Updated clang 3.5 prebuilts for LLVM rebase to r212749.

Change-Id: Ibf52f8d1c037f5ae5c83bd43cf2bf9d3b49faf3c

15 files changed, 68595 insertions, 9520 deletions

diff --git a/bin/clang b/bin/clang
index 5affa96..39c95f8 100755
--- a/bin/clang
+++ b/bin/clang
diff --git a/bin/llvm-as b/bin/llvm-as
index e35ab8b..02980fc 100755
--- a/bin/llvm-as
+++ b/bin/llvm-as
diff --git a/bin/llvm-dis b/bin/llvm-dis
index 5a706d6..03b5fa4 100755
--- a/bin/llvm-dis
+++ b/bin/llvm-dis
diff --git a/bin/llvm-link b/bin/llvm-link
index eb711e4..f210372 100755
--- a/bin/llvm-link
+++ b/bin/llvm-link
diff --git a/lib/clang/3.5/include/Intrin.h b/lib/clang/3.5/include/Intrin.h
index ff6d278..13e105e 100644
--- a/lib/clang/3.5/include/Intrin.h
+++ b/lib/clang/3.5/include/Intrin.h
@@ -30,21 +30,27 @@
 #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);
@@ -223,8 +229,7 @@ static __inline__
 long __cdecl _InterlockedDecrement(long volatile *_Addend);
 static __inline__
 short _InterlockedDecrement16(short volatile *_Addend);
-static __inline__
-long __cdecl _InterlockedExchange(long volatile *_Target, long _Value);
+long _InterlockedExchange(long volatile *_Target, long _Value);
 static __inline__
 short _InterlockedExchange16(short volatile *_Target, short _Value);
 static __inline__
@@ -288,7 +293,9 @@ 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);
@@ -411,7 +418,6 @@ __int64 _InterlockedCompareExchange64_HLERelease(__int64 volatile *, __int64,
                                                  __int64);
 __int64 _InterlockedCompareExchange64_np(__int64 volatile *_Destination,
                                          __int64 _Exchange, __int64 _Comparand);
-static __inline__
 void *_InterlockedCompareExchangePointer(void *volatile *_Destination,
                                          void *_Exchange, void *_Comparand);
 void *_InterlockedCompareExchangePointer_np(void *volatile *_Destination,
@@ -422,7 +428,6 @@ static __inline__
 __int64 _InterlockedExchange64(__int64 volatile *_Target, __int64 _Value);
 static __inline__
 __int64 _InterlockedExchangeAdd64(__int64 volatile *_Addend, __int64 _Value);
-static __inline__
 void *_InterlockedExchangePointer(void *volatile *_Target, void *_Value);
 static __inline__
 __int64 _InterlockedIncrement64(__int64 volatile *_Addend);
@@ -448,8 +453,9 @@ unsigned int __cdecl _readgsbase_u32(void);
 unsigned __int64 __cdecl _readgsbase_u64(void);
 unsigned __int64 _rorx_u64(unsigned __int64, const unsigned int);
 __int64 _sarx_i64(__int64, unsigned int);
-/* FIXME: Need definition for jmp_buf.
-  int __cdecl _setjmpex(jmp_buf); */
+#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 _tzcnt_u64(unsigned __int64);
@@ -575,6 +581,7 @@ _bittestandset(long *a, long b) {
   *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;
@@ -585,6 +592,7 @@ _interlockedbittestandset(long volatile *__BitBase, long __BitPos) {
            : "Ir"(__BitPos));
   return __Res;
 }
+#endif
 #ifdef __x86_64__
 static __inline__ unsigned char __attribute__((__always_inline__, __nodebug__))
 _BitScanForward64(unsigned long *_Index, unsigned __int64 _Mask) {
@@ -785,22 +793,12 @@ _InterlockedExchange16(short volatile *_Target, short _Value) {
   __atomic_exchange(_Target, &_Value, &_Value, 0);
   return _Value;
 }
-static __inline__ long __attribute__((__always_inline__, __nodebug__))
-_InterlockedExchange(long volatile *_Target, long _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;
 }
-static __inline__ void *__attribute__((__always_inline__, __nodebug__))
-_InterlockedExchangePointer(void *volatile *_Target, void *_Value) {
-  __atomic_exchange(_Target, &_Value, &_Value, 0);
-  return _Value;
-}
 #endif
 /*----------------------------------------------------------------------------*\
 |* Interlocked Compare Exchange
@@ -817,14 +815,6 @@ _InterlockedCompareExchange16(short volatile *_Destination,
   __atomic_compare_exchange(_Destination, &_Comparand, &_Exchange, 0, 0, 0);
   return _Comparand;
 }
-#ifdef __x86_64__
-static __inline__ void *__attribute__((__always_inline__, __nodebug__))
-_InterlockedCompareExchangePointer(void *volatile *_Destination,
-                                   void *_Exchange, void *_Comparand) {
-  __atomic_compare_exchange(_Destination, &_Comparand, &_Exchange, 0, 0, 0);
-  return _Comparand;
-}
-#endif
 static __inline__ __int64 __attribute__((__always_inline__, __nodebug__))
 _InterlockedCompareExchange64(__int64 volatile *_Destination,
                               __int64 _Exchange, __int64 _Comparand) {
@@ -834,6 +824,7 @@ _InterlockedCompareExchange64(__int64 volatile *_Destination,
 /*----------------------------------------------------------------------------*\
 |* 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) {
@@ -849,6 +840,7 @@ __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) {
@@ -904,6 +896,7 @@ __readgsword(unsigned long __offset) {
 /*----------------------------------------------------------------------------*\
 |* 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)
@@ -934,6 +927,7 @@ __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) {
@@ -958,6 +952,7 @@ 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])
@@ -978,10 +973,12 @@ 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
@@ -1007,6 +1004,7 @@ static __inline__ void __attribute__((always_inline, __nodebug__))
 __writecr3(unsigned int __cr3_val) {
   __asm__ ("mov %0, %%cr3" : : "q"(__cr3_val) : "memory");
 }
+#endif
 
 #ifdef __cplusplus
 }
diff --git a/lib/clang/3.5/include/altivec.h b/lib/clang/3.5/include/altivec.h
index 74ce08a..7a4a774 100644
--- a/lib/clang/3.5/include/altivec.h
+++ b/lib/clang/3.5/include/altivec.h
@@ -73,6 +73,9 @@ 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
@@ -3485,30 +3488,49 @@ 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 */
@@ -3516,7 +3538,11 @@ vec_mule(vector unsigned short __a, vector unsigned short __b)
 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 */
@@ -3524,7 +3550,11 @@ vec_vmulesb(vector signed char __a, vector signed char __b)
 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 */
@@ -3532,7 +3562,11 @@ vec_vmuleub(vector unsigned char __a, vector unsigned char __b)
 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 */
@@ -3540,7 +3574,11 @@ vec_vmulesh(vector short __a, vector short __b)
 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 */
@@ -3548,25 +3586,41 @@ vec_vmuleuh(vector unsigned short __a, vector unsigned short __b)
 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 */
@@ -3574,7 +3628,11 @@ vec_mulo(vector unsigned short __a, vector unsigned short __b)
 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 */
@@ -3582,7 +3640,11 @@ vec_vmulosb(vector signed char __a, vector signed char __b)
 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 */
@@ -3590,7 +3652,11 @@ vec_vmuloub(vector unsigned char __a, vector unsigned char __b)
 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 */
@@ -3598,7 +3664,11 @@ vec_vmulosh(vector short __a, vector short __b)
 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 */
@@ -4047,52 +4117,91 @@ vec_vor(vector float __a, vector bool int __b)
 
 /* 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 */
@@ -4102,25 +4211,43 @@ vec_pack(vector bool int __a, vector bool int __b)
 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 */
@@ -4130,25 +4257,43 @@ vec_vpkuhum(vector bool short __a, vector bool short __b)
 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 */
@@ -4156,7 +4301,11 @@ vec_vpkuwum(vector bool int __a, vector bool int __b)
 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 */
@@ -4164,7 +4313,11 @@ vec_packpx(vector unsigned int __a, vector unsigned int __b)
 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 */
@@ -4172,25 +4325,41 @@ vec_vpkpx(vector unsigned int __a, vector unsigned int __b)
 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 */
@@ -4198,7 +4367,11 @@ vec_packs(vector unsigned int __a, vector unsigned int __b)
 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 */
@@ -4206,7 +4379,11 @@ vec_vpkshss(vector short __a, vector short __b)
 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 */
@@ -4214,7 +4391,11 @@ vec_vpkuhus(vector unsigned short __a, vector unsigned short __b)
 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 */
@@ -4222,7 +4403,11 @@ vec_vpkswss(vector int __a, vector int __b)
 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 */
@@ -4230,25 +4415,41 @@ vec_vpkuwus(vector unsigned int __a, vector unsigned int __b)
 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 */
@@ -4256,13 +4457,21 @@ vec_packsu(vector unsigned int __a, vector unsigned int __b)
 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 */
@@ -4270,22 +4479,46 @@ vec_vpkshus(vector unsigned short __a, vector unsigned short __b)
 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
@@ -4293,22 +4526,46 @@ 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
@@ -4316,49 +4573,104 @@ 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
 }
 
 /* vec_vperm */
@@ -4366,8 +4678,7 @@ vec_perm(vector float __a, vector float __b, vector unsigned char __c)
 static vector signed char __ATTRS_o_ai
 vec_vperm(vector signed char __a, vector signed char __b, vector unsigned char __c)
 {
-  return (vector signed char)
-           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __c);
+  return vec_perm(__a, __b, __c);
 }
 
 static vector unsigned char __ATTRS_o_ai
@@ -4375,22 +4686,19 @@ vec_vperm(vector unsigned char __a,
           vector unsigned char __b,
           vector unsigned char __c)
 {
-  return (vector unsigned char)
-           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __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 (vector bool char)
-           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __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 (vector short)
-           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __c);
+  return vec_perm(__a, __b, __c);
 }
 
 static vector unsigned short __ATTRS_o_ai
@@ -4398,49 +4706,43 @@ vec_vperm(vector unsigned short __a,
           vector unsigned short __b,
           vector unsigned char __c)
 {
-  return (vector unsigned short)
-           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __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 (vector bool short)
-           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __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 (vector pixel)
-           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __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 (vector int)__builtin_altivec_vperm_4si(__a, __b, __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 (vector unsigned int)
-           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __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 (vector bool int)
-           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __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 (vector float)
-           __builtin_altivec_vperm_4si((vector int)__a, (vector int)__b, __c);
+  return vec_perm(__a, __b, __c);
 }
 
