void)
+{
+  return (__m256d){ 0, 0, 0, 0 };
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_setzero_ps(void)
+{
+  return (__m256){ 0, 0, 0, 0, 0, 0, 0, 0 };
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_setzero_si256(void)
+{
+  return (__m256i){ 0LL, 0LL, 0LL, 0LL };
+}
+
+/* Cast between vector types */
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_castpd_ps(__m256d __a)
+{
+  return (__m256)__a;
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_castpd_si256(__m256d __a)
+{
+  return (__m256i)__a;
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_castps_pd(__m256 __a)
+{
+  return (__m256d)__a;
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_castps_si256(__m256 __a)
+{
+  return (__m256i)__a;
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_castsi256_ps(__m256i __a)
+{
+  return (__m256)__a;
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_castsi256_pd(__m256i __a)
+{
+  return (__m256d)__a;
+}
+
+static __inline __m128d __attribute__((__always_inline__, __nodebug__))
+_mm256_castpd256_pd128(__m256d __a)
+{
+  return __builtin_shufflevector(__a, __a, 0, 1);
+}
+
+static __inline __m128 __attribute__((__always_inline__, __nodebug__))
+_mm256_castps256_ps128(__m256 __a)
+{
+  return __builtin_shufflevector(__a, __a, 0, 1, 2, 3);
+}
+
+static __inline __m128i __attribute__((__always_inline__, __nodebug__))
+_mm256_castsi256_si128(__m256i __a)
+{
+  return __builtin_shufflevector(__a, __a, 0, 1);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_castpd128_pd256(__m128d __a)
+{
+  return __builtin_shufflevector(__a, __a, 0, 1, -1, -1);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_castps128_ps256(__m128 __a)
+{
+  return __builtin_shufflevector(__a, __a, 0, 1, 2, 3, -1, -1, -1, -1);
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_castsi128_si256(__m128i __a)
+{
+  return __builtin_shufflevector(__a, __a, 0, 1, -1, -1);
+}
+
+/* SIMD load ops (unaligned) */
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_loadu2_m128(float const *__addr_hi, float const *__addr_lo)
+{
+  struct __loadu_ps {
+    __m128 __v;
+  } __attribute__((__packed__, __may_alias__));
+
+  __m256 __v256 = _mm256_castps128_ps256(((struct __loadu_ps*)__addr_lo)->__v);
+  return _mm256_insertf128_ps(__v256, ((struct __loadu_ps*)__addr_hi)->__v, 1);
+}
+
+static __inline __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_loadu2_m128d(double const *__addr_hi, double const *__addr_lo)
+{
+  struct __loadu_pd {
+    __m128d __v;
+  } __attribute__((__packed__, __may_alias__));
+  
+  __m256d __v256 = _mm256_castpd128_pd256(((struct __loadu_pd*)__addr_lo)->__v);
+  return _mm256_insertf128_pd(__v256, ((struct __loadu_pd*)__addr_hi)->__v, 1);
+}
+
+static __inline __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_loadu2_m128i(__m128i const *__addr_hi, __m128i const *__addr_lo)
+{
+  struct __loadu_si128 {
+    __m128i __v;
+  } __attribute__((packed, may_alias));
+  __m256i __v256 = _mm256_castsi128_si256(
+    ((struct __loadu_si128*)__addr_lo)->__v);
+  return _mm256_insertf128_si256(__v256,
+                                 ((struct __loadu_si128*)__addr_hi)->__v, 1);
+}
+
+/* SIMD store ops (unaligned) */
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_storeu2_m128(float *__addr_hi, float *__addr_lo, __m256 __a)
+{
+  __m128 __v128;
+
+  __v128 = _mm256_castps256_ps128(__a);
+  __builtin_ia32_storeups(__addr_lo, __v128);
+  __v128 = _mm256_extractf128_ps(__a, 1);
+  __builtin_ia32_storeups(__addr_hi, __v128);
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_storeu2_m128d(double *__addr_hi, double *__addr_lo, __m256d __a)
+{
+  __m128d __v128;
+
+  __v128 = _mm256_castpd256_pd128(__a);
+  __builtin_ia32_storeupd(__addr_lo, __v128);
+  __v128 = _mm256_extractf128_pd(__a, 1);
+  __builtin_ia32_storeupd(__addr_hi, __v128);
+}
+
+static __inline void __attribute__((__always_inline__, __nodebug__))
+_mm256_storeu2_m128i(__m128i *__addr_hi, __m128i *__addr_lo, __m256i __a)
+{
+  __m128i __v128;
+
+  __v128 = _mm256_castsi256_si128(__a);
+  __builtin_ia32_storedqu((char *)__addr_lo, (__v16qi)__v128);
+  __v128 = _mm256_extractf128_si256(__a, 1);
+  __builtin_ia32_storedqu((char *)__addr_hi, (__v16qi)__v128);
+}
+
+#endif /* __AVXINTRIN_H */
diff --git a/lib/clang/3.6/include/bmi2intrin.h b/lib/clang/3.6/include/bmi2intrin.h
new file mode 100644
index 0000000..a05cfad
--- /dev/null
+++ b/lib/clang/3.6/include/bmi2intrin.h
@@ -0,0 +1,94 @@
+/*===---- bmi2intrin.h - BMI2 intrinsics -----------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#if !defined __X86INTRIN_H && !defined __IMMINTRIN_H
+#error "Never use <bmi2intrin.h> directly; include <x86intrin.h> instead."
+#endif
+
+#ifndef __BMI2__
+# error "BMI2 instruction set not enabled"
+#endif /* __BMI2__ */
+
+#ifndef __BMI2INTRIN_H
+#define __BMI2INTRIN_H
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_bzhi_u32(unsigned int __X, unsigned int __Y)
+{
+  return __builtin_ia32_bzhi_si(__X, __Y);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_pdep_u32(unsigned int __X, unsigned int __Y)
+{
+  return __builtin_ia32_pdep_si(__X, __Y);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_pext_u32(unsigned int __X, unsigned int __Y)
+{
+  return __builtin_ia32_pext_si(__X, __Y);
+}
+
+#ifdef  __x86_64__
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+_bzhi_u64(unsigned long long __X, unsigned long long __Y)
+{
+  return __builtin_ia32_bzhi_di(__X, __Y);
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+_pdep_u64(unsigned long long __X, unsigned long long __Y)
+{
+  return __builtin_ia32_pdep_di(__X, __Y);
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+_pext_u64(unsigned long long __X, unsigned long long __Y)
+{
+  return __builtin_ia32_pext_di(__X, __Y);
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+_mulx_u64 (unsigned long long __X, unsigned long long __Y,
+	   unsigned long long *__P)
+{
+  unsigned __int128 __res = (unsigned __int128) __X * __Y;
+  *__P = (unsigned long long) (__res >> 64);
+  return (unsigned long long) __res;
+}
+
+#else /* !__x86_64__ */
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_mulx_u32 (unsigned int __X, unsigned int __Y, unsigned int *__P)
+{
+  unsigned long long __res = (unsigned long long) __X * __Y;
+  *__P = (unsigned int) (__res >> 32);
+  return (unsigned int) __res;
+}
+
+#endif /* !__x86_64__  */
+
+#endif /* __BMI2INTRIN_H */
diff --git a/lib/clang/3.6/include/bmiintrin.h b/lib/clang/3.6/include/bmiintrin.h
new file mode 100644
index 0000000..0e5fd55
--- /dev/null
+++ b/lib/clang/3.6/include/bmiintrin.h
@@ -0,0 +1,148 @@
+/*===---- bmiintrin.h - BMI intrinsics -------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#if !defined __X86INTRIN_H && !defined __IMMINTRIN_H
+#error "Never use <bmiintrin.h> directly; include <x86intrin.h> instead."
+#endif
+
+#ifndef __BMI__
+# error "BMI instruction set not enabled"
+#endif /* __BMI__ */
+
+#ifndef __BMIINTRIN_H
+#define __BMIINTRIN_H
+
+#define _tzcnt_u16(a)     (__tzcnt_u16((a)))
+#define _andn_u32(a, b)   (__andn_u32((a), (b)))
+/* _bextr_u32 != __bextr_u32 */
+#define _blsi_u32(a)      (__blsi_u32((a)))
+#define _blsmsk_u32(a)    (__blsmsk_u32((a)))
+#define _blsr_u32(a)      (__blsr_u32((a)))
+#define _tzcnt_u32(a)     (__tzcnt_u32((a)))
+
+static __inline__ unsigned short __attribute__((__always_inline__, __nodebug__))
+__tzcnt_u16(unsigned short __X)
+{
+  return __X ? __builtin_ctzs(__X) : 16;
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__andn_u32(unsigned int __X, unsigned int __Y)
+{
+  return ~__X & __Y;
+}
+
+/* AMD-specified, double-leading-underscore version of BEXTR */
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__bextr_u32(unsigned int __X, unsigned int __Y)
+{
+  return __builtin_ia32_bextr_u32(__X, __Y);
+}
+
+/* Intel-specified, single-leading-underscore version of BEXTR */
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_bextr_u32(unsigned int __X, unsigned int __Y, unsigned int __Z)
+{
+  return __builtin_ia32_bextr_u32 (__X, ((__Y & 0xff) | ((__Z & 0xff) << 8)));
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__blsi_u32(unsigned int __X)
+{
+  return __X & -__X;
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__blsmsk_u32(unsigned int __X)
+{
+  return __X ^ (__X - 1);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__blsr_u32(unsigned int __X)
+{
+  return __X & (__X - 1);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__tzcnt_u32(unsigned int __X)
+{
+  return __X ? __builtin_ctz(__X) : 32;
+}
+
+#ifdef __x86_64__
+
+#define _andn_u64(a, b)   (__andn_u64((a), (b)))
+/* _bextr_u64 != __bextr_u64 */
+#define _blsi_u64(a)      (__blsi_u64((a)))
+#define _blsmsk_u64(a)    (__blsmsk_u64((a)))
+#define _blsr_u64(a)      (__blsr_u64((a)))
+#define _tzcnt_u64(a)     (__tzcnt_u64((a)))
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+__andn_u64 (unsigned long long __X, unsigned long long __Y)
+{
+  return ~__X & __Y;
+}
+
+/* AMD-specified, double-leading-underscore version of BEXTR */
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+__bextr_u64(unsigned long long __X, unsigned long long __Y)
+{
+  return __builtin_ia32_bextr_u64(__X, __Y);
+}
+
+/* Intel-specified, single-leading-underscore version of BEXTR */
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+_bextr_u64(unsigned long long __X, unsigned int __Y, unsigned int __Z)
+{
+  return __builtin_ia32_bextr_u64 (__X, ((__Y & 0xff) | ((__Z & 0xff) << 8)));
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+__blsi_u64(unsigned long long __X)
+{
+  return __X & -__X;
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+__blsmsk_u64(unsigned long long __X)
+{
+  return __X ^ (__X - 1);
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+__blsr_u64(unsigned long long __X)
+{
+  return __X & (__X - 1);
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+__tzcnt_u64(unsigned long long __X)
+{
+  return __X ? __builtin_ctzll(__X) : 64;
+}
+
+#endif /* __x86_64__ */
+
+#endif /* __BMIINTRIN_H */
diff --git a/lib/clang/3.6/include/cpuid.h b/lib/clang/3.6/include/cpuid.h
new file mode 100644
index 0000000..5da02e0
--- /dev/null
+++ b/lib/clang/3.6/include/cpuid.h
@@ -0,0 +1,209 @@
+/*===---- cpuid.h - X86 cpu model detection --------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#if !(__x86_64__ || __i386__)
+#error this header is for x86 only
+#endif
+
+/* Responses identification request with %eax 0 */
+/* AMD:     "AuthenticAMD" */
+#define signature_AMD_ebx 0x68747541
+#define signature_AMD_edx 0x69746e65
+#define signature_AMD_ecx 0x444d4163
+/* CENTAUR: "CentaurHauls" */
+#define signature_CENTAUR_ebx 0x746e6543
+#define signature_CENTAUR_edx 0x48727561
+#define signature_CENTAUR_ecx 0x736c7561
+/* CYRIX:   "CyrixInstead" */
+#define signature_CYRIX_ebx 0x69727943
+#define signature_CYRIX_edx 0x736e4978
+#define signature_CYRIX_ecx 0x64616574
+/* INTEL:   "GenuineIntel" */
+#define signature_INTEL_ebx 0x756e6547
+#define signature_INTEL_edx 0x49656e69
+#define signature_INTEL_ecx 0x6c65746e
+/* TM1:     "TransmetaCPU" */
+#define signature_TM1_ebx 0x6e617254
+#define signature_TM1_edx 0x74656d73
+#define signature_TM1_ecx 0x55504361
+/* TM2:     "GenuineTMx86" */
+#define signature_TM2_ebx 0x756e6547
+#define signature_TM2_edx 0x54656e69
+#define signature_TM2_ecx 0x3638784d
+/* NSC:     "Geode by NSC" */
+#define signature_NSC_ebx 0x646f6547
+#define signature_NSC_edx 0x43534e20
+#define signature_NSC_ecx 0x79622065
+/* NEXGEN:  "NexGenDriven" */
+#define signature_NEXGEN_ebx 0x4778654e
+#define signature_NEXGEN_edx 0x72446e65
+#define signature_NEXGEN_ecx 0x6e657669
+/* RISE:    "RiseRiseRise" */
+#define signature_RISE_ebx 0x65736952
+#define signature_RISE_edx 0x65736952
+#define signature_RISE_ecx 0x65736952
+/* SIS:     "SiS SiS SiS " */
+#define signature_SIS_ebx 0x20536953
+#define signature_SIS_edx 0x20536953
+#define signature_SIS_ecx 0x20536953
+/* UMC:     "UMC UMC UMC " */
+#define signature_UMC_ebx 0x20434d55
+#define signature_UMC_edx 0x20434d55
+#define signature_UMC_ecx 0x20434d55
+/* VIA:     "VIA VIA VIA " */
+#define signature_VIA_ebx 0x20414956
+#define signature_VIA_edx 0x20414956
+#define signature_VIA_ecx 0x20414956
+/* VORTEX:  "Vortex86 SoC" */
+#define signature_VORTEX_ebx 0x74726f56
+#define signature_VORTEX_edx 0x36387865
+#define signature_VORTEX_ecx 0x436f5320
+
+/* Features in %ecx for level 1 */
+#define bit_SSE3        0x00000001
+#define bit_PCLMULQDQ   0x00000002
+#define bit_DTES64      0x00000004
+#define bit_MONITOR     0x00000008
+#define bit_DSCPL       0x00000010
+#define bit_VMX         0x00000020
+#define bit_SMX         0x00000040
+#define bit_EIST        0x00000080
+#define bit_TM2         0x00000100
+#define bit_SSSE3       0x00000200
+#define bit_CNXTID      0x00000400
+#define bit_FMA         0x00001000
+#define bit_CMPXCHG16B  0x00002000
+#define bit_xTPR        0x00004000
+#define bit_PDCM        0x00008000
+#define bit_PCID        0x00020000
+#define bit_DCA         0x00040000
+#define bit_SSE41       0x00080000
+#define bit_SSE42       0x00100000
+#define bit_x2APIC      0x00200000
+#define bit_MOVBE       0x00400000
+#define bit_POPCNT      0x00800000
+#define bit_TSCDeadline 0x01000000
+#define bit_AESNI       0x02000000
+#define bit_XSAVE       0x04000000
+#define bit_OSXSAVE     0x08000000
+#define bit_AVX         0x10000000
+#define bit_RDRND       0x40000000
+
+/* Features in %edx for level 1 */
+#define bit_FPU         0x00000001
+#define bit_VME         0x00000002
+#define bit_DE          0x00000004
+#define bit_PSE         0x00000008
+#define bit_TSC         0x00000010
+#define bit_MSR         0x00000020
+#define bit_PAE         0x00000040
+#define bit_MCE         0x00000080
+#define bit_CX8         0x00000100
+#define bit_APIC        0x00000200
+#define bit_SEP         0x00000800
+#define bit_MTRR        0x00001000
+#define bit_PGE         0x00002000
+#define bit_MCA         0x00004000
+#define bit_CMOV        0x00008000
+#define bit_PAT         0x00010000
+#define bit_PSE36       0x00020000
+#define bit_PSN         0x00040000
+#define bit_CLFSH       0x00080000
+#define bit_DS          0x00200000
+#define bit_ACPI        0x00400000
+#define bit_MMX         0x00800000
+#define bit_FXSR        0x01000000
+#define bit_FXSAVE      bit_FXSR    /* for gcc compat */
+#define bit_SSE         0x02000000
+#define bit_SSE2        0x04000000
+#define bit_SS          0x08000000
+#define bit_HTT         0x10000000
+#define bit_TM          0x20000000
+#define bit_PBE         0x80000000
+
+/* Features in %ebx for level 7 sub-leaf 0 */
+#define bit_FSGSBASE    0x00000001
+#define bit_SMEP        0x00000080
+#define bit_ENH_MOVSB   0x00000200
+
+#if __i386__
+#define __cpuid(__level, __eax, __ebx, __ecx, __edx) \
+    __asm("cpuid" : "=a"(__eax), "=b" (__ebx), "=c"(__ecx), "=d"(__edx) \
+                  : "0"(__level))
+
+#define __cpuid_count(__level, __count, __eax, __ebx, __ecx, __edx) \
+    __asm("cpuid" : "=a"(__eax), "=b" (__ebx), "=c"(__ecx), "=d"(__edx) \
+                  : "0"(__level), "2"(__count))
+#else
+/* x86-64 uses %rbx as the base register, so preserve it. */
+#define __cpuid(__level, __eax, __ebx, __ecx, __edx) \
+    __asm("  xchgq  %%rbx,%q1\n" \
+          "  cpuid\n" \
+          "  xchgq  %%rbx,%q1" \
+        : "=a"(__eax), "=r" (__ebx), "=c"(__ecx), "=d"(__edx) \
+        : "0"(__level))
+
+#define __cpuid_count(__level, __count, __eax, __ebx, __ecx, __edx) \
+    __asm("  xchgq  %%rbx,%q1\n" \
+          "  cpuid\n" \
+          "  xchgq  %%rbx,%q1" \
+        : "=a"(__eax), "=r" (__ebx), "=c"(__ecx), "=d"(__edx) \
+        : "0"(__level), "2"(__count))
+#endif
+
+static __inline int __get_cpuid (unsigned int __level, unsigned int *__eax,
+                                 unsigned int *__ebx, unsigned int *__ecx,
+                                 unsigned int *__edx) {
+    __cpuid(__level, *__eax, *__ebx, *__ecx, *__edx);
+    return 1;
+}
+
+static __inline int __get_cpuid_max (unsigned int __level, unsigned int *__sig)
+{
+    unsigned int __eax, __ebx, __ecx, __edx;
+#if __i386__
+    int __cpuid_supported;
+
+    __asm("  pushfl\n"
+          "  popl   %%eax\n"
+          "  movl   %%eax,%%ecx\n"
+          "  xorl   $0x00200000,%%eax\n"
+          "  pushl  %%eax\n"
+          "  popfl\n"
+          "  pushfl\n"
+          "  popl   %%eax\n"
+          "  movl   $0,%0\n"
+          "  cmpl   %%eax,%%ecx\n"
+          "  je     1f\n"
+          "  movl   $1,%0\n"
+          "1:"
+        : "=r" (__cpuid_supported) : : "eax", "ecx");
+    if (!__cpuid_supported)
+        return 0;
+#endif
+
+    __cpuid(__level, __eax, __ebx, __ecx, __edx);
+    if (__sig)
+        *__sig = __ebx;
+    return __eax;
+}
diff --git a/lib/clang/3.6/include/emmintrin.h b/lib/clang/3.6/include/emmintrin.h
new file mode 100644
index 0000000..b3f8569
--- /dev/null
+++ b/lib/clang/3.6/include/emmintrin.h
@@ -0,0 +1,1451 @@
+/*===---- emmintrin.h - SSE2 intrinsics ------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __EMMINTRIN_H
+#define __EMMINTRIN_H
+
+#ifndef __SSE2__
+#error "SSE2 instruction set not enabled"
+#else
+
+#include <xmmintrin.h>
+
+typedef double __m128d __attribute__((__vector_size__(16)));
+typedef long long __m128i __attribute__((__vector_size__(16)));
+
+/* Type defines.  */
+typedef double __v2df __attribute__ ((__vector_size__ (16)));
+typedef long long __v2di __attribute__ ((__vector_size__ (16)));
+typedef short __v8hi __attribute__((__vector_size__(16)));
+typedef char __v16qi __attribute__((__vector_size__(16)));
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_add_sd(__m128d __a, __m128d __b)
+{
+  __a[0] += __b[0];
+  return __a;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_add_pd(__m128d __a, __m128d __b)
+{
+  return __a + __b;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_sub_sd(__m128d __a, __m128d __b)
+{
+  __a[0] -= __b[0];
+  return __a;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_sub_pd(__m128d __a, __m128d __b)
+{
+  return __a - __b;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_mul_sd(__m128d __a, __m128d __b)
+{
+  __a[0] *= __b[0];
+  return __a;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_mul_pd(__m128d __a, __m128d __b)
+{
+  return __a * __b;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_div_sd(__m128d __a, __m128d __b)
+{
+  __a[0] /= __b[0];
+  return __a;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_div_pd(__m128d __a, __m128d __b)
+{
+  return __a / __b;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_sqrt_sd(__m128d __a, __m128d __b)
+{
+  __m128d __c = __builtin_ia32_sqrtsd(__b);
+  return (__m128d) { __c[0], __a[1] };
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_sqrt_pd(__m128d __a)
+{
+  return __builtin_ia32_sqrtpd(__a);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_min_sd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_minsd(__a, __b);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_min_pd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_minpd(__a, __b);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_max_sd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_maxsd(__a, __b);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_max_pd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_maxpd(__a, __b);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_and_pd(__m128d __a, __m128d __b)
+{
+  return (__m128d)((__v4si)__a & (__v4si)__b);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_andnot_pd(__m128d __a, __m128d __b)
+{
+  return (__m128d)(~(__v4si)__a & (__v4si)__b);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_or_pd(__m128d __a, __m128d __b)
+{
+  return (__m128d)((__v4si)__a | (__v4si)__b);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_xor_pd(__m128d __a, __m128d __b)
+{
+  return (__m128d)((__v4si)__a ^ (__v4si)__b);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_pd(__m128d __a, __m128d __b)
+{
+  return (__m128d)__builtin_ia32_cmppd(__a, __b, 0);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmplt_pd(__m128d __a, __m128d __b)
+{
+  return (__m128d)__builtin_ia32_cmppd(__a, __b, 1);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmple_pd(__m128d __a, __m128d __b)
+{
+  return (__m128d)__builtin_ia32_cmppd(__a, __b, 2);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_pd(__m128d __a, __m128d __b)
+{
+  return (__m128d)__builtin_ia32_cmppd(__b, __a, 1);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpge_pd(__m128d __a, __m128d __b)
+{
+  return (__m128d)__builtin_ia32_cmppd(__b, __a, 2);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpord_pd(__m128d __a, __m128d __b)
+{
+  return (__m128d)__builtin_ia32_cmppd(__a, __b, 7);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpunord_pd(__m128d __a, __m128d __b)
+{
+  return (__m128d)__builtin_ia32_cmppd(__a, __b, 3);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpneq_pd(__m128d __a, __m128d __b)
+{
+  return (__m128d)__builtin_ia32_cmppd(__a, __b, 4);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnlt_pd(__m128d __a, __m128d __b)
+{
+  return (__m128d)__builtin_ia32_cmppd(__a, __b, 5);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnle_pd(__m128d __a, __m128d __b)
+{
+  return (__m128d)__builtin_ia32_cmppd(__a, __b, 6);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpngt_pd(__m128d __a, __m128d __b)
+{
+  return (__m128d)__builtin_ia32_cmppd(__b, __a, 5);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnge_pd(__m128d __a, __m128d __b)
+{
+  return (__m128d)__builtin_ia32_cmppd(__b, __a, 6);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_sd(__m128d __a, __m128d __b)
+{
+  return (__m128d)__builtin_ia32_cmpsd(__a, __b, 0);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmplt_sd(__m128d __a, __m128d __b)
+{
+  return (__m128d)__builtin_ia32_cmpsd(__a, __b, 1);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmple_sd(__m128d __a, __m128d __b)
+{
+  return (__m128d)__builtin_ia32_cmpsd(__a, __b, 2);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_sd(__m128d __a, __m128d __b)
+{
+  __m128d __c = __builtin_ia32_cmpsd(__b, __a, 1);
+  return (__m128d) { __c[0], __a[1] };
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpge_sd(__m128d __a, __m128d __b)
+{
+  __m128d __c = __builtin_ia32_cmpsd(__b, __a, 2);
+  return (__m128d) { __c[0], __a[1] };
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpord_sd(__m128d __a, __m128d __b)
+{
+  return (__m128d)__builtin_ia32_cmpsd(__a, __b, 7);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpunord_sd(__m128d __a, __m128d __b)
+{
+  return (__m128d)__builtin_ia32_cmpsd(__a, __b, 3);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpneq_sd(__m128d __a, __m128d __b)
+{
+  return (__m128d)__builtin_ia32_cmpsd(__a, __b, 4);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnlt_sd(__m128d __a, __m128d __b)
+{
+  return (__m128d)__builtin_ia32_cmpsd(__a, __b, 5);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnle_sd(__m128d __a, __m128d __b)
+{
+  return (__m128d)__builtin_ia32_cmpsd(__a, __b, 6);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpngt_sd(__m128d __a, __m128d __b)
+{
+  __m128d __c = __builtin_ia32_cmpsd(__b, __a, 5);
+  return (__m128d) { __c[0], __a[1] };
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnge_sd(__m128d __a, __m128d __b)
+{
+  __m128d __c = __builtin_ia32_cmpsd(__b, __a, 6);
+  return (__m128d) { __c[0], __a[1] };
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comieq_sd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_comisdeq(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comilt_sd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_comisdlt(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comile_sd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_comisdle(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comigt_sd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_comisdgt(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comige_sd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_comisdge(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comineq_sd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_comisdneq(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomieq_sd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_ucomisdeq(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomilt_sd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_ucomisdlt(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomile_sd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_ucomisdle(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomigt_sd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_ucomisdgt(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomige_sd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_ucomisdge(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomineq_sd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_ucomisdneq(__a, __b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtpd_ps(__m128d __a)
+{
+  return __builtin_ia32_cvtpd2ps(__a);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cvtps_pd(__m128 __a)
+{
+  return __builtin_ia32_cvtps2pd(__a);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepi32_pd(__m128i __a)
+{
+  return __builtin_ia32_cvtdq2pd((__v4si)__a);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtpd_epi32(__m128d __a)
+{
+  return __builtin_ia32_cvtpd2dq(__a);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsd_si32(__m128d __a)
+{
+  return __builtin_ia32_cvtsd2si(__a);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsd_ss(__m128 __a, __m128d __b)
+{
+  __a[0] = __b[0];
+  return __a;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi32_sd(__m128d __a, int __b)
+{
+  __a[0] = __b;
+  return __a;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cvtss_sd(__m128d __a, __m128 __b)
+{
+  __a[0] = __b[0];
+  return __a;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvttpd_epi32(__m128d __a)
+{
+  return (__m128i)__builtin_ia32_cvttpd2dq(__a);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_cvttsd_si32(__m128d __a)
+{
+  return __a[0];
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtpd_pi32(__m128d __a)
+{
+  return (__m64)__builtin_ia32_cvtpd2pi(__a);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cvttpd_pi32(__m128d __a)
+{
+  return (__m64)__builtin_ia32_cvttpd2pi(__a);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cvtpi32_pd(__m64 __a)
+{
+  return __builtin_ia32_cvtpi2pd((__v2si)__a);
+}
+
+static __inline__ double __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsd_f64(__m128d __a)
+{
+  return __a[0];
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_load_pd(double const *__dp)
+{
+  return *(__m128d*)__dp;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_load1_pd(double const *__dp)
+{
+  struct __mm_load1_pd_struct {
+    double __u;
+  } __attribute__((__packed__, __may_alias__));
+  double __u = ((struct __mm_load1_pd_struct*)__dp)->__u;
+  return (__m128d){ __u, __u };
+}
+
+#define        _mm_load_pd1(dp)        _mm_load1_pd(dp)
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_loadr_pd(double const *__dp)
+{
+  __m128d __u = *(__m128d*)__dp;
+  return __builtin_shufflevector(__u, __u, 1, 0);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_loadu_pd(double const *__dp)
+{
+  struct __loadu_pd {
+    __m128d __v;
+  } __attribute__((packed, may_alias));
+  return ((struct __loadu_pd*)__dp)->__v;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_load_sd(double const *__dp)
+{
+  struct __mm_load_sd_struct {
+    double __u;
+  } __attribute__((__packed__, __may_alias__));
+  double __u = ((struct __mm_load_sd_struct*)__dp)->__u;
+  return (__m128d){ __u, 0 };
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_loadh_pd(__m128d __a, double const *__dp)
+{
+  struct __mm_loadh_pd_struct {
+    double __u;
+  } __attribute__((__packed__, __may_alias__));
+  double __u = ((struct __mm_loadh_pd_struct*)__dp)->__u;
+  return (__m128d){ __a[0], __u };
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_loadl_pd(__m128d __a, double const *__dp)
+{
+  struct __mm_loadl_pd_struct {
+    double __u;
+  } __attribute__((__packed__, __may_alias__));
+  double __u = ((struct __mm_loadl_pd_struct*)__dp)->__u;
+  return (__m128d){ __u, __a[1] };
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_set_sd(double __w)
+{
+  return (__m128d){ __w, 0 };
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_set1_pd(double __w)
+{
+  return (__m128d){ __w, __w };
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_set_pd(double __w, double __x)
+{
+  return (__m128d){ __x, __w };
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_setr_pd(double __w, double __x)
+{
+  return (__m128d){ __w, __x };
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_setzero_pd(void)
+{
+  return (__m128d){ 0, 0 };
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_move_sd(__m128d __a, __m128d __b)
+{
+  return (__m128d){ __b[0], __a[1] };
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_store_sd(double *__dp, __m128d __a)
+{
+  struct __mm_store_sd_struct {
+    double __u;
+  } __attribute__((__packed__, __may_alias__));
+  ((struct __mm_store_sd_struct*)__dp)->__u = __a[0];
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_store1_pd(double *__dp, __m128d __a)
+{
+  struct __mm_store1_pd_struct {
+    double __u[2];
+  } __attribute__((__packed__, __may_alias__));
+  ((struct __mm_store1_pd_struct*)__dp)->__u[0] = __a[0];
+  ((struct __mm_store1_pd_struct*)__dp)->__u[1] = __a[0];
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_store_pd(double *__dp, __m128d __a)
+{
+  *(__m128d *)__dp = __a;
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_storeu_pd(double *__dp, __m128d __a)
+{
+  __builtin_ia32_storeupd(__dp, __a);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_storer_pd(double *__dp, __m128d __a)
+{
+  __a = __builtin_shufflevector(__a, __a, 1, 0);
+  *(__m128d *)__dp = __a;
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_storeh_pd(double *__dp, __m128d __a)
+{
+  struct __mm_storeh_pd_struct {
+    double __u;
+  } __attribute__((__packed__, __may_alias__));
+  ((struct __mm_storeh_pd_struct*)__dp)->__u = __a[1];
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_storel_pd(double *__dp, __m128d __a)
+{
+  struct __mm_storeh_pd_struct {
+    double __u;
+  } __attribute__((__packed__, __may_alias__));
+  ((struct __mm_storeh_pd_struct*)__dp)->__u = __a[0];
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_add_epi8(__m128i __a, __m128i __b)
+{
+  return (__m128i)((__v16qi)__a + (__v16qi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_add_epi16(__m128i __a, __m128i __b)
+{
+  return (__m128i)((__v8hi)__a + (__v8hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_add_epi32(__m128i __a, __m128i __b)
+{
+  return (__m128i)((__v4si)__a + (__v4si)__b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_add_si64(__m64 __a, __m64 __b)
+{
+  return __a + __b;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_add_epi64(__m128i __a, __m128i __b)
+{
+  return __a + __b;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_adds_epi8(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_ia32_paddsb128((__v16qi)__a, (__v16qi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_adds_epi16(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_ia32_paddsw128((__v8hi)__a, (__v8hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_adds_epu8(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_ia32_paddusb128((__v16qi)__a, (__v16qi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_adds_epu16(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_ia32_paddusw128((__v8hi)__a, (__v8hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_avg_epu8(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_ia32_pavgb128((__v16qi)__a, (__v16qi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_avg_epu16(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_ia32_pavgw128((__v8hi)__a, (__v8hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_madd_epi16(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_ia32_pmaddwd128((__v8hi)__a, (__v8hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_max_epi16(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_ia32_pmaxsw128((__v8hi)__a, (__v8hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_max_epu8(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_ia32_pmaxub128((__v16qi)__a, (__v16qi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_min_epi16(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_ia32_pminsw128((__v8hi)__a, (__v8hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_min_epu8(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_ia32_pminub128((__v16qi)__a, (__v16qi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_mulhi_epi16(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_ia32_pmulhw128((__v8hi)__a, (__v8hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_mulhi_epu16(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_ia32_pmulhuw128((__v8hi)__a, (__v8hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_mullo_epi16(__m128i __a, __m128i __b)
+{
+  return (__m128i)((__v8hi)__a * (__v8hi)__b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_mul_su32(__m64 __a, __m64 __b)
+{
+  return __builtin_ia32_pmuludq((__v2si)__a, (__v2si)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_mul_epu32(__m128i __a, __m128i __b)
+{
+  return __builtin_ia32_pmuludq128((__v4si)__a, (__v4si)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sad_epu8(__m128i __a, __m128i __b)
+{
+  return __builtin_ia32_psadbw128((__v16qi)__a, (__v16qi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sub_epi8(__m128i __a, __m128i __b)
+{
+  return (__m128i)((__v16qi)__a - (__v16qi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sub_epi16(__m128i __a, __m128i __b)
+{
+  return (__m128i)((__v8hi)__a - (__v8hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sub_epi32(__m128i __a, __m128i __b)
+{
+  return (__m128i)((__v4si)__a - (__v4si)__b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sub_si64(__m64 __a, __m64 __b)
+{
+  return __a - __b;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sub_epi64(__m128i __a, __m128i __b)
+{
+  return __a - __b;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_subs_epi8(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_ia32_psubsb128((__v16qi)__a, (__v16qi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_subs_epi16(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_ia32_psubsw128((__v8hi)__a, (__v8hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_subs_epu8(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_ia32_psubusb128((__v16qi)__a, (__v16qi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_subs_epu16(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_ia32_psubusw128((__v8hi)__a, (__v8hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_and_si128(__m128i __a, __m128i __b)
+{
+  return __a & __b;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_andnot_si128(__m128i __a, __m128i __b)
+{
+  return ~__a & __b;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_or_si128(__m128i __a, __m128i __b)
+{
+  return __a | __b;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_xor_si128(__m128i __a, __m128i __b)
+{
+  return __a ^ __b;
+}
+
+#define _mm_slli_si128(a, count) __extension__ ({ \
+  _Pragma("clang diagnostic push") _Pragma("clang diagnostic ignored \"-Wshadow\""); \
+  __m128i __a = (a); \
+   _Pragma("clang diagnostic pop"); \
+  (__m128i)__builtin_ia32_pslldqi128(__a, (count)*8); })
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_slli_epi16(__m128i __a, int __count)
+{
+  return (__m128i)__builtin_ia32_psllwi128((__v8hi)__a, __count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sll_epi16(__m128i __a, __m128i __count)
+{
+  return (__m128i)__builtin_ia32_psllw128((__v8hi)__a, (__v8hi)__count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_slli_epi32(__m128i __a, int __count)
+{
+  return (__m128i)__builtin_ia32_pslldi128((__v4si)__a, __count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sll_epi32(__m128i __a, __m128i __count)
+{
+  return (__m128i)__builtin_ia32_pslld128((__v4si)__a, (__v4si)__count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_slli_epi64(__m128i __a, int __count)
+{
+  return __builtin_ia32_psllqi128(__a, __count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sll_epi64(__m128i __a, __m128i __count)
+{
+  return __builtin_ia32_psllq128(__a, __count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srai_epi16(__m128i __a, int __count)
+{
+  return (__m128i)__builtin_ia32_psrawi128((__v8hi)__a, __count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sra_epi16(__m128i __a, __m128i __count)
+{
+  return (__m128i)__builtin_ia32_psraw128((__v8hi)__a, (__v8hi)__count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srai_epi32(__m128i __a, int __count)
+{
+  return (__m128i)__builtin_ia32_psradi128((__v4si)__a, __count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sra_epi32(__m128i __a, __m128i __count)
+{
+  return (__m128i)__builtin_ia32_psrad128((__v4si)__a, (__v4si)__count);
+}
+
+
+#define _mm_srli_si128(a, count) __extension__ ({ \
+  _Pragma("clang diagnostic push") _Pragma("clang diagnostic ignored \"-Wshadow\""); \
+  __m128i __a = (a); \
+  _Pragma("clang diagnostic pop"); \
+  (__m128i)__builtin_ia32_psrldqi128(__a, (count)*8); })
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srli_epi16(__m128i __a, int __count)
+{
+  return (__m128i)__builtin_ia32_psrlwi128((__v8hi)__a, __count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srl_epi16(__m128i __a, __m128i __count)
+{
+  return (__m128i)__builtin_ia32_psrlw128((__v8hi)__a, (__v8hi)__count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srli_epi32(__m128i __a, int __count)
+{
+  return (__m128i)__builtin_ia32_psrldi128((__v4si)__a, __count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srl_epi32(__m128i __a, __m128i __count)
+{
+  return (__m128i)__builtin_ia32_psrld128((__v4si)__a, (__v4si)__count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srli_epi64(__m128i __a, int __count)
+{
+  return __builtin_ia32_psrlqi128(__a, __count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_srl_epi64(__m128i __a, __m128i __count)
+{
+  return __builtin_ia32_psrlq128(__a, __count);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_epi8(__m128i __a, __m128i __b)
+{
+  return (__m128i)((__v16qi)__a == (__v16qi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_epi16(__m128i __a, __m128i __b)
+{
+  return (__m128i)((__v8hi)__a == (__v8hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_epi32(__m128i __a, __m128i __b)
+{
+  return (__m128i)((__v4si)__a == (__v4si)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_epi8(__m128i __a, __m128i __b)
+{
+  /* This function always performs a signed comparison, but __v16qi is a char
+     which may be signed or unsigned. */
+  typedef signed char __v16qs __attribute__((__vector_size__(16)));
+  return (__m128i)((__v16qs)__a > (__v16qs)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_epi16(__m128i __a, __m128i __b)
+{
+  return (__m128i)((__v8hi)__a > (__v8hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_epi32(__m128i __a, __m128i __b)
+{
+  return (__m128i)((__v4si)__a > (__v4si)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmplt_epi8(__m128i __a, __m128i __b)
+{
+  return _mm_cmpgt_epi8(__b, __a);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmplt_epi16(__m128i __a, __m128i __b)
+{
+  return _mm_cmpgt_epi16(__b, __a);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmplt_epi32(__m128i __a, __m128i __b)
+{
+  return _mm_cmpgt_epi32(__b, __a);
+}
+
+#ifdef __x86_64__
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi64_sd(__m128d __a, long long __b)
+{
+  __a[0] = __b;
+  return __a;
+}
+
+static __inline__ long long __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsd_si64(__m128d __a)
+{
+  return __builtin_ia32_cvtsd2si64(__a);
+}
+
+static __inline__ long long __attribute__((__always_inline__, __nodebug__))
+_mm_cvttsd_si64(__m128d __a)
+{
+  return __a[0];
+}
+#endif
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepi32_ps(__m128i __a)
+{
+  return __builtin_ia32_cvtdq2ps((__v4si)__a);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtps_epi32(__m128 __a)
+{
+  return (__m128i)__builtin_ia32_cvtps2dq(__a);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvttps_epi32(__m128 __a)
+{
+  return (__m128i)__builtin_ia32_cvttps2dq(__a);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi32_si128(int __a)
+{
+  return (__m128i)(__v4si){ __a, 0, 0, 0 };
+}
+
+#ifdef __x86_64__
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi64_si128(long long __a)
+{
+  return (__m128i){ __a, 0 };
+}
+#endif
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi128_si32(__m128i __a)
+{
+  __v4si __b = (__v4si)__a;
+  return __b[0];
+}
+
+#ifdef __x86_64__
+static __inline__ long long __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi128_si64(__m128i __a)
+{
+  return __a[0];
+}
+#endif
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_load_si128(__m128i const *__p)
+{
+  return *__p;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_loadu_si128(__m128i const *__p)
+{
+  struct __loadu_si128 {
+    __m128i __v;
+  } __attribute__((packed, may_alias));
+  return ((struct __loadu_si128*)__p)->__v;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_loadl_epi64(__m128i const *__p)
+{
+  struct __mm_loadl_epi64_struct {
+    long long __u;
+  } __attribute__((__packed__, __may_alias__));
+  return (__m128i) { ((struct __mm_loadl_epi64_struct*)__p)->__u, 0};
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_set_epi64x(long long q1, long long q0)
+{
+  return (__m128i){ q0, q1 };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_set_epi64(__m64 q1, __m64 q0)
+{
+  return (__m128i){ (long long)q0, (long long)q1 };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_set_epi32(int i3, int i2, int i1, int i0)
+{
+  return (__m128i)(__v4si){ i0, i1, i2, i3};
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_set_epi16(short w7, short w6, short w5, short w4, short w3, short w2, short w1, short w0)
+{
+  return (__m128i)(__v8hi){ w0, w1, w2, w3, w4, w5, w6, w7 };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_set_epi8(char b15, char b14, char b13, char b12, char b11, char b10, char b9, char b8, char b7, char b6, char b5, char b4, char b3, char b2, char b1, char b0)
+{
+  return (__m128i)(__v16qi){ b0, b1, b2, b3, b4, b5, b6, b7, b8, b9, b10, b11, b12, b13, b14, b15 };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_set1_epi64x(long long __q)
+{
+  return (__m128i){ __q, __q };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_set1_epi64(__m64 __q)
+{
+  return (__m128i){ (long long)__q, (long long)__q };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_set1_epi32(int __i)
+{
+  return (__m128i)(__v4si){ __i, __i, __i, __i };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_set1_epi16(short __w)
+{
+  return (__m128i)(__v8hi){ __w, __w, __w, __w, __w, __w, __w, __w };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_set1_epi8(char __b)
+{
+  return (__m128i)(__v16qi){ __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_setr_epi64(__m64 q0, __m64 q1)
+{
+  return (__m128i){ (long long)q0, (long long)q1 };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_setr_epi32(int i0, int i1, int i2, int i3)
+{
+  return (__m128i)(__v4si){ i0, i1, i2, i3};
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_setr_epi16(short w0, short w1, short w2, short w3, short w4, short w5, short w6, short w7)
+{
+  return (__m128i)(__v8hi){ w0, w1, w2, w3, w4, w5, w6, w7 };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_setr_epi8(char b0, char b1, char b2, char b3, char b4, char b5, char b6, char b7, char b8, char b9, char b10, char b11, char b12, char b13, char b14, char b15)
+{
+  return (__m128i)(__v16qi){ b0, b1, b2, b3, b4, b5, b6, b7, b8, b9, b10, b11, b12, b13, b14, b15 };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_setzero_si128(void)
+{
+  return (__m128i){ 0LL, 0LL };
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_store_si128(__m128i *__p, __m128i __b)
+{
+  *__p = __b;
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_storeu_si128(__m128i *__p, __m128i __b)
+{
+  __builtin_ia32_storedqu((char *)__p, (__v16qi)__b);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_maskmoveu_si128(__m128i __d, __m128i __n, char *__p)
+{
+  __builtin_ia32_maskmovdqu((__v16qi)__d, (__v16qi)__n, __p);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_storel_epi64(__m128i *__p, __m128i __a)
+{
+  struct __mm_storel_epi64_struct {
+    long long __u;
+  } __attribute__((__packed__, __may_alias__));
+  ((struct __mm_storel_epi64_struct*)__p)->__u = __a[0];
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_stream_pd(double *__p, __m128d __a)
+{
+  __builtin_ia32_movntpd(__p, __a);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_stream_si128(__m128i *__p, __m128i __a)
+{
+  __builtin_ia32_movntdq(__p, __a);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_stream_si32(int *__p, int __a)
+{
+  __builtin_ia32_movnti(__p, __a);
+}
+
+#ifdef __x86_64__
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_stream_si64(long long *__p, long long __a)
+{
+  __builtin_ia32_movnti64(__p, __a);
+}
+#endif
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_clflush(void const *__p)
+{
+  __builtin_ia32_clflush(__p);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_lfence(void)
+{
+  __builtin_ia32_lfence();
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_mfence(void)
+{
+  __builtin_ia32_mfence();
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_packs_epi16(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_ia32_packsswb128((__v8hi)__a, (__v8hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_packs_epi32(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_ia32_packssdw128((__v4si)__a, (__v4si)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_packus_epi16(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_ia32_packuswb128((__v8hi)__a, (__v8hi)__b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_extract_epi16(__m128i __a, int __imm)
+{
+  __v8hi __b = (__v8hi)__a;
+  return (unsigned short)__b[__imm & 7];
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_insert_epi16(__m128i __a, int __b, int __imm)
+{
+  __v8hi __c = (__v8hi)__a;
+  __c[__imm & 7] = __b;
+  return (__m128i)__c;
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_movemask_epi8(__m128i __a)
+{
+  return __builtin_ia32_pmovmskb128((__v16qi)__a);
+}
+
+#define _mm_shuffle_epi32(a, imm) __extension__ ({ \
+  _Pragma("clang diagnostic push") _Pragma("clang diagnostic ignored \"-Wshadow\""); \
+  __m128i __a = (a); \
+  _Pragma("clang diagnostic pop"); \
+  (__m128i)__builtin_shufflevector((__v4si)__a, (__v4si) _mm_set1_epi32(0), \
+                                   (imm) & 0x3, ((imm) & 0xc) >> 2, \
+                                   ((imm) & 0x30) >> 4, ((imm) & 0xc0) >> 6); })
+
+#define _mm_shufflelo_epi16(a, imm) __extension__ ({ \
+  _Pragma("clang diagnostic push") _Pragma("clang diagnostic ignored \"-Wshadow\""); \
+  __m128i __a = (a); \
+  _Pragma("clang diagnostic pop"); \
+  (__m128i)__builtin_shufflevector((__v8hi)__a, (__v8hi) _mm_set1_epi16(0), \
+                                   (imm) & 0x3, ((imm) & 0xc) >> 2, \
+                                   ((imm) & 0x30) >> 4, ((imm) & 0xc0) >> 6, \
+                                   4, 5, 6, 7); })
+
+#define _mm_shufflehi_epi16(a, imm) __extension__ ({ \
+  _Pragma("clang diagnostic push") _Pragma("clang diagnostic ignored \"-Wshadow\""); \
+  __m128i __a = (a); \
+  _Pragma("clang diagnostic pop"); \
+  (__m128i)__builtin_shufflevector((__v8hi)__a, (__v8hi) _mm_set1_epi16(0), \
+                                   0, 1, 2, 3, \
+                                   4 + (((imm) & 0x03) >> 0), \
+                                   4 + (((imm) & 0x0c) >> 2), \
+                                   4 + (((imm) & 0x30) >> 4), \
+                                   4 + (((imm) & 0xc0) >> 6)); })
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_unpackhi_epi8(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_shufflevector((__v16qi)__a, (__v16qi)__b, 8, 16+8, 9, 16+9, 10, 16+10, 11, 16+11, 12, 16+12, 13, 16+13, 14, 16+14, 15, 16+15);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_unpackhi_epi16(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_shufflevector((__v8hi)__a, (__v8hi)__b, 4, 8+4, 5, 8+5, 6, 8+6, 7, 8+7);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_unpackhi_epi32(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_shufflevector((__v4si)__a, (__v4si)__b, 2, 4+2, 3, 4+3);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_unpackhi_epi64(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_shufflevector(__a, __b, 1, 2+1);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_unpacklo_epi8(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_shufflevector((__v16qi)__a, (__v16qi)__b, 0, 16+0, 1, 16+1, 2, 16+2, 3, 16+3, 4, 16+4, 5, 16+5, 6, 16+6, 7, 16+7);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_unpacklo_epi16(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_shufflevector((__v8hi)__a, (__v8hi)__b, 0, 8+0, 1, 8+1, 2, 8+2, 3, 8+3);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_unpacklo_epi32(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_shufflevector((__v4si)__a, (__v4si)__b, 0, 4+0, 1, 4+1);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_unpacklo_epi64(__m128i __a, __m128i __b)
+{
+  return (__m128i)__builtin_shufflevector(__a, __b, 0, 2+0);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_movepi64_pi64(__m128i __a)
+{
+  return (__m64)__a[0];
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_movpi64_epi64(__m64 __a)
+{
+  return (__m128i){ (long long)__a, 0 };
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_move_epi64(__m128i __a)
+{
+  return __builtin_shufflevector(__a, (__m128i){ 0 }, 0, 2);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_unpackhi_pd(__m128d __a, __m128d __b)
+{
+  return __builtin_shufflevector(__a, __b, 1, 2+1);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_unpacklo_pd(__m128d __a, __m128d __b)
+{
+  return __builtin_shufflevector(__a, __b, 0, 2+0);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_movemask_pd(__m128d __a)
+{
+  return __builtin_ia32_movmskpd(__a);
+}
+
+#define _mm_shuffle_pd(a, b, i) __extension__ ({ \
+  _Pragma("clang diagnostic push") _Pragma("clang diagnostic ignored \"-Wshadow\""); \
+  __m128d __a = (a); \
+  __m128d __b = (b); \
+  _Pragma("clang diagnostic pop"); \
+  __builtin_shufflevector(__a, __b, (i) & 1, (((i) & 2) >> 1) + 2); })
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_castpd_ps(__m128d __a)
+{
+  return (__m128)__a;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_castpd_si128(__m128d __a)
+{
+  return (__m128i)__a;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_castps_pd(__m128 __a)
+{
+  return (__m128d)__a;
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_castps_si128(__m128 __a)
+{
+  return (__m128i)__a;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_castsi128_ps(__m128i __a)
+{
+  return (__m128)__a;
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_castsi128_pd(__m128i __a)
+{
+  return (__m128d)__a;
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_pause(void)
+{
+  __asm__ volatile ("pause");
+}
+
+#define _MM_SHUFFLE2(x, y) (((x) << 1) | (y))
+
+#endif /* __SSE2__ */
+
+#endif /* __EMMINTRIN_H */
diff --git a/lib/clang/3.6/include/f16cintrin.h b/lib/clang/3.6/include/f16cintrin.h
new file mode 100644
index 0000000..f3614c0
--- /dev/null
+++ b/lib/clang/3.6/include/f16cintrin.h
@@ -0,0 +1,58 @@
+/*===---- f16cintrin.h - F16C intrinsics -----------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#if !defined __X86INTRIN_H && !defined __IMMINTRIN_H
+#error "Never use <f16cintrin.h> directly; include <x86intrin.h> instead."
+#endif
+
+#ifndef __F16C__
+# error "F16C instruction is not enabled"
+#endif /* __F16C__ */
+
+#ifndef __F16CINTRIN_H
+#define __F16CINTRIN_H
+
+typedef float __v8sf __attribute__ ((__vector_size__ (32)));
+typedef float __m256 __attribute__ ((__vector_size__ (32)));
+
+#define _mm_cvtps_ph(a, imm) __extension__ ({ \
+  __m128 __a = (a); \
+ (__m128i)__builtin_ia32_vcvtps2ph((__v4sf)__a, (imm)); })
+
+#define _mm256_cvtps_ph(a, imm) __extension__ ({ \
+  __m256 __a = (a); \
+ (__m128i)__builtin_ia32_vcvtps2ph256((__v8sf)__a, (imm)); })
+
+static __inline __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtph_ps(__m128i __a)
+{
+  return (__m128)__builtin_ia32_vcvtph2ps((__v8hi)__a);
+}
+
+static __inline __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_cvtph_ps(__m128i __a)
+{
+  return (__m256)__builtin_ia32_vcvtph2ps256((__v8hi)__a);
+}
+
+#endif /* __F16CINTRIN_H */
diff --git a/lib/clang/3.6/include/float.h b/lib/clang/3.6/include/float.h
new file mode 100644
index 0000000..238cf76
--- /dev/null
+++ b/lib/clang/3.6/include/float.h
@@ -0,0 +1,124 @@
+/*===---- float.h - Characteristics of floating point types ----------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __FLOAT_H
+#define __FLOAT_H
+
+/* If we're on MinGW, fall back to the system's float.h, which might have
+ * additional definitions provided for Windows.
+ * For more details see http://msdn.microsoft.com/en-us/library/y0ybw9fy.aspx
+ */
+#if (defined(__MINGW32__) || defined(_MSC_VER)) && __STDC_HOSTED__ && \
+    __has_include_next(<float.h>)
+#  include_next <float.h>
+
+/* Undefine anything that we'll be redefining below. */
+#  undef FLT_EVAL_METHOD
+#  undef FLT_ROUNDS
+#  undef FLT_RADIX
+#  undef FLT_MANT_DIG
+#  undef DBL_MANT_DIG
+#  undef LDBL_MANT_DIG
+#  undef DECIMAL_DIG
+#  undef FLT_DIG
+#  undef DBL_DIG
+#  undef LDBL_DIG
+#  undef FLT_MIN_EXP
+#  undef DBL_MIN_EXP
+#  undef LDBL_MIN_EXP
+#  undef FLT_MIN_10_EXP
+#  undef DBL_MIN_10_EXP
+#  undef LDBL_MIN_10_EXP
+#  undef FLT_MAX_EXP
+#  undef DBL_MAX_EXP
+#  undef LDBL_MAX_EXP
+#  undef FLT_MAX_10_EXP
+#  undef DBL_MAX_10_EXP
+#  undef LDBL_MAX_10_EXP
+#  undef FLT_MAX
+#  undef DBL_MAX
+#  undef LDBL_MAX
+#  undef FLT_EPSILON
+#  undef DBL_EPSILON
+#  undef LDBL_EPSILON
+#  undef FLT_MIN
+#  undef DBL_MIN
+#  undef LDBL_MIN
+#  if __STDC_VERSION__ >= 201112L || !defined(__STRICT_ANSI__)
+#    undef FLT_TRUE_MIN
+#    undef DBL_TRUE_MIN
+#    undef LDBL_TRUE_MIN
+#  endif
+#endif
+
+/* Characteristics of floating point types, C99 5.2.4.2.2 */
+
+#define FLT_EVAL_METHOD __FLT_EVAL_METHOD__
+#define FLT_ROUNDS (__builtin_flt_rounds())
+#define FLT_RADIX __FLT_RADIX__
+
+#define FLT_MANT_DIG __FLT_MANT_DIG__
+#define DBL_MANT_DIG __DBL_MANT_DIG__
+#define LDBL_MANT_DIG __LDBL_MANT_DIG__
+
+#define DECIMAL_DIG __DECIMAL_DIG__
+
+#define FLT_DIG __FLT_DIG__
+#define DBL_DIG __DBL_DIG__
+#define LDBL_DIG __LDBL_DIG__
+
+#define FLT_MIN_EXP __FLT_MIN_EXP__
+#define DBL_MIN_EXP __DBL_MIN_EXP__
+#define LDBL_MIN_EXP __LDBL_MIN_EXP__
+
+#define FLT_MIN_10_EXP __FLT_MIN_10_EXP__
+#define DBL_MIN_10_EXP __DBL_MIN_10_EXP__
+#define LDBL_MIN_10_EXP __LDBL_MIN_10_EXP__
+
+#define FLT_MAX_EXP __FLT_MAX_EXP__
+#define DBL_MAX_EXP __DBL_MAX_EXP__
+#define LDBL_MAX_EXP __LDBL_MAX_EXP__
+
+#define FLT_MAX_10_EXP __FLT_MAX_10_EXP__
+#define DBL_MAX_10_EXP __DBL_MAX_10_EXP__
+#define LDBL_MAX_10_EXP __LDBL_MAX_10_EXP__
+
+#define FLT_MAX __FLT_MAX__
+#define DBL_MAX __DBL_MAX__
+#define LDBL_MAX __LDBL_MAX__
+
+#define FLT_EPSILON __FLT_EPSILON__
+#define DBL_EPSILON __DBL_EPSILON__
+#define LDBL_EPSILON __LDBL_EPSILON__
+
+#define FLT_MIN __FLT_MIN__
+#define DBL_MIN __DBL_MIN__
+#define LDBL_MIN __LDBL_MIN__
+
+#if __STDC_VERSION__ >= 201112L || !defined(__STRICT_ANSI__)
+#  define FLT_TRUE_MIN __FLT_DENORM_MIN__
+#  define DBL_TRUE_MIN __DBL_DENORM_MIN__
+#  define LDBL_TRUE_MIN __LDBL_DENORM_MIN__
+#endif
+
+#endif /* __FLOAT_H */
diff --git a/lib/clang/3.6/include/fma4intrin.h b/lib/clang/3.6/include/fma4intrin.h
new file mode 100644
index 0000000..c30920d
--- /dev/null
+++ b/lib/clang/3.6/include/fma4intrin.h
@@ -0,0 +1,231 @@
+/*===---- fma4intrin.h - FMA4 intrinsics -----------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __X86INTRIN_H
+#error "Never use <fma4intrin.h> directly; include <x86intrin.h> instead."
+#endif
+
+#ifndef __FMA4INTRIN_H
+#define __FMA4INTRIN_H
+
+#ifndef __FMA4__
+# error "FMA4 instruction set is not enabled"
+#else
+
+#include <pmmintrin.h>
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_macc_ps(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfmaddps(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_macc_pd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfmaddpd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_macc_ss(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfmaddss(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_macc_sd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfmaddsd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_msub_ps(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfmsubps(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_msub_pd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfmsubpd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_msub_ss(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfmsubss(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_msub_sd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfmsubsd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_nmacc_ps(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfnmaddps(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_nmacc_pd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfnmaddpd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_nmacc_ss(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfnmaddss(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_nmacc_sd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfnmaddsd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_nmsub_ps(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfnmsubps(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_nmsub_pd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfnmsubpd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_nmsub_ss(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfnmsubss(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_nmsub_sd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfnmsubsd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_maddsub_ps(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfmaddsubps(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_maddsub_pd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfmaddsubpd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_msubadd_ps(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfmsubaddps(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_msubadd_pd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfmsubaddpd(__A, __B, __C);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_macc_ps(__m256 __A, __m256 __B, __m256 __C)
+{
+  return (__m256)__builtin_ia32_vfmaddps256(__A, __B, __C);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_macc_pd(__m256d __A, __m256d __B, __m256d __C)
+{
+  return (__m256d)__builtin_ia32_vfmaddpd256(__A, __B, __C);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_msub_ps(__m256 __A, __m256 __B, __m256 __C)
+{
+  return (__m256)__builtin_ia32_vfmsubps256(__A, __B, __C);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_msub_pd(__m256d __A, __m256d __B, __m256d __C)
+{
+  return (__m256d)__builtin_ia32_vfmsubpd256(__A, __B, __C);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_nmacc_ps(__m256 __A, __m256 __B, __m256 __C)
+{
+  return (__m256)__builtin_ia32_vfnmaddps256(__A, __B, __C);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_nmacc_pd(__m256d __A, __m256d __B, __m256d __C)
+{
+  return (__m256d)__builtin_ia32_vfnmaddpd256(__A, __B, __C);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_nmsub_ps(__m256 __A, __m256 __B, __m256 __C)
+{
+  return (__m256)__builtin_ia32_vfnmsubps256(__A, __B, __C);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_nmsub_pd(__m256d __A, __m256d __B, __m256d __C)
+{
+  return (__m256d)__builtin_ia32_vfnmsubpd256(__A, __B, __C);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_maddsub_ps(__m256 __A, __m256 __B, __m256 __C)
+{
+  return (__m256)__builtin_ia32_vfmaddsubps256(__A, __B, __C);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_maddsub_pd(__m256d __A, __m256d __B, __m256d __C)
+{
+  return (__m256d)__builtin_ia32_vfmaddsubpd256(__A, __B, __C);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_msubadd_ps(__m256 __A, __m256 __B, __m256 __C)
+{
+  return (__m256)__builtin_ia32_vfmsubaddps256(__A, __B, __C);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_msubadd_pd(__m256d __A, __m256d __B, __m256d __C)
+{
+  return (__m256d)__builtin_ia32_vfmsubaddpd256(__A, __B, __C);
+}
+
+#endif /* __FMA4__ */
+
+#endif /* __FMA4INTRIN_H */
diff --git a/lib/clang/3.6/include/fmaintrin.h b/lib/clang/3.6/include/fmaintrin.h
new file mode 100644
index 0000000..6bfd5a8
--- /dev/null
+++ b/lib/clang/3.6/include/fmaintrin.h
@@ -0,0 +1,229 @@
+/*===---- fma4intrin.h - FMA4 intrinsics -----------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __IMMINTRIN_H
+#error "Never use <fmaintrin.h> directly; include <immintrin.h> instead."
+#endif
+
+#ifndef __FMAINTRIN_H
+#define __FMAINTRIN_H
+
+#ifndef __FMA__
+# error "FMA instruction set is not enabled"
+#else
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_fmadd_ps(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfmaddps(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_fmadd_pd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfmaddpd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_fmadd_ss(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfmaddss(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_fmadd_sd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfmaddsd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_fmsub_ps(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfmsubps(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_fmsub_pd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfmsubpd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_fmsub_ss(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfmsubss(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_fmsub_sd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfmsubsd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_fnmadd_ps(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfnmaddps(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_fnmadd_pd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfnmaddpd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_fnmadd_ss(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfnmaddss(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_fnmadd_sd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfnmaddsd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_fnmsub_ps(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfnmsubps(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_fnmsub_pd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfnmsubpd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_fnmsub_ss(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfnmsubss(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_fnmsub_sd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfnmsubsd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_fmaddsub_ps(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfmaddsubps(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_fmaddsub_pd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfmaddsubpd(__A, __B, __C);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_fmsubadd_ps(__m128 __A, __m128 __B, __m128 __C)
+{
+  return (__m128)__builtin_ia32_vfmsubaddps(__A, __B, __C);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_fmsubadd_pd(__m128d __A, __m128d __B, __m128d __C)
+{
+  return (__m128d)__builtin_ia32_vfmsubaddpd(__A, __B, __C);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_fmadd_ps(__m256 __A, __m256 __B, __m256 __C)
+{
+  return (__m256)__builtin_ia32_vfmaddps256(__A, __B, __C);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_fmadd_pd(__m256d __A, __m256d __B, __m256d __C)
+{
+  return (__m256d)__builtin_ia32_vfmaddpd256(__A, __B, __C);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_fmsub_ps(__m256 __A, __m256 __B, __m256 __C)
+{
+  return (__m256)__builtin_ia32_vfmsubps256(__A, __B, __C);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_fmsub_pd(__m256d __A, __m256d __B, __m256d __C)
+{
+  return (__m256d)__builtin_ia32_vfmsubpd256(__A, __B, __C);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_fnmadd_ps(__m256 __A, __m256 __B, __m256 __C)
+{
+  return (__m256)__builtin_ia32_vfnmaddps256(__A, __B, __C);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_fnmadd_pd(__m256d __A, __m256d __B, __m256d __C)
+{
+  return (__m256d)__builtin_ia32_vfnmaddpd256(__A, __B, __C);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_fnmsub_ps(__m256 __A, __m256 __B, __m256 __C)
+{
+  return (__m256)__builtin_ia32_vfnmsubps256(__A, __B, __C);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_fnmsub_pd(__m256d __A, __m256d __B, __m256d __C)
+{
+  return (__m256d)__builtin_ia32_vfnmsubpd256(__A, __B, __C);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_fmaddsub_ps(__m256 __A, __m256 __B, __m256 __C)
+{
+  return (__m256)__builtin_ia32_vfmaddsubps256(__A, __B, __C);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_fmaddsub_pd(__m256d __A, __m256d __B, __m256d __C)
+{
+  return (__m256d)__builtin_ia32_vfmaddsubpd256(__A, __B, __C);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_fmsubadd_ps(__m256 __A, __m256 __B, __m256 __C)
+{
+  return (__m256)__builtin_ia32_vfmsubaddps256(__A, __B, __C);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_fmsubadd_pd(__m256d __A, __m256d __B, __m256d __C)
+{
+  return (__m256d)__builtin_ia32_vfmsubaddpd256(__A, __B, __C);
+}
+
+#endif /* __FMA__ */
+
+#endif /* __FMAINTRIN_H */
diff --git a/lib/clang/3.6/include/ia32intrin.h b/lib/clang/3.6/include/ia32intrin.h
new file mode 100644
index 0000000..5adf3f1
--- /dev/null
+++ b/lib/clang/3.6/include/ia32intrin.h
@@ -0,0 +1,101 @@
+/* ===-------- ia32intrin.h ---------------------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __X86INTRIN_H
+#error "Never use <ia32intrin.h> directly; include <x86intrin.h> instead."
+#endif
+
+#ifndef __IA32INTRIN_H
+#define __IA32INTRIN_H
+
+#ifdef __x86_64__
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+__readeflags(void)
+{
+  unsigned long long __res = 0;
+  __asm__ __volatile__ ("pushf\n\t"
+                        "popq %0\n"
+                        :"=r"(__res)
+                        :
+                        :
+                       );
+  return __res;
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+__writeeflags(unsigned long long __f)
+{
+  __asm__ __volatile__ ("pushq %0\n\t"
+                        "popf\n"
+                        :
+                        :"r"(__f)
+                        :"flags"
+                       );
+}
+
+#else /* !__x86_64__ */
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__readeflags(void)
+{
+  unsigned int __res = 0;
+  __asm__ __volatile__ ("pushf\n\t"
+                        "popl %0\n"
+                        :"=r"(__res)
+                        :
+                        :
+                       );
+  return __res;
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+__writeeflags(unsigned int __f)
+{
+  __asm__ __volatile__ ("pushl %0\n\t"
+                        "popf\n"
+                        :
+                        :"r"(__f)
+                        :"flags"
+                       );
+}
+#endif /* !__x86_64__ */
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+__rdpmc(int __A) {
+  return __builtin_ia32_rdpmc(__A);
+}
+
+/* __rdtsc */
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+__rdtsc(void) {
+  return __builtin_ia32_rdtsc();
+}
+
+/* __rdtscp */
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+__rdtscp(unsigned int *__A) {
+  return __builtin_ia32_rdtscp(__A);
+}
+
+#define _rdtsc() __rdtsc()
+
+#endif /* __IA32INTRIN_H */
diff --git a/lib/clang/3.6/include/immintrin.h b/lib/clang/3.6/include/immintrin.h
new file mode 100644
index 0000000..2400fea
--- /dev/null
+++ b/lib/clang/3.6/include/immintrin.h
@@ -0,0 +1,194 @@
+/*===---- immintrin.h - Intel intrinsics -----------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __IMMINTRIN_H
+#define __IMMINTRIN_H
+
+#ifdef __MMX__
+#include <mmintrin.h>
+#endif
+
+#ifdef __SSE__
+#include <xmmintrin.h>
+#endif
+
+#ifdef __SSE2__
+#include <emmintrin.h>
+#endif
+
+#ifdef __SSE3__
+#include <pmmintrin.h>
+#endif
+
+#ifdef __SSSE3__
+#include <tmmintrin.h>
+#endif
+
+#if defined (__SSE4_2__) || defined (__SSE4_1__)
+#include <smmintrin.h>
+#endif
+
+#if defined (__AES__) || defined (__PCLMUL__)
+#include <wmmintrin.h>
+#endif
+
+#ifdef __AVX__
+#include <avxintrin.h>
+#endif
+
+#ifdef __AVX2__
+#include <avx2intrin.h>
+#endif
+
+#ifdef __BMI__
+#include <bmiintrin.h>
+#endif
+
+#ifdef __BMI2__
+#include <bmi2intrin.h>
+#endif
+
+#ifdef __LZCNT__
+#include <lzcntintrin.h>
+#endif
+
+#ifdef __FMA__
+#include <fmaintrin.h>
+#endif
+
+#ifdef __AVX512F__
+#include <avx512fintrin.h>
+#endif
+
+#ifdef __AVX512VL__
+#include <avx512vlintrin.h>
+#endif
+
+#ifdef __AVX512BW__
+#include <avx512bwintrin.h>
+#endif
+
+#if defined (__AVX512VL__) && defined (__AVX512BW__)
+#include <avx512vlbwintrin.h>
+#endif
+
+#ifdef __AVX512ER__
+#include <avx512erintrin.h>
+#endif
+
+#ifdef __RDRND__
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_rdrand16_step(unsigned short *__p)
+{
+  return __builtin_ia32_rdrand16_step(__p);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_rdrand32_step(unsigned int *__p)
+{
+  return __builtin_ia32_rdrand32_step(__p);
+}
+
+#ifdef __x86_64__
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_rdrand64_step(unsigned long long *__p)
+{
+  return __builtin_ia32_rdrand64_step(__p);
+}
+#endif
+#endif /* __RDRND__ */
+
+#ifdef __FSGSBASE__
+#ifdef __x86_64__
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_readfsbase_u32(void)
+{
+  return __builtin_ia32_rdfsbase32();
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+_readfsbase_u64(void)
+{
+  return __builtin_ia32_rdfsbase64();
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_readgsbase_u32(void)
+{
+  return __builtin_ia32_rdgsbase32();
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+_readgsbase_u64(void)
+{
+  return __builtin_ia32_rdgsbase64();
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_writefsbase_u32(unsigned int __V)
+{
+  return __builtin_ia32_wrfsbase32(__V);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_writefsbase_u64(unsigned long long __V)
+{
+  return __builtin_ia32_wrfsbase64(__V);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_writegsbase_u32(unsigned int __V)
+{
+  return __builtin_ia32_wrgsbase32(__V);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_writegsbase_u64(unsigned long long __V)
+{
+  return __builtin_ia32_wrgsbase64(__V);
+}
+#endif
+#endif /* __FSGSBASE__ */
+
+#ifdef __RTM__
+#include <rtmintrin.h>
+#endif
+
+/* FIXME: check __HLE__ as well when HLE is supported. */
+#if defined (__RTM__)
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_xtest(void)
+{
+  return __builtin_ia32_xtest();
+}
+#endif
+
+#ifdef __SHA__
+#include <shaintrin.h>
+#endif
+
+/* Some intrinsics inside adxintrin.h are available only if __ADX__ defined,
+ * whereas others are also available if __ADX__ undefined */
+#include <adxintrin.h>
+
+#endif /* __IMMINTRIN_H */
diff --git a/lib/clang/3.6/include/iso646.h b/lib/clang/3.6/include/iso646.h
new file mode 100644
index 0000000..dca13c5
--- /dev/null
+++ b/lib/clang/3.6/include/iso646.h
@@ -0,0 +1,43 @@
+/*===---- iso646.h - Standard header for alternate spellings of operators---===
+ *
+ * Copyright (c) 2008 Eli Friedman
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __ISO646_H
+#define __ISO646_H
+
+#ifndef __cplusplus
+#define and    &&
+#define and_eq &=
+#define bitand &
+#define bitor  |
+#define compl  ~
+#define not    !
+#define not_eq !=
+#define or     ||
+#define or_eq  |=
+#define xor    ^
+#define xor_eq ^=
+#endif
+
+#endif /* __ISO646_H */
diff --git a/lib/clang/3.6/include/limits.h b/lib/clang/3.6/include/limits.h
new file mode 100644
index 0000000..f04187c
--- /dev/null
+++ b/lib/clang/3.6/include/limits.h
@@ -0,0 +1,118 @@
+/*===---- limits.h - Standard header for integer sizes --------------------===*\
+ *
+ * Copyright (c) 2009 Chris Lattner
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+\*===----------------------------------------------------------------------===*/
+
+#ifndef __CLANG_LIMITS_H
+#define __CLANG_LIMITS_H
+
+/* The system's limits.h may, in turn, try to #include_next GCC's limits.h.