 /* vec_re */
@@ -4922,65 +5224,113 @@ vec_vslw(vector unsigned int __a, vector unsigned int __b)
 static vector signed char __ATTRS_o_ai
 vec_sld(vector signed char __a, vector signed char __b, unsigned char __c)
 {
+#ifdef __LITTLE_ENDIAN__
+  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));
+#else
   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));
+#endif
 }
 
 static vector unsigned char __ATTRS_o_ai
 vec_sld(vector unsigned char __a, vector unsigned char __b, unsigned char __c)
 {
+#ifdef __LITTLE_ENDIAN__
+  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));
+#else
   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));
+#endif
 }
 
 static vector short __ATTRS_o_ai
 vec_sld(vector short __a, vector short __b, unsigned char __c)
 {
+#ifdef __LITTLE_ENDIAN__
+  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));
+#else
   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));
+#endif
 }
 
 static vector unsigned short __ATTRS_o_ai
 vec_sld(vector unsigned short __a, vector unsigned short __b, unsigned char __c)
 {
+#ifdef __LITTLE_ENDIAN__
+  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));
+#else
   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));
+#endif
 }
 
 static vector pixel __ATTRS_o_ai
 vec_sld(vector pixel __a, vector pixel __b, unsigned char __c)
 {
+#ifdef __LITTLE_ENDIAN__
+  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));
+#else
   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));
+#endif
 }
 
 static vector int __ATTRS_o_ai
 vec_sld(vector int __a, vector int __b, unsigned char __c)
 {
+#ifdef __LITTLE_ENDIAN__
+  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));
+#else
   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));
+#endif
 }
 
 static vector unsigned int __ATTRS_o_ai
 vec_sld(vector unsigned int __a, vector unsigned int __b, unsigned char __c)
 {
+#ifdef __LITTLE_ENDIAN__
+  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));
+#else
   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));
+#endif
 }
 
 static vector float __ATTRS_o_ai
 vec_sld(vector float __a, vector float __b, unsigned char __c)
 {
+#ifdef __LITTLE_ENDIAN__
+  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));
+#else
   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));
+#endif
 }
 
 /* vec_vsldoi */
@@ -4988,65 +5338,113 @@ vec_sld(vector float __a, vector float __b, unsigned char __c)
 static vector signed char __ATTRS_o_ai
 vec_vsldoi(vector signed char __a, vector signed char __b, unsigned char __c)
 {
+#ifdef __LITTLE_ENDIAN__
+  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));
+#else
   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));
+#endif
 }
 
 static vector unsigned char __ATTRS_o_ai
 vec_vsldoi(vector unsigned char __a, vector unsigned char __b, unsigned char __c)
 {
+#ifdef __LITTLE_ENDIAN__
+  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));
+#else
   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));
+#endif
 }
 
 static vector short __ATTRS_o_ai
 vec_vsldoi(vector short __a, vector short __b, unsigned char __c)
 {
+#ifdef __LITTLE_ENDIAN__
+  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));
+#else
   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));
+#endif
 }
 
 static vector unsigned short __ATTRS_o_ai
 vec_vsldoi(vector unsigned short __a, vector unsigned short __b, unsigned char __c)
 {
+#ifdef __LITTLE_ENDIAN__
+  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));
+#else
   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));
+#endif
 }
 
 static vector pixel __ATTRS_o_ai
 vec_vsldoi(vector pixel __a, vector pixel __b, unsigned char __c)
 {
+#ifdef __LITTLE_ENDIAN__
+  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));
+#else
   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));
+#endif
 }
 
 static vector int __ATTRS_o_ai
 vec_vsldoi(vector int __a, vector int __b, unsigned char __c)
 {
+#ifdef __LITTLE_ENDIAN__
+  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));
+#else
   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));
+#endif
 }
 
 static vector unsigned int __ATTRS_o_ai
 vec_vsldoi(vector unsigned int __a, vector unsigned int __b, unsigned char __c)
 {
+#ifdef __LITTLE_ENDIAN__
+  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));
+#else
   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));
+#endif
 }
 
 static vector float __ATTRS_o_ai
 vec_vsldoi(vector float __a, vector float __b, unsigned char __c)
 {
+#ifdef __LITTLE_ENDIAN__
+  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));
+#else
   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));
+#endif
 }
 
 /* vec_sll */
@@ -8054,10 +8452,26 @@ vec_vsum4shs(vector signed short __a, vector int __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 */
@@ -8065,15 +8479,41 @@ vec_sum2s(vector int __a, vector int __b)
 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_perm(__b, __b, (vector unsigned char)
+	     (12,13,14,15,0,1,2,3,4,5,6,7,8,9,10,11));
+  __b = __builtin_altivec_vsumsws(__a, __b);
+  return (vector signed int)
+    vec_perm(__b, __b, (vector unsigned char)
+	     (4,5,6,7,8,9,10,11,12,13,14,15,0,1,2,3));
+#else
   return __builtin_altivec_vsumsws(__a, __b);
+#endif
 }
 
 /* vec_vsumsws */
@@ -8081,7 +8521,17 @@ vec_sums(vector signed int __a, vector signed int __b)
 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_perm(__b, __b, (vector unsigned char)
+	     (12,13,14,15,0,1,2,3,4,5,6,7,8,9,10,11));
+  __b = __builtin_altivec_vsumsws(__a, __b);
+  return (vector signed int)
+    vec_perm(__b, __b, (vector unsigned char)
+	     (4,5,6,7,8,9,10,11,12,13,14,15,0,1,2,3));
+#else
   return __builtin_altivec_vsumsws(__a, __b);
+#endif
 }
 
 /* vec_trunc */
@@ -8102,34 +8552,57 @@ vec_vrfiz(vector float __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)
 {
-  return (vector unsigned int)__builtin_altivec_vupkhsh((vector short)__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 */
@@ -8137,13 +8610,21 @@ vec_unpackh(vector pixel __a)
 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 */
@@ -8151,19 +8632,31 @@ vec_vupkhsb(vector bool char __a)
 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)
 {
-  return (vector unsigned int)__builtin_altivec_vupkhsh((vector short)__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 */
@@ -8171,31 +8664,51 @@ vec_vupkhsh(vector pixel __a)
 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)
 {
-  return (vector unsigned int)__builtin_altivec_vupklsh((vector short)__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 */
@@ -8203,13 +8716,21 @@ vec_unpackl(vector pixel __a)
 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 */
@@ -8217,19 +8738,31 @@ vec_vupklsb(vector bool char __a)
 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)
 {
-  return (vector unsigned int)__builtin_altivec_vupklsh((vector short)__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_xor */
diff --git a/lib/clang/3.5/include/arm_acle.h b/lib/clang/3.5/include/arm_acle.h
new file mode 100644
index 0000000..d706745
--- /dev/null
+++ b/lib/clang/3.5/include/arm_acle.h
@@ -0,0 +1,151 @@
+/*===---- 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
+
+/* Miscellaneous data-processing intrinsics */
+
+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) {
+#if __SIZEOF_LONG_LONG__ == 8
+  return __builtin_clzll(t);
+#else
+  return __builtin_clzl(t);
+#endif
+}
+
+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);
+}
+
+
+/*
+ * Saturating intrinsics
+ *
+ * FIXME: Change guard to their corrosponding __ARM_FEATURE flag when Q flag
+ * intrinsics are implemented and the flag is enabled.
+ */
+#if __ARM_32BIT_STATE
+#define __ssat(x, y) __builtin_arm_ssat(x, y)
+#define __usat(x, y) __builtin_arm_usat(x, y)
+
+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
+
+/* 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/3.5/include/arm_neon.h b/lib/clang/3.5/include/arm_neon.h
index f6ea155..7978aa0 100644
--- a/lib/clang/3.5/include/arm_neon.h
+++ b/lib/clang/3.5/include/arm_neon.h
@@ -45,37 +45,37 @@ typedef __uint128_t poly128_t;
 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(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(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;
+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;
+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(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;
+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;
+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 {
@@ -429,9601 +429,67951 @@ typedef struct poly64x2x4_t {
 
 #define __ai static inline __attribute__((__always_inline__, __nodebug__))
 