+   Avert this #include_next madness. */
+#if defined __GNUC__ && !defined _GCC_LIMITS_H_
+#define _GCC_LIMITS_H_
+#endif
+
+/* System headers include a number of constants from POSIX in <limits.h>.
+   Include it if we're hosted. */
+#if __STDC_HOSTED__ && __has_include_next(<limits.h>)
+#include_next <limits.h>
+#endif
+
+/* Many system headers try to "help us out" by defining these.  No really, we
+   know how big each datatype is. */
+#undef  SCHAR_MIN
+#undef  SCHAR_MAX
+#undef  UCHAR_MAX
+#undef  SHRT_MIN
+#undef  SHRT_MAX
+#undef  USHRT_MAX
+#undef  INT_MIN
+#undef  INT_MAX
+#undef  UINT_MAX
+#undef  LONG_MIN
+#undef  LONG_MAX
+#undef  ULONG_MAX
+
+#undef  CHAR_BIT
+#undef  CHAR_MIN
+#undef  CHAR_MAX
+
+/* C90/99 5.2.4.2.1 */
+#define SCHAR_MAX __SCHAR_MAX__
+#define SHRT_MAX  __SHRT_MAX__
+#define INT_MAX   __INT_MAX__
+#define LONG_MAX  __LONG_MAX__
+
+#define SCHAR_MIN (-__SCHAR_MAX__-1)
+#define SHRT_MIN  (-__SHRT_MAX__ -1)
+#define INT_MIN   (-__INT_MAX__  -1)
+#define LONG_MIN  (-__LONG_MAX__ -1L)
+
+#define UCHAR_MAX (__SCHAR_MAX__*2  +1)
+#define USHRT_MAX (__SHRT_MAX__ *2  +1)
+#define UINT_MAX  (__INT_MAX__  *2U +1U)
+#define ULONG_MAX (__LONG_MAX__ *2UL+1UL)
+
+#ifndef MB_LEN_MAX
+#define MB_LEN_MAX 1
+#endif
+
+#define CHAR_BIT  __CHAR_BIT__
+
+#ifdef __CHAR_UNSIGNED__  /* -funsigned-char */
+#define CHAR_MIN 0
+#define CHAR_MAX UCHAR_MAX
+#else
+#define CHAR_MIN SCHAR_MIN
+#define CHAR_MAX __SCHAR_MAX__
+#endif
+
+/* C99 5.2.4.2.1: Added long long.
+   C++11 18.3.3.2: same contents as the Standard C Library header <limits.h>.
+ */
+#if __STDC_VERSION__ >= 199901L || __cplusplus >= 201103L
+
+#undef  LLONG_MIN
+#undef  LLONG_MAX
+#undef  ULLONG_MAX
+
+#define LLONG_MAX  __LONG_LONG_MAX__
+#define LLONG_MIN  (-__LONG_LONG_MAX__-1LL)
+#define ULLONG_MAX (__LONG_LONG_MAX__*2ULL+1ULL)
+#endif
+
+/* LONG_LONG_MIN/LONG_LONG_MAX/ULONG_LONG_MAX are a GNU extension.  It's too bad
+   that we don't have something like #pragma poison that could be used to
+   deprecate a macro - the code should just use LLONG_MAX and friends.
+ */
+#if defined(__GNU_LIBRARY__) ? defined(__USE_GNU) : !defined(__STRICT_ANSI__)
+
+#undef   LONG_LONG_MIN
+#undef   LONG_LONG_MAX
+#undef   ULONG_LONG_MAX
+
+#define LONG_LONG_MAX  __LONG_LONG_MAX__
+#define LONG_LONG_MIN  (-__LONG_LONG_MAX__-1LL)
+#define ULONG_LONG_MAX (__LONG_LONG_MAX__*2ULL+1ULL)
+#endif
+
+#endif /* __CLANG_LIMITS_H */
diff --git a/lib/clang/3.6/include/lzcntintrin.h b/lib/clang/3.6/include/lzcntintrin.h
new file mode 100644
index 0000000..35d6659
--- /dev/null
+++ b/lib/clang/3.6/include/lzcntintrin.h
@@ -0,0 +1,67 @@
+/*===---- lzcntintrin.h - LZCNT intrinsics ---------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#if !defined __X86INTRIN_H && !defined __IMMINTRIN_H
+#error "Never use <lzcntintrin.h> directly; include <x86intrin.h> instead."
+#endif
+
+#ifndef __LZCNT__
+# error "LZCNT instruction is not enabled"
+#endif /* __LZCNT__ */
+
+#ifndef __LZCNTINTRIN_H
+#define __LZCNTINTRIN_H
+
+static __inline__ unsigned short __attribute__((__always_inline__, __nodebug__))
+__lzcnt16(unsigned short __X)
+{
+  return __X ? __builtin_clzs(__X) : 16;
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__lzcnt32(unsigned int __X)
+{
+  return __X ? __builtin_clz(__X) : 32;
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_lzcnt_u32(unsigned int __X)
+{
+  return __X ? __builtin_clz(__X) : 32;
+}
+
+#ifdef __x86_64__
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+__lzcnt64(unsigned long long __X)
+{
+  return __X ? __builtin_clzll(__X) : 64;
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+_lzcnt_u64(unsigned long long __X)
+{
+  return __X ? __builtin_clzll(__X) : 64;
+}
+#endif
+
+#endif /* __LZCNTINTRIN_H */
diff --git a/lib/clang/3.6/include/mm3dnow.h b/lib/clang/3.6/include/mm3dnow.h
new file mode 100644
index 0000000..5242d99
--- /dev/null
+++ b/lib/clang/3.6/include/mm3dnow.h
@@ -0,0 +1,162 @@
+/*===---- mm3dnow.h - 3DNow! intrinsics ------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef _MM3DNOW_H_INCLUDED
+#define _MM3DNOW_H_INCLUDED
+
+#include <mmintrin.h>
+#include <prfchwintrin.h>
+
+typedef float __v2sf __attribute__((__vector_size__(8)));
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_m_femms() {
+  __builtin_ia32_femms();
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pavgusb(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pavgusb((__v8qi)__m1, (__v8qi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pf2id(__m64 __m) {
+  return (__m64)__builtin_ia32_pf2id((__v2sf)__m);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfacc(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfacc((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfadd(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfadd((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfcmpeq(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfcmpeq((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfcmpge(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfcmpge((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfcmpgt(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfcmpgt((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfmax(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfmax((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfmin(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfmin((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfmul(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfmul((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfrcp(__m64 __m) {
+  return (__m64)__builtin_ia32_pfrcp((__v2sf)__m);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfrcpit1(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfrcpit1((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfrcpit2(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfrcpit2((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfrsqrt(__m64 __m) {
+  return (__m64)__builtin_ia32_pfrsqrt((__v2sf)__m);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfrsqrtit1(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfrsqit1((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfsub(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfsub((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfsubr(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfsubr((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pi2fd(__m64 __m) {
+  return (__m64)__builtin_ia32_pi2fd((__v2si)__m);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pmulhrw(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pmulhrw((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pf2iw(__m64 __m) {
+  return (__m64)__builtin_ia32_pf2iw((__v2sf)__m);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfnacc(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfnacc((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pfpnacc(__m64 __m1, __m64 __m2) {
+  return (__m64)__builtin_ia32_pfpnacc((__v2sf)__m1, (__v2sf)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pi2fw(__m64 __m) {
+  return (__m64)__builtin_ia32_pi2fw((__v2si)__m);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pswapdsf(__m64 __m) {
+  return (__m64)__builtin_ia32_pswapdsf((__v2sf)__m);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_m_pswapdsi(__m64 __m) {
+  return (__m64)__builtin_ia32_pswapdsi((__v2si)__m);
+}
+
+#endif
diff --git a/lib/clang/3.6/include/mm_malloc.h b/lib/clang/3.6/include/mm_malloc.h
new file mode 100644
index 0000000..305afd3
--- /dev/null
+++ b/lib/clang/3.6/include/mm_malloc.h
@@ -0,0 +1,75 @@
+/*===---- mm_malloc.h - Allocating and Freeing Aligned Memory Blocks -------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __MM_MALLOC_H
+#define __MM_MALLOC_H
+
+#include <stdlib.h>
+
+#ifdef _WIN32
+#include <malloc.h>
+#else
+#ifndef __cplusplus
+extern int posix_memalign(void **__memptr, size_t __alignment, size_t __size);
+#else
+// Some systems (e.g. those with GNU libc) declare posix_memalign with an
+// exception specifier. Via an "egregious workaround" in
+// Sema::CheckEquivalentExceptionSpec, Clang accepts the following as a valid
+// redeclaration of glibc's declaration.
+extern "C" int posix_memalign(void **__memptr, size_t __alignment, size_t __size);
+#endif
+#endif
+
+#if !(defined(_WIN32) && defined(_mm_malloc))
+static __inline__ void *__attribute__((__always_inline__, __nodebug__,
+                                       __malloc__))
+_mm_malloc(size_t __size, size_t __align)
+{
+  if (__align == 1) {
+    return malloc(__size);
+  }
+
+  if (!(__align & (__align - 1)) && __align < sizeof(void *))
+    __align = sizeof(void *);
+
+  void *__mallocedMemory;
+#if defined(__MINGW32__)
+  __mallocedMemory = __mingw_aligned_malloc(__size, __align);
+#elif defined(_WIN32)
+  __mallocedMemory = _aligned_malloc(__size, __align);
+#else
+  if (posix_memalign(&__mallocedMemory, __align, __size))
+    return 0;
+#endif
+
+  return __mallocedMemory;
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_free(void *__p)
+{
+  free(__p);
+}
+#endif
+
+#endif /* __MM_MALLOC_H */
diff --git a/lib/clang/3.6/include/mmintrin.h b/lib/clang/3.6/include/mmintrin.h
new file mode 100644
index 0000000..986870a
--- /dev/null
+++ b/lib/clang/3.6/include/mmintrin.h
@@ -0,0 +1,503 @@
+/*===---- mmintrin.h - MMX intrinsics --------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __MMINTRIN_H
+#define __MMINTRIN_H
+
+#ifndef __MMX__
+#error "MMX instruction set not enabled"
+#else
+
+typedef long long __m64 __attribute__((__vector_size__(8)));
+
+typedef int __v2si __attribute__((__vector_size__(8)));
+typedef short __v4hi __attribute__((__vector_size__(8)));
+typedef char __v8qi __attribute__((__vector_size__(8)));
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_empty(void)
+{
+    __builtin_ia32_emms();
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi32_si64(int __i)
+{
+    return (__m64)__builtin_ia32_vec_init_v2si(__i, 0);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi64_si32(__m64 __m)
+{
+    return __builtin_ia32_vec_ext_v2si((__v2si)__m, 0);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi64_m64(long long __i)
+{
+    return (__m64)__i;
+}
+
+static __inline__ long long __attribute__((__always_inline__, __nodebug__))
+_mm_cvtm64_si64(__m64 __m)
+{
+    return (long long)__m;
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_packs_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_packsswb((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_packs_pi32(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_packssdw((__v2si)__m1, (__v2si)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_packs_pu16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_packuswb((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_unpackhi_pi8(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_punpckhbw((__v8qi)__m1, (__v8qi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_unpackhi_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_punpckhwd((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_unpackhi_pi32(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_punpckhdq((__v2si)__m1, (__v2si)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_unpacklo_pi8(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_punpcklbw((__v8qi)__m1, (__v8qi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_unpacklo_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_punpcklwd((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_unpacklo_pi32(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_punpckldq((__v2si)__m1, (__v2si)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_add_pi8(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_paddb((__v8qi)__m1, (__v8qi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_add_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_paddw((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_add_pi32(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_paddd((__v2si)__m1, (__v2si)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_adds_pi8(__m64 __m1, __m64 __m2) 
+{
+    return (__m64)__builtin_ia32_paddsb((__v8qi)__m1, (__v8qi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_adds_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_paddsw((__v4hi)__m1, (__v4hi)__m2);    
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_adds_pu8(__m64 __m1, __m64 __m2) 
+{
+    return (__m64)__builtin_ia32_paddusb((__v8qi)__m1, (__v8qi)__m2);
+}
+ 
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_adds_pu16(__m64 __m1, __m64 __m2) 
+{
+    return (__m64)__builtin_ia32_paddusw((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sub_pi8(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_psubb((__v8qi)__m1, (__v8qi)__m2);
+}
+ 
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sub_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_psubw((__v4hi)__m1, (__v4hi)__m2);
+}
+ 
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sub_pi32(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_psubd((__v2si)__m1, (__v2si)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_subs_pi8(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_psubsb((__v8qi)__m1, (__v8qi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_subs_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_psubsw((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_subs_pu8(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_psubusb((__v8qi)__m1, (__v8qi)__m2);
+}
+ 
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_subs_pu16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_psubusw((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_madd_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_pmaddwd((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_mulhi_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_pmulhw((__v4hi)__m1, (__v4hi)__m2);
+}
+ 
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_mullo_pi16(__m64 __m1, __m64 __m2) 
+{
+    return (__m64)__builtin_ia32_pmullw((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sll_pi16(__m64 __m, __m64 __count)
+{
+    return (__m64)__builtin_ia32_psllw((__v4hi)__m, __count);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_slli_pi16(__m64 __m, int __count)
+{
+    return (__m64)__builtin_ia32_psllwi((__v4hi)__m, __count);    
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sll_pi32(__m64 __m, __m64 __count)
+{
+    return (__m64)__builtin_ia32_pslld((__v2si)__m, __count);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_slli_pi32(__m64 __m, int __count)
+{
+    return (__m64)__builtin_ia32_pslldi((__v2si)__m, __count);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sll_si64(__m64 __m, __m64 __count)
+{
+    return (__m64)__builtin_ia32_psllq(__m, __count);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_slli_si64(__m64 __m, int __count)
+{
+    return (__m64)__builtin_ia32_psllqi(__m, __count);    
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sra_pi16(__m64 __m, __m64 __count)
+{
+    return (__m64)__builtin_ia32_psraw((__v4hi)__m, __count);    
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_srai_pi16(__m64 __m, int __count)
+{
+    return (__m64)__builtin_ia32_psrawi((__v4hi)__m, __count);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sra_pi32(__m64 __m, __m64 __count)
+{
+    return (__m64)__builtin_ia32_psrad((__v2si)__m, __count);    
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_srai_pi32(__m64 __m, int __count)
+{
+    return (__m64)__builtin_ia32_psradi((__v2si)__m, __count);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_srl_pi16(__m64 __m, __m64 __count)
+{
+    return (__m64)__builtin_ia32_psrlw((__v4hi)__m, __count);    
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_srli_pi16(__m64 __m, int __count)
+{
+    return (__m64)__builtin_ia32_psrlwi((__v4hi)__m, __count);    
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_srl_pi32(__m64 __m, __m64 __count)
+{
+    return (__m64)__builtin_ia32_psrld((__v2si)__m, __count);       
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_srli_pi32(__m64 __m, int __count)
+{
+    return (__m64)__builtin_ia32_psrldi((__v2si)__m, __count);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_srl_si64(__m64 __m, __m64 __count)
+{
+    return (__m64)__builtin_ia32_psrlq(__m, __count);    
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_srli_si64(__m64 __m, int __count)
+{
+    return (__m64)__builtin_ia32_psrlqi(__m, __count);    
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_and_si64(__m64 __m1, __m64 __m2)
+{
+    return __builtin_ia32_pand(__m1, __m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_andnot_si64(__m64 __m1, __m64 __m2)
+{
+    return __builtin_ia32_pandn(__m1, __m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_or_si64(__m64 __m1, __m64 __m2)
+{
+    return __builtin_ia32_por(__m1, __m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_xor_si64(__m64 __m1, __m64 __m2)
+{
+    return __builtin_ia32_pxor(__m1, __m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_pi8(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_pcmpeqb((__v8qi)__m1, (__v8qi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_pcmpeqw((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_pi32(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_pcmpeqd((__v2si)__m1, (__v2si)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_pi8(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_pcmpgtb((__v8qi)__m1, (__v8qi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_pi16(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_pcmpgtw((__v4hi)__m1, (__v4hi)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_pi32(__m64 __m1, __m64 __m2)
+{
+    return (__m64)__builtin_ia32_pcmpgtd((__v2si)__m1, (__v2si)__m2);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_setzero_si64(void)
+{
+    return (__m64){ 0LL };
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_set_pi32(int __i1, int __i0)
+{
+    return (__m64)__builtin_ia32_vec_init_v2si(__i0, __i1);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_set_pi16(short __s3, short __s2, short __s1, short __s0)
+{
+    return (__m64)__builtin_ia32_vec_init_v4hi(__s0, __s1, __s2, __s3);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_set_pi8(char __b7, char __b6, char __b5, char __b4, char __b3, char __b2,
+            char __b1, char __b0)
+{
+    return (__m64)__builtin_ia32_vec_init_v8qi(__b0, __b1, __b2, __b3,
+                                               __b4, __b5, __b6, __b7);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_set1_pi32(int __i)
+{
+    return _mm_set_pi32(__i, __i);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_set1_pi16(short __w)
+{
+    return _mm_set_pi16(__w, __w, __w, __w);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_set1_pi8(char __b)
+{
+    return _mm_set_pi8(__b, __b, __b, __b, __b, __b, __b, __b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_setr_pi32(int __i0, int __i1)
+{
+    return _mm_set_pi32(__i1, __i0);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_setr_pi16(short __w0, short __w1, short __w2, short __w3)
+{
+    return _mm_set_pi16(__w3, __w2, __w1, __w0);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_setr_pi8(char __b0, char __b1, char __b2, char __b3, char __b4, char __b5,
+             char __b6, char __b7)
+{
+    return _mm_set_pi8(__b7, __b6, __b5, __b4, __b3, __b2, __b1, __b0);
+}
+
+
+/* Aliases for compatibility. */
+#define _m_empty _mm_empty
+#define _m_from_int _mm_cvtsi32_si64
+#define _m_to_int _mm_cvtsi64_si32
+#define _m_packsswb _mm_packs_pi16
+#define _m_packssdw _mm_packs_pi32
+#define _m_packuswb _mm_packs_pu16
+#define _m_punpckhbw _mm_unpackhi_pi8
+#define _m_punpckhwd _mm_unpackhi_pi16
+#define _m_punpckhdq _mm_unpackhi_pi32
+#define _m_punpcklbw _mm_unpacklo_pi8
+#define _m_punpcklwd _mm_unpacklo_pi16
+#define _m_punpckldq _mm_unpacklo_pi32
+#define _m_paddb _mm_add_pi8
+#define _m_paddw _mm_add_pi16
+#define _m_paddd _mm_add_pi32
+#define _m_paddsb _mm_adds_pi8
+#define _m_paddsw _mm_adds_pi16
+#define _m_paddusb _mm_adds_pu8
+#define _m_paddusw _mm_adds_pu16
+#define _m_psubb _mm_sub_pi8
+#define _m_psubw _mm_sub_pi16
+#define _m_psubd _mm_sub_pi32
+#define _m_psubsb _mm_subs_pi8
+#define _m_psubsw _mm_subs_pi16
+#define _m_psubusb _mm_subs_pu8
+#define _m_psubusw _mm_subs_pu16
+#define _m_pmaddwd _mm_madd_pi16
+#define _m_pmulhw _mm_mulhi_pi16
+#define _m_pmullw _mm_mullo_pi16
+#define _m_psllw _mm_sll_pi16
+#define _m_psllwi _mm_slli_pi16
+#define _m_pslld _mm_sll_pi32
+#define _m_pslldi _mm_slli_pi32
+#define _m_psllq _mm_sll_si64
+#define _m_psllqi _mm_slli_si64
+#define _m_psraw _mm_sra_pi16
+#define _m_psrawi _mm_srai_pi16
+#define _m_psrad _mm_sra_pi32
+#define _m_psradi _mm_srai_pi32
+#define _m_psrlw _mm_srl_pi16
+#define _m_psrlwi _mm_srli_pi16
+#define _m_psrld _mm_srl_pi32
+#define _m_psrldi _mm_srli_pi32
+#define _m_psrlq _mm_srl_si64
+#define _m_psrlqi _mm_srli_si64
+#define _m_pand _mm_and_si64
+#define _m_pandn _mm_andnot_si64
+#define _m_por _mm_or_si64
+#define _m_pxor _mm_xor_si64
+#define _m_pcmpeqb _mm_cmpeq_pi8
+#define _m_pcmpeqw _mm_cmpeq_pi16
+#define _m_pcmpeqd _mm_cmpeq_pi32
+#define _m_pcmpgtb _mm_cmpgt_pi8
+#define _m_pcmpgtw _mm_cmpgt_pi16
+#define _m_pcmpgtd _mm_cmpgt_pi32
+
+#endif /* __MMX__ */
+
+#endif /* __MMINTRIN_H */
+
diff --git a/lib/clang/3.6/include/module.modulemap b/lib/clang/3.6/include/module.modulemap
new file mode 100644
index 0000000..3c42477
--- /dev/null
+++ b/lib/clang/3.6/include/module.modulemap
@@ -0,0 +1,176 @@
+module _Builtin_intrinsics [system] {
+  explicit module altivec {
+    requires altivec
+    header "altivec.h"
+  }
+
+  explicit module arm {
+    requires arm
+
+    explicit module acle {
+      header "arm_acle.h"
+      export *
+    }
+
+    explicit module neon {
+      requires neon
+      header "arm_neon.h"
+      export *
+    }
+  }
+
+  explicit module intel {
+    requires x86
+    export *
+
+    header "immintrin.h"
+    header "x86intrin.h"
+
+    explicit module mm_malloc {
+      header "mm_malloc.h"
+      export * // note: for <stdlib.h> dependency
+    }
+
+    explicit module cpuid {
+      requires x86
+      header "cpuid.h"
+    }
+
+    explicit module mmx {
+      requires mmx
+      header "mmintrin.h"
+    }
+
+    explicit module f16c {
+      requires f16c
+      header "f16cintrin.h"
+    }
+
+    explicit module sse {
+      requires sse
+      export mmx
+      export * // note: for hackish <emmintrin.h> dependency
+      header "xmmintrin.h"
+    }
+
+    explicit module sse2 {
+      requires sse2
+      export sse
+      header "emmintrin.h"
+    }
+
+    explicit module sse3 {
+      requires sse3
+      export sse2
+      header "pmmintrin.h"
+    }
+
+    explicit module ssse3 {
+      requires ssse3
+      export sse3
+      header "tmmintrin.h"
+    }
+
+    explicit module sse4_1 {
+      requires sse41
+      export ssse3
+      header "smmintrin.h"
+    }
+
+    explicit module sse4_2 {
+      requires sse42
+      export sse4_1
+      header "nmmintrin.h"
+    }
+
+    explicit module sse4a {
+      requires sse4a
+      export sse3
+      header "ammintrin.h"
+    }
+
+    explicit module avx {
+      requires avx
+      export sse4_2
+      header "avxintrin.h"
+    }
+
+    explicit module avx2 {
+      requires avx2
+      export avx
+      header "avx2intrin.h"
+    }
+
+    explicit module avx512f {
+      requires avx512f
+      export avx2
+      header "avx512fintrin.h"
+    }
+
+    explicit module avx512er {
+      requires avx512er
+      header "avx512erintrin.h"
+    }
+
+    explicit module bmi {
+      requires bmi
+      header "bmiintrin.h"
+    }
+
+    explicit module bmi2 {
+      requires bmi2
+      header "bmi2intrin.h"
+    }
+
+    explicit module fma {
+      requires fma
+      header "fmaintrin.h"
+    }
+
+    explicit module fma4 {
+      requires fma4
+      export sse3
+      header "fma4intrin.h"
+    }
+
+    explicit module lzcnt {
+      requires lzcnt
+      header "lzcntintrin.h"
+    }
+
+    explicit module popcnt {
+      requires popcnt
+      header "popcntintrin.h"
+    }
+
+    explicit module mm3dnow {
+      requires mm3dnow
+      header "mm3dnow.h"
+    }
+
+    explicit module xop {
+      requires xop
+      export fma4
+      header "xopintrin.h"
+    }
+
+    explicit module aes_pclmul {
+      requires aes, pclmul
+      header "wmmintrin.h"
+    }
+
+    explicit module aes {
+      requires aes
+      header "__wmmintrin_aes.h"
+    }
+
+    explicit module pclmul {
+      requires pclmul
+      header "__wmmintrin_pclmul.h"
+    }
+  }
+}
+
+module _Builtin_stddef_max_align_t [system] [extern_c] {
+  header "__stddef_max_align_t.h"
+}
diff --git a/lib/clang/3.6/include/nmmintrin.h b/lib/clang/3.6/include/nmmintrin.h
new file mode 100644
index 0000000..f12622d
--- /dev/null
+++ b/lib/clang/3.6/include/nmmintrin.h
@@ -0,0 +1,35 @@
+/*===---- nmmintrin.h - SSE4 intrinsics ------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef _NMMINTRIN_H
+#define _NMMINTRIN_H
+
+#ifndef __SSE4_2__
+#error "SSE4.2 instruction set not enabled"
+#else
+
+/* To match expectations of gcc we put the sse4.2 definitions into smmintrin.h,
+   just include it now then.  */
+#include <smmintrin.h>
+#endif /* __SSE4_2__ */
+#endif /* _NMMINTRIN_H */
diff --git a/lib/clang/3.6/include/pmmintrin.h b/lib/clang/3.6/include/pmmintrin.h
new file mode 100644
index 0000000..6f1fc32
--- /dev/null
+++ b/lib/clang/3.6/include/pmmintrin.h
@@ -0,0 +1,117 @@
+/*===---- pmmintrin.h - SSE3 intrinsics ------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+ 
+#ifndef __PMMINTRIN_H
+#define __PMMINTRIN_H
+
+#ifndef __SSE3__
+#error "SSE3 instruction set not enabled"
+#else
+
+#include <emmintrin.h>
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_lddqu_si128(__m128i const *__p)
+{
+  return (__m128i)__builtin_ia32_lddqu((char const *)__p);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_addsub_ps(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_addsubps(__a, __b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_hadd_ps(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_haddps(__a, __b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_hsub_ps(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_hsubps(__a, __b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_movehdup_ps(__m128 __a)
+{
+  return __builtin_shufflevector(__a, __a, 1, 1, 3, 3);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_moveldup_ps(__m128 __a)
+{
+  return __builtin_shufflevector(__a, __a, 0, 0, 2, 2);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_addsub_pd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_addsubpd(__a, __b);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_hadd_pd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_haddpd(__a, __b);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_hsub_pd(__m128d __a, __m128d __b)
+{
+  return __builtin_ia32_hsubpd(__a, __b);
+}
+
+#define        _mm_loaddup_pd(dp)        _mm_load1_pd(dp)
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_movedup_pd(__m128d __a)
+{
+  return __builtin_shufflevector(__a, __a, 0, 0);
+}
+
+#define _MM_DENORMALS_ZERO_ON   (0x0040)
+#define _MM_DENORMALS_ZERO_OFF  (0x0000)
+
+#define _MM_DENORMALS_ZERO_MASK (0x0040)
+
+#define _MM_GET_DENORMALS_ZERO_MODE() (_mm_getcsr() & _MM_DENORMALS_ZERO_MASK)
+#define _MM_SET_DENORMALS_ZERO_MODE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_DENORMALS_ZERO_MASK) | (x)))
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_monitor(void const *__p, unsigned __extensions, unsigned __hints)
+{
+  __builtin_ia32_monitor((void *)__p, __extensions, __hints);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_mwait(unsigned __extensions, unsigned __hints)
+{
+  __builtin_ia32_mwait(__extensions, __hints);
+}
+
+#endif /* __SSE3__ */
+
+#endif /* __PMMINTRIN_H */
diff --git a/lib/clang/3.6/include/popcntintrin.h b/lib/clang/3.6/include/popcntintrin.h
new file mode 100644
index 0000000..d439daa
--- /dev/null
+++ b/lib/clang/3.6/include/popcntintrin.h
@@ -0,0 +1,45 @@
+/*===---- popcntintrin.h - POPCNT intrinsics -------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __POPCNT__
+#error "POPCNT instruction set not enabled"
+#endif
+
+#ifndef _POPCNTINTRIN_H
+#define _POPCNTINTRIN_H
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_popcnt_u32(unsigned int __A)
+{
+  return __builtin_popcount(__A);
+}
+
+#ifdef __x86_64__
+static __inline__ long long __attribute__((__always_inline__, __nodebug__))
+_mm_popcnt_u64(unsigned long long __A)
+{
+  return __builtin_popcountll(__A);
+}
+#endif /* __x86_64__ */
+
+#endif /* _POPCNTINTRIN_H */
diff --git a/lib/clang/3.6/include/prfchwintrin.h b/lib/clang/3.6/include/prfchwintrin.h
new file mode 100644
index 0000000..9825bd8
--- /dev/null
+++ b/lib/clang/3.6/include/prfchwintrin.h
@@ -0,0 +1,39 @@
+/*===---- prfchwintrin.h - PREFETCHW intrinsic -----------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#if !defined(__X86INTRIN_H) && !defined(_MM3DNOW_H_INCLUDED)
+#error "Never use <prfchwintrin.h> directly; include <x86intrin.h> or <mm3dnow.h> instead."
+#endif
+
+#ifndef __PRFCHWINTRIN_H
+#define __PRFCHWINTRIN_H
+
+#if defined(__PRFCHW__) || defined(__3dNOW__)
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_m_prefetchw(void *__P)
+{
+  __builtin_prefetch (__P, 1, 3 /* _MM_HINT_T0 */);
+}
+#endif
+
+#endif /* __PRFCHWINTRIN_H */
diff --git a/lib/clang/3.6/include/rdseedintrin.h b/lib/clang/3.6/include/rdseedintrin.h
new file mode 100644
index 0000000..0fef1fa
--- /dev/null
+++ b/lib/clang/3.6/include/rdseedintrin.h
@@ -0,0 +1,52 @@
+/*===---- rdseedintrin.h - RDSEED intrinsics -------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __X86INTRIN_H
+#error "Never use <rdseedintrin.h> directly; include <x86intrin.h> instead."
+#endif
+
+#ifndef __RDSEEDINTRIN_H
+#define __RDSEEDINTRIN_H
+
+#ifdef __RDSEED__
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_rdseed16_step(unsigned short *__p)
+{
+  return __builtin_ia32_rdseed16_step(__p);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_rdseed32_step(unsigned int *__p)
+{
+  return __builtin_ia32_rdseed32_step(__p);
+}
+
+#ifdef __x86_64__
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_rdseed64_step(unsigned long long *__p)
+{
+  return __builtin_ia32_rdseed64_step(__p);
+}
+#endif
+#endif /* __RDSEED__ */
+#endif /* __RDSEEDINTRIN_H */
diff --git a/lib/clang/3.6/include/rtmintrin.h b/lib/clang/3.6/include/rtmintrin.h
new file mode 100644
index 0000000..26149ca
--- /dev/null
+++ b/lib/clang/3.6/include/rtmintrin.h
@@ -0,0 +1,54 @@
+/*===---- rtmintrin.h - RTM intrinsics -------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __IMMINTRIN_H
+#error "Never use <rtmintrin.h> directly; include <immintrin.h> instead."
+#endif
+
+#ifndef __RTMINTRIN_H
+#define __RTMINTRIN_H
+
+#define _XBEGIN_STARTED   (~0u)
+#define _XABORT_EXPLICIT  (1 << 0)
+#define _XABORT_RETRY     (1 << 1)
+#define _XABORT_CONFLICT  (1 << 2)
+#define _XABORT_CAPACITY  (1 << 3)
+#define _XABORT_DEBUG     (1 << 4)
+#define _XABORT_NESTED    (1 << 5)
+#define _XABORT_CODE(x)   (((x) >> 24) & 0xFF)
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_xbegin(void)
+{
+  return __builtin_ia32_xbegin();
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_xend(void)
+{
+  __builtin_ia32_xend();
+}
+
+#define _xabort(imm) __builtin_ia32_xabort((imm))
+
+#endif /* __RTMINTRIN_H */
diff --git a/lib/clang/3.6/include/shaintrin.h b/lib/clang/3.6/include/shaintrin.h
new file mode 100644
index 0000000..66ed055
--- /dev/null
+++ b/lib/clang/3.6/include/shaintrin.h
@@ -0,0 +1,74 @@
+/*===---- shaintrin.h - SHA intrinsics -------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __IMMINTRIN_H
+#error "Never use <shaintrin.h> directly; include <immintrin.h> instead."