-__ai float32x2_t vfma_f32(float32x2_t __a, float32x2_t __b, float32x2_t __c) {
-  return (float32x2_t)__builtin_neon_vfma_v((int8x8_t)__a, (int8x8_t)__b, (int8x8_t)__c, 9); }
-__ai float32x4_t vfmaq_f32(float32x4_t __a, float32x4_t __b, float32x4_t __c) {
-  return (float32x4_t)__builtin_neon_vfmaq_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 41); }
-
-__ai int8x8_t vqmovn_s16(int16x8_t __a) {
-  return (int8x8_t)__builtin_neon_vqmovn_v((int8x16_t)__a, 0); }
-__ai int16x4_t vqmovn_s32(int32x4_t __a) {
-  return (int16x4_t)__builtin_neon_vqmovn_v((int8x16_t)__a, 1); }
-__ai int32x2_t vqmovn_s64(int64x2_t __a) {
-  return (int32x2_t)__builtin_neon_vqmovn_v((int8x16_t)__a, 2); }
-__ai uint8x8_t vqmovn_u16(uint16x8_t __a) {
-  return (uint8x8_t)__builtin_neon_vqmovn_v((int8x16_t)__a, 16); }
-__ai uint16x4_t vqmovn_u32(uint32x4_t __a) {
-  return (uint16x4_t)__builtin_neon_vqmovn_v((int8x16_t)__a, 17); }
-__ai uint32x2_t vqmovn_u64(uint64x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vqmovn_v((int8x16_t)__a, 18); }
-
-__ai uint8x8_t vqmovun_s16(int16x8_t __a) {
-  return (uint8x8_t)__builtin_neon_vqmovun_v((int8x16_t)__a, 16); }
-__ai uint16x4_t vqmovun_s32(int32x4_t __a) {
-  return (uint16x4_t)__builtin_neon_vqmovun_v((int8x16_t)__a, 17); }
-__ai uint32x2_t vqmovun_s64(int64x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vqmovun_v((int8x16_t)__a, 18); }
-
-__ai int8x8_t vabd_s8(int8x8_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vabd_v(__a, __b, 0); }
-__ai int16x4_t vabd_s16(int16x4_t __a, int16x4_t __b) {
-  return (int16x4_t)__builtin_neon_vabd_v((int8x8_t)__a, (int8x8_t)__b, 1); }
-__ai int32x2_t vabd_s32(int32x2_t __a, int32x2_t __b) {
-  return (int32x2_t)__builtin_neon_vabd_v((int8x8_t)__a, (int8x8_t)__b, 2); }
-__ai uint8x8_t vabd_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vabd_v((int8x8_t)__a, (int8x8_t)__b, 16); }
-__ai uint16x4_t vabd_u16(uint16x4_t __a, uint16x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vabd_v((int8x8_t)__a, (int8x8_t)__b, 17); }
-__ai uint32x2_t vabd_u32(uint32x2_t __a, uint32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vabd_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai float32x2_t vabd_f32(float32x2_t __a, float32x2_t __b) {
-  return (float32x2_t)__builtin_neon_vabd_v((int8x8_t)__a, (int8x8_t)__b, 9); }
-__ai int8x16_t vabdq_s8(int8x16_t __a, int8x16_t __b) {
-  return (int8x16_t)__builtin_neon_vabdq_v(__a, __b, 32); }
-__ai int16x8_t vabdq_s16(int16x8_t __a, int16x8_t __b) {
-  return (int16x8_t)__builtin_neon_vabdq_v((int8x16_t)__a, (int8x16_t)__b, 33); }
-__ai int32x4_t vabdq_s32(int32x4_t __a, int32x4_t __b) {
-  return (int32x4_t)__builtin_neon_vabdq_v((int8x16_t)__a, (int8x16_t)__b, 34); }
-__ai uint8x16_t vabdq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vabdq_v((int8x16_t)__a, (int8x16_t)__b, 48); }
-__ai uint16x8_t vabdq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint16x8_t)__builtin_neon_vabdq_v((int8x16_t)__a, (int8x16_t)__b, 49); }
-__ai uint32x4_t vabdq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vabdq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-__ai float32x4_t vabdq_f32(float32x4_t __a, float32x4_t __b) {
-  return (float32x4_t)__builtin_neon_vabdq_v((int8x16_t)__a, (int8x16_t)__b, 41); }
-
-__ai int16x8_t vmovl_s8(int8x8_t __a) {
-  return (int16x8_t)__builtin_neon_vmovl_v(__a, 33); }
-__ai int32x4_t vmovl_s16(int16x4_t __a) {
-  return (int32x4_t)__builtin_neon_vmovl_v((int8x8_t)__a, 34); }
-__ai int64x2_t vmovl_s32(int32x2_t __a) {
-  return (int64x2_t)__builtin_neon_vmovl_v((int8x8_t)__a, 35); }
-__ai uint16x8_t vmovl_u8(uint8x8_t __a) {
-  return (uint16x8_t)__builtin_neon_vmovl_v((int8x8_t)__a, 49); }
-__ai uint32x4_t vmovl_u16(uint16x4_t __a) {
-  return (uint32x4_t)__builtin_neon_vmovl_v((int8x8_t)__a, 50); }
-__ai uint64x2_t vmovl_u32(uint32x2_t __a) {
-  return (uint64x2_t)__builtin_neon_vmovl_v((int8x8_t)__a, 51); }
-
-__ai int16x8_t vabdl_s8(int8x8_t __a, int8x8_t __b) {
-  return (int16x8_t)vmovl_u8((uint8x8_t)vabd_s8(__a, __b)); }
-__ai int32x4_t vabdl_s16(int16x4_t __a, int16x4_t __b) {
-  return (int32x4_t)vmovl_u16((uint16x4_t)vabd_s16(__a, __b)); }
-__ai int64x2_t vabdl_s32(int32x2_t __a, int32x2_t __b) {
-  return (int64x2_t)vmovl_u32((uint32x2_t)vabd_s32(__a, __b)); }
-__ai uint16x8_t vabdl_u8(uint8x8_t __a, uint8x8_t __b) {
-  return vmovl_u8(vabd_u8(__a, __b)); }
-__ai uint32x4_t vabdl_u16(uint16x4_t __a, uint16x4_t __b) {
-  return vmovl_u16(vabd_u16(__a, __b)); }
-__ai uint64x2_t vabdl_u32(uint32x2_t __a, uint32x2_t __b) {
-  return vmovl_u32(vabd_u32(__a, __b)); }
-
-__ai int8x8_t vget_high_s8(int8x16_t __a) {
-  return __builtin_shufflevector(__a, __a, 8, 9, 10, 11, 12, 13, 14, 15); }
-__ai int16x4_t vget_high_s16(int16x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 4, 5, 6, 7); }
-__ai int32x2_t vget_high_s32(int32x4_t __a) {
-  return __builtin_shufflevector(__a, __a, 2, 3); }
-__ai int64x1_t vget_high_s64(int64x2_t __a) {
-  return __builtin_shufflevector(__a, __a, 1); }
-__ai float16x4_t vget_high_f16(float16x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 4, 5, 6, 7); }
-__ai float32x2_t vget_high_f32(float32x4_t __a) {
-  return __builtin_shufflevector(__a, __a, 2, 3); }
-__ai uint8x8_t vget_high_u8(uint8x16_t __a) {
-  return __builtin_shufflevector(__a, __a, 8, 9, 10, 11, 12, 13, 14, 15); }
-__ai uint16x4_t vget_high_u16(uint16x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 4, 5, 6, 7); }
-__ai uint32x2_t vget_high_u32(uint32x4_t __a) {
-  return __builtin_shufflevector(__a, __a, 2, 3); }
-__ai uint64x1_t vget_high_u64(uint64x2_t __a) {
-  return __builtin_shufflevector(__a, __a, 1); }
-__ai poly8x8_t vget_high_p8(poly8x16_t __a) {
-  return __builtin_shufflevector(__a, __a, 8, 9, 10, 11, 12, 13, 14, 15); }
-__ai poly16x4_t vget_high_p16(poly16x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 4, 5, 6, 7); }
-
-__ai int8x16_t vcombine_s8(int8x8_t __a, int8x8_t __b) {
-  return (int8x16_t)__builtin_shufflevector((int64x1_t)__a, (int64x1_t)__b, 0, 1); }
-__ai int16x8_t vcombine_s16(int16x4_t __a, int16x4_t __b) {
-  return (int16x8_t)__builtin_shufflevector((int64x1_t)__a, (int64x1_t)__b, 0, 1); }
-__ai int32x4_t vcombine_s32(int32x2_t __a, int32x2_t __b) {
-  return (int32x4_t)__builtin_shufflevector((int64x1_t)__a, (int64x1_t)__b, 0, 1); }
-__ai int64x2_t vcombine_s64(int64x1_t __a, int64x1_t __b) {
-  return (int64x2_t)__builtin_shufflevector((int64x1_t)__a, (int64x1_t)__b, 0, 1); }
-__ai float16x8_t vcombine_f16(float16x4_t __a, float16x4_t __b) {
-  return (float16x8_t)__builtin_shufflevector((int64x1_t)__a, (int64x1_t)__b, 0, 1); }
-__ai float32x4_t vcombine_f32(float32x2_t __a, float32x2_t __b) {
-  return (float32x4_t)__builtin_shufflevector((int64x1_t)__a, (int64x1_t)__b, 0, 1); }
-__ai uint8x16_t vcombine_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint8x16_t)__builtin_shufflevector((int64x1_t)__a, (int64x1_t)__b, 0, 1); }
-__ai uint16x8_t vcombine_u16(uint16x4_t __a, uint16x4_t __b) {
-  return (uint16x8_t)__builtin_shufflevector((int64x1_t)__a, (int64x1_t)__b, 0, 1); }
-__ai uint32x4_t vcombine_u32(uint32x2_t __a, uint32x2_t __b) {
-  return (uint32x4_t)__builtin_shufflevector((int64x1_t)__a, (int64x1_t)__b, 0, 1); }
-__ai uint64x2_t vcombine_u64(uint64x1_t __a, uint64x1_t __b) {
-  return (uint64x2_t)__builtin_shufflevector((int64x1_t)__a, (int64x1_t)__b, 0, 1); }
-__ai poly8x16_t vcombine_p8(poly8x8_t __a, poly8x8_t __b) {
-  return (poly8x16_t)__builtin_shufflevector((int64x1_t)__a, (int64x1_t)__b, 0, 1); }
-__ai poly16x8_t vcombine_p16(poly16x4_t __a, poly16x4_t __b) {
-  return (poly16x8_t)__builtin_shufflevector((int64x1_t)__a, (int64x1_t)__b, 0, 1); }
-
-#define vshll_n_s8(a, __b) __extension__ ({ \
-  int8x8_t __a = (a); \
-  (int16x8_t)__builtin_neon_vshll_n_v(__a, __b, 33); })
-#define vshll_n_s16(a, __b) __extension__ ({ \
-  int16x4_t __a = (a); \
-  (int32x4_t)__builtin_neon_vshll_n_v((int8x8_t)__a, __b, 34); })
-#define vshll_n_s32(a, __b) __extension__ ({ \
-  int32x2_t __a = (a); \
-  (int64x2_t)__builtin_neon_vshll_n_v((int8x8_t)__a, __b, 35); })
-#define vshll_n_u8(a, __b) __extension__ ({ \
-  uint8x8_t __a = (a); \
-  (uint16x8_t)__builtin_neon_vshll_n_v((int8x8_t)__a, __b, 49); })
-#define vshll_n_u16(a, __b) __extension__ ({ \
-  uint16x4_t __a = (a); \
-  (uint32x4_t)__builtin_neon_vshll_n_v((int8x8_t)__a, __b, 50); })
-#define vshll_n_u32(a, __b) __extension__ ({ \
-  uint32x2_t __a = (a); \
-  (uint64x2_t)__builtin_neon_vshll_n_v((int8x8_t)__a, __b, 51); })
+#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
 
-#if defined(__aarch64__)
-__ai int16x8_t vmovl_high_s8(int8x16_t __a) {
-  int8x8_t __a1 = vget_high_s8(__a);
-  return (int16x8_t)vshll_n_s8(__a1, 0); }
-__ai int32x4_t vmovl_high_s16(int16x8_t __a) {
-  int16x4_t __a1 = vget_high_s16(__a);
-  return (int32x4_t)vshll_n_s16(__a1, 0); }
-__ai int64x2_t vmovl_high_s32(int32x4_t __a) {
-  int32x2_t __a1 = vget_high_s32(__a);
-  return (int64x2_t)vshll_n_s32(__a1, 0); }
-__ai uint16x8_t vmovl_high_u8(uint8x16_t __a) {
-  uint8x8_t __a1 = vget_high_u8(__a);
-  return (uint16x8_t)vshll_n_u8(__a1, 0); }
-__ai uint32x4_t vmovl_high_u16(uint16x8_t __a) {
-  uint16x4_t __a1 = vget_high_u16(__a);
-  return (uint32x4_t)vshll_n_u16(__a1, 0); }
-__ai uint64x2_t vmovl_high_u32(uint32x4_t __a) {
-  uint32x2_t __a1 = vget_high_u32(__a);
-  return (uint64x2_t)vshll_n_u32(__a1, 0); }
-
-#endif
-
-__ai int16x8_t vmull_s8(int8x8_t __a, int8x8_t __b) {
-  return (int16x8_t)__builtin_neon_vmull_v(__a, __b, 33); }
-__ai int32x4_t vmull_s16(int16x4_t __a, int16x4_t __b) {
-  return (int32x4_t)__builtin_neon_vmull_v((int8x8_t)__a, (int8x8_t)__b, 34); }
-__ai int64x2_t vmull_s32(int32x2_t __a, int32x2_t __b) {
-  return (int64x2_t)__builtin_neon_vmull_v((int8x8_t)__a, (int8x8_t)__b, 35); }
-__ai uint16x8_t vmull_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint16x8_t)__builtin_neon_vmull_v((int8x8_t)__a, (int8x8_t)__b, 49); }
-__ai uint32x4_t vmull_u16(uint16x4_t __a, uint16x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vmull_v((int8x8_t)__a, (int8x8_t)__b, 50); }
-__ai uint64x2_t vmull_u32(uint32x2_t __a, uint32x2_t __b) {
-  return (uint64x2_t)__builtin_neon_vmull_v((int8x8_t)__a, (int8x8_t)__b, 51); }
-__ai poly16x8_t vmull_p8(poly8x8_t __a, poly8x8_t __b) {
-  return (poly16x8_t)__builtin_neon_vmull_v((int8x8_t)__a, (int8x8_t)__b, 37); }
-
-__ai int32x4_t vmlal_n_s16(int32x4_t __a, int16x4_t __b, int16_t __c) {
-  return __a + vmull_s16(__b, (int16x4_t){ __c, __c, __c, __c }); }
-__ai int64x2_t vmlal_n_s32(int64x2_t __a, int32x2_t __b, int32_t __c) {
-  return __a + vmull_s32(__b, (int32x2_t){ __c, __c }); }
-__ai uint32x4_t vmlal_n_u16(uint32x4_t __a, uint16x4_t __b, uint16_t __c) {
-  return __a + vmull_u16(__b, (uint16x4_t){ __c, __c, __c, __c }); }
-__ai uint64x2_t vmlal_n_u32(uint64x2_t __a, uint32x2_t __b, uint32_t __c) {
-  return __a + vmull_u32(__b, (uint32x2_t){ __c, __c }); }
-
-__ai int32x4_t vmlsl_n_s16(int32x4_t __a, int16x4_t __b, int16_t __c) {
-  return __a - vmull_s16(__b, (int16x4_t){ __c, __c, __c, __c }); }
-__ai int64x2_t vmlsl_n_s32(int64x2_t __a, int32x2_t __b, int32_t __c) {
-  return __a - vmull_s32(__b, (int32x2_t){ __c, __c }); }
-__ai uint32x4_t vmlsl_n_u16(uint32x4_t __a, uint16x4_t __b, uint16_t __c) {
-  return __a - vmull_u16(__b, (uint16x4_t){ __c, __c, __c, __c }); }
-__ai uint64x2_t vmlsl_n_u32(uint64x2_t __a, uint32x2_t __b, uint32_t __c) {
-  return __a - vmull_u32(__b, (uint32x2_t){ __c, __c }); }
-
-__ai int32x4_t vmull_n_s16(int16x4_t __a, int16_t __b) {
-  return (int32x4_t)__builtin_neon_vmull_v((int8x8_t)__a, (int8x8_t)(int16x4_t){ __b, __b, __b, __b }, 34); }
-__ai int64x2_t vmull_n_s32(int32x2_t __a, int32_t __b) {
-  return (int64x2_t)__builtin_neon_vmull_v((int8x8_t)__a, (int8x8_t)(int32x2_t){ __b, __b }, 35); }
-__ai uint32x4_t vmull_n_u16(uint16x4_t __a, uint16_t __b) {
-  return (uint32x4_t)__builtin_neon_vmull_v((int8x8_t)__a, (int8x8_t)(uint16x4_t){ __b, __b, __b, __b }, 50); }
-__ai uint64x2_t vmull_n_u32(uint32x2_t __a, uint32_t __b) {
-  return (uint64x2_t)__builtin_neon_vmull_v((int8x8_t)__a, (int8x8_t)(uint32x2_t){ __b, __b }, 51); }
+#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
 