+#endif
+
+#ifndef __SHAINTRIN_H
+#define __SHAINTRIN_H
+
+#if !defined (__SHA__)
+#  error "SHA instructions not enabled"
+#endif
+
+#define _mm_sha1rnds4_epu32(V1, V2, M) __extension__ ({ \
+  __builtin_ia32_sha1rnds4((V1), (V2), (M)); })
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sha1nexte_epu32(__m128i __X, __m128i __Y)
+{
+  return __builtin_ia32_sha1nexte(__X, __Y);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sha1msg1_epu32(__m128i __X, __m128i __Y)
+{
+  return __builtin_ia32_sha1msg1(__X, __Y);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sha1msg2_epu32(__m128i __X, __m128i __Y)
+{
+  return __builtin_ia32_sha1msg2(__X, __Y);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sha256rnds2_epu32(__m128i __X, __m128i __Y, __m128i __Z)
+{
+  return __builtin_ia32_sha256rnds2(__X, __Y, __Z);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sha256msg1_epu32(__m128i __X, __m128i __Y)
+{
+  return __builtin_ia32_sha256msg1(__X, __Y);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sha256msg2_epu32(__m128i __X, __m128i __Y)
+{
+  return __builtin_ia32_sha256msg2(__X, __Y);
+}
+
+#endif /* __SHAINTRIN_H */
diff --git a/lib/clang/3.6/include/smmintrin.h b/lib/clang/3.6/include/smmintrin.h
new file mode 100644
index 0000000..6e35734
--- /dev/null
+++ b/lib/clang/3.6/include/smmintrin.h
@@ -0,0 +1,482 @@
+/*===---- smmintrin.h - SSE4 intrinsics ------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef _SMMINTRIN_H
+#define _SMMINTRIN_H
+
+#ifndef __SSE4_1__
+#error "SSE4.1 instruction set not enabled"
+#else
+
+#include <tmmintrin.h>
+
+/* SSE4 Rounding macros. */
+#define _MM_FROUND_TO_NEAREST_INT    0x00
+#define _MM_FROUND_TO_NEG_INF        0x01
+#define _MM_FROUND_TO_POS_INF        0x02
+#define _MM_FROUND_TO_ZERO           0x03
+#define _MM_FROUND_CUR_DIRECTION     0x04
+
+#define _MM_FROUND_RAISE_EXC         0x00
+#define _MM_FROUND_NO_EXC            0x08
+
+#define _MM_FROUND_NINT      (_MM_FROUND_RAISE_EXC | _MM_FROUND_TO_NEAREST_INT)
+#define _MM_FROUND_FLOOR     (_MM_FROUND_RAISE_EXC | _MM_FROUND_TO_NEG_INF)
+#define _MM_FROUND_CEIL      (_MM_FROUND_RAISE_EXC | _MM_FROUND_TO_POS_INF)
+#define _MM_FROUND_TRUNC     (_MM_FROUND_RAISE_EXC | _MM_FROUND_TO_ZERO)
+#define _MM_FROUND_RINT      (_MM_FROUND_RAISE_EXC | _MM_FROUND_CUR_DIRECTION)
+#define _MM_FROUND_NEARBYINT (_MM_FROUND_NO_EXC | _MM_FROUND_CUR_DIRECTION)
+
+#define _mm_ceil_ps(X)       _mm_round_ps((X), _MM_FROUND_CEIL)
+#define _mm_ceil_pd(X)       _mm_round_pd((X), _MM_FROUND_CEIL)
+#define _mm_ceil_ss(X, Y)    _mm_round_ss((X), (Y), _MM_FROUND_CEIL)
+#define _mm_ceil_sd(X, Y)    _mm_round_sd((X), (Y), _MM_FROUND_CEIL)
+
+#define _mm_floor_ps(X)      _mm_round_ps((X), _MM_FROUND_FLOOR)
+#define _mm_floor_pd(X)      _mm_round_pd((X), _MM_FROUND_FLOOR)
+#define _mm_floor_ss(X, Y)   _mm_round_ss((X), (Y), _MM_FROUND_FLOOR)
+#define _mm_floor_sd(X, Y)   _mm_round_sd((X), (Y), _MM_FROUND_FLOOR)
+
+#define _mm_round_ps(X, M) __extension__ ({ \
+  __m128 __X = (X); \
+  (__m128) __builtin_ia32_roundps((__v4sf)__X, (M)); })
+
+#define _mm_round_ss(X, Y, M) __extension__ ({ \
+  __m128 __X = (X); \
+  __m128 __Y = (Y); \
+  (__m128) __builtin_ia32_roundss((__v4sf)__X, (__v4sf)__Y, (M)); })
+
+#define _mm_round_pd(X, M) __extension__ ({ \
+  __m128d __X = (X); \
+  (__m128d) __builtin_ia32_roundpd((__v2df)__X, (M)); })
+
+#define _mm_round_sd(X, Y, M) __extension__ ({ \
+  __m128d __X = (X); \
+  __m128d __Y = (Y); \
+  (__m128d) __builtin_ia32_roundsd((__v2df)__X, (__v2df)__Y, (M)); })
+
+/* SSE4 Packed Blending Intrinsics.  */
+#define _mm_blend_pd(V1, V2, M) __extension__ ({ \
+  __m128d __V1 = (V1); \
+  __m128d __V2 = (V2); \
+  (__m128d)__builtin_shufflevector((__v2df)__V1, (__v2df)__V2, \
+                                   (((M) & 0x01) ? 2 : 0), \
+                                   (((M) & 0x02) ? 3 : 1)); })
+
+#define _mm_blend_ps(V1, V2, M) __extension__ ({ \
+  __m128 __V1 = (V1); \
+  __m128 __V2 = (V2); \
+  (__m128)__builtin_shufflevector((__v4sf)__V1, (__v4sf)__V2, \
+                                  (((M) & 0x01) ? 4 : 0), \
+                                  (((M) & 0x02) ? 5 : 1), \
+                                  (((M) & 0x04) ? 6 : 2), \
+                                  (((M) & 0x08) ? 7 : 3)); })
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_blendv_pd (__m128d __V1, __m128d __V2, __m128d __M)
+{
+  return (__m128d) __builtin_ia32_blendvpd ((__v2df)__V1, (__v2df)__V2,
+                                            (__v2df)__M);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_blendv_ps (__m128 __V1, __m128 __V2, __m128 __M)
+{
+  return (__m128) __builtin_ia32_blendvps ((__v4sf)__V1, (__v4sf)__V2,
+                                           (__v4sf)__M);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_blendv_epi8 (__m128i __V1, __m128i __V2, __m128i __M)
+{
+  return (__m128i) __builtin_ia32_pblendvb128 ((__v16qi)__V1, (__v16qi)__V2,
+                                               (__v16qi)__M);
+}
+
+#define _mm_blend_epi16(V1, V2, M) __extension__ ({ \
+  __m128i __V1 = (V1); \
+  __m128i __V2 = (V2); \
+  (__m128i)__builtin_shufflevector((__v8hi)__V1, (__v8hi)__V2, \
+                                   (((M) & 0x01) ?  8 : 0), \
+                                   (((M) & 0x02) ?  9 : 1), \
+                                   (((M) & 0x04) ? 10 : 2), \
+                                   (((M) & 0x08) ? 11 : 3), \
+                                   (((M) & 0x10) ? 12 : 4), \
+                                   (((M) & 0x20) ? 13 : 5), \
+                                   (((M) & 0x40) ? 14 : 6), \
+                                   (((M) & 0x80) ? 15 : 7)); })
+
+/* SSE4 Dword Multiply Instructions.  */
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_mullo_epi32 (__m128i __V1, __m128i __V2)
+{
+  return (__m128i) ((__v4si)__V1 * (__v4si)__V2);
+}
+
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_mul_epi32 (__m128i __V1, __m128i __V2)
+{
+  return (__m128i) __builtin_ia32_pmuldq128 ((__v4si)__V1, (__v4si)__V2);
+}
+
+/* SSE4 Floating Point Dot Product Instructions.  */
+#define _mm_dp_ps(X, Y, M) __extension__ ({ \
+  __m128 __X = (X); \
+  __m128 __Y = (Y); \
+  (__m128) __builtin_ia32_dpps((__v4sf)__X, (__v4sf)__Y, (M)); })
+
+#define _mm_dp_pd(X, Y, M) __extension__ ({\
+  __m128d __X = (X); \
+  __m128d __Y = (Y); \
+  (__m128d) __builtin_ia32_dppd((__v2df)__X, (__v2df)__Y, (M)); })
+
+/* SSE4 Streaming Load Hint Instruction.  */
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_stream_load_si128 (__m128i *__V)
+{
+  return (__m128i) __builtin_ia32_movntdqa ((__v2di *) __V);
+}
+
+/* SSE4 Packed Integer Min/Max Instructions.  */
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_min_epi8 (__m128i __V1, __m128i __V2)
+{
+  return (__m128i) __builtin_ia32_pminsb128 ((__v16qi) __V1, (__v16qi) __V2);
+}
+
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_max_epi8 (__m128i __V1, __m128i __V2)
+{
+  return (__m128i) __builtin_ia32_pmaxsb128 ((__v16qi) __V1, (__v16qi) __V2);
+}
+
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_min_epu16 (__m128i __V1, __m128i __V2)
+{
+  return (__m128i) __builtin_ia32_pminuw128 ((__v8hi) __V1, (__v8hi) __V2);
+}
+
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_max_epu16 (__m128i __V1, __m128i __V2)
+{
+  return (__m128i) __builtin_ia32_pmaxuw128 ((__v8hi) __V1, (__v8hi) __V2);
+}
+
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_min_epi32 (__m128i __V1, __m128i __V2)
+{
+  return (__m128i) __builtin_ia32_pminsd128 ((__v4si) __V1, (__v4si) __V2);
+}
+
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_max_epi32 (__m128i __V1, __m128i __V2)
+{
+  return (__m128i) __builtin_ia32_pmaxsd128 ((__v4si) __V1, (__v4si) __V2);
+}
+
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_min_epu32 (__m128i __V1, __m128i __V2)
+{
+  return (__m128i) __builtin_ia32_pminud128((__v4si) __V1, (__v4si) __V2);
+}
+
+static __inline__  __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_max_epu32 (__m128i __V1, __m128i __V2)
+{
+  return (__m128i) __builtin_ia32_pmaxud128((__v4si) __V1, (__v4si) __V2);
+}
+
+/* SSE4 Insertion and Extraction from XMM Register Instructions.  */
+#define _mm_insert_ps(X, Y, N) __builtin_ia32_insertps128((X), (Y), (N))
+#define _mm_extract_ps(X, N) (__extension__                      \
+                              ({ union { int __i; float __f; } __t;  \
+                                 __v4sf __a = (__v4sf)(X);       \
+                                 __t.__f = __a[(N) & 3];                 \
+                                 __t.__i;}))
+
+/* Miscellaneous insert and extract macros.  */
+/* Extract a single-precision float from X at index N into D.  */
+#define _MM_EXTRACT_FLOAT(D, X, N) (__extension__ ({ __v4sf __a = (__v4sf)(X); \
+                                                    (D) = __a[N]; }))
+                                                    
+/* Or together 2 sets of indexes (X and Y) with the zeroing bits (Z) to create
+   an index suitable for _mm_insert_ps.  */
+#define _MM_MK_INSERTPS_NDX(X, Y, Z) (((X) << 6) | ((Y) << 4) | (Z))
+                                           
+/* Extract a float from X at index N into the first index of the return.  */
+#define _MM_PICK_OUT_PS(X, N) _mm_insert_ps (_mm_setzero_ps(), (X),   \
+                                             _MM_MK_INSERTPS_NDX((N), 0, 0x0e))
+                                             
+/* Insert int into packed integer array at index.  */
+#define _mm_insert_epi8(X, I, N) (__extension__ ({ __v16qi __a = (__v16qi)(X); \
+                                                   __a[(N) & 15] = (I);             \
+                                                   __a;}))
+#define _mm_insert_epi32(X, I, N) (__extension__ ({ __v4si __a = (__v4si)(X); \
+                                                    __a[(N) & 3] = (I);           \
+                                                    __a;}))
+#ifdef __x86_64__
+#define _mm_insert_epi64(X, I, N) (__extension__ ({ __v2di __a = (__v2di)(X); \
+                                                    __a[(N) & 1] = (I);           \
+                                                    __a;}))
+#endif /* __x86_64__ */
+
+/* Extract int from packed integer array at index.  This returns the element
+ * as a zero extended value, so it is unsigned.
+ */
+#define _mm_extract_epi8(X, N) (__extension__ ({ __v16qi __a = (__v16qi)(X); \
+                                                 (int)(unsigned char) \
+                                                     __a[(N) & 15];}))
+#define _mm_extract_epi32(X, N) (__extension__ ({ __v4si __a = (__v4si)(X); \
+                                                  __a[(N) & 3];}))
+#ifdef __x86_64__
+#define _mm_extract_epi64(X, N) (__extension__ ({ __v2di __a = (__v2di)(X); \
+                                                  __a[(N) & 1];}))
+#endif /* __x86_64 */
+
+/* SSE4 128-bit Packed Integer Comparisons.  */
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_testz_si128(__m128i __M, __m128i __V)
+{
+  return __builtin_ia32_ptestz128((__v2di)__M, (__v2di)__V);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_testc_si128(__m128i __M, __m128i __V)
+{
+  return __builtin_ia32_ptestc128((__v2di)__M, (__v2di)__V);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_testnzc_si128(__m128i __M, __m128i __V)
+{
+  return __builtin_ia32_ptestnzc128((__v2di)__M, (__v2di)__V);
+}
+
+#define _mm_test_all_ones(V) _mm_testc_si128((V), _mm_cmpeq_epi32((V), (V)))
+#define _mm_test_mix_ones_zeros(M, V) _mm_testnzc_si128((M), (V))
+#define _mm_test_all_zeros(M, V) _mm_testz_si128 ((M), (V))
+
+/* SSE4 64-bit Packed Integer Comparisons.  */
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_epi64(__m128i __V1, __m128i __V2)
+{
+  return (__m128i)((__v2di)__V1 == (__v2di)__V2);
+}
+
+/* SSE4 Packed Integer Sign-Extension.  */
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepi8_epi16(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovsxbw128((__v16qi) __V);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepi8_epi32(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovsxbd128((__v16qi) __V);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepi8_epi64(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovsxbq128((__v16qi) __V);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepi16_epi32(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovsxwd128((__v8hi) __V); 
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepi16_epi64(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovsxwq128((__v8hi)__V);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepi32_epi64(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovsxdq128((__v4si)__V);
+}
+
+/* SSE4 Packed Integer Zero-Extension.  */
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepu8_epi16(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovzxbw128((__v16qi) __V);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepu8_epi32(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovzxbd128((__v16qi)__V);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepu8_epi64(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovzxbq128((__v16qi)__V);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepu16_epi32(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovzxwd128((__v8hi)__V);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepu16_epi64(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovzxwq128((__v8hi)__V);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cvtepu32_epi64(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_pmovzxdq128((__v4si)__V);
+}
+
+/* SSE4 Pack with Unsigned Saturation.  */
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_packus_epi32(__m128i __V1, __m128i __V2)
+{
+  return (__m128i) __builtin_ia32_packusdw128((__v4si)__V1, (__v4si)__V2);
+}
+
+/* SSE4 Multiple Packed Sums of Absolute Difference.  */
+#define _mm_mpsadbw_epu8(X, Y, M) __extension__ ({ \
+  __m128i __X = (X); \
+  __m128i __Y = (Y); \
+  (__m128i) __builtin_ia32_mpsadbw128((__v16qi)__X, (__v16qi)__Y, (M)); })
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_minpos_epu16(__m128i __V)
+{
+  return (__m128i) __builtin_ia32_phminposuw128((__v8hi)__V);
+}
+
+/* These definitions are normally in nmmintrin.h, but gcc puts them in here
+   so we'll do the same.  */
+#ifdef __SSE4_2__
+
+/* These specify the type of data that we're comparing.  */
+#define _SIDD_UBYTE_OPS                 0x00
+#define _SIDD_UWORD_OPS                 0x01
+#define _SIDD_SBYTE_OPS                 0x02
+#define _SIDD_SWORD_OPS                 0x03
+
+/* These specify the type of comparison operation.  */
+#define _SIDD_CMP_EQUAL_ANY             0x00
+#define _SIDD_CMP_RANGES                0x04
+#define _SIDD_CMP_EQUAL_EACH            0x08
+#define _SIDD_CMP_EQUAL_ORDERED         0x0c
+
+/* These macros specify the polarity of the operation.  */
+#define _SIDD_POSITIVE_POLARITY         0x00
+#define _SIDD_NEGATIVE_POLARITY         0x10
+#define _SIDD_MASKED_POSITIVE_POLARITY  0x20
+#define _SIDD_MASKED_NEGATIVE_POLARITY  0x30
+
+/* These macros are used in _mm_cmpXstri() to specify the return.  */
+#define _SIDD_LEAST_SIGNIFICANT         0x00
+#define _SIDD_MOST_SIGNIFICANT          0x40
+
+/* These macros are used in _mm_cmpXstri() to specify the return.  */
+#define _SIDD_BIT_MASK                  0x00
+#define _SIDD_UNIT_MASK                 0x40
+
+/* SSE4.2 Packed Comparison Intrinsics.  */
+#define _mm_cmpistrm(A, B, M) __builtin_ia32_pcmpistrm128((A), (B), (M))
+#define _mm_cmpistri(A, B, M) __builtin_ia32_pcmpistri128((A), (B), (M))
+
+#define _mm_cmpestrm(A, LA, B, LB, M) \
+     __builtin_ia32_pcmpestrm128((A), (LA), (B), (LB), (M))
+#define _mm_cmpestri(A, LA, B, LB, M) \
+     __builtin_ia32_pcmpestri128((A), (LA), (B), (LB), (M))
+     
+/* SSE4.2 Packed Comparison Intrinsics and EFlag Reading.  */
+#define _mm_cmpistra(A, B, M) \
+     __builtin_ia32_pcmpistria128((A), (B), (M))
+#define _mm_cmpistrc(A, B, M) \
+     __builtin_ia32_pcmpistric128((A), (B), (M))
+#define _mm_cmpistro(A, B, M) \
+     __builtin_ia32_pcmpistrio128((A), (B), (M))
+#define _mm_cmpistrs(A, B, M) \
+     __builtin_ia32_pcmpistris128((A), (B), (M))
+#define _mm_cmpistrz(A, B, M) \
+     __builtin_ia32_pcmpistriz128((A), (B), (M))
+
+#define _mm_cmpestra(A, LA, B, LB, M) \
+     __builtin_ia32_pcmpestria128((A), (LA), (B), (LB), (M))
+#define _mm_cmpestrc(A, LA, B, LB, M) \
+     __builtin_ia32_pcmpestric128((A), (LA), (B), (LB), (M))
+#define _mm_cmpestro(A, LA, B, LB, M) \
+     __builtin_ia32_pcmpestrio128((A), (LA), (B), (LB), (M))
+#define _mm_cmpestrs(A, LA, B, LB, M) \
+     __builtin_ia32_pcmpestris128((A), (LA), (B), (LB), (M))
+#define _mm_cmpestrz(A, LA, B, LB, M) \
+     __builtin_ia32_pcmpestriz128((A), (LA), (B), (LB), (M))
+
+/* SSE4.2 Compare Packed Data -- Greater Than.  */
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_epi64(__m128i __V1, __m128i __V2)
+{
+  return (__m128i)((__v2di)__V1 > (__v2di)__V2);
+}
+
+/* SSE4.2 Accumulate CRC32.  */
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_mm_crc32_u8(unsigned int __C, unsigned char __D)
+{
+  return __builtin_ia32_crc32qi(__C, __D);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_mm_crc32_u16(unsigned int __C, unsigned short __D)
+{
+  return __builtin_ia32_crc32hi(__C, __D);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_mm_crc32_u32(unsigned int __C, unsigned int __D)
+{
+  return __builtin_ia32_crc32si(__C, __D);
+}
+
+#ifdef __x86_64__
+static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
+_mm_crc32_u64(unsigned long long __C, unsigned long long __D)
+{
+  return __builtin_ia32_crc32di(__C, __D);
+}
+#endif /* __x86_64__ */
+
+#ifdef __POPCNT__
+#include <popcntintrin.h>
+#endif
+
+#endif /* __SSE4_2__ */
+#endif /* __SSE4_1__ */
+
+#endif /* _SMMINTRIN_H */
diff --git a/lib/clang/3.6/include/stdalign.h b/lib/clang/3.6/include/stdalign.h
new file mode 100644
index 0000000..3738d12
--- /dev/null
+++ b/lib/clang/3.6/include/stdalign.h
@@ -0,0 +1,35 @@
+/*===---- stdalign.h - Standard header for alignment ------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __STDALIGN_H
+#define __STDALIGN_H
+
+#ifndef __cplusplus
+#define alignas _Alignas
+#define alignof _Alignof
+#endif
+
+#define __alignas_is_defined 1
+#define __alignof_is_defined 1
+
+#endif /* __STDALIGN_H */
diff --git a/lib/clang/3.6/include/stdarg.h b/lib/clang/3.6/include/stdarg.h
new file mode 100644
index 0000000..a57e183
--- /dev/null
+++ b/lib/clang/3.6/include/stdarg.h
@@ -0,0 +1,52 @@
+/*===---- stdarg.h - Variable argument handling ----------------------------===
+ *
+ * Copyright (c) 2008 Eli Friedman
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __STDARG_H
+#define __STDARG_H
+
+#ifndef _VA_LIST
+typedef __builtin_va_list va_list;
+#define _VA_LIST
+#endif
+#define va_start(ap, param) __builtin_va_start(ap, param)
+#define va_end(ap)          __builtin_va_end(ap)
+#define va_arg(ap, type)    __builtin_va_arg(ap, type)
+
+/* GCC always defines __va_copy, but does not define va_copy unless in c99 mode
+ * or -ansi is not specified, since it was not part of C90.
+ */
+#define __va_copy(d,s) __builtin_va_copy(d,s)
+
+#if __STDC_VERSION__ >= 199901L || __cplusplus >= 201103L || !defined(__STRICT_ANSI__)
+#define va_copy(dest, src)  __builtin_va_copy(dest, src)
+#endif
+
+/* Hack required to make standard headers work, at least on Ubuntu */
+#ifndef __GNUC_VA_LIST
+#define __GNUC_VA_LIST 1
+#endif
+typedef __builtin_va_list __gnuc_va_list;
+
+#endif /* __STDARG_H */
diff --git a/lib/clang/3.6/include/stdatomic.h b/lib/clang/3.6/include/stdatomic.h
new file mode 100644
index 0000000..2c4f1ce
--- /dev/null
+++ b/lib/clang/3.6/include/stdatomic.h
@@ -0,0 +1,582 @@
+/*-
+ * Copyright (c) 2011 Ed Schouten <ed@FreeBSD.org>
+ *                    David Chisnall <theraven@FreeBSD.org>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ * $FreeBSD$
+ */
+
+#ifndef _STDATOMIC_H_
+#define	_STDATOMIC_H_
+
+#include <sys/cdefs.h>
+
+#if defined(__GNUC__) && !defined(__GNUC_PREREQ)
+/* Duplicate definition here, since the mingw sys/cdefs.h omits the  */
+/* definition, and this needs to be usable there.                    */
+#define	__GNUC_PREREQ(x, y)    \
+	((__GNUC__ == (x) && __GNUC_MINOR__ >= (y)) || (__GNUC__ > (x)))
+#endif /* __GNUC__ && ... */
+
+#if defined(__cplusplus) && __cplusplus >= 201103L && defined(_USING_LIBCXX)
+# ifdef __clang__
+#  if __has_feature(cxx_atomic)
+#   define _STDATOMIC_HAVE_ATOMIC
+#  endif
+# else /* gcc */
+#  if __GNUC_PREREQ(4, 7)
+#   define _STDATOMIC_HAVE_ATOMIC
+#  endif
+# endif
+#endif
+
+#ifdef _STDATOMIC_HAVE_ATOMIC
+
+/* We have a usable C++ <atomic>; use it instead.  */
+
+#include <atomic>
+
+#undef _Atomic
+        /* Also defined by <atomic> for gcc.  But not used in macros. */
+        /* Also a clang intrinsic.                                    */
+        /* Should not be used by client code before this file is      */
+        /* included.  The definitions in <atomic> themselves see      */
+        /* the old definition, as they should.                        */
+        /* Client code sees the following definition.                 */
+
+#define _Atomic(t) std::atomic<t>
+
+using std::atomic_is_lock_free;
+using std::atomic_init;
+using std::atomic_store;
+using std::atomic_store_explicit;
+using std::atomic_load;
+using std::atomic_load_explicit;
+using std::atomic_exchange;
+using std::atomic_exchange_explicit;
+using std::atomic_compare_exchange_strong;
+using std::atomic_compare_exchange_strong_explicit;
+using std::atomic_compare_exchange_weak;
+using std::atomic_compare_exchange_weak_explicit;
+using std::atomic_fetch_add;
+using std::atomic_fetch_add_explicit;
+using std::atomic_fetch_sub;
+using std::atomic_fetch_sub_explicit;
+using std::atomic_fetch_or;
+using std::atomic_fetch_or_explicit;
+using std::atomic_fetch_xor;
+using std::atomic_fetch_xor_explicit;
+using std::atomic_fetch_and;
+using std::atomic_fetch_and_explicit;
+using std::atomic_thread_fence;
+using std::atomic_signal_fence;
+
+using std::memory_order;
+using std::memory_order_relaxed;
+using std::memory_order_consume;
+using std::memory_order_acquire;
+using std::memory_order_release;
+using std::memory_order_acq_rel;
+using std::memory_order_seq_cst;
+
+using std::atomic_bool;
+using std::atomic_char;
+using std::atomic_schar;
+using std::atomic_uchar;
+using std::atomic_short;
+using std::atomic_ushort;
+using std::atomic_int;
+using std::atomic_uint;
+using std::atomic_long;
+using std::atomic_ulong;
+using std::atomic_llong;
+using std::atomic_ullong;
+using std::atomic_char16_t;
+using std::atomic_char32_t;
+using std::atomic_wchar_t;
+using std::atomic_int_least8_t;
+using std::atomic_uint_least8_t;
+using std::atomic_int_least16_t;
+using std::atomic_uint_least16_t;
+using std::atomic_int_least32_t;
+using std::atomic_uint_least32_t;
+using std::atomic_int_least64_t;
+using std::atomic_uint_least64_t;
+using std::atomic_int_fast8_t;
+using std::atomic_uint_fast8_t;
+using std::atomic_int_fast16_t;
+using std::atomic_uint_fast16_t;
+using std::atomic_int_fast32_t;
+using std::atomic_uint_fast32_t;
+using std::atomic_int_fast64_t;
+using std::atomic_uint_fast64_t;
+using std::atomic_intptr_t;
+using std::atomic_uintptr_t;
+using std::atomic_size_t;
+using std::atomic_ptrdiff_t;
+using std::atomic_intmax_t;
+using std::atomic_uintmax_t;
+
+#else /* <atomic> unavailable, possibly because this is C, not C++ */
+
+#include <sys/types.h>
+#include <stdbool.h>
+
+/*
+ * C: Do it ourselves.
+ * Note that the runtime representation defined here should be compatible
+ * with the C++ one, i.e. an _Atomic(T) needs to contain the same
+ * bits as a T.
+ */
+
+#include <stddef.h>  /* For ptrdiff_t.                          */
+#include <stdint.h>  /* TODO: Should pollute namespace less.    */
+#if __STDC_VERSION__ >= 201112L
+# include <uchar.h>  /* For char16_t and char32_t.              */
+#endif
+
+#ifdef __clang__
+# if __has_extension(c_atomic) || __has_extension(cxx_atomic)
+#  define       __CLANG_ATOMICS
+# else
+#  error "stdatomic.h does not support your compiler"
+# endif
+# if __has_builtin(__sync_swap)
+#  define __HAS_BUILTIN_SYNC_SWAP
+# endif
+#else
+# if __GNUC_PREREQ(4, 7)
+#  define	__GNUC_ATOMICS
+# else
+#  define	__SYNC_ATOMICS
+#  ifdef __cplusplus
+#   define       __ATOMICS_AVOID_DOT_INIT
+#  endif
+# endif
+#endif
+
+/*
+ * 7.17.1 Atomic lock-free macros.
+ */
+
+#ifdef __GCC_ATOMIC_BOOL_LOCK_FREE
+#define	ATOMIC_BOOL_LOCK_FREE		__GCC_ATOMIC_BOOL_LOCK_FREE
+#elif defined(__SYNC_ATOMICS)
+#define	ATOMIC_BOOL_LOCK_FREE           2 /* For all modern platforms */
+#endif
+#ifdef __GCC_ATOMIC_CHAR_LOCK_FREE
+#define	ATOMIC_CHAR_LOCK_FREE		__GCC_ATOMIC_CHAR_LOCK_FREE
+#elif defined(__SYNC_ATOMICS)
+#define	ATOMIC_CHAR_LOCK_FREE           2
+#endif
+#ifdef __GCC_ATOMIC_CHAR16_T_LOCK_FREE
+#define	ATOMIC_CHAR16_T_LOCK_FREE	__GCC_ATOMIC_CHAR16_T_LOCK_FREE
+#elif defined(__SYNC_ATOMICS)
+#define	ATOMIC_CHAR16_T_LOCK_FREE       2
+#endif
+#ifdef __GCC_ATOMIC_CHAR32_T_LOCK_FREE
+#define	ATOMIC_CHAR32_T_LOCK_FREE	__GCC_ATOMIC_CHAR32_T_LOCK_FREE
+#elif defined(__SYNC_ATOMICS)
+#define	ATOMIC_CHAR32_T_LOCK_FREE       2
+#endif
+#ifdef __GCC_ATOMIC_WCHAR_T_LOCK_FREE
+#define	ATOMIC_WCHAR_T_LOCK_FREE	__GCC_ATOMIC_WCHAR_T_LOCK_FREE
+#elif defined(__SYNC_ATOMICS)
+#define	ATOMIC_WCHAR_T_LOCK_FREE        2
+#endif
+#ifdef __GCC_ATOMIC_SHORT_LOCK_FREE
+#define	ATOMIC_SHORT_LOCK_FREE		__GCC_ATOMIC_SHORT_LOCK_FREE
+#elif defined(__SYNC_ATOMICS)
+#define	ATOMIC_SHORT_LOCK_FREE          2
+#endif
+#ifdef __GCC_ATOMIC_INT_LOCK_FREE
+#define	ATOMIC_INT_LOCK_FREE		__GCC_ATOMIC_INT_LOCK_FREE
+#elif defined(__SYNC_ATOMICS)
+#define	ATOMIC_INT_LOCK_FREE            2
+#endif
+#ifdef __GCC_ATOMIC_LONG_LOCK_FREE
+#define	ATOMIC_LONG_LOCK_FREE		__GCC_ATOMIC_LONG_LOCK_FREE
+#elif defined(__SYNC_ATOMICS)
+#define	ATOMIC_LONG_LOCK_FREE           2
+#endif
+#ifdef __GCC_ATOMIC_LLONG_LOCK_FREE
+#define	ATOMIC_LLONG_LOCK_FREE		__GCC_ATOMIC_LLONG_LOCK_FREE
+#elif defined(__SYNC_ATOMICS)
+#define	ATOMIC_LLONG_LOCK_FREE          1 /* maybe */
+#endif
+#ifdef __GCC_ATOMIC_POINTER_LOCK_FREE
+#define	ATOMIC_POINTER_LOCK_FREE	__GCC_ATOMIC_POINTER_LOCK_FREE
+#elif defined(__SYNC_ATOMICS)
+#define	ATOMIC_POINTER_LOCK_FREE        2
+#endif
+
+/*
+ * 7.17.2 Initialization.
+ */
+
+#if defined(__CLANG_ATOMICS)
+#define	ATOMIC_VAR_INIT(value)		(value)
+#define	atomic_init(obj, value)		__c11_atomic_init(obj, value)
+#else
+#ifdef __ATOMICS_AVOID_DOT_INIT
+#define	ATOMIC_VAR_INIT(value)		{ value }
+#else
+#define	ATOMIC_VAR_INIT(value)		{ .__val = (value) }
+#endif
+#define	atomic_init(obj, value)		((void)((obj)->__val = (value)))
+#endif
+
+/*
+ * Clang and recent GCC both provide predefined macros for the memory
+ * orderings.  If we are using a compiler that doesn't define them, use the
+ * clang values - these will be ignored in the fallback path.
+ */
+
+#ifndef __ATOMIC_RELAXED
+#define __ATOMIC_RELAXED		0
+#endif
+#ifndef __ATOMIC_CONSUME
+#define __ATOMIC_CONSUME		1
+#endif
+#ifndef __ATOMIC_ACQUIRE
+#define __ATOMIC_ACQUIRE		2
+#endif
+#ifndef __ATOMIC_RELEASE
+#define __ATOMIC_RELEASE		3
+#endif
+#ifndef __ATOMIC_ACQ_REL
+#define __ATOMIC_ACQ_REL		4
+#endif
+#ifndef __ATOMIC_SEQ_CST
+#define __ATOMIC_SEQ_CST		5
+#endif
+
+/*
+ * 7.17.3 Order and consistency.
+ *
+ * The memory_order_* constants that denote the barrier behaviour of the
+ * atomic operations.
+ * The enum values must be identical to those used by the
+ * C++ <atomic> header.
+ */
+
+typedef enum {
+	memory_order_relaxed = __ATOMIC_RELAXED,
+	memory_order_consume = __ATOMIC_CONSUME,
+	memory_order_acquire = __ATOMIC_ACQUIRE,
+	memory_order_release = __ATOMIC_RELEASE,
+	memory_order_acq_rel = __ATOMIC_ACQ_REL,
+	memory_order_seq_cst = __ATOMIC_SEQ_CST
+} memory_order;
+
+/*
+ * 7.17.4 Fences.
+ */
+
+static __inline void
+atomic_thread_fence(memory_order __order __attribute__((unused)))
+{
+
+#ifdef __CLANG_ATOMICS
+	__c11_atomic_thread_fence(__order);
+#elif defined(__GNUC_ATOMICS)
+	__atomic_thread_fence(__order);
+#else
+	__sync_synchronize();
+#endif
+}
+
+static __inline void
+atomic_signal_fence(memory_order __order __attribute__((unused)))
+{
+
+#ifdef __CLANG_ATOMICS
+	__c11_atomic_signal_fence(__order);
+#elif defined(__GNUC_ATOMICS)
+	__atomic_signal_fence(__order);
+#else
+	__asm volatile ("" ::: "memory");
+#endif
+}
+
+/*
+ * 7.17.5 Lock-free property.
+ */
+
+#if defined(_KERNEL)
+/* Atomics in kernelspace are always lock-free. */
+#define	atomic_is_lock_free(obj) \
+	((void)(obj), (_Bool)1)
+#elif defined(__CLANG_ATOMICS)
+#define	atomic_is_lock_free(obj) \
+	__c11_atomic_is_lock_free(sizeof(*(obj)))
+#elif defined(__GNUC_ATOMICS)
+#define	atomic_is_lock_free(obj) \
+	__atomic_is_lock_free(sizeof((obj)->__val), &(obj)->__val)
+#else
+#define	atomic_is_lock_free(obj) \
+	((void)(obj), sizeof((obj)->__val) <= sizeof(void *))
+#endif
+
+/*
+ * 7.17.6 Atomic integer types.
+ */
+
+#ifndef __CLANG_ATOMICS
+/*
+ * No native support for _Atomic(). Place object in structure to prevent
+ * most forms of direct non-atomic access.
+ */
+#define _Atomic(T)              struct { T volatile __val; }
+#endif
+
+typedef _Atomic(bool)			atomic_bool;
+typedef _Atomic(char)			atomic_char;
+typedef _Atomic(signed char)		atomic_schar;
+typedef _Atomic(unsigned char)		atomic_uchar;
+typedef _Atomic(short)			atomic_short;
+typedef _Atomic(unsigned short)		atomic_ushort;
+typedef _Atomic(int)			atomic_int;
+typedef _Atomic(unsigned int)		atomic_uint;
+typedef _Atomic(long)			atomic_long;
+typedef _Atomic(unsigned long)		atomic_ulong;
+typedef _Atomic(long long)		atomic_llong;
+typedef _Atomic(unsigned long long)	atomic_ullong;
+#if __STDC_VERSION__ >= 201112L || __cplusplus >= 201103L
+  typedef _Atomic(char16_t)		atomic_char16_t;
+  typedef _Atomic(char32_t)		atomic_char32_t;
+#endif
+typedef _Atomic(wchar_t)		atomic_wchar_t;
+typedef _Atomic(int_least8_t)		atomic_int_least8_t;
+typedef _Atomic(uint_least8_t)	atomic_uint_least8_t;
+typedef _Atomic(int_least16_t)	atomic_int_least16_t;
+typedef _Atomic(uint_least16_t)	atomic_uint_least16_t;
+typedef _Atomic(int_least32_t)	atomic_int_least32_t;
+typedef _Atomic(uint_least32_t)	atomic_uint_least32_t;
+typedef _Atomic(int_least64_t)	atomic_int_least64_t;
+typedef _Atomic(uint_least64_t)	atomic_uint_least64_t;
+typedef _Atomic(int_fast8_t)		atomic_int_fast8_t;
+typedef _Atomic(uint_fast8_t)		atomic_uint_fast8_t;
+typedef _Atomic(int_fast16_t)		atomic_int_fast16_t;
+typedef _Atomic(uint_fast16_t)	atomic_uint_fast16_t;
+typedef _Atomic(int_fast32_t)		atomic_int_fast32_t;
+typedef _Atomic(uint_fast32_t)	atomic_uint_fast32_t;
+typedef _Atomic(int_fast64_t)		atomic_int_fast64_t;
+typedef _Atomic(uint_fast64_t)	atomic_uint_fast64_t;
+typedef _Atomic(intptr_t)		atomic_intptr_t;
+typedef _Atomic(uintptr_t)		atomic_uintptr_t;
+typedef _Atomic(size_t)		atomic_size_t;
+typedef _Atomic(ptrdiff_t)		atomic_ptrdiff_t;
+typedef _Atomic(intmax_t)		atomic_intmax_t;
+typedef _Atomic(uintmax_t)		atomic_uintmax_t;
+
+/*
+ * 7.17.7 Operations on atomic types.
+ */
+
+/*
+ * Compiler-specific operations.
+ */
+
+#if defined(__CLANG_ATOMICS)
+#define	atomic_compare_exchange_strong_explicit(object, expected,	\
+    desired, success, failure)						\
+	__c11_atomic_compare_exchange_strong(object, expected, desired,	\
+	    success, failure)
+#define	atomic_compare_exchange_weak_explicit(object, expected,		\
+    desired, success, failure)						\
+	__c11_atomic_compare_exchange_weak(object, expected, desired,	\
+	    success, failure)
+#define	atomic_exchange_explicit(object, desired, order)		\
+	__c11_atomic_exchange(object, desired, order)
+#define	atomic_fetch_add_explicit(object, operand, order)		\
+	__c11_atomic_fetch_add(object, operand, order)
+#define	atomic_fetch_and_explicit(object, operand, order)		\
+	__c11_atomic_fetch_and(object, operand, order)
+#define	atomic_fetch_or_explicit(object, operand, order)		\
+	__c11_atomic_fetch_or(object, operand, order)
+#define	atomic_fetch_sub_explicit(object, operand, order)		\
+	__c11_atomic_fetch_sub(object, operand, order)
+#define	atomic_fetch_xor_explicit(object, operand, order)		\
+	__c11_atomic_fetch_xor(object, operand, order)
+#define	atomic_load_explicit(object, order)				\
+	__c11_atomic_load(object, order)
+#define	atomic_store_explicit(object, desired, order)			\
+	__c11_atomic_store(object, desired, order)
+#elif defined(__GNUC_ATOMICS)
+#define	atomic_compare_exchange_strong_explicit(object, expected,	\
+    desired, success, failure)						\
+	__atomic_compare_exchange_n(&(object)->__val, expected,		\
+	    desired, 0, success, failure)
+#define	atomic_compare_exchange_weak_explicit(object, expected,		\
+    desired, success, failure)						\
+	__atomic_compare_exchange_n(&(object)->__val, expected,		\
+	    desired, 1, success, failure)
+#define	atomic_exchange_explicit(object, desired, order)		\
+	__atomic_exchange_n(&(object)->__val, desired, order)
+#define	atomic_fetch_add_explicit(object, operand, order)		\
+	__atomic_fetch_add(&(object)->__val, operand, order)
+#define	atomic_fetch_and_explicit(object, operand, order)		\
+	__atomic_fetch_and(&(object)->__val, operand, order)
+#define	atomic_fetch_or_explicit(object, operand, order)		\
+	__atomic_fetch_or(&(object)->__val, operand, order)
+#define	atomic_fetch_sub_explicit(object, operand, order)		\
+	__atomic_fetch_sub(&(object)->__val, operand, order)
+#define	atomic_fetch_xor_explicit(object, operand, order)		\
+	__atomic_fetch_xor(&(object)->__val, operand, order)
+#define	atomic_load_explicit(object, order)				\
+	__atomic_load_n(&(object)->__val, order)
+#define	atomic_store_explicit(object, desired, order)			\
+	__atomic_store_n(&(object)->__val, desired, order)
+#else
+#define	__atomic_apply_stride(object, operand) \
+	(((__typeof__((object)->__val))0) + (operand))
+#define	atomic_compare_exchange_strong_explicit(object, expected,	\
+    desired, success, failure)	__extension__ ({			\
+	__typeof__(expected) __ep = (expected);				\
+	__typeof__(*__ep) __e = *__ep;					\
+	(void)(success); (void)(failure);				\
+	(bool)((*__ep = __sync_val_compare_and_swap(&(object)->__val,	\
+	    __e, desired)) == __e);					\
+})
+#define	atomic_compare_exchange_weak_explicit(object, expected,		\
+    desired, success, failure)						\
+	atomic_compare_exchange_strong_explicit(object, expected,	\
+		desired, success, failure)
+#ifdef __HAS_BUILTIN_SYNC_SWAP
+/* Clang provides a full-barrier atomic exchange - use it if available. */
+#define	atomic_exchange_explicit(object, desired, order)		\
+	((void)(order), __sync_swap(&(object)->__val, desired))
+#else
+/*
+ * __sync_lock_test_and_set() is only an acquire barrier in theory (although in
+ * practice it is usually a full barrier) so we need an explicit barrier before
+ * it.
+ */
+#define	atomic_exchange_explicit(object, desired, order)		\
+__extension__ ({							\
+	__typeof__(object) __o = (object);				\
+	__typeof__(desired) __d = (desired);				\
+	(void)(order);							\
+	__sync_synchronize();						\
+	__sync_lock_test_and_set(&(__o)->__val, __d);			\
+})
+#endif
+#define	atomic_fetch_add_explicit(object, operand, order)		\
+	((void)(order), __sync_fetch_and_add(&(object)->__val,		\
+	    __atomic_apply_stride(object, operand)))
+#define	atomic_fetch_and_explicit(object, operand, order)		\
+	((void)(order), __sync_fetch_and_and(&(object)->__val, operand))
+#define	atomic_fetch_or_explicit(object, operand, order)		\
+	((void)(order), __sync_fetch_and_or(&(object)->__val, operand))
+#define	atomic_fetch_sub_explicit(object, operand, order)		\
+	((void)(order), __sync_fetch_and_sub(&(object)->__val,		\
+	    __atomic_apply_stride(object, operand)))
+#define	atomic_fetch_xor_explicit(object, operand, order)		\
+	((void)(order), __sync_fetch_and_xor(&(object)->__val, operand))
+#define	atomic_load_explicit(object, order)				\
+	((void)(order), __sync_fetch_and_add(&(object)->__val, 0))
+#define	atomic_store_explicit(object, desired, order)			\
+	((void)atomic_exchange_explicit(object, desired, order))
+#endif
+
+/*
+ * Convenience functions.
+ *
+ * Don't provide these in kernel space. In kernel space, we should be
+ * disciplined enough to always provide explicit barriers.
+ */
+
+#ifndef _KERNEL
+#define	atomic_compare_exchange_strong(object, expected, desired)	\
+	atomic_compare_exchange_strong_explicit(object, expected,	\
+	    desired, memory_order_seq_cst, memory_order_seq_cst)
+#define	atomic_compare_exchange_weak(object, expected, desired)		\
+	atomic_compare_exchange_weak_explicit(object, expected,		\
+	    desired, memory_order_seq_cst, memory_order_seq_cst)
+#define	atomic_exchange(object, desired)				\
+	atomic_exchange_explicit(object, desired, memory_order_seq_cst)
+#define	atomic_fetch_add(object, operand)				\
+	atomic_fetch_add_explicit(object, operand, memory_order_seq_cst)
+#define	atomic_fetch_and(object, operand)				\
+	atomic_fetch_and_explicit(object, operand, memory_order_seq_cst)
+#define	atomic_fetch_or(object, operand)				\
+	atomic_fetch_or_explicit(object, operand, memory_order_seq_cst)
+#define	atomic_fetch_sub(object, operand)				\
+	atomic_fetch_sub_explicit(object, operand, memory_order_seq_cst)
+#define	atomic_fetch_xor(object, operand)				\
+	atomic_fetch_xor_explicit(object, operand, memory_order_seq_cst)
+#define	atomic_load(object)						\
+	atomic_load_explicit(object, memory_order_seq_cst)
+#define	atomic_store(object, desired)					\
+	atomic_store_explicit(object, desired, memory_order_seq_cst)
+#endif /* !_KERNEL */
+
+/*
+ * 7.17.8 Atomic flag type and operations.
+ *
+ * XXX: Assume atomic_bool can be used as an atomic_flag. Is there some
+ * kind of compiler built-in type we could use?
+ */
+
+typedef struct {
+	atomic_bool	__flag;
+} atomic_flag;
+
+#define	ATOMIC_FLAG_INIT		{ ATOMIC_VAR_INIT(false) }
+
+static __inline bool
+atomic_flag_test_and_set_explicit(volatile atomic_flag *__object,
+    memory_order __order)
+{
+	return (atomic_exchange_explicit(&__object->__flag, 1, __order));
+}
+
+static __inline void
+atomic_flag_clear_explicit(volatile atomic_flag *__object, memory_order __order)
+{
+
+	atomic_store_explicit(&__object->__flag, 0, __order);
+}
+
+#ifndef _KERNEL
+static __inline bool
+atomic_flag_test_and_set(volatile atomic_flag *__object)
+{
+
+	return (atomic_flag_test_and_set_explicit(__object,
+	    memory_order_seq_cst));
+}
+
+static __inline void
+atomic_flag_clear(volatile atomic_flag *__object)
+{
+
+	atomic_flag_clear_explicit(__object, memory_order_seq_cst);
+}
+#endif /* !_KERNEL */
+
+#endif /* <atomic> unavailable */
+
+#endif /* !_STDATOMIC_H_ */
diff --git a/lib/clang/3.6/include/stdbool.h b/lib/clang/3.6/include/stdbool.h
new file mode 100644
index 0000000..0467893
--- /dev/null
+++ b/lib/clang/3.6/include/stdbool.h
@@ -0,0 +1,44 @@
+/*===---- stdbool.h - Standard header for booleans -------------------------===
+ *
+ * Copyright (c) 2008 Eli Friedman
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __STDBOOL_H
+#define __STDBOOL_H
+
+/* Don't define bool, true, and false in C++, except as a GNU extension. */
+#ifndef __cplusplus
+#define bool _Bool
+#define true 1
+#define false 0
+#elif defined(__GNUC__) && !defined(__STRICT_ANSI__)
+/* Define _Bool, bool, false, true as a GNU extension. */
+#define _Bool bool
+#define bool  bool
+#define false false
+#define true  true
+#endif
+
+#define __bool_true_false_are_defined 1
+
+#endif /* __STDBOOL_H */
diff --git a/lib/clang/3.6/include/stddef.h b/lib/clang/3.6/include/stddef.h
new file mode 100644
index 0000000..7354996
--- /dev/null
+++ b/lib/clang/3.6/include/stddef.h
@@ -0,0 +1,137 @@
+/*===---- stddef.h - Basic type definitions --------------------------------===
+ *
+ * Copyright (c) 2008 Eli Friedman
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#if !defined(__STDDEF_H) || defined(__need_ptrdiff_t) ||                       \
+    defined(__need_size_t) || defined(__need_wchar_t) ||                       \
+    defined(__need_NULL) || defined(__need_wint_t)
+
+#if !defined(__need_ptrdiff_t) && !defined(__need_size_t) &&                   \
+    !defined(__need_wchar_t) && !defined(__need_NULL) &&                       \
+    !defined(__need_wint_t)
+/* Always define miscellaneous pieces when modules are available. */
+#if !__has_feature(modules)
+#define __STDDEF_H
+#endif
+#define __need_ptrdiff_t
+#define __need_size_t
+#define __need_wchar_t
+#define __need_NULL
+#define __need_STDDEF_H_misc
+/* __need_wint_t is intentionally not defined here. */
+#endif
+
+#if defined(__need_ptrdiff_t)
+#if !defined(_PTRDIFF_T) || __has_feature(modules)
+/* Always define ptrdiff_t when modules are available. */
+#if !__has_feature(modules)
+#define _PTRDIFF_T
+#endif
+typedef __PTRDIFF_TYPE__ ptrdiff_t;
+#endif
+#undef __need_ptrdiff_t
+#endif /* defined(__need_ptrdiff_t) */
+
+#if defined(__need_size_t)
+#if !defined(_SIZE_T) || __has_feature(modules)
+/* Always define size_t when modules are available. */
+#if !__has_feature(modules)
+#define _SIZE_T
+#endif
+typedef __SIZE_TYPE__ size_t;
+#endif
+#undef __need_size_t
+#endif /*defined(__need_size_t) */
+
+#if defined(__need_STDDEF_H_misc)
+/* ISO9899:2011 7.20 (C11 Annex K): Define rsize_t if __STDC_WANT_LIB_EXT1__ is
+ * enabled. */
+#if (defined(__STDC_WANT_LIB_EXT1__) && __STDC_WANT_LIB_EXT1__ >= 1 && \
+     !defined(_RSIZE_T)) || __has_feature(modules)
+/* Always define rsize_t when modules are available. */
+#if !__has_feature(modules)
+#define _RSIZE_T
+#endif
+typedef __SIZE_TYPE__ rsize_t;
+#endif
+#endif /* defined(__need_STDDEF_H_misc) */
+
+#if defined(__need_wchar_t)
+#ifndef __cplusplus
+/* Always define wchar_t when modules are available. */
+#if !defined(_WCHAR_T) || __has_feature(modules)
+#if !__has_feature(modules)
+#define _WCHAR_T
+#if defined(_MSC_EXTENSIONS)
+#define _WCHAR_T_DEFINED
+#endif
+#endif
+typedef __WCHAR_TYPE__ wchar_t;
+#endif
+#endif
+#undef __need_wchar_t
+#endif /* defined(__need_wchar_t) */
+
+#if defined(__need_NULL)
+#undef NULL
+#ifdef __cplusplus
+#  if !defined(__MINGW32__) && !defined(_MSC_VER)
+#    define NULL __null
+#  else
+#    define NULL 0
+#  endif
+#else
+#  define NULL ((void*)0)
+#endif
+#ifdef __cplusplus
+#if defined(_MSC_EXTENSIONS) && defined(_NATIVE_NULLPTR_SUPPORTED)
+namespace std { typedef decltype(nullptr) nullptr_t; }
+using ::std::nullptr_t;
+#endif
+#endif
+#undef __need_NULL
+#endif /* defined(__need_NULL) */
+
+#if defined(__need_STDDEF_H_misc)
+#if __STDC_VERSION__ >= 201112L || __cplusplus >= 201103L
+#include "__stddef_max_align_t.h"
+#endif
+#define offsetof(t, d) __builtin_offsetof(t, d)
+#undef __need_STDDEF_H_misc
+#endif  /* defined(__need_STDDEF_H_misc) */
+
+/* Some C libraries expect to see a wint_t here. Others (notably MinGW) will use
+__WINT_TYPE__ directly; accommodate both by requiring __need_wint_t */
+#if defined(__need_wint_t)
+/* Always define wint_t when modules are available. */
+#if !defined(_WINT_T) || __has_feature(modules)
+#if !__has_feature(modules)
+#define _WINT_T
+#endif
+typedef __WINT_TYPE__ wint_t;
+#endif
+#undef __need_wint_t
+#endif /* __need_wint_t */
+
+#endif
diff --git a/lib/clang/3.6/include/stdint.h b/lib/clang/3.6/include/stdint.h
new file mode 100644
index 0000000..0303db9
--- /dev/null
+++ b/lib/clang/3.6/include/stdint.h
@@ -0,0 +1,707 @@
+/*===---- stdint.h - Standard header for sized integer types --------------===*\
+ *
+ * Copyright (c) 2009 Chris Lattner
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+\*===----------------------------------------------------------------------===*/
+
+#ifndef __CLANG_STDINT_H
+#define __CLANG_STDINT_H
+
+/* If we're hosted, fall back to the system's stdint.h, which might have
+ * additional definitions.
+ */
+#if __STDC_HOSTED__ && __has_include_next(<stdint.h>)
+
+// C99 7.18.3 Limits of other integer types
+//
+//  Footnote 219, 220: C++ implementations should define these macros only when
+//  __STDC_LIMIT_MACROS is defined before <stdint.h> is included.
+//
+//  Footnote 222: C++ implementations should define these macros only when
+//  __STDC_CONSTANT_MACROS is defined before <stdint.h> is included.
+//
+// C++11 [cstdint.syn]p2:
+//
+//  The macros defined by <cstdint> are provided unconditionally. In particular,
+//  the symbols __STDC_LIMIT_MACROS and __STDC_CONSTANT_MACROS (mentioned in
+//  footnotes 219, 220, and 222 in the C standard) play no role in C++.
+//
+// C11 removed the problematic footnotes.
+//
+// Work around this inconsistency by always defining those macros in C++ mode,
+// so that a C library implementation which follows the C99 standard can be
+// used in C++.
+# ifdef __cplusplus
+#  if !defined(__STDC_LIMIT_MACROS)
+#   define __STDC_LIMIT_MACROS
+#   define __STDC_LIMIT_MACROS_DEFINED_BY_CLANG
+#  endif
+#  if !defined(__STDC_CONSTANT_MACROS)
+#   define __STDC_CONSTANT_MACROS
+#   define __STDC_CONSTANT_MACROS_DEFINED_BY_CLANG
+#  endif
+# endif
+
+# include_next <stdint.h>
+
+# ifdef __STDC_LIMIT_MACROS_DEFINED_BY_CLANG
+#  undef __STDC_LIMIT_MACROS
+#  undef __STDC_LIMIT_MACROS_DEFINED_BY_CLANG
+# endif
+# ifdef __STDC_CONSTANT_MACROS_DEFINED_BY_CLANG
+#  undef __STDC_CONSTANT_MACROS
+#  undef __STDC_CONSTANT_MACROS_DEFINED_BY_CLANG
+# endif
+
+#else
+
+/* C99 7.18.1.1 Exact-width integer types.
+ * C99 7.18.1.2 Minimum-width integer types.
+ * C99 7.18.1.3 Fastest minimum-width integer types.
+ *
+ * The standard requires that exact-width type be defined for 8-, 16-, 32-, and 
+ * 64-bit types if they are implemented. Other exact width types are optional.
+ * This implementation defines an exact-width types for every integer width
+ * that is represented in the standard integer types.
+ *
+ * The standard also requires minimum-width types be defined for 8-, 16-, 32-,
+ * and 64-bit widths regardless of whether there are corresponding exact-width
+ * types. 
+ *
+ * To accommodate targets that are missing types that are exactly 8, 16, 32, or
+ * 64 bits wide, this implementation takes an approach of cascading
+ * redefintions, redefining __int_leastN_t to successively smaller exact-width
+ * types. It is therefore important that the types are defined in order of
+ * descending widths.
+ *
+ * We currently assume that the minimum-width types and the fastest
+ * minimum-width types are the same. This is allowed by the standard, but is
+ * suboptimal.
+ *
+ * In violation of the standard, some targets do not implement a type that is
+ * wide enough to represent all of the required widths (8-, 16-, 32-, 64-bit).  
+ * To accommodate these targets, a required minimum-width type is only
+ * defined if there exists an exact-width type of equal or greater width.
+ */
+
+#ifdef __INT64_TYPE__
+# ifndef __int8_t_defined /* glibc sys/types.h also defines int64_t*/
+typedef __INT64_TYPE__ int64_t;
+# endif /* __int8_t_defined */
+typedef __UINT64_TYPE__ uint64_t;
+# define __int_least64_t int64_t
+# define __uint_least64_t uint64_t
+# define __int_least32_t int64_t
+# define __uint_least32_t uint64_t
+# define __int_least16_t int64_t
+# define __uint_least16_t uint64_t
+# define __int_least8_t int64_t
+# define __uint_least8_t uint64_t
+#endif /* __INT64_TYPE__ */
+
+#ifdef __int_least64_t
+typedef __int_least64_t int_least64_t;
+typedef __uint_least64_t uint_least64_t;
+typedef __int_least64_t int_fast64_t;
+typedef __uint_least64_t uint_fast64_t;
+#endif /* __int_least64_t */
+
+#ifdef __INT56_TYPE__
+typedef __INT56_TYPE__ int56_t;
+typedef __UINT56_TYPE__ uint56_t;
+typedef int56_t int_least56_t;
+typedef uint56_t uint_least56_t;
+typedef int56_t int_fast56_t;
+typedef uint56_t uint_fast56_t;
+# define __int_least32_t int56_t
+# define __uint_least32_t uint56_t
+# define __int_least16_t int56_t
+# define __uint_least16_t uint56_t
+# define __int_least8_t int56_t
+# define __uint_least8_t uint56_t
+#endif /* __INT56_TYPE__ */
+
+
+#ifdef __INT48_TYPE__
+typedef __INT48_TYPE__ int48_t;
+typedef __UINT48_TYPE__ uint48_t;
+typedef int48_t int_least48_t;
+typedef uint48_t uint_least48_t;
+typedef int48_t int_fast48_t;
+typedef uint48_t uint_fast48_t;
+# define __int_least32_t int48_t
+# define __uint_least32_t uint48_t
+# define __int_least16_t int48_t
+# define __uint_least16_t uint48_t
+# define __int_least8_t int48_t
+# define __uint_least8_t uint48_t
+#endif /* __INT48_TYPE__ */
+
+
+#ifdef __INT40_TYPE__
+typedef __INT40_TYPE__ int40_t;
+typedef __UINT40_TYPE__ uint40_t;
+typedef int40_t int_least40_t;
+typedef uint40_t uint_least40_t;
+typedef int40_t int_fast40_t;
+typedef uint40_t uint_fast40_t;
+# define __int_least32_t int40_t
+# define __uint_least32_t uint40_t
+# define __int_least16_t int40_t
+# define __uint_least16_t uint40_t
+# define __int_least8_t int40_t
+# define __uint_least8_t uint40_t
+#endif /* __INT40_TYPE__ */
+
+
+#ifdef __INT32_TYPE__
+
+# ifndef __int8_t_defined /* glibc sys/types.h also defines int32_t*/
+typedef __INT32_TYPE__ int32_t;
+# endif /* __int8_t_defined */
+
+# ifndef __uint32_t_defined  /* more glibc compatibility */
+# define __uint32_t_defined
+typedef __UINT32_TYPE__ uint32_t;
+# endif /* __uint32_t_defined */
+
+# define __int_least32_t int32_t
+# define __uint_least32_t uint32_t
+# define __int_least16_t int32_t
+# define __uint_least16_t uint32_t
+# define __int_least8_t int32_t
+# define __uint_least8_t uint32_t
+#endif /* __INT32_TYPE__ */
+
+#ifdef __int_least32_t
+typedef __int_least32_t int_least32_t;
+typedef __uint_least32_t uint_least32_t;
+typedef __int_least32_t int_fast32_t;
+typedef __uint_least32_t uint_fast32_t;
+#endif /* __int_least32_t */
+
+#ifdef __INT24_TYPE__
+typedef __INT24_TYPE__ int24_t;
+typedef __UINT24_TYPE__ uint24_t;
+typedef int24_t int_least24_t;
+typedef uint24_t uint_least24_t;
+typedef int24_t int_fast24_t;
+typedef uint24_t uint_fast24_t;
+# define __int_least16_t int24_t
+# define __uint_least16_t uint24_t
+# define __int_least8_t int24_t
+# define __uint_least8_t uint24_t
+#endif /* __INT24_TYPE__ */
+
+#ifdef __INT16_TYPE__
+#ifndef __int8_t_defined /* glibc sys/types.h also defines int16_t*/
+typedef __INT16_TYPE__ int16_t;
+#endif /* __int8_t_defined */
+typedef __UINT16_TYPE__ uint16_t;
+# define __int_least16_t int16_t
+# define __uint_least16_t uint16_t
+# define __int_least8_t int16_t
+# define __uint_least8_t uint16_t
+#endif /* __INT16_TYPE__ */
+
+#ifdef __int_least16_t
+typedef __int_least16_t int_least16_t;
+typedef __uint_least16_t uint_least16_t;
+typedef __int_least16_t int_fast16_t;
+typedef __uint_least16_t uint_fast16_t;
+#endif /* __int_least16_t */
+
+
+#ifdef __INT8_TYPE__
+#ifndef __int8_t_defined  /* glibc sys/types.h also defines int8_t*/
+typedef __INT8_TYPE__ int8_t;
+#endif /* __int8_t_defined */
+typedef __UINT8_TYPE__ uint8_t;
+# define __int_least8_t int8_t
+# define __uint_least8_t uint8_t
+#endif /* __INT8_TYPE__ */
+
+#ifdef __int_least8_t
+typedef __int_least8_t int_least8_t;
+typedef __uint_least8_t uint_least8_t;
+typedef __int_least8_t int_fast8_t;
+typedef __uint_least8_t uint_fast8_t;
+#endif /* __int_least8_t */
+
+/* prevent glibc sys/types.h from defining conflicting types */
+#ifndef __int8_t_defined  
+# define __int8_t_defined
+#endif /* __int8_t_defined */
+
+/* C99 7.18.1.4 Integer types capable of holding object pointers.
+ */
+#define __stdint_join3(a,b,c) a ## b ## c
+
+#define  __intn_t(n) __stdint_join3( int, n, _t)
+#define __uintn_t(n) __stdint_join3(uint, n, _t)
+
+#ifndef _INTPTR_T
+#ifndef __intptr_t_defined
+typedef  __intn_t(__INTPTR_WIDTH__)  intptr_t;
+#define __intptr_t_defined
+#define _INTPTR_T
+#endif
+#endif
+
+#ifndef _UINTPTR_T
+typedef __uintn_t(__INTPTR_WIDTH__) uintptr_t;
+#define _UINTPTR_T
+#endif
+
+/* C99 7.18.1.5 Greatest-width integer types.
+ */
+typedef __INTMAX_TYPE__  intmax_t;
+typedef __UINTMAX_TYPE__ uintmax_t;
+
+/* C99 7.18.4 Macros for minimum-width integer constants.
+ *
+ * The standard requires that integer constant macros be defined for all the
+ * minimum-width types defined above. As 8-, 16-, 32-, and 64-bit minimum-width
+ * types are required, the corresponding integer constant macros are defined 
+ * here. This implementation also defines minimum-width types for every other
+ * integer width that the target implements, so corresponding macros are 
+ * defined below, too.
+ *
+ * These macros are defined using the same successive-shrinking approach as
+ * the type definitions above. It is likewise important that macros are defined
+ * in order of decending width.
+ *
+ * Note that C++ should not check __STDC_CONSTANT_MACROS here, contrary to the
+ * claims of the C standard (see C++ 18.3.1p2, [cstdint.syn]).
+ */
+
+#define __int_c_join(a, b) a ## b
+#define __int_c(v, suffix) __int_c_join(v, suffix)
+#define __uint_c(v, suffix) __int_c_join(v##U, suffix)
+
+
+#ifdef __INT64_TYPE__
+# ifdef __INT64_C_SUFFIX__
+#  define __int64_c_suffix __INT64_C_SUFFIX__
+#  define __int32_c_suffix __INT64_C_SUFFIX__
+#  define __int16_c_suffix __INT64_C_SUFFIX__
+#  define  __int8_c_suffix __INT64_C_SUFFIX__
+# else
+#  undef __int64_c_suffix
+#  undef __int32_c_suffix
+#  undef __int16_c_suffix
+#  undef  __int8_c_suffix
+# endif /* __INT64_C_SUFFIX__ */
+#endif /* __INT64_TYPE__ */
+
+#ifdef __int_least64_t
+# ifdef __int64_c_suffix
+#  define INT64_C(v) __int_c(v, __int64_c_suffix)
+#  define UINT64_C(v) __uint_c(v, __int64_c_suffix)
+# else
+#  define INT64_C(v) v
+#  define UINT64_C(v) v ## U
+# endif /* __int64_c_suffix */
+#endif /* __int_least64_t */
+
+
+#ifdef __INT56_TYPE__
+# ifdef __INT56_C_SUFFIX__
+#  define INT56_C(v) __int_c(v, __INT56_C_SUFFIX__)
+#  define UINT56_C(v) __uint_c(v, __INT56_C_SUFFIX__)
+#  define __int32_c_suffix __INT56_C_SUFFIX__
+#  define __int16_c_suffix __INT56_C_SUFFIX__
+#  define __int8_c_suffix  __INT56_C_SUFFIX__
+# else
+#  define INT56_C(v) v
+#  define UINT56_C(v) v ## U
+#  undef __int32_c_suffix
+#  undef __int16_c_suffix
+#  undef  __int8_c_suffix
+# endif /* __INT56_C_SUFFIX__ */
+#endif /* __INT56_TYPE__ */
+
+
+#ifdef __INT48_TYPE__
+# ifdef __INT48_C_SUFFIX__
+#  define INT48_C(v) __int_c(v, __INT48_C_SUFFIX__)
+#  define UINT48_C(v) __uint_c(v, __INT48_C_SUFFIX__)
+#  define __int32_c_suffix __INT48_C_SUFFIX__
+#  define __int16_c_suffix __INT48_C_SUFFIX__
+#  define __int8_c_suffix  __INT48_C_SUFFIX__
+# else
+#  define INT48_C(v) v
+#  define UINT48_C(v) v ## U
+#  undef __int32_c_suffix
+#  undef __int16_c_suffix
+#  undef  __int8_c_suffix
+# endif /* __INT48_C_SUFFIX__ */
+#endif /* __INT48_TYPE__ */
+
+
+#ifdef __INT40_TYPE__
+# ifdef __INT40_C_SUFFIX__
+#  define INT40_C(v) __int_c(v, __INT40_C_SUFFIX__)
+#  define UINT40_C(v) __uint_c(v, __INT40_C_SUFFIX__)
+#  define __int32_c_suffix __INT40_C_SUFFIX__
+#  define __int16_c_suffix __INT40_C_SUFFIX__
+#  define __int8_c_suffix  __INT40_C_SUFFIX__
+# else
+#  define INT40_C(v) v
+#  define UINT40_C(v) v ## U
+#  undef __int32_c_suffix
+#  undef __int16_c_suffix
+#  undef  __int8_c_suffix
+# endif /* __INT40_C_SUFFIX__ */
+#endif /* __INT40_TYPE__ */
+
+
+#ifdef __INT32_TYPE__
+# ifdef __INT32_C_SUFFIX__
+#  define __int32_c_suffix __INT32_C_SUFFIX__
+#  define __int16_c_suffix __INT32_C_SUFFIX__
+#  define __int8_c_suffix  __INT32_C_SUFFIX__
+#else
+#  undef __int32_c_suffix
+#  undef __int16_c_suffix
+#  undef  __int8_c_suffix
+# endif /* __INT32_C_SUFFIX__ */
+#endif /* __INT32_TYPE__ */
+
+#ifdef __int_least32_t
+# ifdef __int32_c_suffix
+#  define INT32_C(v) __int_c(v, __int32_c_suffix)
+#  define UINT32_C(v) __uint_c(v, __int32_c_suffix)
+# else
+#  define INT32_C(v) v
+#  define UINT32_C(v) v ## U
+# endif /* __int32_c_suffix */
+#endif /* __int_least32_t */
+
+
+#ifdef __INT24_TYPE__
+# ifdef __INT24_C_SUFFIX__
+#  define INT24_C(v) __int_c(v, __INT24_C_SUFFIX__)
+#  define UINT24_C(v) __uint_c(v, __INT24_C_SUFFIX__)
+#  define __int16_c_suffix __INT24_C_SUFFIX__
+#  define __int8_c_suffix  __INT24_C_SUFFIX__
+# else
+#  define INT24_C(v) v
+#  define UINT24_C(v) v ## U
+#  undef __int16_c_suffix
+#  undef  __int8_c_suffix
+# endif /* __INT24_C_SUFFIX__ */
+#endif /* __INT24_TYPE__ */
+
+
+#ifdef __INT16_TYPE__
+# ifdef __INT16_C_SUFFIX__
+#  define __int16_c_suffix __INT16_C_SUFFIX__
+#  define __int8_c_suffix  __INT16_C_SUFFIX__
+#else
+#  undef __int16_c_suffix
+#  undef  __int8_c_suffix
+# endif /* __INT16_C_SUFFIX__ */
+#endif /* __INT16_TYPE__ */
+
+#ifdef __int_least16_t
+# ifdef __int16_c_suffix
+#  define INT16_C(v) __int_c(v, __int16_c_suffix)
+#  define UINT16_C(v) __uint_c(v, __int16_c_suffix)
+# else
+#  define INT16_C(v) v
+#  define UINT16_C(v) v ## U
+# endif /* __int16_c_suffix */
+#endif /* __int_least16_t */
+
+
+#ifdef __INT8_TYPE__
+# ifdef __INT8_C_SUFFIX__
+#  define __int8_c_suffix __INT8_C_SUFFIX__
+#else
+#  undef  __int8_c_suffix
+# endif /* __INT8_C_SUFFIX__ */
+#endif /* __INT8_TYPE__ */
+
+#ifdef __int_least8_t
+# ifdef __int8_c_suffix
+#  define INT8_C(v) __int_c(v, __int8_c_suffix)
+#  define UINT8_C(v) __uint_c(v, __int8_c_suffix)
+# else
+#  define INT8_C(v) v
+#  define UINT8_C(v) v ## U
+# endif /* __int8_c_suffix */
+#endif /* __int_least8_t */
+
+
+/* C99 7.18.2.1 Limits of exact-width integer types. 
+ * C99 7.18.2.2 Limits of minimum-width integer types.
+ * C99 7.18.2.3 Limits of fastest minimum-width integer types.
+ *
+ * The presence of limit macros are completely optional in C99.  This
+ * implementation defines limits for all of the types (exact- and
+ * minimum-width) that it defines above, using the limits of the minimum-width
+ * type for any types that do not have exact-width representations.
+ *
+ * As in the type definitions, this section takes an approach of
+ * successive-shrinking to determine which limits to use for the standard (8,
+ * 16, 32, 64) bit widths when they don't have exact representations. It is
+ * therefore important that the defintions be kept in order of decending
+ * widths.
+ *
+ * Note that C++ should not check __STDC_LIMIT_MACROS here, contrary to the
+ * claims of the C standard (see C++ 18.3.1p2, [cstdint.syn]).