-#if defined(__aarch64__)
-__ai poly128_t vmull_p64(poly64_t __a, poly64_t __b) {
-  return (poly128_t)__builtin_neon_vmull_p64(__a, __b); }
+#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
 
-__ai int32x4_t vqdmlal_n_s16(int32x4_t __a, int16x4_t __b, int16_t __c) {
-  return (int32x4_t)__builtin_neon_vqdmlal_v((int8x16_t)__a, (int8x8_t)__b, (int8x8_t)(int16x4_t){ __c, __c, __c, __c }, 34); }
-__ai int64x2_t vqdmlal_n_s32(int64x2_t __a, int32x2_t __b, int32_t __c) {
-  return (int64x2_t)__builtin_neon_vqdmlal_v((int8x16_t)__a, (int8x8_t)__b, (int8x8_t)(int32x2_t){ __c, __c }, 35); }
+#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
 
-__ai int32x4_t vqdmlsl_n_s16(int32x4_t __a, int16x4_t __b, int16_t __c) {
-  return (int32x4_t)__builtin_neon_vqdmlsl_v((int8x16_t)__a, (int8x8_t)__b, (int8x8_t)(int16x4_t){ __c, __c, __c, __c }, 34); }
-__ai int64x2_t vqdmlsl_n_s32(int64x2_t __a, int32x2_t __b, int32_t __c) {
-  return (int64x2_t)__builtin_neon_vqdmlsl_v((int8x16_t)__a, (int8x8_t)__b, (int8x8_t)(int32x2_t){ __c, __c }, 35); }
+#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
 
-__ai int32x4_t vqdmull_n_s16(int16x4_t __a, int16_t __b) {
-  return (int32x4_t)__builtin_neon_vqdmull_v((int8x8_t)__a, (int8x8_t)(int16x4_t){ __b, __b, __b, __b }, 34); }
-__ai int64x2_t vqdmull_n_s32(int32x2_t __a, int32_t __b) {
-  return (int64x2_t)__builtin_neon_vqdmull_v((int8x8_t)__a, (int8x8_t)(int32x2_t){ __b, __b }, 35); }
+#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
 
-#if defined(__aarch64__)
-__ai float64x1_t vabd_f64(float64x1_t __a, float64x1_t __b) {
-  return (float64x1_t)__builtin_neon_vabd_v((int8x8_t)__a, (int8x8_t)__b, 10); }
-__ai float64x2_t vabdq_f64(float64x2_t __a, float64x2_t __b) {
-  return (float64x2_t)__builtin_neon_vabdq_v((int8x16_t)__a, (int8x16_t)__b, 42); }
-
-__ai float64x1_t vabs_f64(float64x1_t __a) {
-  return (float64x1_t)__builtin_neon_vabs_v((int8x8_t)__a, 10); }
-__ai float64x2_t vabsq_f64(float64x2_t __a) {
-  return (float64x2_t)__builtin_neon_vabsq_v((int8x16_t)__a, 42); }
-__ai int64x1_t vabs_s64(int64x1_t __a) {
-  return (int64x1_t)__builtin_neon_vabs_v((int8x8_t)__a, 3); }
-__ai int64x2_t vabsq_s64(int64x2_t __a) {
-  return (int64x2_t)__builtin_neon_vabsq_v((int8x16_t)__a, 35); }
-
-__ai float64x1_t vadd_f64(float64x1_t __a, float64x1_t __b) {
-  return __a + __b; }
-__ai float64x2_t vaddq_f64(float64x2_t __a, float64x2_t __b) {
-  return __a + __b; }
-
-__ai int8x16_t vpaddq_s8(int8x16_t __a, int8x16_t __b) {
-  return (int8x16_t)__builtin_neon_vpaddq_v(__a, __b, 32); }
-__ai int16x8_t vpaddq_s16(int16x8_t __a, int16x8_t __b) {
-  return (int16x8_t)__builtin_neon_vpaddq_v((int8x16_t)__a, (int8x16_t)__b, 33); }
-__ai int32x4_t vpaddq_s32(int32x4_t __a, int32x4_t __b) {
-  return (int32x4_t)__builtin_neon_vpaddq_v((int8x16_t)__a, (int8x16_t)__b, 34); }
-__ai int64x2_t vpaddq_s64(int64x2_t __a, int64x2_t __b) {
-  return (int64x2_t)__builtin_neon_vpaddq_v((int8x16_t)__a, (int8x16_t)__b, 35); }
-__ai uint8x16_t vpaddq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vpaddq_v((int8x16_t)__a, (int8x16_t)__b, 48); }
-__ai uint16x8_t vpaddq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint16x8_t)__builtin_neon_vpaddq_v((int8x16_t)__a, (int8x16_t)__b, 49); }
-__ai uint32x4_t vpaddq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vpaddq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-__ai uint64x2_t vpaddq_u64(uint64x2_t __a, uint64x2_t __b) {
-  return (uint64x2_t)__builtin_neon_vpaddq_v((int8x16_t)__a, (int8x16_t)__b, 51); }
-__ai float32x4_t vpaddq_f32(float32x4_t __a, float32x4_t __b) {
-  return (float32x4_t)__builtin_neon_vpaddq_v((int8x16_t)__a, (int8x16_t)__b, 41); }
-__ai float64x2_t vpaddq_f64(float64x2_t __a, float64x2_t __b) {
-  return (float64x2_t)__builtin_neon_vpaddq_v((int8x16_t)__a, (int8x16_t)__b, 42); }
+#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
 
-#if __ARM_FEATURE_CRYPTO
-__ai uint8x16_t vaesdq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vaesdq_v((int8x16_t)__a, (int8x16_t)__b, 48); }
+#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
 
-__ai uint8x16_t vaeseq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vaeseq_v((int8x16_t)__a, (int8x16_t)__b, 48); }
+#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
 
-__ai uint8x16_t vaesimcq_u8(uint8x16_t __a) {
-  return (uint8x16_t)__builtin_neon_vaesimcq_v((int8x16_t)__a, 48); }
+#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
 
-__ai uint8x16_t vaesmcq_u8(uint8x16_t __a) {
-  return (uint8x16_t)__builtin_neon_vaesmcq_v((int8x16_t)__a, 48); }
+#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
 