+ */
+
+#ifdef __INT64_TYPE__
+# define INT64_MAX           INT64_C( 9223372036854775807)
+# define INT64_MIN         (-INT64_C( 9223372036854775807)-1)
+# define UINT64_MAX         UINT64_C(18446744073709551615)
+# define __INT_LEAST64_MIN   INT64_MIN
+# define __INT_LEAST64_MAX   INT64_MAX
+# define __UINT_LEAST64_MAX UINT64_MAX
+# define __INT_LEAST32_MIN   INT64_MIN
+# define __INT_LEAST32_MAX   INT64_MAX
+# define __UINT_LEAST32_MAX UINT64_MAX
+# define __INT_LEAST16_MIN   INT64_MIN
+# define __INT_LEAST16_MAX   INT64_MAX
+# define __UINT_LEAST16_MAX UINT64_MAX
+# define __INT_LEAST8_MIN    INT64_MIN
+# define __INT_LEAST8_MAX    INT64_MAX
+# define __UINT_LEAST8_MAX  UINT64_MAX
+#endif /* __INT64_TYPE__ */
+
+#ifdef __INT_LEAST64_MIN
+# define INT_LEAST64_MIN   __INT_LEAST64_MIN
+# define INT_LEAST64_MAX   __INT_LEAST64_MAX
+# define UINT_LEAST64_MAX __UINT_LEAST64_MAX
+# define INT_FAST64_MIN    __INT_LEAST64_MIN
+# define INT_FAST64_MAX    __INT_LEAST64_MAX
+# define UINT_FAST64_MAX  __UINT_LEAST64_MAX
+#endif /* __INT_LEAST64_MIN */
+
+
+#ifdef __INT56_TYPE__
+# define INT56_MAX           INT56_C(36028797018963967)
+# define INT56_MIN         (-INT56_C(36028797018963967)-1)
+# define UINT56_MAX         UINT56_C(72057594037927935)
+# define INT_LEAST56_MIN     INT56_MIN
+# define INT_LEAST56_MAX     INT56_MAX
+# define UINT_LEAST56_MAX   UINT56_MAX
+# define INT_FAST56_MIN      INT56_MIN
+# define INT_FAST56_MAX      INT56_MAX
+# define UINT_FAST56_MAX    UINT56_MAX
+# define __INT_LEAST32_MIN   INT56_MIN
+# define __INT_LEAST32_MAX   INT56_MAX
+# define __UINT_LEAST32_MAX UINT56_MAX
+# define __INT_LEAST16_MIN   INT56_MIN
+# define __INT_LEAST16_MAX   INT56_MAX
+# define __UINT_LEAST16_MAX UINT56_MAX
+# define __INT_LEAST8_MIN    INT56_MIN
+# define __INT_LEAST8_MAX    INT56_MAX
+# define __UINT_LEAST8_MAX  UINT56_MAX
+#endif /* __INT56_TYPE__ */
+
+
+#ifdef __INT48_TYPE__
+# define INT48_MAX           INT48_C(140737488355327)
+# define INT48_MIN         (-INT48_C(140737488355327)-1)
+# define UINT48_MAX         UINT48_C(281474976710655)
+# define INT_LEAST48_MIN     INT48_MIN
+# define INT_LEAST48_MAX     INT48_MAX
+# define UINT_LEAST48_MAX   UINT48_MAX
+# define INT_FAST48_MIN      INT48_MIN
+# define INT_FAST48_MAX      INT48_MAX
+# define UINT_FAST48_MAX    UINT48_MAX
+# define __INT_LEAST32_MIN   INT48_MIN
+# define __INT_LEAST32_MAX   INT48_MAX
+# define __UINT_LEAST32_MAX UINT48_MAX
+# define __INT_LEAST16_MIN   INT48_MIN
+# define __INT_LEAST16_MAX   INT48_MAX
+# define __UINT_LEAST16_MAX UINT48_MAX
+# define __INT_LEAST8_MIN    INT48_MIN
+# define __INT_LEAST8_MAX    INT48_MAX
+# define __UINT_LEAST8_MAX  UINT48_MAX
+#endif /* __INT48_TYPE__ */
+
+
+#ifdef __INT40_TYPE__
+# define INT40_MAX           INT40_C(549755813887)
+# define INT40_MIN         (-INT40_C(549755813887)-1)
+# define UINT40_MAX         UINT40_C(1099511627775)
+# define INT_LEAST40_MIN     INT40_MIN
+# define INT_LEAST40_MAX     INT40_MAX
+# define UINT_LEAST40_MAX   UINT40_MAX
+# define INT_FAST40_MIN      INT40_MIN
+# define INT_FAST40_MAX      INT40_MAX
+# define UINT_FAST40_MAX    UINT40_MAX
+# define __INT_LEAST32_MIN   INT40_MIN
+# define __INT_LEAST32_MAX   INT40_MAX
+# define __UINT_LEAST32_MAX UINT40_MAX
+# define __INT_LEAST16_MIN   INT40_MIN
+# define __INT_LEAST16_MAX   INT40_MAX
+# define __UINT_LEAST16_MAX UINT40_MAX
+# define __INT_LEAST8_MIN    INT40_MIN
+# define __INT_LEAST8_MAX    INT40_MAX
+# define __UINT_LEAST8_MAX  UINT40_MAX
+#endif /* __INT40_TYPE__ */
+
+
+#ifdef __INT32_TYPE__
+# define INT32_MAX           INT32_C(2147483647)
+# define INT32_MIN         (-INT32_C(2147483647)-1)
+# define UINT32_MAX         UINT32_C(4294967295)
+# define __INT_LEAST32_MIN   INT32_MIN
+# define __INT_LEAST32_MAX   INT32_MAX
+# define __UINT_LEAST32_MAX UINT32_MAX
+# define __INT_LEAST16_MIN   INT32_MIN
+# define __INT_LEAST16_MAX   INT32_MAX
+# define __UINT_LEAST16_MAX UINT32_MAX
+# define __INT_LEAST8_MIN    INT32_MIN
+# define __INT_LEAST8_MAX    INT32_MAX
+# define __UINT_LEAST8_MAX  UINT32_MAX
+#endif /* __INT32_TYPE__ */
+
+#ifdef __INT_LEAST32_MIN
+# define INT_LEAST32_MIN   __INT_LEAST32_MIN
+# define INT_LEAST32_MAX   __INT_LEAST32_MAX
+# define UINT_LEAST32_MAX __UINT_LEAST32_MAX
+# define INT_FAST32_MIN    __INT_LEAST32_MIN
+# define INT_FAST32_MAX    __INT_LEAST32_MAX
+# define UINT_FAST32_MAX  __UINT_LEAST32_MAX
+#endif /* __INT_LEAST32_MIN */
+
+
+#ifdef __INT24_TYPE__
+# define INT24_MAX           INT24_C(8388607)
+# define INT24_MIN         (-INT24_C(8388607)-1)
+# define UINT24_MAX         UINT24_C(16777215)
+# define INT_LEAST24_MIN     INT24_MIN
+# define INT_LEAST24_MAX     INT24_MAX
+# define UINT_LEAST24_MAX   UINT24_MAX
+# define INT_FAST24_MIN      INT24_MIN
+# define INT_FAST24_MAX      INT24_MAX
+# define UINT_FAST24_MAX    UINT24_MAX
+# define __INT_LEAST16_MIN   INT24_MIN
+# define __INT_LEAST16_MAX   INT24_MAX
+# define __UINT_LEAST16_MAX UINT24_MAX
+# define __INT_LEAST8_MIN    INT24_MIN
+# define __INT_LEAST8_MAX    INT24_MAX
+# define __UINT_LEAST8_MAX  UINT24_MAX
+#endif /* __INT24_TYPE__ */
+
+
+#ifdef __INT16_TYPE__
+#define INT16_MAX            INT16_C(32767)
+#define INT16_MIN          (-INT16_C(32767)-1)
+#define UINT16_MAX          UINT16_C(65535)
+# define __INT_LEAST16_MIN   INT16_MIN
+# define __INT_LEAST16_MAX   INT16_MAX
+# define __UINT_LEAST16_MAX UINT16_MAX
+# define __INT_LEAST8_MIN    INT16_MIN
+# define __INT_LEAST8_MAX    INT16_MAX
+# define __UINT_LEAST8_MAX  UINT16_MAX
+#endif /* __INT16_TYPE__ */
+
+#ifdef __INT_LEAST16_MIN
+# define INT_LEAST16_MIN   __INT_LEAST16_MIN
+# define INT_LEAST16_MAX   __INT_LEAST16_MAX
+# define UINT_LEAST16_MAX __UINT_LEAST16_MAX
+# define INT_FAST16_MIN    __INT_LEAST16_MIN
+# define INT_FAST16_MAX    __INT_LEAST16_MAX
+# define UINT_FAST16_MAX  __UINT_LEAST16_MAX
+#endif /* __INT_LEAST16_MIN */
+
+
+#ifdef __INT8_TYPE__
+# define INT8_MAX            INT8_C(127)
+# define INT8_MIN          (-INT8_C(127)-1)
+# define UINT8_MAX          UINT8_C(255)
+# define __INT_LEAST8_MIN    INT8_MIN
+# define __INT_LEAST8_MAX    INT8_MAX
+# define __UINT_LEAST8_MAX  UINT8_MAX
+#endif /* __INT8_TYPE__ */
+
+#ifdef __INT_LEAST8_MIN
+# define INT_LEAST8_MIN   __INT_LEAST8_MIN
+# define INT_LEAST8_MAX   __INT_LEAST8_MAX
+# define UINT_LEAST8_MAX __UINT_LEAST8_MAX
+# define INT_FAST8_MIN    __INT_LEAST8_MIN
+# define INT_FAST8_MAX    __INT_LEAST8_MAX
+# define UINT_FAST8_MAX  __UINT_LEAST8_MAX
+#endif /* __INT_LEAST8_MIN */
+
+/* Some utility macros */
+#define  __INTN_MIN(n)  __stdint_join3( INT, n, _MIN)
+#define  __INTN_MAX(n)  __stdint_join3( INT, n, _MAX)
+#define __UINTN_MAX(n)  __stdint_join3(UINT, n, _MAX)
+#define  __INTN_C(n, v) __stdint_join3( INT, n, _C(v))
+#define __UINTN_C(n, v) __stdint_join3(UINT, n, _C(v))
+
+/* C99 7.18.2.4 Limits of integer types capable of holding object pointers. */
+/* C99 7.18.3 Limits of other integer types. */
+
+#define  INTPTR_MIN  __INTN_MIN(__INTPTR_WIDTH__)
+#define  INTPTR_MAX  __INTN_MAX(__INTPTR_WIDTH__)
+#define UINTPTR_MAX __UINTN_MAX(__INTPTR_WIDTH__)
+#define PTRDIFF_MIN  __INTN_MIN(__PTRDIFF_WIDTH__)
+#define PTRDIFF_MAX  __INTN_MAX(__PTRDIFF_WIDTH__)
+#define    SIZE_MAX __UINTN_MAX(__SIZE_WIDTH__)
+
+/* ISO9899:2011 7.20 (C11 Annex K): Define RSIZE_MAX if __STDC_WANT_LIB_EXT1__
+ * is enabled. */
+#if defined(__STDC_WANT_LIB_EXT1__) && __STDC_WANT_LIB_EXT1__ >= 1
+#define   RSIZE_MAX            (SIZE_MAX >> 1)
+#endif
+
+/* C99 7.18.2.5 Limits of greatest-width integer types. */
+#define INTMAX_MIN   __INTN_MIN(__INTMAX_WIDTH__)
+#define INTMAX_MAX   __INTN_MAX(__INTMAX_WIDTH__)
+#define UINTMAX_MAX __UINTN_MAX(__INTMAX_WIDTH__)
+
+/* C99 7.18.3 Limits of other integer types. */
+#define SIG_ATOMIC_MIN __INTN_MIN(__SIG_ATOMIC_WIDTH__)
+#define SIG_ATOMIC_MAX __INTN_MAX(__SIG_ATOMIC_WIDTH__)
+#ifdef __WINT_UNSIGNED__
+# define WINT_MIN       __UINTN_C(__WINT_WIDTH__, 0)
+# define WINT_MAX       __UINTN_MAX(__WINT_WIDTH__)
+#else
+# define WINT_MIN       __INTN_MIN(__WINT_WIDTH__)
+# define WINT_MAX       __INTN_MAX(__WINT_WIDTH__)
+#endif
+
+#ifndef WCHAR_MAX
+# define WCHAR_MAX __WCHAR_MAX__
+#endif
+#ifndef WCHAR_MIN
+# if __WCHAR_MAX__ == __INTN_MAX(__WCHAR_WIDTH__)
+#  define WCHAR_MIN __INTN_MIN(__WCHAR_WIDTH__)
+# else
+#  define WCHAR_MIN __UINTN_C(__WCHAR_WIDTH__, 0)
+# endif
+#endif
+
+/* 7.18.4.2 Macros for greatest-width integer constants. */
+#define INTMAX_C(v)   __INTN_C(__INTMAX_WIDTH__, v)
+#define UINTMAX_C(v) __UINTN_C(__INTMAX_WIDTH__, v)
+
+#endif /* __STDC_HOSTED__ */
+#endif /* __CLANG_STDINT_H */
diff --git a/lib/clang/3.6/include/stdnoreturn.h b/lib/clang/3.6/include/stdnoreturn.h
new file mode 100644
index 0000000..a7a301d
--- /dev/null
+++ b/lib/clang/3.6/include/stdnoreturn.h
@@ -0,0 +1,30 @@
+/*===---- stdnoreturn.h - Standard header for noreturn macro ---------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __STDNORETURN_H
+#define __STDNORETURN_H
+
+#define noreturn _Noreturn
+#define __noreturn_is_defined 1
+
+#endif /* __STDNORETURN_H */
diff --git a/lib/clang/3.6/include/tbmintrin.h b/lib/clang/3.6/include/tbmintrin.h
new file mode 100644
index 0000000..f95e34f
--- /dev/null
+++ b/lib/clang/3.6/include/tbmintrin.h
@@ -0,0 +1,158 @@
+/*===---- tbmintrin.h - TBM intrinsics -------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __TBM__
+#error "TBM instruction set is not enabled"
+#endif
+
+#ifndef __X86INTRIN_H
+#error "Never use <tbmintrin.h> directly; include <x86intrin.h> instead."
+#endif
+
+#ifndef __TBMINTRIN_H
+#define __TBMINTRIN_H
+
+#define __bextri_u32(a, b) (__builtin_ia32_bextri_u32((a), (b)))
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__blcfill_u32(unsigned int a)
+{
+  return a & (a + 1);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__blci_u32(unsigned int a)
+{
+  return a | ~(a + 1);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__blcic_u32(unsigned int a)
+{
+  return ~a & (a + 1);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__blcmsk_u32(unsigned int a)
+{
+  return a ^ (a + 1);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__blcs_u32(unsigned int a)
+{
+  return a | (a + 1);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__blsfill_u32(unsigned int a)
+{
+  return a | (a - 1);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__blsic_u32(unsigned int a)
+{
+  return ~a | (a - 1);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__t1mskc_u32(unsigned int a)
+{
+  return ~a | (a + 1);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+__tzmsk_u32(unsigned int a)
+{
+  return ~a & (a - 1);
+}
+
+#ifdef __x86_64__
+#define __bextri_u64(a, b) (__builtin_ia32_bextri_u64((a), (int)(b)))
+
+static __inline__ unsigned long long __attribute__((__always_inline__,
+                                                    __nodebug__))
+__blcfill_u64(unsigned long long a)
+{
+  return a & (a + 1);
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__,
+                                                    __nodebug__))
+__blci_u64(unsigned long long a)
+{
+  return a | ~(a + 1);
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__,
+                                                    __nodebug__))
+__blcic_u64(unsigned long long a)
+{
+  return ~a & (a + 1);
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__,
+                                                    __nodebug__))
+__blcmsk_u64(unsigned long long a)
+{
+  return a ^ (a + 1);
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__,
+                                                    __nodebug__))
+__blcs_u64(unsigned long long a)
+{
+  return a | (a + 1);
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__,
+                                                    __nodebug__))
+__blsfill_u64(unsigned long long a)
+{
+  return a | (a - 1);
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__,
+                                                    __nodebug__))
+__blsic_u64(unsigned long long a)
+{
+  return ~a | (a - 1);
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__,
+                                                    __nodebug__))
+__t1mskc_u64(unsigned long long a)
+{
+  return ~a | (a + 1);
+}
+
+static __inline__ unsigned long long __attribute__((__always_inline__,
+                                                    __nodebug__))
+__tzmsk_u64(unsigned long long a)
+{
+  return ~a & (a - 1);
+}
+#endif
+
+#endif /* __TBMINTRIN_H */
diff --git a/lib/clang/3.6/include/tgmath.h b/lib/clang/3.6/include/tgmath.h
new file mode 100644
index 0000000..a48e267
--- /dev/null
+++ b/lib/clang/3.6/include/tgmath.h
@@ -0,0 +1,1374 @@
+/*===---- tgmath.h - Standard header for type generic math ----------------===*\
+ *
+ * Copyright (c) 2009 Howard Hinnant
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+\*===----------------------------------------------------------------------===*/
+
+#ifndef __TGMATH_H
+#define __TGMATH_H
+
+/* C99 7.22 Type-generic math <tgmath.h>. */
+#include <math.h>
+
+/* C++ handles type genericity with overloading in math.h. */
+#ifndef __cplusplus
+#include <complex.h>
+
+#define _TG_ATTRSp __attribute__((__overloadable__))
+#define _TG_ATTRS __attribute__((__overloadable__, __always_inline__))
+
+// promotion
+
+typedef void _Argument_type_is_not_arithmetic;
+static _Argument_type_is_not_arithmetic __tg_promote(...)
+  __attribute__((__unavailable__,__overloadable__));
+static double               _TG_ATTRSp __tg_promote(int);
+static double               _TG_ATTRSp __tg_promote(unsigned int);
+static double               _TG_ATTRSp __tg_promote(long);
+static double               _TG_ATTRSp __tg_promote(unsigned long);
+static double               _TG_ATTRSp __tg_promote(long long);
+static double               _TG_ATTRSp __tg_promote(unsigned long long);
+static float                _TG_ATTRSp __tg_promote(float);
+static double               _TG_ATTRSp __tg_promote(double);
+static long double          _TG_ATTRSp __tg_promote(long double);
+static float _Complex       _TG_ATTRSp __tg_promote(float _Complex);
+static double _Complex      _TG_ATTRSp __tg_promote(double _Complex);
+static long double _Complex _TG_ATTRSp __tg_promote(long double _Complex);
+
+#define __tg_promote1(__x)           (__typeof__(__tg_promote(__x)))
+#define __tg_promote2(__x, __y)      (__typeof__(__tg_promote(__x) + \
+                                                 __tg_promote(__y)))
+#define __tg_promote3(__x, __y, __z) (__typeof__(__tg_promote(__x) + \
+                                                 __tg_promote(__y) + \
+                                                 __tg_promote(__z)))
+
+// acos
+
+static float
+    _TG_ATTRS
+    __tg_acos(float __x) {return acosf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_acos(double __x) {return acos(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_acos(long double __x) {return acosl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_acos(float _Complex __x) {return cacosf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_acos(double _Complex __x) {return cacos(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_acos(long double _Complex __x) {return cacosl(__x);}
+
+#undef acos
+#define acos(__x) __tg_acos(__tg_promote1((__x))(__x))
+
+// asin
+
+static float
+    _TG_ATTRS
+    __tg_asin(float __x) {return asinf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_asin(double __x) {return asin(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_asin(long double __x) {return asinl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_asin(float _Complex __x) {return casinf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_asin(double _Complex __x) {return casin(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_asin(long double _Complex __x) {return casinl(__x);}
+
+#undef asin
+#define asin(__x) __tg_asin(__tg_promote1((__x))(__x))
+
+// atan
+
+static float
+    _TG_ATTRS
+    __tg_atan(float __x) {return atanf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_atan(double __x) {return atan(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_atan(long double __x) {return atanl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_atan(float _Complex __x) {return catanf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_atan(double _Complex __x) {return catan(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_atan(long double _Complex __x) {return catanl(__x);}
+
+#undef atan
+#define atan(__x) __tg_atan(__tg_promote1((__x))(__x))
+
+// acosh
+
+static float
+    _TG_ATTRS
+    __tg_acosh(float __x) {return acoshf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_acosh(double __x) {return acosh(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_acosh(long double __x) {return acoshl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_acosh(float _Complex __x) {return cacoshf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_acosh(double _Complex __x) {return cacosh(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_acosh(long double _Complex __x) {return cacoshl(__x);}
+
+#undef acosh
+#define acosh(__x) __tg_acosh(__tg_promote1((__x))(__x))
+
+// asinh
+
+static float
+    _TG_ATTRS
+    __tg_asinh(float __x) {return asinhf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_asinh(double __x) {return asinh(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_asinh(long double __x) {return asinhl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_asinh(float _Complex __x) {return casinhf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_asinh(double _Complex __x) {return casinh(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_asinh(long double _Complex __x) {return casinhl(__x);}
+
+#undef asinh
+#define asinh(__x) __tg_asinh(__tg_promote1((__x))(__x))
+
+// atanh
+
+static float
+    _TG_ATTRS
+    __tg_atanh(float __x) {return atanhf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_atanh(double __x) {return atanh(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_atanh(long double __x) {return atanhl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_atanh(float _Complex __x) {return catanhf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_atanh(double _Complex __x) {return catanh(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_atanh(long double _Complex __x) {return catanhl(__x);}
+
+#undef atanh
+#define atanh(__x) __tg_atanh(__tg_promote1((__x))(__x))
+
+// cos
+
+static float
+    _TG_ATTRS
+    __tg_cos(float __x) {return cosf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_cos(double __x) {return cos(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_cos(long double __x) {return cosl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_cos(float _Complex __x) {return ccosf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_cos(double _Complex __x) {return ccos(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_cos(long double _Complex __x) {return ccosl(__x);}
+
+#undef cos
+#define cos(__x) __tg_cos(__tg_promote1((__x))(__x))
+
+// sin
+
+static float
+    _TG_ATTRS
+    __tg_sin(float __x) {return sinf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_sin(double __x) {return sin(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_sin(long double __x) {return sinl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_sin(float _Complex __x) {return csinf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_sin(double _Complex __x) {return csin(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_sin(long double _Complex __x) {return csinl(__x);}
+
+#undef sin
+#define sin(__x) __tg_sin(__tg_promote1((__x))(__x))
+
+// tan
+
+static float
+    _TG_ATTRS
+    __tg_tan(float __x) {return tanf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_tan(double __x) {return tan(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_tan(long double __x) {return tanl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_tan(float _Complex __x) {return ctanf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_tan(double _Complex __x) {return ctan(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_tan(long double _Complex __x) {return ctanl(__x);}
+
+#undef tan
+#define tan(__x) __tg_tan(__tg_promote1((__x))(__x))
+
+// cosh
+
+static float
+    _TG_ATTRS
+    __tg_cosh(float __x) {return coshf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_cosh(double __x) {return cosh(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_cosh(long double __x) {return coshl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_cosh(float _Complex __x) {return ccoshf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_cosh(double _Complex __x) {return ccosh(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_cosh(long double _Complex __x) {return ccoshl(__x);}
+
+#undef cosh
+#define cosh(__x) __tg_cosh(__tg_promote1((__x))(__x))
+
+// sinh
+
+static float
+    _TG_ATTRS
+    __tg_sinh(float __x) {return sinhf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_sinh(double __x) {return sinh(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_sinh(long double __x) {return sinhl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_sinh(float _Complex __x) {return csinhf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_sinh(double _Complex __x) {return csinh(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_sinh(long double _Complex __x) {return csinhl(__x);}
+
+#undef sinh
+#define sinh(__x) __tg_sinh(__tg_promote1((__x))(__x))
+
+// tanh
+
+static float
+    _TG_ATTRS
+    __tg_tanh(float __x) {return tanhf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_tanh(double __x) {return tanh(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_tanh(long double __x) {return tanhl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_tanh(float _Complex __x) {return ctanhf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_tanh(double _Complex __x) {return ctanh(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_tanh(long double _Complex __x) {return ctanhl(__x);}
+
+#undef tanh
+#define tanh(__x) __tg_tanh(__tg_promote1((__x))(__x))
+
+// exp
+
+static float
+    _TG_ATTRS
+    __tg_exp(float __x) {return expf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_exp(double __x) {return exp(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_exp(long double __x) {return expl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_exp(float _Complex __x) {return cexpf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_exp(double _Complex __x) {return cexp(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_exp(long double _Complex __x) {return cexpl(__x);}
+
+#undef exp
+#define exp(__x) __tg_exp(__tg_promote1((__x))(__x))
+
+// log
+
+static float
+    _TG_ATTRS
+    __tg_log(float __x) {return logf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_log(double __x) {return log(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_log(long double __x) {return logl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_log(float _Complex __x) {return clogf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_log(double _Complex __x) {return clog(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_log(long double _Complex __x) {return clogl(__x);}
+
+#undef log
+#define log(__x) __tg_log(__tg_promote1((__x))(__x))
+
+// pow
+
+static float
+    _TG_ATTRS
+    __tg_pow(float __x, float __y) {return powf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_pow(double __x, double __y) {return pow(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_pow(long double __x, long double __y) {return powl(__x, __y);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_pow(float _Complex __x, float _Complex __y) {return cpowf(__x, __y);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_pow(double _Complex __x, double _Complex __y) {return cpow(__x, __y);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_pow(long double _Complex __x, long double _Complex __y) 
+    {return cpowl(__x, __y);}
+
+#undef pow
+#define pow(__x, __y) __tg_pow(__tg_promote2((__x), (__y))(__x), \
+                               __tg_promote2((__x), (__y))(__y))
+
+// sqrt
+
+static float
+    _TG_ATTRS
+    __tg_sqrt(float __x) {return sqrtf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_sqrt(double __x) {return sqrt(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_sqrt(long double __x) {return sqrtl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_sqrt(float _Complex __x) {return csqrtf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_sqrt(double _Complex __x) {return csqrt(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_sqrt(long double _Complex __x) {return csqrtl(__x);}
+
+#undef sqrt
+#define sqrt(__x) __tg_sqrt(__tg_promote1((__x))(__x))
+
+// fabs
+
+static float
+    _TG_ATTRS
+    __tg_fabs(float __x) {return fabsf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_fabs(double __x) {return fabs(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_fabs(long double __x) {return fabsl(__x);}
+
+static float
+    _TG_ATTRS
+    __tg_fabs(float _Complex __x) {return cabsf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_fabs(double _Complex __x) {return cabs(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_fabs(long double _Complex __x) {return cabsl(__x);}
+
+#undef fabs
+#define fabs(__x) __tg_fabs(__tg_promote1((__x))(__x))
+
+// atan2
+
+static float
+    _TG_ATTRS
+    __tg_atan2(float __x, float __y) {return atan2f(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_atan2(double __x, double __y) {return atan2(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_atan2(long double __x, long double __y) {return atan2l(__x, __y);}
+
+#undef atan2
+#define atan2(__x, __y) __tg_atan2(__tg_promote2((__x), (__y))(__x), \
+                                   __tg_promote2((__x), (__y))(__y))
+
+// cbrt
+
+static float
+    _TG_ATTRS
+    __tg_cbrt(float __x) {return cbrtf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_cbrt(double __x) {return cbrt(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_cbrt(long double __x) {return cbrtl(__x);}
+
+#undef cbrt
+#define cbrt(__x) __tg_cbrt(__tg_promote1((__x))(__x))
+
+// ceil
+
+static float
+    _TG_ATTRS
+    __tg_ceil(float __x) {return ceilf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_ceil(double __x) {return ceil(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_ceil(long double __x) {return ceill(__x);}
+
+#undef ceil
+#define ceil(__x) __tg_ceil(__tg_promote1((__x))(__x))
+
+// copysign
+
+static float
+    _TG_ATTRS
+    __tg_copysign(float __x, float __y) {return copysignf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_copysign(double __x, double __y) {return copysign(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_copysign(long double __x, long double __y) {return copysignl(__x, __y);}
+
+#undef copysign
+#define copysign(__x, __y) __tg_copysign(__tg_promote2((__x), (__y))(__x), \
+                                         __tg_promote2((__x), (__y))(__y))
+
+// erf
+
+static float
+    _TG_ATTRS
+    __tg_erf(float __x) {return erff(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_erf(double __x) {return erf(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_erf(long double __x) {return erfl(__x);}
+
+#undef erf
+#define erf(__x) __tg_erf(__tg_promote1((__x))(__x))
+
+// erfc
+
+static float
+    _TG_ATTRS
+    __tg_erfc(float __x) {return erfcf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_erfc(double __x) {return erfc(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_erfc(long double __x) {return erfcl(__x);}
+
+#undef erfc
+#define erfc(__x) __tg_erfc(__tg_promote1((__x))(__x))
+
+// exp2
+
+static float
+    _TG_ATTRS
+    __tg_exp2(float __x) {return exp2f(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_exp2(double __x) {return exp2(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_exp2(long double __x) {return exp2l(__x);}
+
+#undef exp2
+#define exp2(__x) __tg_exp2(__tg_promote1((__x))(__x))
+
+// expm1
+
+static float
+    _TG_ATTRS
+    __tg_expm1(float __x) {return expm1f(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_expm1(double __x) {return expm1(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_expm1(long double __x) {return expm1l(__x);}
+
+#undef expm1
+#define expm1(__x) __tg_expm1(__tg_promote1((__x))(__x))
+
+// fdim
+
+static float
+    _TG_ATTRS
+    __tg_fdim(float __x, float __y) {return fdimf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_fdim(double __x, double __y) {return fdim(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_fdim(long double __x, long double __y) {return fdiml(__x, __y);}
+
+#undef fdim
+#define fdim(__x, __y) __tg_fdim(__tg_promote2((__x), (__y))(__x), \
+                                 __tg_promote2((__x), (__y))(__y))
+
+// floor
+
+static float
+    _TG_ATTRS
+    __tg_floor(float __x) {return floorf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_floor(double __x) {return floor(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_floor(long double __x) {return floorl(__x);}
+
+#undef floor
+#define floor(__x) __tg_floor(__tg_promote1((__x))(__x))
+
+// fma
+
+static float
+    _TG_ATTRS
+    __tg_fma(float __x, float __y, float __z)
+    {return fmaf(__x, __y, __z);}
+
+static double
+    _TG_ATTRS
+    __tg_fma(double __x, double __y, double __z)
+    {return fma(__x, __y, __z);}
+
+static long double
+    _TG_ATTRS
+    __tg_fma(long double __x,long double __y, long double __z)
+    {return fmal(__x, __y, __z);}
+
+#undef fma
+#define fma(__x, __y, __z)                                \
+        __tg_fma(__tg_promote3((__x), (__y), (__z))(__x), \
+                 __tg_promote3((__x), (__y), (__z))(__y), \
+                 __tg_promote3((__x), (__y), (__z))(__z))
+
+// fmax
+
+static float
+    _TG_ATTRS
+    __tg_fmax(float __x, float __y) {return fmaxf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_fmax(double __x, double __y) {return fmax(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_fmax(long double __x, long double __y) {return fmaxl(__x, __y);}
+
+#undef fmax
+#define fmax(__x, __y) __tg_fmax(__tg_promote2((__x), (__y))(__x), \
+                                 __tg_promote2((__x), (__y))(__y))
+
+// fmin
+
+static float
+    _TG_ATTRS
+    __tg_fmin(float __x, float __y) {return fminf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_fmin(double __x, double __y) {return fmin(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_fmin(long double __x, long double __y) {return fminl(__x, __y);}
+
+#undef fmin
+#define fmin(__x, __y) __tg_fmin(__tg_promote2((__x), (__y))(__x), \
+                                 __tg_promote2((__x), (__y))(__y))
+
+// fmod
+
+static float
+    _TG_ATTRS
+    __tg_fmod(float __x, float __y) {return fmodf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_fmod(double __x, double __y) {return fmod(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_fmod(long double __x, long double __y) {return fmodl(__x, __y);}
+
+#undef fmod
+#define fmod(__x, __y) __tg_fmod(__tg_promote2((__x), (__y))(__x), \
+                                 __tg_promote2((__x), (__y))(__y))
+
+// frexp
+
+static float
+    _TG_ATTRS
+    __tg_frexp(float __x, int* __y) {return frexpf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_frexp(double __x, int* __y) {return frexp(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_frexp(long double __x, int* __y) {return frexpl(__x, __y);}
+
+#undef frexp
+#define frexp(__x, __y) __tg_frexp(__tg_promote1((__x))(__x), __y)
+
+// hypot
+
+static float
+    _TG_ATTRS
+    __tg_hypot(float __x, float __y) {return hypotf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_hypot(double __x, double __y) {return hypot(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_hypot(long double __x, long double __y) {return hypotl(__x, __y);}
+
+#undef hypot
+#define hypot(__x, __y) __tg_hypot(__tg_promote2((__x), (__y))(__x), \
+                                   __tg_promote2((__x), (__y))(__y))
+
+// ilogb
+
+static int
+    _TG_ATTRS
+    __tg_ilogb(float __x) {return ilogbf(__x);}
+
+static int
+    _TG_ATTRS
+    __tg_ilogb(double __x) {return ilogb(__x);}
+
+static int
+    _TG_ATTRS
+    __tg_ilogb(long double __x) {return ilogbl(__x);}
+
+#undef ilogb
+#define ilogb(__x) __tg_ilogb(__tg_promote1((__x))(__x))
+
+// ldexp
+
+static float
+    _TG_ATTRS
+    __tg_ldexp(float __x, int __y) {return ldexpf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_ldexp(double __x, int __y) {return ldexp(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_ldexp(long double __x, int __y) {return ldexpl(__x, __y);}
+
+#undef ldexp
+#define ldexp(__x, __y) __tg_ldexp(__tg_promote1((__x))(__x), __y)
+
+// lgamma
+
+static float
+    _TG_ATTRS
+    __tg_lgamma(float __x) {return lgammaf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_lgamma(double __x) {return lgamma(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_lgamma(long double __x) {return lgammal(__x);}
+
+#undef lgamma
+#define lgamma(__x) __tg_lgamma(__tg_promote1((__x))(__x))
+
+// llrint
+
+static long long
+    _TG_ATTRS
+    __tg_llrint(float __x) {return llrintf(__x);}
+
+static long long
+    _TG_ATTRS
+    __tg_llrint(double __x) {return llrint(__x);}
+
+static long long
+    _TG_ATTRS
+    __tg_llrint(long double __x) {return llrintl(__x);}
+
+#undef llrint
+#define llrint(__x) __tg_llrint(__tg_promote1((__x))(__x))
+
+// llround
+
+static long long
+    _TG_ATTRS
+    __tg_llround(float __x) {return llroundf(__x);}
+
+static long long
+    _TG_ATTRS
+    __tg_llround(double __x) {return llround(__x);}
+
+static long long
+    _TG_ATTRS
+    __tg_llround(long double __x) {return llroundl(__x);}
+
+#undef llround
+#define llround(__x) __tg_llround(__tg_promote1((__x))(__x))
+
+// log10
+
+static float
+    _TG_ATTRS
+    __tg_log10(float __x) {return log10f(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_log10(double __x) {return log10(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_log10(long double __x) {return log10l(__x);}
+
+#undef log10
+#define log10(__x) __tg_log10(__tg_promote1((__x))(__x))
+
+// log1p
+
+static float
+    _TG_ATTRS
+    __tg_log1p(float __x) {return log1pf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_log1p(double __x) {return log1p(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_log1p(long double __x) {return log1pl(__x);}
+
+#undef log1p
+#define log1p(__x) __tg_log1p(__tg_promote1((__x))(__x))
+
+// log2
+
+static float
+    _TG_ATTRS
+    __tg_log2(float __x) {return log2f(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_log2(double __x) {return log2(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_log2(long double __x) {return log2l(__x);}
+
+#undef log2
+#define log2(__x) __tg_log2(__tg_promote1((__x))(__x))
+
+// logb
+
+static float
+    _TG_ATTRS
+    __tg_logb(float __x) {return logbf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_logb(double __x) {return logb(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_logb(long double __x) {return logbl(__x);}
+
+#undef logb
+#define logb(__x) __tg_logb(__tg_promote1((__x))(__x))
+
+// lrint
+
+static long
+    _TG_ATTRS
+    __tg_lrint(float __x) {return lrintf(__x);}
+
+static long
+    _TG_ATTRS
+    __tg_lrint(double __x) {return lrint(__x);}
+
+static long
+    _TG_ATTRS
+    __tg_lrint(long double __x) {return lrintl(__x);}
+
+#undef lrint
+#define lrint(__x) __tg_lrint(__tg_promote1((__x))(__x))
+
+// lround
+
+static long
+    _TG_ATTRS
+    __tg_lround(float __x) {return lroundf(__x);}
+
+static long
+    _TG_ATTRS
+    __tg_lround(double __x) {return lround(__x);}
+
+static long
+    _TG_ATTRS
+    __tg_lround(long double __x) {return lroundl(__x);}
+
+#undef lround
+#define lround(__x) __tg_lround(__tg_promote1((__x))(__x))
+
+// nearbyint
+
+static float
+    _TG_ATTRS
+    __tg_nearbyint(float __x) {return nearbyintf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_nearbyint(double __x) {return nearbyint(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_nearbyint(long double __x) {return nearbyintl(__x);}
+
+#undef nearbyint
+#define nearbyint(__x) __tg_nearbyint(__tg_promote1((__x))(__x))
+
+// nextafter
+
+static float
+    _TG_ATTRS
+    __tg_nextafter(float __x, float __y) {return nextafterf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_nextafter(double __x, double __y) {return nextafter(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_nextafter(long double __x, long double __y) {return nextafterl(__x, __y);}
+
+#undef nextafter
+#define nextafter(__x, __y) __tg_nextafter(__tg_promote2((__x), (__y))(__x), \
+                                           __tg_promote2((__x), (__y))(__y))
+
+// nexttoward
+
+static float
+    _TG_ATTRS
+    __tg_nexttoward(float __x, long double __y) {return nexttowardf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_nexttoward(double __x, long double __y) {return nexttoward(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_nexttoward(long double __x, long double __y) {return nexttowardl(__x, __y);}
+
+#undef nexttoward
+#define nexttoward(__x, __y) __tg_nexttoward(__tg_promote1((__x))(__x), (__y))
+
+// remainder
+
+static float
+    _TG_ATTRS
+    __tg_remainder(float __x, float __y) {return remainderf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_remainder(double __x, double __y) {return remainder(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_remainder(long double __x, long double __y) {return remainderl(__x, __y);}
+
+#undef remainder
+#define remainder(__x, __y) __tg_remainder(__tg_promote2((__x), (__y))(__x), \
+                                           __tg_promote2((__x), (__y))(__y))
+
+// remquo
+
+static float
+    _TG_ATTRS
+    __tg_remquo(float __x, float __y, int* __z)
+    {return remquof(__x, __y, __z);}
+
+static double
+    _TG_ATTRS
+    __tg_remquo(double __x, double __y, int* __z)
+    {return remquo(__x, __y, __z);}
+
+static long double
+    _TG_ATTRS
+    __tg_remquo(long double __x,long double __y, int* __z)
+    {return remquol(__x, __y, __z);}
+
+#undef remquo
+#define remquo(__x, __y, __z)                         \
+        __tg_remquo(__tg_promote2((__x), (__y))(__x), \
+                    __tg_promote2((__x), (__y))(__y), \
+                    (__z))
+
+// rint
+
+static float
+    _TG_ATTRS
+    __tg_rint(float __x) {return rintf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_rint(double __x) {return rint(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_rint(long double __x) {return rintl(__x);}
+
+#undef rint
+#define rint(__x) __tg_rint(__tg_promote1((__x))(__x))
+
+// round
+
+static float
+    _TG_ATTRS
+    __tg_round(float __x) {return roundf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_round(double __x) {return round(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_round(long double __x) {return roundl(__x);}
+
+#undef round
+#define round(__x) __tg_round(__tg_promote1((__x))(__x))
+
+// scalbn
+
+static float
+    _TG_ATTRS
+    __tg_scalbn(float __x, int __y) {return scalbnf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_scalbn(double __x, int __y) {return scalbn(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_scalbn(long double __x, int __y) {return scalbnl(__x, __y);}
+
+#undef scalbn
+#define scalbn(__x, __y) __tg_scalbn(__tg_promote1((__x))(__x), __y)
+
+// scalbln
+
+static float
+    _TG_ATTRS
+    __tg_scalbln(float __x, long __y) {return scalblnf(__x, __y);}
+
+static double
+    _TG_ATTRS
+    __tg_scalbln(double __x, long __y) {return scalbln(__x, __y);}
+
+static long double
+    _TG_ATTRS
+    __tg_scalbln(long double __x, long __y) {return scalblnl(__x, __y);}
+
+#undef scalbln
+#define scalbln(__x, __y) __tg_scalbln(__tg_promote1((__x))(__x), __y)
+
+// tgamma
+
+static float
+    _TG_ATTRS
+    __tg_tgamma(float __x) {return tgammaf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_tgamma(double __x) {return tgamma(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_tgamma(long double __x) {return tgammal(__x);}
+
+#undef tgamma
+#define tgamma(__x) __tg_tgamma(__tg_promote1((__x))(__x))
+
+// trunc
+
+static float
+    _TG_ATTRS
+    __tg_trunc(float __x) {return truncf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_trunc(double __x) {return trunc(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_trunc(long double __x) {return truncl(__x);}
+
+#undef trunc
+#define trunc(__x) __tg_trunc(__tg_promote1((__x))(__x))
+
+// carg
+
+static float
+    _TG_ATTRS
+    __tg_carg(float __x) {return atan2f(0.F, __x);}
+
+static double
+    _TG_ATTRS
+    __tg_carg(double __x) {return atan2(0., __x);}
+
+static long double
+    _TG_ATTRS
+    __tg_carg(long double __x) {return atan2l(0.L, __x);}
+
+static float
+    _TG_ATTRS
+    __tg_carg(float _Complex __x) {return cargf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_carg(double _Complex __x) {return carg(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_carg(long double _Complex __x) {return cargl(__x);}
+
+#undef carg
+#define carg(__x) __tg_carg(__tg_promote1((__x))(__x))
+
+// cimag
+
+static float
+    _TG_ATTRS
+    __tg_cimag(float __x) {return 0;}
+
+static double
+    _TG_ATTRS
+    __tg_cimag(double __x) {return 0;}
+
+static long double
+    _TG_ATTRS
+    __tg_cimag(long double __x) {return 0;}
+
+static float
+    _TG_ATTRS
+    __tg_cimag(float _Complex __x) {return cimagf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_cimag(double _Complex __x) {return cimag(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_cimag(long double _Complex __x) {return cimagl(__x);}
+
+#undef cimag
+#define cimag(__x) __tg_cimag(__tg_promote1((__x))(__x))
+
+// conj
+
+static float _Complex
+    _TG_ATTRS
+    __tg_conj(float __x) {return __x;}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_conj(double __x) {return __x;}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_conj(long double __x) {return __x;}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_conj(float _Complex __x) {return conjf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_conj(double _Complex __x) {return conj(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_conj(long double _Complex __x) {return conjl(__x);}
+
+#undef conj
+#define conj(__x) __tg_conj(__tg_promote1((__x))(__x))
+
+// cproj
+
+static float _Complex
+    _TG_ATTRS
+    __tg_cproj(float __x) {return cprojf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_cproj(double __x) {return cproj(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_cproj(long double __x) {return cprojl(__x);}
+
+static float _Complex
+    _TG_ATTRS
+    __tg_cproj(float _Complex __x) {return cprojf(__x);}
+
+static double _Complex
+    _TG_ATTRS
+    __tg_cproj(double _Complex __x) {return cproj(__x);}
+
+static long double _Complex
+    _TG_ATTRS
+    __tg_cproj(long double _Complex __x) {return cprojl(__x);}
+
+#undef cproj
+#define cproj(__x) __tg_cproj(__tg_promote1((__x))(__x))
+
+// creal
+
+static float
+    _TG_ATTRS
+    __tg_creal(float __x) {return __x;}
+
+static double
+    _TG_ATTRS
+    __tg_creal(double __x) {return __x;}
+
+static long double
+    _TG_ATTRS
+    __tg_creal(long double __x) {return __x;}
+
+static float
+    _TG_ATTRS
+    __tg_creal(float _Complex __x) {return crealf(__x);}
+
+static double
+    _TG_ATTRS
+    __tg_creal(double _Complex __x) {return creal(__x);}
+
+static long double
+    _TG_ATTRS
+    __tg_creal(long double _Complex __x) {return creall(__x);}
+
+#undef creal
+#define creal(__x) __tg_creal(__tg_promote1((__x))(__x))
+
+#undef _TG_ATTRSp
+#undef _TG_ATTRS
+
+#endif /* __cplusplus */
+#endif /* __TGMATH_H */
diff --git a/lib/clang/3.6/include/tmmintrin.h b/lib/clang/3.6/include/tmmintrin.h
new file mode 100644
index 0000000..4238f5b
--- /dev/null
+++ b/lib/clang/3.6/include/tmmintrin.h
@@ -0,0 +1,225 @@
+/*===---- tmmintrin.h - SSSE3 intrinsics -----------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+ 
+#ifndef __TMMINTRIN_H
+#define __TMMINTRIN_H
+
+#ifndef __SSSE3__
+#error "SSSE3 instruction set not enabled"
+#else
+
+#include <pmmintrin.h>
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_abs_pi8(__m64 __a)
+{
+    return (__m64)__builtin_ia32_pabsb((__v8qi)__a);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_abs_epi8(__m128i __a)
+{
+    return (__m128i)__builtin_ia32_pabsb128((__v16qi)__a);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_abs_pi16(__m64 __a)
+{
+    return (__m64)__builtin_ia32_pabsw((__v4hi)__a);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_abs_epi16(__m128i __a)
+{
+    return (__m128i)__builtin_ia32_pabsw128((__v8hi)__a);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_abs_pi32(__m64 __a)
+{
+    return (__m64)__builtin_ia32_pabsd((__v2si)__a);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_abs_epi32(__m128i __a)
+{
+    return (__m128i)__builtin_ia32_pabsd128((__v4si)__a);
+}
+
+#define _mm_alignr_epi8(a, b, n) __extension__ ({ \
+  __m128i __a = (a); \
+  __m128i __b = (b); \
+  (__m128i)__builtin_ia32_palignr128((__v16qi)__a, (__v16qi)__b, (n)); })
+
+#define _mm_alignr_pi8(a, b, n) __extension__ ({ \
+  __m64 __a = (a); \
+  __m64 __b = (b); \
+  (__m64)__builtin_ia32_palignr((__v8qi)__a, (__v8qi)__b, (n)); })
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_hadd_epi16(__m128i __a, __m128i __b)
+{
+    return (__m128i)__builtin_ia32_phaddw128((__v8hi)__a, (__v8hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_hadd_epi32(__m128i __a, __m128i __b)
+{
+    return (__m128i)__builtin_ia32_phaddd128((__v4si)__a, (__v4si)__b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_hadd_pi16(__m64 __a, __m64 __b)
+{
+    return (__m64)__builtin_ia32_phaddw((__v4hi)__a, (__v4hi)__b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_hadd_pi32(__m64 __a, __m64 __b)
+{
+    return (__m64)__builtin_ia32_phaddd((__v2si)__a, (__v2si)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_hadds_epi16(__m128i __a, __m128i __b)
+{
+    return (__m128i)__builtin_ia32_phaddsw128((__v8hi)__a, (__v8hi)__b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_hadds_pi16(__m64 __a, __m64 __b)
+{
+    return (__m64)__builtin_ia32_phaddsw((__v4hi)__a, (__v4hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_hsub_epi16(__m128i __a, __m128i __b)
+{
+    return (__m128i)__builtin_ia32_phsubw128((__v8hi)__a, (__v8hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_hsub_epi32(__m128i __a, __m128i __b)
+{
+    return (__m128i)__builtin_ia32_phsubd128((__v4si)__a, (__v4si)__b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_hsub_pi16(__m64 __a, __m64 __b)
+{
+    return (__m64)__builtin_ia32_phsubw((__v4hi)__a, (__v4hi)__b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_hsub_pi32(__m64 __a, __m64 __b)
+{
+    return (__m64)__builtin_ia32_phsubd((__v2si)__a, (__v2si)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_hsubs_epi16(__m128i __a, __m128i __b)
+{
+    return (__m128i)__builtin_ia32_phsubsw128((__v8hi)__a, (__v8hi)__b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_hsubs_pi16(__m64 __a, __m64 __b)
+{
+    return (__m64)__builtin_ia32_phsubsw((__v4hi)__a, (__v4hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_maddubs_epi16(__m128i __a, __m128i __b)
+{
+    return (__m128i)__builtin_ia32_pmaddubsw128((__v16qi)__a, (__v16qi)__b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_maddubs_pi16(__m64 __a, __m64 __b)
+{
+    return (__m64)__builtin_ia32_pmaddubsw((__v8qi)__a, (__v8qi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_mulhrs_epi16(__m128i __a, __m128i __b)
+{
+    return (__m128i)__builtin_ia32_pmulhrsw128((__v8hi)__a, (__v8hi)__b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_mulhrs_pi16(__m64 __a, __m64 __b)
+{
+    return (__m64)__builtin_ia32_pmulhrsw((__v4hi)__a, (__v4hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_shuffle_epi8(__m128i __a, __m128i __b)
+{
+    return (__m128i)__builtin_ia32_pshufb128((__v16qi)__a, (__v16qi)__b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_shuffle_pi8(__m64 __a, __m64 __b)
+{
+    return (__m64)__builtin_ia32_pshufb((__v8qi)__a, (__v8qi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sign_epi8(__m128i __a, __m128i __b)
+{
+    return (__m128i)__builtin_ia32_psignb128((__v16qi)__a, (__v16qi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sign_epi16(__m128i __a, __m128i __b)
+{
+    return (__m128i)__builtin_ia32_psignw128((__v8hi)__a, (__v8hi)__b);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sign_epi32(__m128i __a, __m128i __b)
+{
+    return (__m128i)__builtin_ia32_psignd128((__v4si)__a, (__v4si)__b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sign_pi8(__m64 __a, __m64 __b)
+{
+    return (__m64)__builtin_ia32_psignb((__v8qi)__a, (__v8qi)__b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sign_pi16(__m64 __a, __m64 __b)
+{
+    return (__m64)__builtin_ia32_psignw((__v4hi)__a, (__v4hi)__b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sign_pi32(__m64 __a, __m64 __b)
+{
+    return (__m64)__builtin_ia32_psignd((__v2si)__a, (__v2si)__b);
+}
+
+#endif /* __SSSE3__ */
+
+#endif /* __TMMINTRIN_H */
diff --git a/lib/clang/3.6/include/unwind.h b/lib/clang/3.6/include/unwind.h
new file mode 100644
index 0000000..685c1df
--- /dev/null
+++ b/lib/clang/3.6/include/unwind.h
@@ -0,0 +1,280 @@
+/*===---- unwind.h - Stack unwinding ----------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+/* See "Data Definitions for libgcc_s" in the Linux Standard Base.*/
+
+#ifndef __CLANG_UNWIND_H
+#define __CLANG_UNWIND_H
+
+#if __has_include_next(<unwind.h>)
+/* Darwin (from 11.x on) and libunwind provide an unwind.h. If that's available,
+ * use it. libunwind wraps some of its definitions in #ifdef _GNU_SOURCE,
+ * so define that around the include.*/
+# ifndef _GNU_SOURCE
+#  define _SHOULD_UNDEFINE_GNU_SOURCE
+#  define _GNU_SOURCE
+# endif
+// libunwind's unwind.h reflects the current visibility.  However, Mozilla
+// builds with -fvisibility=hidden and relies on gcc's unwind.h to reset the
+// visibility to default and export its contents.  gcc also allows users to
+// override its override by #defining HIDE_EXPORTS (but note, this only obeys
+// the user's -fvisibility setting; it doesn't hide any exports on its own).  We
+// imitate gcc's header here:
+# ifdef HIDE_EXPORTS
+#  include_next <unwind.h>
+# else
+#  pragma GCC visibility push(default)
+#  include_next <unwind.h>
+#  pragma GCC visibility pop
+# endif
+# ifdef _SHOULD_UNDEFINE_GNU_SOURCE
+#  undef _GNU_SOURCE
+#  undef _SHOULD_UNDEFINE_GNU_SOURCE
+# endif
+#else
+
+#include <stdint.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/* It is a bit strange for a header to play with the visibility of the
+   symbols it declares, but this matches gcc's behavior and some programs
+   depend on it */
+#ifndef HIDE_EXPORTS
+#pragma GCC visibility push(default)
+#endif
+
+typedef uintptr_t _Unwind_Word;
+typedef intptr_t _Unwind_Sword;
+typedef uintptr_t _Unwind_Ptr;
+typedef uintptr_t _Unwind_Internal_Ptr;
+typedef uint64_t _Unwind_Exception_Class;
+
+typedef intptr_t _sleb128_t;
+typedef uintptr_t _uleb128_t;
+
+struct _Unwind_Context;
+struct _Unwind_Exception;
+typedef enum {
+  _URC_NO_REASON = 0,
+  _URC_FOREIGN_EXCEPTION_CAUGHT = 1,
+
+  _URC_FATAL_PHASE2_ERROR = 2,
+  _URC_FATAL_PHASE1_ERROR = 3,
+  _URC_NORMAL_STOP = 4,
+
+  _URC_END_OF_STACK = 5,
+  _URC_HANDLER_FOUND = 6,
+  _URC_INSTALL_CONTEXT = 7,
+  _URC_CONTINUE_UNWIND = 8
+} _Unwind_Reason_Code;
+
+typedef enum {
+  _UA_SEARCH_PHASE = 1,
+  _UA_CLEANUP_PHASE = 2,
+
+  _UA_HANDLER_FRAME = 4,
+  _UA_FORCE_UNWIND = 8,
+  _UA_END_OF_STACK = 16 /* gcc extension to C++ ABI */
+} _Unwind_Action;
+
+typedef void (*_Unwind_Exception_Cleanup_Fn)(_Unwind_Reason_Code,
+                                             struct _Unwind_Exception *);
+
+struct _Unwind_Exception {
+  _Unwind_Exception_Class exception_class;
+  _Unwind_Exception_Cleanup_Fn exception_cleanup;
+  _Unwind_Word private_1;
+  _Unwind_Word private_2;
+  /* The Itanium ABI requires that _Unwind_Exception objects are "double-word
+   * aligned".  GCC has interpreted this to mean "use the maximum useful
+   * alignment for the target"; so do we. */
+} __attribute__((__aligned__));
+
+typedef _Unwind_Reason_Code (*_Unwind_Stop_Fn)(int, _Unwind_Action,
+                                               _Unwind_Exception_Class,
+                                               struct _Unwind_Exception *,
+                                               struct _Unwind_Context *,
+                                               void *);
+
+typedef _Unwind_Reason_Code (*_Unwind_Personality_Fn)(
+    int, _Unwind_Action, _Unwind_Exception_Class, struct _Unwind_Exception *,
+    struct _Unwind_Context *);
+typedef _Unwind_Personality_Fn __personality_routine;
+
+typedef _Unwind_Reason_Code (*_Unwind_Trace_Fn)(struct _Unwind_Context *,
+                                                void *);
+
+#if defined(__arm__) && !defined(__APPLE__)
+
+typedef enum {
+  _UVRSC_CORE = 0,        /* integer register */
+  _UVRSC_VFP = 1,         /* vfp */
+  _UVRSC_WMMXD = 3,       /* Intel WMMX data register */
+  _UVRSC_WMMXC = 4        /* Intel WMMX control register */
+} _Unwind_VRS_RegClass;
+
+typedef enum {
+  _UVRSD_UINT32 = 0,
+  _UVRSD_VFPX = 1,
+  _UVRSD_UINT64 = 3,
+  _UVRSD_FLOAT = 4,
+  _UVRSD_DOUBLE = 5
+} _Unwind_VRS_DataRepresentation;
+
+typedef enum {
+  _UVRSR_OK = 0,
+  _UVRSR_NOT_IMPLEMENTED = 1,
+  _UVRSR_FAILED = 2
+} _Unwind_VRS_Result;
+
+_Unwind_VRS_Result _Unwind_VRS_Get(struct _Unwind_Context *__context,
+  _Unwind_VRS_RegClass __regclass,
+  uint32_t __regno,
+  _Unwind_VRS_DataRepresentation __representation,
+  void *__valuep);
+
+_Unwind_VRS_Result _Unwind_VRS_Set(struct _Unwind_Context *__context,
+  _Unwind_VRS_RegClass __regclass,
+  uint32_t __regno,
+  _Unwind_VRS_DataRepresentation __representation,
+  void *__valuep);
+
+static __inline__
+_Unwind_Word _Unwind_GetGR(struct _Unwind_Context *__context, int __index) {
+  _Unwind_Word __value;
+  _Unwind_VRS_Get(__context, _UVRSC_CORE, __index, _UVRSD_UINT32, &__value);
+  return __value;
+}
+
+static __inline__
+void _Unwind_SetGR(struct _Unwind_Context *__context, int __index,
+                   _Unwind_Word __value) {
+  _Unwind_VRS_Set(__context, _UVRSC_CORE, __index, _UVRSD_UINT32, &__value);
+}
+
+static __inline__
+_Unwind_Word _Unwind_GetIP(struct _Unwind_Context *__context) {
+  _Unwind_Word __ip = _Unwind_GetGR(__context, 15);
+  return __ip & ~(_Unwind_Word)(0x1); /* Remove thumb mode bit. */
+}
+
+static __inline__
+void _Unwind_SetIP(struct _Unwind_Context *__context, _Unwind_Word __value) {
+  _Unwind_Word __thumb_mode_bit = _Unwind_GetGR(__context, 15) & 0x1;
+  _Unwind_SetGR(__context, 15, __value | __thumb_mode_bit);
+}
+#else
+_Unwind_Word _Unwind_GetGR(struct _Unwind_Context *, int);
+void _Unwind_SetGR(struct _Unwind_Context *, int, _Unwind_Word);
+
+_Unwind_Word _Unwind_GetIP(struct _Unwind_Context *);
+void _Unwind_SetIP(struct _Unwind_Context *, _Unwind_Word);
+#endif
+
+
+_Unwind_Word _Unwind_GetIPInfo(struct _Unwind_Context *, int *);
+
+_Unwind_Word _Unwind_GetCFA(struct _Unwind_Context *);
+
+void *_Unwind_GetLanguageSpecificData(struct _Unwind_Context *);
+
+_Unwind_Ptr _Unwind_GetRegionStart(struct _Unwind_Context *);
+
+/* DWARF EH functions; currently not available on Darwin/ARM */
+#if !defined(__APPLE__) || !defined(__arm__)
+
+_Unwind_Reason_Code _Unwind_RaiseException(struct _Unwind_Exception *);
+_Unwind_Reason_Code _Unwind_ForcedUnwind(struct _Unwind_Exception *,
+                                         _Unwind_Stop_Fn, void *);
+void _Unwind_DeleteException(struct _Unwind_Exception *);
+void _Unwind_Resume(struct _Unwind_Exception *);
+_Unwind_Reason_Code _Unwind_Resume_or_Rethrow(struct _Unwind_Exception *);
+
+#endif
+
+_Unwind_Reason_Code _Unwind_Backtrace(_Unwind_Trace_Fn, void *);
+
+/* setjmp(3)/longjmp(3) stuff */
+typedef struct SjLj_Function_Context *_Unwind_FunctionContext_t;
+
+void _Unwind_SjLj_Register(_Unwind_FunctionContext_t);
+void _Unwind_SjLj_Unregister(_Unwind_FunctionContext_t);
+_Unwind_Reason_Code _Unwind_SjLj_RaiseException(struct _Unwind_Exception *);
+_Unwind_Reason_Code _Unwind_SjLj_ForcedUnwind(struct _Unwind_Exception *,
+                                              _Unwind_Stop_Fn, void *);
+void _Unwind_SjLj_Resume(struct _Unwind_Exception *);
+_Unwind_Reason_Code _Unwind_SjLj_Resume_or_Rethrow(struct _Unwind_Exception *);
+
+void *_Unwind_FindEnclosingFunction(void *);
+
+#ifdef __APPLE__
+
+_Unwind_Ptr _Unwind_GetDataRelBase(struct _Unwind_Context *)
+    __attribute__((unavailable));
+_Unwind_Ptr _Unwind_GetTextRelBase(struct _Unwind_Context *)
+    __attribute__((unavailable));
+
+/* Darwin-specific functions */
+void __register_frame(const void *);
+void __deregister_frame(const void *);
+
+struct dwarf_eh_bases {
+  uintptr_t tbase;
+  uintptr_t dbase;
+  uintptr_t func;
+};
+void *_Unwind_Find_FDE(const void *, struct dwarf_eh_bases *);
+
+void __register_frame_info_bases(const void *, void *, void *, void *)
+  __attribute__((unavailable));
+void __register_frame_info(const void *, void *) __attribute__((unavailable));
+void __register_frame_info_table_bases(const void *, void*, void *, void *)
+  __attribute__((unavailable));
+void __register_frame_info_table(const void *, void *)
+  __attribute__((unavailable));
+void __register_frame_table(const void *) __attribute__((unavailable));
+void __deregister_frame_info(const void *) __attribute__((unavailable));
+void __deregister_frame_info_bases(const void *)__attribute__((unavailable));
+
+#else
+
+_Unwind_Ptr _Unwind_GetDataRelBase(struct _Unwind_Context *);
+_Unwind_Ptr _Unwind_GetTextRelBase(struct _Unwind_Context *);
+
+#endif
+
+
+#ifndef HIDE_EXPORTS
+#pragma GCC visibility pop
+#endif
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif
+
+#endif /* __CLANG_UNWIND_H */
diff --git a/lib/clang/3.6/include/vadefs.h b/lib/clang/3.6/include/vadefs.h
new file mode 100644
index 0000000..7fe9a74
--- /dev/null
+++ b/lib/clang/3.6/include/vadefs.h
@@ -0,0 +1,65 @@
+/* ===-------- vadefs.h ---------------------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+/* Only include this if we are aiming for MSVC compatibility. */
+#ifndef _MSC_VER
+#include_next <vadefs.h>
+#else
+
+#ifndef __clang_vadefs_h
+#define __clang_vadefs_h
+
+#include_next <vadefs.h>
+
+/* Override macros from vadefs.h with definitions that work with Clang. */
+#ifdef _crt_va_start
+#undef _crt_va_start
+#define _crt_va_start(ap, param) __builtin_va_start(ap, param)
+#endif
+#ifdef _crt_va_end
+#undef _crt_va_end
+#define _crt_va_end(ap)          __builtin_va_end(ap)
+#endif
+#ifdef _crt_va_arg
+#undef _crt_va_arg
+#define _crt_va_arg(ap, type)    __builtin_va_arg(ap, type)
+#endif
+
+/* VS 2015 switched to double underscore names, which is an improvement, but now
+ * we have to intercept those names too.
+ */
+#ifdef __crt_va_start
+#undef __crt_va_start
+#define __crt_va_start(ap, param) __builtin_va_start(ap, param)
+#endif
+#ifdef __crt_va_end
+#undef __crt_va_end
+#define __crt_va_end(ap)          __builtin_va_end(ap)
+#endif
+#ifdef __crt_va_arg
+#undef __crt_va_arg
+#define __crt_va_arg(ap, type)    __builtin_va_arg(ap, type)
+#endif
+
+#endif
+#endif
diff --git a/lib/clang/3.6/include/varargs.h b/lib/clang/3.6/include/varargs.h
new file mode 100644
index 0000000..b5477d0
--- /dev/null
+++ b/lib/clang/3.6/include/varargs.h
@@ -0,0 +1,26 @@
+/*===---- varargs.h - Variable argument handling -------------------------------------===
+*
+* Permission is hereby granted, free of charge, to any person obtaining a copy
+* of this software and associated documentation files (the "Software"), to deal
+* in the Software without restriction, including without limitation the rights
+* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+* copies of the Software, and to permit persons to whom the Software is
+* furnished to do so, subject to the following conditions:
+*
+* The above copyright notice and this permission notice shall be included in
+* all copies or substantial portions of the Software.
+*
+* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+* THE SOFTWARE.
+*
+*===-----------------------------------------------------------------------===
+*/
+#ifndef __VARARGS_H
+#define __VARARGS_H
+  #error "Please use <stdarg.h> instead of <varargs.h>"
+#endif
diff --git a/lib/clang/3.6/include/wmmintrin.h b/lib/clang/3.6/include/wmmintrin.h
new file mode 100644
index 0000000..369e3c2
--- /dev/null
+++ b/lib/clang/3.6/include/wmmintrin.h
@@ -0,0 +1,42 @@
+/*===---- wmmintrin.h - AES intrinsics ------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef _WMMINTRIN_H
+#define _WMMINTRIN_H
+
+#include <emmintrin.h>
+
+#if !defined (__AES__) && !defined (__PCLMUL__)
+# error "AES/PCLMUL instructions not enabled"
+#else
+
+#ifdef __AES__
+#include <__wmmintrin_aes.h>
+#endif /* __AES__ */
+
+#ifdef __PCLMUL__
+#include <__wmmintrin_pclmul.h>
+#endif /* __PCLMUL__ */
+
+#endif /* __AES__ || __PCLMUL__ */
+#endif /* _WMMINTRIN_H */
diff --git a/lib/clang/3.6/include/x86intrin.h b/lib/clang/3.6/include/x86intrin.h
new file mode 100644
index 0000000..21a43da
--- /dev/null
+++ b/lib/clang/3.6/include/x86intrin.h
@@ -0,0 +1,81 @@
+/*===---- x86intrin.h - X86 intrinsics -------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __X86INTRIN_H
+#define __X86INTRIN_H
+
+#include <ia32intrin.h>
+
+#include <immintrin.h>
+
+#ifdef __3dNOW__
+#include <mm3dnow.h>
+#endif
+
+#ifdef __BMI__
+#include <bmiintrin.h>
+#endif
+
+#ifdef __BMI2__
+#include <bmi2intrin.h>
+#endif
+
+#ifdef __LZCNT__
+#include <lzcntintrin.h>
+#endif
+
+#ifdef __POPCNT__
+#include <popcntintrin.h>
+#endif
+
+#ifdef __RDSEED__
+#include <rdseedintrin.h>
+#endif
+
+#ifdef __PRFCHW__
+#include <prfchwintrin.h>
+#endif
+
+#ifdef __SSE4A__
+#include <ammintrin.h>
+#endif
+
+#ifdef __FMA4__
+#include <fma4intrin.h>
+#endif
+
+#ifdef __XOP__
+#include <xopintrin.h>
+#endif
+
+#ifdef __TBM__
+#include <tbmintrin.h>
+#endif
+
+#ifdef __F16C__
+#include <f16cintrin.h>
+#endif
+
+/* FIXME: LWP */
+
+#endif /* __X86INTRIN_H */
diff --git a/lib/clang/3.6/include/xmmintrin.h b/lib/clang/3.6/include/xmmintrin.h
new file mode 100644
index 0000000..c9befcb
--- /dev/null
+++ b/lib/clang/3.6/include/xmmintrin.h
@@ -0,0 +1,1003 @@
+/*===---- xmmintrin.h - SSE intrinsics -------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+ 
+#ifndef __XMMINTRIN_H
+#define __XMMINTRIN_H
+ 
+#ifndef __SSE__
+#error "SSE instruction set not enabled"
+#else
+
+#include <mmintrin.h>
+
+typedef int __v4si __attribute__((__vector_size__(16)));
+typedef float __v4sf __attribute__((__vector_size__(16)));
+typedef float __m128 __attribute__((__vector_size__(16)));
+
+/* This header should only be included in a hosted environment as it depends on
+ * a standard library to provide allocation routines. */
+#if __STDC_HOSTED__
+#include <mm_malloc.h>
+#endif
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_add_ss(__m128 __a, __m128 __b)
+{
+  __a[0] += __b[0];
+  return __a;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_add_ps(__m128 __a, __m128 __b)
+{
+  return __a + __b;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_sub_ss(__m128 __a, __m128 __b)
+{
+  __a[0] -= __b[0];
+  return __a;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_sub_ps(__m128 __a, __m128 __b)
+{
+  return __a - __b;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_mul_ss(__m128 __a, __m128 __b)
+{
+  __a[0] *= __b[0];
+  return __a;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_mul_ps(__m128 __a, __m128 __b)
+{
+  return __a * __b;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_div_ss(__m128 __a, __m128 __b)
+{
+  __a[0] /= __b[0];
+  return __a;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_div_ps(__m128 __a, __m128 __b)
+{
+  return __a / __b;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_sqrt_ss(__m128 __a)
+{
+  __m128 __c = __builtin_ia32_sqrtss(__a);
+  return (__m128) { __c[0], __a[1], __a[2], __a[3] };
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_sqrt_ps(__m128 __a)
+{
+  return __builtin_ia32_sqrtps(__a);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_rcp_ss(__m128 __a)
+{
+  __m128 __c = __builtin_ia32_rcpss(__a);
+  return (__m128) { __c[0], __a[1], __a[2], __a[3] };
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_rcp_ps(__m128 __a)
+{
+  return __builtin_ia32_rcpps(__a);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_rsqrt_ss(__m128 __a)
+{
+  __m128 __c = __builtin_ia32_rsqrtss(__a);
+  return (__m128) { __c[0], __a[1], __a[2], __a[3] };
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_rsqrt_ps(__m128 __a)
+{
+  return __builtin_ia32_rsqrtps(__a);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_min_ss(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_minss(__a, __b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_min_ps(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_minps(__a, __b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_max_ss(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_maxss(__a, __b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_max_ps(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_maxps(__a, __b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_and_ps(__m128 __a, __m128 __b)
+{
+  return (__m128)((__v4si)__a & (__v4si)__b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_andnot_ps(__m128 __a, __m128 __b)
+{
+  return (__m128)(~(__v4si)__a & (__v4si)__b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_or_ps(__m128 __a, __m128 __b)
+{
+  return (__m128)((__v4si)__a | (__v4si)__b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_xor_ps(__m128 __a, __m128 __b)
+{
+  return (__m128)((__v4si)__a ^ (__v4si)__b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_ss(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_ia32_cmpss(__a, __b, 0);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpeq_ps(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_ia32_cmpps(__a, __b, 0);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmplt_ss(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_ia32_cmpss(__a, __b, 1);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmplt_ps(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_ia32_cmpps(__a, __b, 1);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmple_ss(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_ia32_cmpss(__a, __b, 2);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmple_ps(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_ia32_cmpps(__a, __b, 2);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_ss(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_shufflevector(__a,
+                                         __builtin_ia32_cmpss(__b, __a, 1),
+                                         4, 1, 2, 3);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpgt_ps(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_ia32_cmpps(__b, __a, 1);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpge_ss(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_shufflevector(__a,
+                                         __builtin_ia32_cmpss(__b, __a, 2),
+                                         4, 1, 2, 3);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpge_ps(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_ia32_cmpps(__b, __a, 2);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpneq_ss(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_ia32_cmpss(__a, __b, 4);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpneq_ps(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_ia32_cmpps(__a, __b, 4);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnlt_ss(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_ia32_cmpss(__a, __b, 5);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnlt_ps(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_ia32_cmpps(__a, __b, 5);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnle_ss(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_ia32_cmpss(__a, __b, 6);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnle_ps(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_ia32_cmpps(__a, __b, 6);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpngt_ss(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_shufflevector(__a,
+                                         __builtin_ia32_cmpss(__b, __a, 5),
+                                         4, 1, 2, 3);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpngt_ps(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_ia32_cmpps(__b, __a, 5);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnge_ss(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_shufflevector(__a,
+                                         __builtin_ia32_cmpss(__b, __a, 6),
+                                         4, 1, 2, 3);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpnge_ps(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_ia32_cmpps(__b, __a, 6);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpord_ss(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_ia32_cmpss(__a, __b, 7);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpord_ps(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_ia32_cmpps(__a, __b, 7);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpunord_ss(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_ia32_cmpss(__a, __b, 3);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cmpunord_ps(__m128 __a, __m128 __b)
+{
+  return (__m128)__builtin_ia32_cmpps(__a, __b, 3);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comieq_ss(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_comieq(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comilt_ss(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_comilt(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comile_ss(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_comile(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comigt_ss(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_comigt(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comige_ss(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_comige(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_comineq_ss(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_comineq(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomieq_ss(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_ucomieq(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomilt_ss(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_ucomilt(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomile_ss(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_ucomile(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomigt_ss(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_ucomigt(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomige_ss(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_ucomige(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_ucomineq_ss(__m128 __a, __m128 __b)
+{
+  return __builtin_ia32_ucomineq(__a, __b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_cvtss_si32(__m128 __a)
+{
+  return __builtin_ia32_cvtss2si(__a);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_cvt_ss2si(__m128 __a)
+{
+  return _mm_cvtss_si32(__a);
+}
+
+#ifdef __x86_64__
+
+static __inline__ long long __attribute__((__always_inline__, __nodebug__))
+_mm_cvtss_si64(__m128 __a)
+{
+  return __builtin_ia32_cvtss2si64(__a);
+}
+
+#endif
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtps_pi32(__m128 __a)
+{
+  return (__m64)__builtin_ia32_cvtps2pi(__a);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cvt_ps2pi(__m128 __a)
+{
+  return _mm_cvtps_pi32(__a);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_cvttss_si32(__m128 __a)
+{
+  return __a[0];
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_cvtt_ss2si(__m128 __a)
+{
+  return _mm_cvttss_si32(__a);
+}
+
+static __inline__ long long __attribute__((__always_inline__, __nodebug__))
+_mm_cvttss_si64(__m128 __a)
+{
+  return __a[0];
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cvttps_pi32(__m128 __a)
+{
+  return (__m64)__builtin_ia32_cvttps2pi(__a);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtt_ps2pi(__m128 __a)
+{