-#if defined(__aarch64__)
-__ai float64x1_t vbsl_f64(uint64x1_t __a, float64x1_t __b, float64x1_t __c) {
-  return (float64x1_t)__builtin_neon_vbsl_v((int8x8_t)__a, (int8x8_t)__b, (int8x8_t)__c, 10); }
-__ai poly64x1_t vbsl_p64(uint64x1_t __a, poly64x1_t __b, poly64x1_t __c) {
-  return (poly64x1_t)__builtin_neon_vbsl_v((int8x8_t)__a, (int8x8_t)__b, (int8x8_t)__c, 6); }
-__ai float64x2_t vbslq_f64(uint64x2_t __a, float64x2_t __b, float64x2_t __c) {
-  return (float64x2_t)__builtin_neon_vbslq_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 42); }
-__ai poly64x2_t vbslq_p64(uint64x2_t __a, poly64x2_t __b, poly64x2_t __c) {
-  return (poly64x2_t)__builtin_neon_vbslq_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 38); }
-
-__ai uint64x1_t vceq_s64(int64x1_t __a, int64x1_t __b) {
-  return (uint64x1_t)(__a == __b); }
-__ai uint64x1_t vceq_u64(uint64x1_t __a, uint64x1_t __b) {
-  return (uint64x1_t)(__a == __b); }
-__ai uint64x1_t vceq_f64(float64x1_t __a, float64x1_t __b) {
-  return (uint64x1_t)(__a == __b); }
-__ai uint64x2_t vceqq_f64(float64x2_t __a, float64x2_t __b) {
-  return (uint64x2_t)(__a == __b); }
-__ai uint64x2_t vceqq_s64(int64x2_t __a, int64x2_t __b) {
-  return (uint64x2_t)(__a == __b); }
-__ai uint64x2_t vceqq_u64(uint64x2_t __a, uint64x2_t __b) {
-  return (uint64x2_t)(__a == __b); }
-__ai uint64x1_t vceq_p64(poly64x1_t __a, poly64x1_t __b) {
-  return (uint64x1_t)(__a == __b); }
-__ai uint64x2_t vceqq_p64(poly64x2_t __a, poly64x2_t __b) {
-  return (uint64x2_t)(__a == __b); }
-
-__ai uint64x1_t vcge_s64(int64x1_t __a, int64x1_t __b) {
-  return (uint64x1_t)(__a >= __b); }
-__ai uint64x1_t vcge_u64(uint64x1_t __a, uint64x1_t __b) {
-  return (uint64x1_t)(__a >= __b); }
-__ai uint64x1_t vcge_f64(float64x1_t __a, float64x1_t __b) {
-  return (uint64x1_t)(__a >= __b); }
-__ai uint64x2_t vcgeq_f64(float64x2_t __a, float64x2_t __b) {
-  return (uint64x2_t)(__a >= __b); }
-__ai uint64x2_t vcgeq_s64(int64x2_t __a, int64x2_t __b) {
-  return (uint64x2_t)(__a >= __b); }
-__ai uint64x2_t vcgeq_u64(uint64x2_t __a, uint64x2_t __b) {
-  return (uint64x2_t)(__a >= __b); }
-
-__ai uint64x1_t vcgt_s64(int64x1_t __a, int64x1_t __b) {
-  return (uint64x1_t)(__a > __b); }
-__ai uint64x1_t vcgt_u64(uint64x1_t __a, uint64x1_t __b) {
-  return (uint64x1_t)(__a > __b); }
-__ai uint64x1_t vcgt_f64(float64x1_t __a, float64x1_t __b) {
-  return (uint64x1_t)(__a > __b); }
-__ai uint64x2_t vcgtq_f64(float64x2_t __a, float64x2_t __b) {
-  return (uint64x2_t)(__a > __b); }
-__ai uint64x2_t vcgtq_s64(int64x2_t __a, int64x2_t __b) {
-  return (uint64x2_t)(__a > __b); }
-__ai uint64x2_t vcgtq_u64(uint64x2_t __a, uint64x2_t __b) {
-  return (uint64x2_t)(__a > __b); }
-
-__ai uint64x1_t vcle_s64(int64x1_t __a, int64x1_t __b) {
-  return (uint64x1_t)(__a <= __b); }
-__ai uint64x1_t vcle_u64(uint64x1_t __a, uint64x1_t __b) {
-  return (uint64x1_t)(__a <= __b); }
-__ai uint64x1_t vcle_f64(float64x1_t __a, float64x1_t __b) {
-  return (uint64x1_t)(__a <= __b); }
-__ai uint64x2_t vcleq_f64(float64x2_t __a, float64x2_t __b) {
-  return (uint64x2_t)(__a <= __b); }
-__ai uint64x2_t vcleq_s64(int64x2_t __a, int64x2_t __b) {
-  return (uint64x2_t)(__a <= __b); }
-__ai uint64x2_t vcleq_u64(uint64x2_t __a, uint64x2_t __b) {
-  return (uint64x2_t)(__a <= __b); }
-
-__ai uint64x1_t vclt_s64(int64x1_t __a, int64x1_t __b) {
-  return (uint64x1_t)(__a < __b); }
-__ai uint64x1_t vclt_u64(uint64x1_t __a, uint64x1_t __b) {
-  return (uint64x1_t)(__a < __b); }
-__ai uint64x1_t vclt_f64(float64x1_t __a, float64x1_t __b) {
-  return (uint64x1_t)(__a < __b); }
-__ai uint64x2_t vcltq_f64(float64x2_t __a, float64x2_t __b) {
-  return (uint64x2_t)(__a < __b); }
-__ai uint64x2_t vcltq_s64(int64x2_t __a, int64x2_t __b) {
-  return (uint64x2_t)(__a < __b); }
-__ai uint64x2_t vcltq_u64(uint64x2_t __a, uint64x2_t __b) {
-  return (uint64x2_t)(__a < __b); }
-
-__ai uint8x8_t vceqz_s8(int8x8_t __a) {
-  return (uint8x8_t)__builtin_neon_vceqz_v(__a, 16); }
-__ai uint16x4_t vceqz_s16(int16x4_t __a) {
-  return (uint16x4_t)__builtin_neon_vceqz_v((int8x8_t)__a, 17); }
-__ai uint32x2_t vceqz_s32(int32x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vceqz_v((int8x8_t)__a, 18); }
-__ai uint64x1_t vceqz_s64(int64x1_t __a) {
-  return (uint64x1_t)__builtin_neon_vceqz_v((int8x8_t)__a, 19); }
-__ai uint32x2_t vceqz_f32(float32x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vceqz_v((int8x8_t)__a, 18); }
-__ai uint8x8_t vceqz_u8(uint8x8_t __a) {
-  return (uint8x8_t)__builtin_neon_vceqz_v((int8x8_t)__a, 16); }
-__ai uint16x4_t vceqz_u16(uint16x4_t __a) {
-  return (uint16x4_t)__builtin_neon_vceqz_v((int8x8_t)__a, 17); }
-__ai uint32x2_t vceqz_u32(uint32x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vceqz_v((int8x8_t)__a, 18); }
-__ai uint64x1_t vceqz_u64(uint64x1_t __a) {
-  return (uint64x1_t)__builtin_neon_vceqz_v((int8x8_t)__a, 19); }
-__ai uint8x8_t vceqz_p8(poly8x8_t __a) {
-  return (uint8x8_t)__builtin_neon_vceqz_v((int8x8_t)__a, 16); }
-__ai uint16x4_t vceqz_p16(poly16x4_t __a) {
-  return (uint16x4_t)__builtin_neon_vceqz_v((int8x8_t)__a, 17); }
-__ai uint64x1_t vceqz_p64(poly64x1_t __a) {
-  return (uint64x1_t)__builtin_neon_vceqz_v((int8x8_t)__a, 19); }
-__ai uint8x16_t vceqzq_s8(int8x16_t __a) {
-  return (uint8x16_t)__builtin_neon_vceqzq_v(__a, 48); }
-__ai uint16x8_t vceqzq_s16(int16x8_t __a) {
-  return (uint16x8_t)__builtin_neon_vceqzq_v((int8x16_t)__a, 49); }
-__ai uint32x4_t vceqzq_s32(int32x4_t __a) {
-  return (uint32x4_t)__builtin_neon_vceqzq_v((int8x16_t)__a, 50); }
-__ai uint64x2_t vceqzq_s64(int64x2_t __a) {
-  return (uint64x2_t)__builtin_neon_vceqzq_v((int8x16_t)__a, 51); }
-__ai uint32x4_t vceqzq_f32(float32x4_t __a) {
-  return (uint32x4_t)__builtin_neon_vceqzq_v((int8x16_t)__a, 50); }
-__ai uint8x16_t vceqzq_u8(uint8x16_t __a) {
-  return (uint8x16_t)__builtin_neon_vceqzq_v((int8x16_t)__a, 48); }
-__ai uint16x8_t vceqzq_u16(uint16x8_t __a) {
-  return (uint16x8_t)__builtin_neon_vceqzq_v((int8x16_t)__a, 49); }
-__ai uint32x4_t vceqzq_u32(uint32x4_t __a) {
-  return (uint32x4_t)__builtin_neon_vceqzq_v((int8x16_t)__a, 50); }
-__ai uint64x2_t vceqzq_u64(uint64x2_t __a) {
-  return (uint64x2_t)__builtin_neon_vceqzq_v((int8x16_t)__a, 51); }
-__ai uint8x16_t vceqzq_p8(poly8x16_t __a) {
-  return (uint8x16_t)__builtin_neon_vceqzq_v((int8x16_t)__a, 48); }
-__ai uint16x8_t vceqzq_p16(poly16x8_t __a) {
-  return (uint16x8_t)__builtin_neon_vceqzq_v((int8x16_t)__a, 49); }
-__ai uint64x1_t vceqz_f64(float64x1_t __a) {
-  return (uint64x1_t)__builtin_neon_vceqz_v((int8x8_t)__a, 19); }
-__ai uint64x2_t vceqzq_f64(float64x2_t __a) {
-  return (uint64x2_t)__builtin_neon_vceqzq_v((int8x16_t)__a, 51); }
-__ai uint64x2_t vceqzq_p64(poly64x2_t __a) {
-  return (uint64x2_t)__builtin_neon_vceqzq_v((int8x16_t)__a, 51); }
-
-__ai uint8x8_t vcgez_s8(int8x8_t __a) {
-  return (uint8x8_t)__builtin_neon_vcgez_v(__a, 16); }
-__ai uint16x4_t vcgez_s16(int16x4_t __a) {
-  return (uint16x4_t)__builtin_neon_vcgez_v((int8x8_t)__a, 17); }
-__ai uint32x2_t vcgez_s32(int32x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vcgez_v((int8x8_t)__a, 18); }
-__ai uint64x1_t vcgez_s64(int64x1_t __a) {
-  return (uint64x1_t)__builtin_neon_vcgez_v((int8x8_t)__a, 19); }
-__ai uint32x2_t vcgez_f32(float32x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vcgez_v((int8x8_t)__a, 18); }
-__ai uint64x1_t vcgez_f64(float64x1_t __a) {
-  return (uint64x1_t)__builtin_neon_vcgez_v((int8x8_t)__a, 19); }
-__ai uint8x16_t vcgezq_s8(int8x16_t __a) {
-  return (uint8x16_t)__builtin_neon_vcgezq_v(__a, 48); }
-__ai uint16x8_t vcgezq_s16(int16x8_t __a) {
-  return (uint16x8_t)__builtin_neon_vcgezq_v((int8x16_t)__a, 49); }
-__ai uint32x4_t vcgezq_s32(int32x4_t __a) {
-  return (uint32x4_t)__builtin_neon_vcgezq_v((int8x16_t)__a, 50); }
-__ai uint64x2_t vcgezq_s64(int64x2_t __a) {
-  return (uint64x2_t)__builtin_neon_vcgezq_v((int8x16_t)__a, 51); }
-__ai uint32x4_t vcgezq_f32(float32x4_t __a) {
-  return (uint32x4_t)__builtin_neon_vcgezq_v((int8x16_t)__a, 50); }
-__ai uint64x2_t vcgezq_f64(float64x2_t __a) {
-  return (uint64x2_t)__builtin_neon_vcgezq_v((int8x16_t)__a, 51); }
-
-__ai uint8x8_t vcgtz_s8(int8x8_t __a) {
-  return (uint8x8_t)__builtin_neon_vcgtz_v(__a, 16); }
-__ai uint16x4_t vcgtz_s16(int16x4_t __a) {
-  return (uint16x4_t)__builtin_neon_vcgtz_v((int8x8_t)__a, 17); }
-__ai uint32x2_t vcgtz_s32(int32x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vcgtz_v((int8x8_t)__a, 18); }
-__ai uint64x1_t vcgtz_s64(int64x1_t __a) {
-  return (uint64x1_t)__builtin_neon_vcgtz_v((int8x8_t)__a, 19); }
-__ai uint32x2_t vcgtz_f32(float32x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vcgtz_v((int8x8_t)__a, 18); }
-__ai uint64x1_t vcgtz_f64(float64x1_t __a) {
-  return (uint64x1_t)__builtin_neon_vcgtz_v((int8x8_t)__a, 19); }
-__ai uint8x16_t vcgtzq_s8(int8x16_t __a) {
-  return (uint8x16_t)__builtin_neon_vcgtzq_v(__a, 48); }
-__ai uint16x8_t vcgtzq_s16(int16x8_t __a) {
-  return (uint16x8_t)__builtin_neon_vcgtzq_v((int8x16_t)__a, 49); }
-__ai uint32x4_t vcgtzq_s32(int32x4_t __a) {
-  return (uint32x4_t)__builtin_neon_vcgtzq_v((int8x16_t)__a, 50); }
-__ai uint64x2_t vcgtzq_s64(int64x2_t __a) {
-  return (uint64x2_t)__builtin_neon_vcgtzq_v((int8x16_t)__a, 51); }
-__ai uint32x4_t vcgtzq_f32(float32x4_t __a) {
-  return (uint32x4_t)__builtin_neon_vcgtzq_v((int8x16_t)__a, 50); }
-__ai uint64x2_t vcgtzq_f64(float64x2_t __a) {
-  return (uint64x2_t)__builtin_neon_vcgtzq_v((int8x16_t)__a, 51); }
-
-__ai uint8x8_t vclez_s8(int8x8_t __a) {