+  return _mm_cvttps_pi32(__a);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi32_ss(__m128 __a, int __b)
+{
+  __a[0] = __b;
+  return __a;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvt_si2ss(__m128 __a, int __b)
+{
+  return _mm_cvtsi32_ss(__a, __b);
+}
+
+#ifdef __x86_64__
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtsi64_ss(__m128 __a, long long __b)
+{
+  __a[0] = __b;
+  return __a;
+}
+
+#endif
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtpi32_ps(__m128 __a, __m64 __b)
+{
+  return __builtin_ia32_cvtpi2ps(__a, (__v2si)__b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvt_pi2ps(__m128 __a, __m64 __b)
+{
+  return _mm_cvtpi32_ps(__a, __b);
+}
+
+static __inline__ float __attribute__((__always_inline__, __nodebug__))
+_mm_cvtss_f32(__m128 __a)
+{
+  return __a[0];
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_loadh_pi(__m128 __a, const __m64 *__p)
+{
+  typedef float __mm_loadh_pi_v2f32 __attribute__((__vector_size__(8)));
+  struct __mm_loadh_pi_struct {
+    __mm_loadh_pi_v2f32 __u;
+  } __attribute__((__packed__, __may_alias__));
+  __mm_loadh_pi_v2f32 __b = ((struct __mm_loadh_pi_struct*)__p)->__u;
+  __m128 __bb = __builtin_shufflevector(__b, __b, 0, 1, 0, 1);
+  return __builtin_shufflevector(__a, __bb, 0, 1, 4, 5);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_loadl_pi(__m128 __a, const __m64 *__p)
+{
+  typedef float __mm_loadl_pi_v2f32 __attribute__((__vector_size__(8)));
+  struct __mm_loadl_pi_struct {
+    __mm_loadl_pi_v2f32 __u;
+  } __attribute__((__packed__, __may_alias__));
+  __mm_loadl_pi_v2f32 __b = ((struct __mm_loadl_pi_struct*)__p)->__u;
+  __m128 __bb = __builtin_shufflevector(__b, __b, 0, 1, 0, 1);
+  return __builtin_shufflevector(__a, __bb, 4, 5, 2, 3);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_load_ss(const float *__p)
+{
+  struct __mm_load_ss_struct {
+    float __u;
+  } __attribute__((__packed__, __may_alias__));
+  float __u = ((struct __mm_load_ss_struct*)__p)->__u;
+  return (__m128){ __u, 0, 0, 0 };
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_load1_ps(const float *__p)
+{
+  struct __mm_load1_ps_struct {
+    float __u;
+  } __attribute__((__packed__, __may_alias__));
+  float __u = ((struct __mm_load1_ps_struct*)__p)->__u;
+  return (__m128){ __u, __u, __u, __u };
+}
+
+#define        _mm_load_ps1(p) _mm_load1_ps(p)
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_load_ps(const float *__p)
+{
+  return *(__m128*)__p;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_loadu_ps(const float *__p)
+{
+  struct __loadu_ps {
+    __m128 __v;
+  } __attribute__((__packed__, __may_alias__));
+  return ((struct __loadu_ps*)__p)->__v;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_loadr_ps(const float *__p)
+{
+  __m128 __a = _mm_load_ps(__p);
+  return __builtin_shufflevector(__a, __a, 3, 2, 1, 0);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_set_ss(float __w)
+{
+  return (__m128){ __w, 0, 0, 0 };
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_set1_ps(float __w)
+{
+  return (__m128){ __w, __w, __w, __w };
+}
+
+/* Microsoft specific. */
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_set_ps1(float __w)
+{
+    return _mm_set1_ps(__w);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_set_ps(float __z, float __y, float __x, float __w)
+{
+  return (__m128){ __w, __x, __y, __z };
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_setr_ps(float __z, float __y, float __x, float __w)
+{
+  return (__m128){ __z, __y, __x, __w };
+}
+
+static __inline__ __m128 __attribute__((__always_inline__))
+_mm_setzero_ps(void)
+{
+  return (__m128){ 0, 0, 0, 0 };
+}
+
+static __inline__ void __attribute__((__always_inline__))
+_mm_storeh_pi(__m64 *__p, __m128 __a)
+{
+  __builtin_ia32_storehps((__v2si *)__p, __a);
+}
+
+static __inline__ void __attribute__((__always_inline__))
+_mm_storel_pi(__m64 *__p, __m128 __a)
+{
+  __builtin_ia32_storelps((__v2si *)__p, __a);
+}
+
+static __inline__ void __attribute__((__always_inline__))
+_mm_store_ss(float *__p, __m128 __a)
+{
+  struct __mm_store_ss_struct {
+    float __u;
+  } __attribute__((__packed__, __may_alias__));
+  ((struct __mm_store_ss_struct*)__p)->__u = __a[0];
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_storeu_ps(float *__p, __m128 __a)
+{
+  __builtin_ia32_storeups(__p, __a);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_store1_ps(float *__p, __m128 __a)
+{
+  __a = __builtin_shufflevector(__a, __a, 0, 0, 0, 0);
+  _mm_storeu_ps(__p, __a);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_store_ps1(float *__p, __m128 __a)
+{
+    return _mm_store1_ps(__p, __a);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_store_ps(float *__p, __m128 __a)
+{
+  *(__m128 *)__p = __a;
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_storer_ps(float *__p, __m128 __a)
+{
+  __a = __builtin_shufflevector(__a, __a, 3, 2, 1, 0);
+  _mm_store_ps(__p, __a);
+}
+
+#define _MM_HINT_T0 3
+#define _MM_HINT_T1 2
+#define _MM_HINT_T2 1
+#define _MM_HINT_NTA 0
+
+#ifndef _MSC_VER
+/* FIXME: We have to #define this because "sel" must be a constant integer, and
+   Sema doesn't do any form of constant propagation yet. */
+
+#define _mm_prefetch(a, sel) (__builtin_prefetch((void *)(a), 0, (sel)))
+#endif
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_stream_pi(__m64 *__p, __m64 __a)
+{
+  __builtin_ia32_movntq(__p, __a);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_stream_ps(float *__p, __m128 __a)
+{
+  __builtin_ia32_movntps(__p, __a);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_sfence(void)
+{
+  __builtin_ia32_sfence();
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_extract_pi16(__m64 __a, int __n)
+{
+  __v4hi __b = (__v4hi)__a;
+  return (unsigned short)__b[__n & 3];
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_insert_pi16(__m64 __a, int __d, int __n)
+{
+   __v4hi __b = (__v4hi)__a;
+   __b[__n & 3] = __d;
+   return (__m64)__b;
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_max_pi16(__m64 __a, __m64 __b)
+{
+  return (__m64)__builtin_ia32_pmaxsw((__v4hi)__a, (__v4hi)__b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_max_pu8(__m64 __a, __m64 __b)
+{
+  return (__m64)__builtin_ia32_pmaxub((__v8qi)__a, (__v8qi)__b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_min_pi16(__m64 __a, __m64 __b)
+{
+  return (__m64)__builtin_ia32_pminsw((__v4hi)__a, (__v4hi)__b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_min_pu8(__m64 __a, __m64 __b)
+{
+  return (__m64)__builtin_ia32_pminub((__v8qi)__a, (__v8qi)__b);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_movemask_pi8(__m64 __a)
+{
+  return __builtin_ia32_pmovmskb((__v8qi)__a);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_mulhi_pu16(__m64 __a, __m64 __b)
+{
+  return (__m64)__builtin_ia32_pmulhuw((__v4hi)__a, (__v4hi)__b);
+}
+
+#define _mm_shuffle_pi16(a, n) __extension__ ({ \
+  __m64 __a = (a); \
+  (__m64)__builtin_ia32_pshufw((__v4hi)__a, (n)); })
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_maskmove_si64(__m64 __d, __m64 __n, char *__p)
+{
+  __builtin_ia32_maskmovq((__v8qi)__d, (__v8qi)__n, __p);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_avg_pu8(__m64 __a, __m64 __b)
+{
+  return (__m64)__builtin_ia32_pavgb((__v8qi)__a, (__v8qi)__b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_avg_pu16(__m64 __a, __m64 __b)
+{
+  return (__m64)__builtin_ia32_pavgw((__v4hi)__a, (__v4hi)__b);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_sad_pu8(__m64 __a, __m64 __b)
+{
+  return (__m64)__builtin_ia32_psadbw((__v8qi)__a, (__v8qi)__b);
+}
+
+static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
+_mm_getcsr(void)
+{
+  return __builtin_ia32_stmxcsr();
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__))
+_mm_setcsr(unsigned int __i)
+{
+  __builtin_ia32_ldmxcsr(__i);
+}
+
+#define _mm_shuffle_ps(a, b, mask) __extension__ ({ \
+  __m128 __a = (a); \
+  __m128 __b = (b); \
+  (__m128)__builtin_shufflevector((__v4sf)__a, (__v4sf)__b, \
+                                  (mask) & 0x3, ((mask) & 0xc) >> 2, \
+                                  (((mask) & 0x30) >> 4) + 4, \
+                                  (((mask) & 0xc0) >> 6) + 4); })
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_unpackhi_ps(__m128 __a, __m128 __b)
+{
+  return __builtin_shufflevector(__a, __b, 2, 6, 3, 7);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_unpacklo_ps(__m128 __a, __m128 __b)
+{
+  return __builtin_shufflevector(__a, __b, 0, 4, 1, 5);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_move_ss(__m128 __a, __m128 __b)
+{
+  return __builtin_shufflevector(__a, __b, 4, 1, 2, 3);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_movehl_ps(__m128 __a, __m128 __b)
+{
+  return __builtin_shufflevector(__a, __b, 6, 7, 2, 3);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_movelh_ps(__m128 __a, __m128 __b)
+{
+  return __builtin_shufflevector(__a, __b, 0, 1, 4, 5);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtpi16_ps(__m64 __a)
+{
+  __m64 __b, __c;
+  __m128 __r;
+
+  __b = _mm_setzero_si64();
+  __b = _mm_cmpgt_pi16(__b, __a);
+  __c = _mm_unpackhi_pi16(__a, __b);
+  __r = _mm_setzero_ps();
+  __r = _mm_cvtpi32_ps(__r, __c);
+  __r = _mm_movelh_ps(__r, __r);
+  __c = _mm_unpacklo_pi16(__a, __b);
+  __r = _mm_cvtpi32_ps(__r, __c);
+
+  return __r;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtpu16_ps(__m64 __a)
+{
+  __m64 __b, __c;
+  __m128 __r;
+
+  __b = _mm_setzero_si64();
+  __c = _mm_unpackhi_pi16(__a, __b);
+  __r = _mm_setzero_ps();
+  __r = _mm_cvtpi32_ps(__r, __c);
+  __r = _mm_movelh_ps(__r, __r);
+  __c = _mm_unpacklo_pi16(__a, __b);
+  __r = _mm_cvtpi32_ps(__r, __c);
+
+  return __r;
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtpi8_ps(__m64 __a)
+{
+  __m64 __b;
+  
+  __b = _mm_setzero_si64();
+  __b = _mm_cmpgt_pi8(__b, __a);
+  __b = _mm_unpacklo_pi8(__a, __b);
+
+  return _mm_cvtpi16_ps(__b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtpu8_ps(__m64 __a)
+{
+  __m64 __b;
+  
+  __b = _mm_setzero_si64();
+  __b = _mm_unpacklo_pi8(__a, __b);
+
+  return _mm_cvtpi16_ps(__b);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtpi32x2_ps(__m64 __a, __m64 __b)
+{
+  __m128 __c;
+  
+  __c = _mm_setzero_ps();
+  __c = _mm_cvtpi32_ps(__c, __b);
+  __c = _mm_movelh_ps(__c, __c);
+
+  return _mm_cvtpi32_ps(__c, __a);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtps_pi16(__m128 __a)
+{
+  __m64 __b, __c;
+  
+  __b = _mm_cvtps_pi32(__a);
+  __a = _mm_movehl_ps(__a, __a);
+  __c = _mm_cvtps_pi32(__a);
+  
+  return _mm_packs_pi32(__b, __c);
+}
+
+static __inline__ __m64 __attribute__((__always_inline__, __nodebug__))
+_mm_cvtps_pi8(__m128 __a)
+{
+  __m64 __b, __c;
+  
+  __b = _mm_cvtps_pi16(__a);
+  __c = _mm_setzero_si64();
+  
+  return _mm_packs_pi16(__b, __c);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__))
+_mm_movemask_ps(__m128 __a)
+{
+  return __builtin_ia32_movmskps(__a);
+}
+
+#define _MM_SHUFFLE(z, y, x, w) (((z) << 6) | ((y) << 4) | ((x) << 2) | (w))
+
+#define _MM_EXCEPT_INVALID    (0x0001)
+#define _MM_EXCEPT_DENORM     (0x0002)
+#define _MM_EXCEPT_DIV_ZERO   (0x0004)
+#define _MM_EXCEPT_OVERFLOW   (0x0008)
+#define _MM_EXCEPT_UNDERFLOW  (0x0010)
+#define _MM_EXCEPT_INEXACT    (0x0020)
+#define _MM_EXCEPT_MASK       (0x003f)
+
+#define _MM_MASK_INVALID      (0x0080)
+#define _MM_MASK_DENORM       (0x0100)
+#define _MM_MASK_DIV_ZERO     (0x0200)
+#define _MM_MASK_OVERFLOW     (0x0400)
+#define _MM_MASK_UNDERFLOW    (0x0800)
+#define _MM_MASK_INEXACT      (0x1000)
+#define _MM_MASK_MASK         (0x1f80)
+
+#define _MM_ROUND_NEAREST     (0x0000)
+#define _MM_ROUND_DOWN        (0x2000)
+#define _MM_ROUND_UP          (0x4000)
+#define _MM_ROUND_TOWARD_ZERO (0x6000)
+#define _MM_ROUND_MASK        (0x6000)
+
+#define _MM_FLUSH_ZERO_MASK   (0x8000)
+#define _MM_FLUSH_ZERO_ON     (0x8000)
+#define _MM_FLUSH_ZERO_OFF    (0x0000)
+
+#define _MM_GET_EXCEPTION_MASK() (_mm_getcsr() & _MM_MASK_MASK)
+#define _MM_GET_EXCEPTION_STATE() (_mm_getcsr() & _MM_EXCEPT_MASK)
+#define _MM_GET_FLUSH_ZERO_MODE() (_mm_getcsr() & _MM_FLUSH_ZERO_MASK)
+#define _MM_GET_ROUNDING_MODE() (_mm_getcsr() & _MM_ROUND_MASK)
+
+#define _MM_SET_EXCEPTION_MASK(x) (_mm_setcsr((_mm_getcsr() & ~_MM_MASK_MASK) | (x)))
+#define _MM_SET_EXCEPTION_STATE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_EXCEPT_MASK) | (x)))
+#define _MM_SET_FLUSH_ZERO_MODE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_FLUSH_ZERO_MASK) | (x)))
+#define _MM_SET_ROUNDING_MODE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_ROUND_MASK) | (x)))
+
+#define _MM_TRANSPOSE4_PS(row0, row1, row2, row3) \
+do { \
+  __m128 tmp3, tmp2, tmp1, tmp0; \
+  tmp0 = _mm_unpacklo_ps((row0), (row1)); \
+  tmp2 = _mm_unpacklo_ps((row2), (row3)); \
+  tmp1 = _mm_unpackhi_ps((row0), (row1)); \
+  tmp3 = _mm_unpackhi_ps((row2), (row3)); \
+  (row0) = _mm_movelh_ps(tmp0, tmp2); \
+  (row1) = _mm_movehl_ps(tmp2, tmp0); \
+  (row2) = _mm_movelh_ps(tmp1, tmp3); \
+  (row3) = _mm_movehl_ps(tmp3, tmp1); \
+} while (0)
+
+/* Aliases for compatibility. */
+#define _m_pextrw _mm_extract_pi16
+#define _m_pinsrw _mm_insert_pi16
+#define _m_pmaxsw _mm_max_pi16
+#define _m_pmaxub _mm_max_pu8
+#define _m_pminsw _mm_min_pi16
+#define _m_pminub _mm_min_pu8
+#define _m_pmovmskb _mm_movemask_pi8
+#define _m_pmulhuw _mm_mulhi_pu16
+#define _m_pshufw _mm_shuffle_pi16
+#define _m_maskmovq _mm_maskmove_si64
+#define _m_pavgb _mm_avg_pu8
+#define _m_pavgw _mm_avg_pu16
+#define _m_psadbw _mm_sad_pu8
+#define _m_ _mm_
+#define _m_ _mm_
+
+/* Ugly hack for backwards-compatibility (compatible with gcc) */
+#ifdef __SSE2__
+#include <emmintrin.h>
+#endif
+
+#endif /* __SSE__ */
+
+#endif /* __XMMINTRIN_H */
diff --git a/lib/clang/3.6/include/xopintrin.h b/lib/clang/3.6/include/xopintrin.h
new file mode 100644
index 0000000..cc94ca0
--- /dev/null
+++ b/lib/clang/3.6/include/xopintrin.h
@@ -0,0 +1,804 @@
+/*===---- xopintrin.h - XOP intrinsics -------------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __X86INTRIN_H
+#error "Never use <xopintrin.h> directly; include <x86intrin.h> instead."
+#endif
+
+#ifndef __XOPINTRIN_H
+#define __XOPINTRIN_H
+
+#ifndef __XOP__
+# error "XOP instruction set is not enabled"
+#else
+
+#include <fma4intrin.h>
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_maccs_epi16(__m128i __A, __m128i __B, __m128i __C)
+{
+  return (__m128i)__builtin_ia32_vpmacssww((__v8hi)__A, (__v8hi)__B, (__v8hi)__C);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_macc_epi16(__m128i __A, __m128i __B, __m128i __C)
+{
+  return (__m128i)__builtin_ia32_vpmacsww((__v8hi)__A, (__v8hi)__B, (__v8hi)__C);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_maccsd_epi16(__m128i __A, __m128i __B, __m128i __C)
+{
+  return (__m128i)__builtin_ia32_vpmacsswd((__v8hi)__A, (__v8hi)__B, (__v4si)__C);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_maccd_epi16(__m128i __A, __m128i __B, __m128i __C)
+{
+  return (__m128i)__builtin_ia32_vpmacswd((__v8hi)__A, (__v8hi)__B, (__v4si)__C);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_maccs_epi32(__m128i __A, __m128i __B, __m128i __C)
+{
+  return (__m128i)__builtin_ia32_vpmacssdd((__v4si)__A, (__v4si)__B, (__v4si)__C);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_macc_epi32(__m128i __A, __m128i __B, __m128i __C)
+{
+  return (__m128i)__builtin_ia32_vpmacsdd((__v4si)__A, (__v4si)__B, (__v4si)__C);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_maccslo_epi32(__m128i __A, __m128i __B, __m128i __C)
+{
+  return (__m128i)__builtin_ia32_vpmacssdql((__v4si)__A, (__v4si)__B, (__v2di)__C);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_macclo_epi32(__m128i __A, __m128i __B, __m128i __C)
+{
+  return (__m128i)__builtin_ia32_vpmacsdql((__v4si)__A, (__v4si)__B, (__v2di)__C);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_maccshi_epi32(__m128i __A, __m128i __B, __m128i __C)
+{
+  return (__m128i)__builtin_ia32_vpmacssdqh((__v4si)__A, (__v4si)__B, (__v2di)__C);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_macchi_epi32(__m128i __A, __m128i __B, __m128i __C)
+{
+  return (__m128i)__builtin_ia32_vpmacsdqh((__v4si)__A, (__v4si)__B, (__v2di)__C);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_maddsd_epi16(__m128i __A, __m128i __B, __m128i __C)
+{
+  return (__m128i)__builtin_ia32_vpmadcsswd((__v8hi)__A, (__v8hi)__B, (__v4si)__C);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_maddd_epi16(__m128i __A, __m128i __B, __m128i __C)
+{
+  return (__m128i)__builtin_ia32_vpmadcswd((__v8hi)__A, (__v8hi)__B, (__v4si)__C);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_haddw_epi8(__m128i __A)
+{
+  return (__m128i)__builtin_ia32_vphaddbw((__v16qi)__A);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_haddd_epi8(__m128i __A)
+{
+  return (__m128i)__builtin_ia32_vphaddbd((__v16qi)__A);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_haddq_epi8(__m128i __A)
+{
+  return (__m128i)__builtin_ia32_vphaddbq((__v16qi)__A);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_haddd_epi16(__m128i __A)
+{
+  return (__m128i)__builtin_ia32_vphaddwd((__v8hi)__A);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_haddq_epi16(__m128i __A)
+{
+  return (__m128i)__builtin_ia32_vphaddwq((__v8hi)__A);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_haddq_epi32(__m128i __A)
+{
+  return (__m128i)__builtin_ia32_vphadddq((__v4si)__A);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_haddw_epu8(__m128i __A)
+{
+  return (__m128i)__builtin_ia32_vphaddubw((__v16qi)__A);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_haddd_epu8(__m128i __A)
+{
+  return (__m128i)__builtin_ia32_vphaddubd((__v16qi)__A);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_haddq_epu8(__m128i __A)
+{
+  return (__m128i)__builtin_ia32_vphaddubq((__v16qi)__A);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_haddd_epu16(__m128i __A)
+{
+  return (__m128i)__builtin_ia32_vphadduwd((__v8hi)__A);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_haddq_epu16(__m128i __A)
+{
+  return (__m128i)__builtin_ia32_vphadduwq((__v8hi)__A);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_haddq_epu32(__m128i __A)
+{
+  return (__m128i)__builtin_ia32_vphaddudq((__v4si)__A);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_hsubw_epi8(__m128i __A)
+{
+  return (__m128i)__builtin_ia32_vphsubbw((__v16qi)__A);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_hsubd_epi16(__m128i __A)
+{
+  return (__m128i)__builtin_ia32_vphsubwd((__v8hi)__A);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_hsubq_epi32(__m128i __A)
+{
+  return (__m128i)__builtin_ia32_vphsubdq((__v4si)__A);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_cmov_si128(__m128i __A, __m128i __B, __m128i __C)
+{
+  return (__m128i)__builtin_ia32_vpcmov(__A, __B, __C);
+}
+
+static __inline__ __m256i __attribute__((__always_inline__, __nodebug__))
+_mm256_cmov_si256(__m256i __A, __m256i __B, __m256i __C)
+{
+  return (__m256i)__builtin_ia32_vpcmov_256(__A, __B, __C);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_perm_epi8(__m128i __A, __m128i __B, __m128i __C)
+{
+  return (__m128i)__builtin_ia32_vpperm((__v16qi)__A, (__v16qi)__B, (__v16qi)__C);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_rot_epi8(__m128i __A, __m128i __B)
+{
+  return (__m128i)__builtin_ia32_vprotb((__v16qi)__A, (__v16qi)__B);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_rot_epi16(__m128i __A, __m128i __B)
+{
+  return (__m128i)__builtin_ia32_vprotw((__v8hi)__A, (__v8hi)__B);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_rot_epi32(__m128i __A, __m128i __B)
+{
+  return (__m128i)__builtin_ia32_vprotd((__v4si)__A, (__v4si)__B);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_rot_epi64(__m128i __A, __m128i __B)
+{
+  return (__m128i)__builtin_ia32_vprotq((__v2di)__A, (__v2di)__B);
+}
+
+#define _mm_roti_epi8(A, N) __extension__ ({ \
+  __m128i __A = (A); \
+  (__m128i)__builtin_ia32_vprotbi((__v16qi)__A, (N)); })
+
+#define _mm_roti_epi16(A, N) __extension__ ({ \
+  __m128i __A = (A); \
+  (__m128i)__builtin_ia32_vprotwi((__v8hi)__A, (N)); })
+
+#define _mm_roti_epi32(A, N) __extension__ ({ \
+  __m128i __A = (A); \
+  (__m128i)__builtin_ia32_vprotdi((__v4si)__A, (N)); })
+
+#define _mm_roti_epi64(A, N) __extension__ ({ \
+  __m128i __A = (A); \
+  (__m128i)__builtin_ia32_vprotqi((__v2di)__A, (N)); })
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_shl_epi8(__m128i __A, __m128i __B)
+{
+  return (__m128i)__builtin_ia32_vpshlb((__v16qi)__A, (__v16qi)__B);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_shl_epi16(__m128i __A, __m128i __B)
+{
+  return (__m128i)__builtin_ia32_vpshlw((__v8hi)__A, (__v8hi)__B);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_shl_epi32(__m128i __A, __m128i __B)
+{
+  return (__m128i)__builtin_ia32_vpshld((__v4si)__A, (__v4si)__B);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_shl_epi64(__m128i __A, __m128i __B)
+{
+  return (__m128i)__builtin_ia32_vpshlq((__v2di)__A, (__v2di)__B);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sha_epi8(__m128i __A, __m128i __B)
+{
+  return (__m128i)__builtin_ia32_vpshab((__v16qi)__A, (__v16qi)__B);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sha_epi16(__m128i __A, __m128i __B)
+{
+  return (__m128i)__builtin_ia32_vpshaw((__v8hi)__A, (__v8hi)__B);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sha_epi32(__m128i __A, __m128i __B)
+{
+  return (__m128i)__builtin_ia32_vpshad((__v4si)__A, (__v4si)__B);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_sha_epi64(__m128i __A, __m128i __B)
+{
+  return (__m128i)__builtin_ia32_vpshaq((__v2di)__A, (__v2di)__B);
+}
+
+#define _mm_com_epu8(A, B, N) __extension__ ({ \
+  __m128i __A = (A); \
+  __m128i __B = (B); \
+  (__m128i)__builtin_ia32_vpcomub((__v16qi)__A, (__v16qi)__B, (N)); })
+
+#define _mm_com_epu16(A, B, N) __extension__ ({ \
+  __m128i __A = (A); \
+  __m128i __B = (B); \
+  (__m128i)__builtin_ia32_vpcomuw((__v8hi)__A, (__v8hi)__B, (N)); })
+
+#define _mm_com_epu32(A, B, N) __extension__ ({ \
+  __m128i __A = (A); \
+  __m128i __B = (B); \
+  (__m128i)__builtin_ia32_vpcomud((__v4si)__A, (__v4si)__B, (N)); })
+
+#define _mm_com_epu64(A, B, N) __extension__ ({ \
+  __m128i __A = (A); \
+  __m128i __B = (B); \
+  (__m128i)__builtin_ia32_vpcomuq((__v2di)__A, (__v2di)__B, (N)); })
+
+#define _mm_com_epi8(A, B, N) __extension__ ({ \
+  __m128i __A = (A); \
+  __m128i __B = (B); \
+  (__m128i)__builtin_ia32_vpcomb((__v16qi)__A, (__v16qi)__B, (N)); })
+
+#define _mm_com_epi16(A, B, N) __extension__ ({ \
+  __m128i __A = (A); \
+  __m128i __B = (B); \
+  (__m128i)__builtin_ia32_vpcomw((__v8hi)__A, (__v8hi)__B, (N)); })
+
+#define _mm_com_epi32(A, B, N) __extension__ ({ \
+  __m128i __A = (A); \
+  __m128i __B = (B); \
+  (__m128i)__builtin_ia32_vpcomd((__v4si)__A, (__v4si)__B, (N)); })
+
+#define _mm_com_epi64(A, B, N) __extension__ ({ \
+  __m128i __A = (A); \
+  __m128i __B = (B); \
+  (__m128i)__builtin_ia32_vpcomq((__v2di)__A, (__v2di)__B, (N)); })
+
+#define _MM_PCOMCTRL_LT    0
+#define _MM_PCOMCTRL_LE    1
+#define _MM_PCOMCTRL_GT    2
+#define _MM_PCOMCTRL_GE    3
+#define _MM_PCOMCTRL_EQ    4
+#define _MM_PCOMCTRL_NEQ   5
+#define _MM_PCOMCTRL_FALSE 6
+#define _MM_PCOMCTRL_TRUE  7
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comlt_epu8(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu8(__A, __B, _MM_PCOMCTRL_LT);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comle_epu8(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu8(__A, __B, _MM_PCOMCTRL_LE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comgt_epu8(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu8(__A, __B, _MM_PCOMCTRL_GT);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comge_epu8(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu8(__A, __B, _MM_PCOMCTRL_GE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comeq_epu8(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu8(__A, __B, _MM_PCOMCTRL_EQ);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comneq_epu8(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu8(__A, __B, _MM_PCOMCTRL_NEQ);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comfalse_epu8(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu8(__A, __B, _MM_PCOMCTRL_FALSE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comtrue_epu8(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu8(__A, __B, _MM_PCOMCTRL_TRUE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comlt_epu16(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu16(__A, __B, _MM_PCOMCTRL_LT);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comle_epu16(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu16(__A, __B, _MM_PCOMCTRL_LE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comgt_epu16(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu16(__A, __B, _MM_PCOMCTRL_GT);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comge_epu16(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu16(__A, __B, _MM_PCOMCTRL_GE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comeq_epu16(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu16(__A, __B, _MM_PCOMCTRL_EQ);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comneq_epu16(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu16(__A, __B, _MM_PCOMCTRL_NEQ);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comfalse_epu16(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu16(__A, __B, _MM_PCOMCTRL_FALSE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comtrue_epu16(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu16(__A, __B, _MM_PCOMCTRL_TRUE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comlt_epu32(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu32(__A, __B, _MM_PCOMCTRL_LT);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comle_epu32(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu32(__A, __B, _MM_PCOMCTRL_LE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comgt_epu32(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu32(__A, __B, _MM_PCOMCTRL_GT);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comge_epu32(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu32(__A, __B, _MM_PCOMCTRL_GE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comeq_epu32(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu32(__A, __B, _MM_PCOMCTRL_EQ);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comneq_epu32(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu32(__A, __B, _MM_PCOMCTRL_NEQ);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comfalse_epu32(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu32(__A, __B, _MM_PCOMCTRL_FALSE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comtrue_epu32(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu32(__A, __B, _MM_PCOMCTRL_TRUE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comlt_epu64(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu64(__A, __B, _MM_PCOMCTRL_LT);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comle_epu64(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu64(__A, __B, _MM_PCOMCTRL_LE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comgt_epu64(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu64(__A, __B, _MM_PCOMCTRL_GT);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comge_epu64(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu64(__A, __B, _MM_PCOMCTRL_GE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comeq_epu64(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu64(__A, __B, _MM_PCOMCTRL_EQ);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comneq_epu64(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu64(__A, __B, _MM_PCOMCTRL_NEQ);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comfalse_epu64(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu64(__A, __B, _MM_PCOMCTRL_FALSE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comtrue_epu64(__m128i __A, __m128i __B)
+{
+  return _mm_com_epu64(__A, __B, _MM_PCOMCTRL_TRUE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comlt_epi8(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi8(__A, __B, _MM_PCOMCTRL_LT);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comle_epi8(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi8(__A, __B, _MM_PCOMCTRL_LE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comgt_epi8(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi8(__A, __B, _MM_PCOMCTRL_GT);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comge_epi8(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi8(__A, __B, _MM_PCOMCTRL_GE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comeq_epi8(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi8(__A, __B, _MM_PCOMCTRL_EQ);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comneq_epi8(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi8(__A, __B, _MM_PCOMCTRL_NEQ);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comfalse_epi8(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi8(__A, __B, _MM_PCOMCTRL_FALSE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comtrue_epi8(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi8(__A, __B, _MM_PCOMCTRL_TRUE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comlt_epi16(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi16(__A, __B, _MM_PCOMCTRL_LT);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comle_epi16(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi16(__A, __B, _MM_PCOMCTRL_LE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comgt_epi16(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi16(__A, __B, _MM_PCOMCTRL_GT);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comge_epi16(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi16(__A, __B, _MM_PCOMCTRL_GE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comeq_epi16(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi16(__A, __B, _MM_PCOMCTRL_EQ);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comneq_epi16(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi16(__A, __B, _MM_PCOMCTRL_NEQ);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comfalse_epi16(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi16(__A, __B, _MM_PCOMCTRL_FALSE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comtrue_epi16(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi16(__A, __B, _MM_PCOMCTRL_TRUE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comlt_epi32(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi32(__A, __B, _MM_PCOMCTRL_LT);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comle_epi32(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi32(__A, __B, _MM_PCOMCTRL_LE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comgt_epi32(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi32(__A, __B, _MM_PCOMCTRL_GT);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comge_epi32(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi32(__A, __B, _MM_PCOMCTRL_GE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comeq_epi32(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi32(__A, __B, _MM_PCOMCTRL_EQ);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comneq_epi32(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi32(__A, __B, _MM_PCOMCTRL_NEQ);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comfalse_epi32(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi32(__A, __B, _MM_PCOMCTRL_FALSE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comtrue_epi32(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi32(__A, __B, _MM_PCOMCTRL_TRUE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comlt_epi64(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi64(__A, __B, _MM_PCOMCTRL_LT);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comle_epi64(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi64(__A, __B, _MM_PCOMCTRL_LE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comgt_epi64(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi64(__A, __B, _MM_PCOMCTRL_GT);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comge_epi64(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi64(__A, __B, _MM_PCOMCTRL_GE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comeq_epi64(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi64(__A, __B, _MM_PCOMCTRL_EQ);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comneq_epi64(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi64(__A, __B, _MM_PCOMCTRL_NEQ);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comfalse_epi64(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi64(__A, __B, _MM_PCOMCTRL_FALSE);
+}
+
+static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
+_mm_comtrue_epi64(__m128i __A, __m128i __B)
+{
+  return _mm_com_epi64(__A, __B, _MM_PCOMCTRL_TRUE);
+}
+
+#define _mm_permute2_pd(X, Y, C, I) __extension__ ({ \
+  __m128d __X = (X); \
+  __m128d __Y = (Y); \
+  __m128i __C = (C); \
+  (__m128d)__builtin_ia32_vpermil2pd((__v2df)__X, (__v2df)__Y, \
+                                     (__v2di)__C, (I)); })
+
+#define _mm256_permute2_pd(X, Y, C, I) __extension__ ({ \
+  __m256d __X = (X); \
+  __m256d __Y = (Y); \
+  __m256i __C = (C); \
+  (__m256d)__builtin_ia32_vpermil2pd256((__v4df)__X, (__v4df)__Y, \
+                                        (__v4di)__C, (I)); })
+
+#define _mm_permute2_ps(X, Y, C, I) __extension__ ({ \
+  __m128 __X = (X); \
+  __m128 __Y = (Y); \
+  __m128i __C = (C); \
+  (__m128)__builtin_ia32_vpermil2ps((__v4sf)__X, (__v4sf)__Y, \
+                                    (__v4si)__C, (I)); })
+
+#define _mm256_permute2_ps(X, Y, C, I) __extension__ ({ \
+  __m256 __X = (X); \
+  __m256 __Y = (Y); \
+  __m256i __C = (C); \
+  (__m256)__builtin_ia32_vpermil2ps256((__v8sf)__X, (__v8sf)__Y, \
+                                       (__v8si)__C, (I)); })
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_frcz_ss(__m128 __A)
+{
+  return (__m128)__builtin_ia32_vfrczss((__v4sf)__A);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_frcz_sd(__m128d __A)
+{
+  return (__m128d)__builtin_ia32_vfrczsd((__v2df)__A);
+}
+
+static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
+_mm_frcz_ps(__m128 __A)
+{
+  return (__m128)__builtin_ia32_vfrczps((__v4sf)__A);
+}
+
+static __inline__ __m128d __attribute__((__always_inline__, __nodebug__))
+_mm_frcz_pd(__m128d __A)
+{
+  return (__m128d)__builtin_ia32_vfrczpd((__v2df)__A);
+}
+
+static __inline__ __m256 __attribute__((__always_inline__, __nodebug__))
+_mm256_frcz_ps(__m256 __A)
+{
+  return (__m256)__builtin_ia32_vfrczps256((__v8sf)__A);
+}
+
+static __inline__ __m256d __attribute__((__always_inline__, __nodebug__))
+_mm256_frcz_pd(__m256d __A)
+{
+  return (__m256d)__builtin_ia32_vfrczpd256((__v4df)__A);
+}
+
+#endif /* __XOP__ */
+
+#endif /* __XOPINTRIN_H */
diff --git a/lib/libc++.dylib b/lib/libc++.dylib
new file mode 100755
index 0000000..5732e72
--- /dev/null
+++ b/lib/libc++.dylib
diff --git a/lib64/libc++.dylib b/lib64/libc++.dylib
new file mode 100755
index 0000000..300f9c7
--- /dev/null
+++ b/lib64/libc++.dylib
diff --git a/update.sh b/update.sh
new file mode 100755
index 0000000..50d097f
--- /dev/null
+++ b/update.sh
@@ -0,0 +1,34 @@
+#!/bin/bash
+
+# Copy binaries
+for b in bin/*; do
+  file=`basename $b`
+  # Don't copy symlinks like clang++
+  if test -h $b; then
+    echo Skipping $file
+  else
+    echo Copying $file
+    cp -a `find ${ANDROID_HOST_OUT}/bin -name $file` $b
+    strip $b
+  fi
+done
+
+# Copy libraries
+echo Copying libc++.dylib
+cp -a ${ANDROID_HOST_OUT}/lib/libc++.dylib lib/
+cp -a ${ANDROID_HOST_OUT}/lib64/libc++.dylib lib64/
+
+# Copy header files
+rm -rf lib/clang/*/include/*
+for i in `find ${ANDROID_BUILD_TOP}/external/clang/lib/Headers -mindepth 1 ! -name \*.mk -a ! -name Makefile -a ! -name CMakeLists.txt`; do
+  echo Copying `basename $i`
+  cp -a $i lib/clang/*/include/
+done
+
+# Copy over stdatomic.h from bionic
+echo Copying stdatomic.h
+cp -a ${ANDROID_BUILD_TOP}/bionic/libc/include/stdatomic.h lib/clang/*/include/
+
+echo Copying arm_neon.h
+cp -a `find ${ANDROID_PRODUCT_OUT} -name arm_neon.h` lib/clang/*/include
+