-  return (uint8x8_t)__builtin_neon_vclez_v(__a, 16); }
-__ai uint16x4_t vclez_s16(int16x4_t __a) {
-  return (uint16x4_t)__builtin_neon_vclez_v((int8x8_t)__a, 17); }
-__ai uint32x2_t vclez_s32(int32x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vclez_v((int8x8_t)__a, 18); }
-__ai uint64x1_t vclez_s64(int64x1_t __a) {
-  return (uint64x1_t)__builtin_neon_vclez_v((int8x8_t)__a, 19); }
-__ai uint32x2_t vclez_f32(float32x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vclez_v((int8x8_t)__a, 18); }
-__ai uint64x1_t vclez_f64(float64x1_t __a) {
-  return (uint64x1_t)__builtin_neon_vclez_v((int8x8_t)__a, 19); }
-__ai uint8x16_t vclezq_s8(int8x16_t __a) {
-  return (uint8x16_t)__builtin_neon_vclezq_v(__a, 48); }
-__ai uint16x8_t vclezq_s16(int16x8_t __a) {
-  return (uint16x8_t)__builtin_neon_vclezq_v((int8x16_t)__a, 49); }
-__ai uint32x4_t vclezq_s32(int32x4_t __a) {
-  return (uint32x4_t)__builtin_neon_vclezq_v((int8x16_t)__a, 50); }
-__ai uint64x2_t vclezq_s64(int64x2_t __a) {
-  return (uint64x2_t)__builtin_neon_vclezq_v((int8x16_t)__a, 51); }
-__ai uint32x4_t vclezq_f32(float32x4_t __a) {
-  return (uint32x4_t)__builtin_neon_vclezq_v((int8x16_t)__a, 50); }
-__ai uint64x2_t vclezq_f64(float64x2_t __a) {
-  return (uint64x2_t)__builtin_neon_vclezq_v((int8x16_t)__a, 51); }
-
-__ai uint8x8_t vcltz_s8(int8x8_t __a) {
-  return (uint8x8_t)__builtin_neon_vcltz_v(__a, 16); }
-__ai uint16x4_t vcltz_s16(int16x4_t __a) {
-  return (uint16x4_t)__builtin_neon_vcltz_v((int8x8_t)__a, 17); }
-__ai uint32x2_t vcltz_s32(int32x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vcltz_v((int8x8_t)__a, 18); }
-__ai uint64x1_t vcltz_s64(int64x1_t __a) {
-  return (uint64x1_t)__builtin_neon_vcltz_v((int8x8_t)__a, 19); }
-__ai uint32x2_t vcltz_f32(float32x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vcltz_v((int8x8_t)__a, 18); }
-__ai uint64x1_t vcltz_f64(float64x1_t __a) {
-  return (uint64x1_t)__builtin_neon_vcltz_v((int8x8_t)__a, 19); }
-__ai uint8x16_t vcltzq_s8(int8x16_t __a) {
-  return (uint8x16_t)__builtin_neon_vcltzq_v(__a, 48); }
-__ai uint16x8_t vcltzq_s16(int16x8_t __a) {
-  return (uint16x8_t)__builtin_neon_vcltzq_v((int8x16_t)__a, 49); }
-__ai uint32x4_t vcltzq_s32(int32x4_t __a) {
-  return (uint32x4_t)__builtin_neon_vcltzq_v((int8x16_t)__a, 50); }
-__ai uint64x2_t vcltzq_s64(int64x2_t __a) {
-  return (uint64x2_t)__builtin_neon_vcltzq_v((int8x16_t)__a, 51); }
-__ai uint32x4_t vcltzq_f32(float32x4_t __a) {
-  return (uint32x4_t)__builtin_neon_vcltzq_v((int8x16_t)__a, 50); }
-__ai uint64x2_t vcltzq_f64(float64x2_t __a) {
-  return (uint64x2_t)__builtin_neon_vcltzq_v((int8x16_t)__a, 51); }
-
-__ai uint64x1_t vtst_s64(int64x1_t __a, int64x1_t __b) {
-  return (uint64x1_t)__builtin_neon_vtst_v((int8x8_t)__a, (int8x8_t)__b, 19); }
-__ai uint64x1_t vtst_u64(uint64x1_t __a, uint64x1_t __b) {
-  return (uint64x1_t)__builtin_neon_vtst_v((int8x8_t)__a, (int8x8_t)__b, 19); }
-__ai uint64x1_t vtst_p64(poly64x1_t __a, poly64x1_t __b) {
-  return (uint64x1_t)__builtin_neon_vtst_v((int8x8_t)__a, (int8x8_t)__b, 19); }
-__ai uint64x2_t vtstq_s64(int64x2_t __a, int64x2_t __b) {
-  return (uint64x2_t)__builtin_neon_vtstq_v((int8x16_t)__a, (int8x16_t)__b, 51); }
-__ai uint64x2_t vtstq_u64(uint64x2_t __a, uint64x2_t __b) {
-  return (uint64x2_t)__builtin_neon_vtstq_v((int8x16_t)__a, (int8x16_t)__b, 51); }
-__ai uint64x2_t vtstq_p64(poly64x2_t __a, poly64x2_t __b) {
-  return (uint64x2_t)__builtin_neon_vtstq_v((int8x16_t)__a, (int8x16_t)__b, 51); }
-
-__ai float64x2_t vcombine_f64(float64x1_t __a, float64x1_t __b) {
-  return (float64x2_t)__builtin_shufflevector((int64x1_t)__a, (int64x1_t)__b, 0, 1); }
-__ai poly64x2_t vcombine_p64(poly64x1_t __a, poly64x1_t __b) {
-  return (poly64x2_t)__builtin_shufflevector((int64x1_t)__a, (int64x1_t)__b, 0, 1); }
-
-#define vcopyq_lane_s8(a, __b, c, __d) __extension__ ({ \
-  int8x16_t __a = (a); int8x8_t __c = (c); \
-  int8x16_t __a1 = __a; \
-  int8x8_t __c1 = __c; \
-  int8_t __c2 = vget_lane_s8(__c1, __d); \
-  vsetq_lane_s8(__c2, __a1, __b); })
-#define vcopyq_lane_s16(a, __b, c, __d) __extension__ ({ \
-  int16x8_t __a = (a); int16x4_t __c = (c); \
-  int16x8_t __a1 = __a; \
-  int16x4_t __c1 = __c; \
-  int16_t __c2 = vget_lane_s16(__c1, __d); \
-  vsetq_lane_s16(__c2, __a1, __b); })
-#define vcopyq_lane_s32(a, __b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int32x2_t __c = (c); \
-  int32x4_t __a1 = __a; \
-  int32x2_t __c1 = __c; \
-  int32_t __c2 = vget_lane_s32(__c1, __d); \
-  vsetq_lane_s32(__c2, __a1, __b); })
-#define vcopyq_lane_s64(a, __b, c, __d) __extension__ ({ \
-  int64x2_t __a = (a); int64x1_t __c = (c); \
-  int64x2_t __a1 = __a; \
-  int64x1_t __c1 = __c; \
-  int64_t __c2 = vget_lane_s64(__c1, __d); \
-  vsetq_lane_s64(__c2, __a1, __b); })
-#define vcopyq_lane_u8(a, __b, c, __d) __extension__ ({ \
-  uint8x16_t __a = (a); uint8x8_t __c = (c); \
-  uint8x16_t __a1 = __a; \
-  uint8x8_t __c1 = __c; \
-  uint8_t __c2 = vget_lane_u8(__c1, __d); \
-  vsetq_lane_u8(__c2, __a1, __b); })
-#define vcopyq_lane_u16(a, __b, c, __d) __extension__ ({ \
-  uint16x8_t __a = (a); uint16x4_t __c = (c); \
-  uint16x8_t __a1 = __a; \
-  uint16x4_t __c1 = __c; \
-  uint16_t __c2 = vget_lane_u16(__c1, __d); \
-  vsetq_lane_u16(__c2, __a1, __b); })
-#define vcopyq_lane_u32(a, __b, c, __d) __extension__ ({ \
-  uint32x4_t __a = (a); uint32x2_t __c = (c); \
-  uint32x4_t __a1 = __a; \
-  uint32x2_t __c1 = __c; \
-  uint32_t __c2 = vget_lane_u32(__c1, __d); \
-  vsetq_lane_u32(__c2, __a1, __b); })
-#define vcopyq_lane_u64(a, __b, c, __d) __extension__ ({ \
-  uint64x2_t __a = (a); uint64x1_t __c = (c); \
-  uint64x2_t __a1 = __a; \
-  uint64x1_t __c1 = __c; \
-  uint64_t __c2 = vget_lane_u64(__c1, __d); \
-  vsetq_lane_u64(__c2, __a1, __b); })
-#define vcopyq_lane_p8(a, __b, c, __d) __extension__ ({ \
-  poly8x16_t __a = (a); poly8x8_t __c = (c); \
-  poly8x16_t __a1 = __a; \
-  poly8x8_t __c1 = __c; \
-  poly8_t __c2 = vget_lane_p8(__c1, __d); \
-  vsetq_lane_p8(__c2, __a1, __b); })
-#define vcopyq_lane_p16(a, __b, c, __d) __extension__ ({ \
-  poly16x8_t __a = (a); poly16x4_t __c = (c); \
-  poly16x8_t __a1 = __a; \
-  poly16x4_t __c1 = __c; \
-  poly16_t __c2 = vget_lane_p16(__c1, __d); \
-  vsetq_lane_p16(__c2, __a1, __b); })
-#define vcopyq_lane_f32(a, __b, c, __d) __extension__ ({ \
-  float32x4_t __a = (a); float32x2_t __c = (c); \
-  float32x4_t __a1 = __a; \
-  float32x2_t __c1 = __c; \
-  float32_t __c2 = vget_lane_f32(__c1, __d); \
-  vsetq_lane_f32(__c2, __a1, __b); })
-#define vcopyq_lane_f64(a, __b, c, __d) __extension__ ({ \
-  float64x2_t __a = (a); float64x1_t __c = (c); \
-  float64x2_t __a1 = __a; \
-  float64x1_t __c1 = __c; \
-  float64_t __c2 = vget_lane_f64(__c1, __d); \
-  vsetq_lane_f64(__c2, __a1, __b); })
-#define vcopyq_lane_p64(a, __b, c, __d) __extension__ ({ \
-  poly64x2_t __a = (a); poly64x1_t __c = (c); \
-  poly64x2_t __a1 = __a; \
-  poly64x1_t __c1 = __c; \
-  poly64_t __c2 = vget_lane_p64(__c1, __d); \
-  vsetq_lane_p64(__c2, __a1, __b); })
-
-#define vcopyq_laneq_s8(a, __b, c, __d) __extension__ ({ \
-  int8x16_t __a = (a); int8x16_t __c = (c); \
-  int8x16_t __a1 = __a; \
-  int8x16_t __c1 = __c; \
-  int8_t __c2 = vgetq_lane_s8(__c1, __d); \
-  vsetq_lane_s8(__c2, __a1, __b); })
-#define vcopyq_laneq_s16(a, __b, c, __d) __extension__ ({ \
-  int16x8_t __a = (a); int16x8_t __c = (c); \
-  int16x8_t __a1 = __a; \
-  int16x8_t __c1 = __c; \
-  int16_t __c2 = vgetq_lane_s16(__c1, __d); \
-  vsetq_lane_s16(__c2, __a1, __b); })
-#define vcopyq_laneq_s32(a, __b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int32x4_t __c = (c); \
-  int32x4_t __a1 = __a; \
-  int32x4_t __c1 = __c; \
-  int32_t __c2 = vgetq_lane_s32(__c1, __d); \
-  vsetq_lane_s32(__c2, __a1, __b); })
-#define vcopyq_laneq_s64(a, __b, c, __d) __extension__ ({ \
-  int64x2_t __a = (a); int64x2_t __c = (c); \
-  int64x2_t __a1 = __a; \
-  int64x2_t __c1 = __c; \
-  int64_t __c2 = vgetq_lane_s64(__c1, __d); \
-  vsetq_lane_s64(__c2, __a1, __b); })
-#define vcopyq_laneq_u8(a, __b, c, __d) __extension__ ({ \
-  uint8x16_t __a = (a); uint8x16_t __c = (c); \
-  uint8x16_t __a1 = __a; \
-  uint8x16_t __c1 = __c; \
-  uint8_t __c2 = vgetq_lane_u8(__c1, __d); \
-  vsetq_lane_u8(__c2, __a1, __b); })
-#define vcopyq_laneq_u16(a, __b, c, __d) __extension__ ({ \
-  uint16x8_t __a = (a); uint16x8_t __c = (c); \
-  uint16x8_t __a1 = __a; \
-  uint16x8_t __c1 = __c; \
-  uint16_t __c2 = vgetq_lane_u16(__c1, __d); \
-  vsetq_lane_u16(__c2, __a1, __b); })
-#define vcopyq_laneq_u32(a, __b, c, __d) __extension__ ({ \
-  uint32x4_t __a = (a); uint32x4_t __c = (c); \
-  uint32x4_t __a1 = __a; \
-  uint32x4_t __c1 = __c; \
-  uint32_t __c2 = vgetq_lane_u32(__c1, __d); \
-  vsetq_lane_u32(__c2, __a1, __b); })
-#define vcopyq_laneq_u64(a, __b, c, __d) __extension__ ({ \
-  uint64x2_t __a = (a); uint64x2_t __c = (c); \
-  uint64x2_t __a1 = __a; \
-  uint64x2_t __c1 = __c; \
-  uint64_t __c2 = vgetq_lane_u64(__c1, __d); \
-  vsetq_lane_u64(__c2, __a1, __b); })
-#define vcopyq_laneq_p8(a, __b, c, __d) __extension__ ({ \
-  poly8x16_t __a = (a); poly8x16_t __c = (c); \
-  poly8x16_t __a1 = __a; \
-  poly8x16_t __c1 = __c; \
-  poly8_t __c2 = vgetq_lane_p8(__c1, __d); \
-  vsetq_lane_p8(__c2, __a1, __b); })
-#define vcopyq_laneq_p16(a, __b, c, __d) __extension__ ({ \
-  poly16x8_t __a = (a); poly16x8_t __c = (c); \
-  poly16x8_t __a1 = __a; \
-  poly16x8_t __c1 = __c; \
-  poly16_t __c2 = vgetq_lane_p16(__c1, __d); \
-  vsetq_lane_p16(__c2, __a1, __b); })
-#define vcopyq_laneq_f32(a, __b, c, __d) __extension__ ({ \
-  float32x4_t __a = (a); float32x4_t __c = (c); \
-  float32x4_t __a1 = __a; \
-  float32x4_t __c1 = __c; \
-  float32_t __c2 = vgetq_lane_f32(__c1, __d); \
-  vsetq_lane_f32(__c2, __a1, __b); })
-#define vcopyq_laneq_f64(a, __b, c, __d) __extension__ ({ \
-  float64x2_t __a = (a); float64x2_t __c = (c); \
-  float64x2_t __a1 = __a; \
-  float64x2_t __c1 = __c; \
-  float64_t __c2 = vgetq_lane_f64(__c1, __d); \
-  vsetq_lane_f64(__c2, __a1, __b); })
-#define vcopyq_laneq_p64(a, __b, c, __d) __extension__ ({ \
-  poly64x2_t __a = (a); poly64x2_t __c = (c); \
-  poly64x2_t __a1 = __a; \
-  poly64x2_t __c1 = __c; \
-  poly64_t __c2 = vgetq_lane_p64(__c1, __d); \
-  vsetq_lane_p64(__c2, __a1, __b); })
-
-#define vcopy_lane_s8(a, __b, c, __d) __extension__ ({ \
-  int8x8_t __a = (a); int8x8_t __c = (c); \
-  int8x8_t __a1 = __a; \
-  int8x8_t __c1 = __c; \
-  int8_t __c2 = vget_lane_s8(__c1, __d); \
-  vset_lane_s8(__c2, __a1, __b); })
-#define vcopy_lane_s16(a, __b, c, __d) __extension__ ({ \
-  int16x4_t __a = (a); int16x4_t __c = (c); \
-  int16x4_t __a1 = __a; \
-  int16x4_t __c1 = __c; \
-  int16_t __c2 = vget_lane_s16(__c1, __d); \
-  vset_lane_s16(__c2, __a1, __b); })
-#define vcopy_lane_s32(a, __b, c, __d) __extension__ ({ \
-  int32x2_t __a = (a); int32x2_t __c = (c); \
-  int32x2_t __a1 = __a; \
-  int32x2_t __c1 = __c; \
-  int32_t __c2 = vget_lane_s32(__c1, __d); \
-  vset_lane_s32(__c2, __a1, __b); })
-#define vcopy_lane_s64(a, __b, c, __d) __extension__ ({ \
-  int64x1_t __a = (a); int64x1_t __c = (c); \
-  int64x1_t __a1 = __a; \
-  int64x1_t __c1 = __c; \
-  int64_t __c2 = vget_lane_s64(__c1, __d); \
-  vset_lane_s64(__c2, __a1, __b); })
-#define vcopy_lane_u8(a, __b, c, __d) __extension__ ({ \
-  uint8x8_t __a = (a); uint8x8_t __c = (c); \
-  uint8x8_t __a1 = __a; \
-  uint8x8_t __c1 = __c; \
-  uint8_t __c2 = vget_lane_u8(__c1, __d); \
-  vset_lane_u8(__c2, __a1, __b); })
-#define vcopy_lane_u16(a, __b, c, __d) __extension__ ({ \
-  uint16x4_t __a = (a); uint16x4_t __c = (c); \
-  uint16x4_t __a1 = __a; \
-  uint16x4_t __c1 = __c; \
-  uint16_t __c2 = vget_lane_u16(__c1, __d); \
-  vset_lane_u16(__c2, __a1, __b); })
-#define vcopy_lane_u32(a, __b, c, __d) __extension__ ({ \
-  uint32x2_t __a = (a); uint32x2_t __c = (c); \
-  uint32x2_t __a1 = __a; \
-  uint32x2_t __c1 = __c; \
-  uint32_t __c2 = vget_lane_u32(__c1, __d); \
-  vset_lane_u32(__c2, __a1, __b); })
-#define vcopy_lane_u64(a, __b, c, __d) __extension__ ({ \
-  uint64x1_t __a = (a); uint64x1_t __c = (c); \
-  uint64x1_t __a1 = __a; \
-  uint64x1_t __c1 = __c; \
-  uint64_t __c2 = vget_lane_u64(__c1, __d); \
-  vset_lane_u64(__c2, __a1, __b); })
-#define vcopy_lane_p8(a, __b, c, __d) __extension__ ({ \
-  poly8x8_t __a = (a); poly8x8_t __c = (c); \
-  poly8x8_t __a1 = __a; \
-  poly8x8_t __c1 = __c; \
-  poly8_t __c2 = vget_lane_p8(__c1, __d); \
-  vset_lane_p8(__c2, __a1, __b); })
-#define vcopy_lane_p16(a, __b, c, __d) __extension__ ({ \
-  poly16x4_t __a = (a); poly16x4_t __c = (c); \
-  poly16x4_t __a1 = __a; \
-  poly16x4_t __c1 = __c; \
-  poly16_t __c2 = vget_lane_p16(__c1, __d); \
-  vset_lane_p16(__c2, __a1, __b); })
-#define vcopy_lane_p64(a, __b, c, __d) __extension__ ({ \
-  poly64x1_t __a = (a); poly64x1_t __c = (c); \
-  poly64x1_t __a1 = __a; \
-  poly64x1_t __c1 = __c; \
-  poly64_t __c2 = vget_lane_p64(__c1, __d); \
-  vset_lane_p64(__c2, __a1, __b); })
-#define vcopy_lane_f32(a, __b, c, __d) __extension__ ({ \
-  float32x2_t __a = (a); float32x2_t __c = (c); \
-  float32x2_t __a1 = __a; \
-  float32x2_t __c1 = __c; \
-  float32_t __c2 = vget_lane_f32(__c1, __d); \
-  vset_lane_f32(__c2, __a1, __b); })
-#define vcopy_lane_f64(a, __b, c, __d) __extension__ ({ \
-  float64x1_t __a = (a); float64x1_t __c = (c); \
-  float64x1_t __a1 = __a; \
-  float64x1_t __c1 = __c; \
-  float64_t __c2 = vget_lane_f64(__c1, __d); \
-  vset_lane_f64(__c2, __a1, __b); })
-
-#define vcopy_laneq_s8(a, __b, c, __d) __extension__ ({ \
-  int8x8_t __a = (a); int8x16_t __c = (c); \
-  int8x8_t __a1 = __a; \
-  int8x16_t __c1 = __c; \
-  int8_t __c2 = vgetq_lane_s8(__c1, __d); \
-  vset_lane_s8(__c2, __a1, __b); })
-#define vcopy_laneq_s16(a, __b, c, __d) __extension__ ({ \
-  int16x4_t __a = (a); int16x8_t __c = (c); \
-  int16x4_t __a1 = __a; \
-  int16x8_t __c1 = __c; \
-  int16_t __c2 = vgetq_lane_s16(__c1, __d); \
-  vset_lane_s16(__c2, __a1, __b); })
-#define vcopy_laneq_s32(a, __b, c, __d) __extension__ ({ \
-  int32x2_t __a = (a); int32x4_t __c = (c); \
-  int32x2_t __a1 = __a; \
-  int32x4_t __c1 = __c; \
-  int32_t __c2 = vgetq_lane_s32(__c1, __d); \
-  vset_lane_s32(__c2, __a1, __b); })
-#define vcopy_laneq_s64(a, __b, c, __d) __extension__ ({ \
-  int64x1_t __a = (a); int64x2_t __c = (c); \
-  int64x1_t __a1 = __a; \
-  int64x2_t __c1 = __c; \
-  int64_t __c2 = vgetq_lane_s64(__c1, __d); \
-  vset_lane_s64(__c2, __a1, __b); })
-#define vcopy_laneq_p8(a, __b, c, __d) __extension__ ({ \
-  poly8x8_t __a = (a); poly8x16_t __c = (c); \
-  poly8x8_t __a1 = __a; \
-  poly8x16_t __c1 = __c; \
-  poly8_t __c2 = vgetq_lane_p8(__c1, __d); \
-  vset_lane_p8(__c2, __a1, __b); })
-#define vcopy_laneq_p16(a, __b, c, __d) __extension__ ({ \
-  poly16x4_t __a = (a); poly16x8_t __c = (c); \
-  poly16x4_t __a1 = __a; \
-  poly16x8_t __c1 = __c; \
-  poly16_t __c2 = vgetq_lane_p16(__c1, __d); \
-  vset_lane_p16(__c2, __a1, __b); })
-#define vcopy_laneq_p64(a, __b, c, __d) __extension__ ({ \
-  poly64x1_t __a = (a); poly64x2_t __c = (c); \
-  poly64x1_t __a1 = __a; \
-  poly64x2_t __c1 = __c; \
-  poly64_t __c2 = vgetq_lane_p64(__c1, __d); \
-  vset_lane_p64(__c2, __a1, __b); })
-#define vcopy_laneq_u8(a, __b, c, __d) __extension__ ({ \
-  uint8x8_t __a = (a); uint8x16_t __c = (c); \
-  uint8x8_t __a1 = __a; \
-  uint8x16_t __c1 = __c; \
-  uint8_t __c2 = vgetq_lane_u8(__c1, __d); \
-  vset_lane_u8(__c2, __a1, __b); })
-#define vcopy_laneq_u16(a, __b, c, __d) __extension__ ({ \
-  uint16x4_t __a = (a); uint16x8_t __c = (c); \
-  uint16x4_t __a1 = __a; \
-  uint16x8_t __c1 = __c; \
-  uint16_t __c2 = vgetq_lane_u16(__c1, __d); \
-  vset_lane_u16(__c2, __a1, __b); })
-#define vcopy_laneq_u32(a, __b, c, __d) __extension__ ({ \
-  uint32x2_t __a = (a); uint32x4_t __c = (c); \
-  uint32x2_t __a1 = __a; \
-  uint32x4_t __c1 = __c; \
-  uint32_t __c2 = vgetq_lane_u32(__c1, __d); \
-  vset_lane_u32(__c2, __a1, __b); })
-#define vcopy_laneq_u64(a, __b, c, __d) __extension__ ({ \
-  uint64x1_t __a = (a); uint64x2_t __c = (c); \
-  uint64x1_t __a1 = __a; \
-  uint64x2_t __c1 = __c; \
-  uint64_t __c2 = vgetq_lane_u64(__c1, __d); \
-  vset_lane_u64(__c2, __a1, __b); })
-#define vcopy_laneq_f32(a, __b, c, __d) __extension__ ({ \
-  float32x2_t __a = (a); float32x4_t __c = (c); \
-  float32x2_t __a1 = __a; \
-  float32x4_t __c1 = __c; \
-  float32_t __c2 = vgetq_lane_f32(__c1, __d); \
-  vset_lane_f32(__c2, __a1, __b); })
-#define vcopy_laneq_f64(a, __b, c, __d) __extension__ ({ \
-  float64x1_t __a = (a); float64x2_t __c = (c); \
-  float64x1_t __a1 = __a; \
-  float64x2_t __c1 = __c; \
-  float64_t __c2 = vgetq_lane_f64(__c1, __d); \
-  vset_lane_f64(__c2, __a1, __b); })
-
-__ai float64x1_t vcreate_f64(uint64_t __a) {
-  return (float64x1_t)__a; }
-__ai poly64x1_t vcreate_p64(uint64_t __a) {
-  return (poly64x1_t)__a; }
-
-#define vcvt_n_f64_s64(a, __b) __extension__ ({ \
-  int64x1_t __a = (a); \
-  (float64x1_t)__builtin_neon_vcvt_n_f64_v((int8x8_t)__a, __b, 3); })
-#define vcvt_n_f64_u64(a, __b) __extension__ ({ \
-  uint64x1_t __a = (a); \
-  (float64x1_t)__builtin_neon_vcvt_n_f64_v((int8x8_t)__a, __b, 19); })
-#define vcvtq_n_f64_s64(a, __b) __extension__ ({ \
-  int64x2_t __a = (a); \
-  (float64x2_t)__builtin_neon_vcvtq_n_f64_v((int8x16_t)__a, __b, 35); })
-#define vcvtq_n_f64_u64(a, __b) __extension__ ({ \
-  uint64x2_t __a = (a); \
-  (float64x2_t)__builtin_neon_vcvtq_n_f64_v((int8x16_t)__a, __b, 51); })
-
-__ai float64x1_t vdup_n_f64(float64_t __a) {
-  return (float64x1_t){ __a }; }
-__ai float64x2_t vdupq_n_f64(float64_t __a) {
-  return (float64x2_t){ __a, __a }; }
-__ai poly64x1_t vdup_n_p64(poly64_t __a) {
-  return (poly64x1_t){ __a }; }
-__ai poly64x2_t vdupq_n_p64(poly64_t __a) {
-  return (poly64x2_t){ __a, __a }; }
-
-__ai uint64x1_t vcage_f64(float64x1_t __a, float64x1_t __b) {
-  return (uint64x1_t)__builtin_neon_vcage_v((int8x8_t)__a, (int8x8_t)__b, 19); }
-__ai uint64x2_t vcageq_f64(float64x2_t __a, float64x2_t __b) {
-  return (uint64x2_t)__builtin_neon_vcageq_v((int8x16_t)__a, (int8x16_t)__b, 51); }
-
-__ai uint64x1_t vcagt_f64(float64x1_t __a, float64x1_t __b) {
-  return (uint64x1_t)__builtin_neon_vcagt_v((int8x8_t)__a, (int8x8_t)__b, 19); }
-__ai uint64x2_t vcagtq_f64(float64x2_t __a, float64x2_t __b) {
-  return (uint64x2_t)__builtin_neon_vcagtq_v((int8x16_t)__a, (int8x16_t)__b, 51); }
-
-__ai uint64x1_t vcale_f64(float64x1_t __a, float64x1_t __b) {
-  return (uint64x1_t)__builtin_neon_vcale_v((int8x8_t)__a, (int8x8_t)__b, 19); }
-__ai uint64x2_t vcaleq_f64(float64x2_t __a, float64x2_t __b) {
-  return (uint64x2_t)__builtin_neon_vcaleq_v((int8x16_t)__a, (int8x16_t)__b, 51); }
-
-__ai uint64x1_t vcalt_f64(float64x1_t __a, float64x1_t __b) {
-  return (uint64x1_t)__builtin_neon_vcalt_v((int8x8_t)__a, (int8x8_t)__b, 19); }
-__ai uint64x2_t vcaltq_f64(float64x2_t __a, float64x2_t __b) {
-  return (uint64x2_t)__builtin_neon_vcaltq_v((int8x16_t)__a, (int8x16_t)__b, 51); }
+#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
 
-#if __ARM_ARCH >= 8
-__ai int32x2_t vcvta_s32_f32(float32x2_t __a) {
-  return (int32x2_t)__builtin_neon_vcvta_s32_v((int8x8_t)__a, 2); }
-__ai int32x4_t vcvtaq_s32_f32(float32x4_t __a) {
-  return (int32x4_t)__builtin_neon_vcvtaq_s32_v((int8x16_t)__a, 34); }
+#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
 
-#if __ARM_ARCH >= 8 && defined(__aarch64__)
-__ai int64x1_t vcvta_s64_f64(float64x1_t __a) {
-  return (int64x1_t)__builtin_neon_vcvta_s64_v((int8x8_t)__a, 3); }
-__ai int64x2_t vcvtaq_s64_f64(float64x2_t __a) {
-  return (int64x2_t)__builtin_neon_vcvtaq_s64_v((int8x16_t)__a, 35); }
+#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
 
-#if __ARM_ARCH >= 8
-__ai uint32x2_t vcvta_u32_f32(float32x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vcvta_u32_v((int8x8_t)__a, 18); }
-__ai uint32x4_t vcvtaq_u32_f32(float32x4_t __a) {
-  return (uint32x4_t)__builtin_neon_vcvtaq_u32_v((int8x16_t)__a, 50); }
+#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
 
-#if __ARM_ARCH >= 8 && defined(__aarch64__)
-__ai uint64x1_t vcvta_u64_f64(float64x1_t __a) {
-  return (uint64x1_t)__builtin_neon_vcvta_u64_v((int8x8_t)__a, 19); }
-__ai uint64x2_t vcvtaq_u64_f64(float64x2_t __a) {
-  return (uint64x2_t)__builtin_neon_vcvtaq_u64_v((int8x16_t)__a, 51); }
+#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
 
-#if __ARM_ARCH >= 8
-__ai int32x2_t vcvtm_s32_f32(float32x2_t __a) {
-  return (int32x2_t)__builtin_neon_vcvtm_s32_v((int8x8_t)__a, 2); }
-__ai int32x4_t vcvtmq_s32_f32(float32x4_t __a) {
-  return (int32x4_t)__builtin_neon_vcvtmq_s32_v((int8x16_t)__a, 34); }
+#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
 
-#if __ARM_ARCH >= 8 && defined(__aarch64__)
-__ai int64x1_t vcvtm_s64_f64(float64x1_t __a) {
-  return (int64x1_t)__builtin_neon_vcvtm_s64_v((int8x8_t)__a, 3); }
-__ai int64x2_t vcvtmq_s64_f64(float64x2_t __a) {
-  return (int64x2_t)__builtin_neon_vcvtmq_s64_v((int8x16_t)__a, 35); }
+#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
 
-#if __ARM_ARCH >= 8
-__ai uint32x2_t vcvtm_u32_f32(float32x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vcvtm_u32_v((int8x8_t)__a, 18); }
-__ai uint32x4_t vcvtmq_u32_f32(float32x4_t __a) {
-  return (uint32x4_t)__builtin_neon_vcvtmq_u32_v((int8x16_t)__a, 50); }
+#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
 
-#if __ARM_ARCH >= 8 && defined(__aarch64__)
-__ai uint64x1_t vcvtm_u64_f64(float64x1_t __a) {
-  return (uint64x1_t)__builtin_neon_vcvtm_u64_v((int8x8_t)__a, 19); }
-__ai uint64x2_t vcvtmq_u64_f64(float64x2_t __a) {
-  return (uint64x2_t)__builtin_neon_vcvtmq_u64_v((int8x16_t)__a, 51); }
+#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
 
-#if __ARM_ARCH >= 8
-__ai int32x2_t vcvtn_s32_f32(float32x2_t __a) {
-  return (int32x2_t)__builtin_neon_vcvtn_s32_v((int8x8_t)__a, 2); }
-__ai int32x4_t vcvtnq_s32_f32(float32x4_t __a) {
-  return (int32x4_t)__builtin_neon_vcvtnq_s32_v((int8x16_t)__a, 34); }
+#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
 
-#if __ARM_ARCH >= 8 && defined(__aarch64__)
-__ai int64x1_t vcvtn_s64_f64(float64x1_t __a) {
-  return (int64x1_t)__builtin_neon_vcvtn_s64_v((int8x8_t)__a, 3); }
-__ai int64x2_t vcvtnq_s64_f64(float64x2_t __a) {
-  return (int64x2_t)__builtin_neon_vcvtnq_s64_v((int8x16_t)__a, 35); }
+#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
 
-#if __ARM_ARCH >= 8
-__ai uint32x2_t vcvtn_u32_f32(float32x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vcvtn_u32_v((int8x8_t)__a, 18); }
-__ai uint32x4_t vcvtnq_u32_f32(float32x4_t __a) {
-  return (uint32x4_t)__builtin_neon_vcvtnq_u32_v((int8x16_t)__a, 50); }
+#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
 
-#if __ARM_ARCH >= 8 && defined(__aarch64__)
-__ai uint64x1_t vcvtn_u64_f64(float64x1_t __a) {
-  return (uint64x1_t)__builtin_neon_vcvtn_u64_v((int8x8_t)__a, 19); }
-__ai uint64x2_t vcvtnq_u64_f64(float64x2_t __a) {
-  return (uint64x2_t)__builtin_neon_vcvtnq_u64_v((int8x16_t)__a, 51); }
+#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
 
-#if __ARM_ARCH >= 8
-__ai int32x2_t vcvtp_s32_f32(float32x2_t __a) {
-  return (int32x2_t)__builtin_neon_vcvtp_s32_v((int8x8_t)__a, 2); }
-__ai int32x4_t vcvtpq_s32_f32(float32x4_t __a) {
-  return (int32x4_t)__builtin_neon_vcvtpq_s32_v((int8x16_t)__a, 34); }
+#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
 
-#if __ARM_ARCH >= 8 && defined(__aarch64__)
-__ai int64x1_t vcvtp_s64_f64(float64x1_t __a) {
-  return (int64x1_t)__builtin_neon_vcvtp_s64_v((int8x8_t)__a, 3); }
-__ai int64x2_t vcvtpq_s64_f64(float64x2_t __a) {
-  return (int64x2_t)__builtin_neon_vcvtpq_s64_v((int8x16_t)__a, 35); }
+#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
-__ai uint32x2_t vcvtp_u32_f32(float32x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vcvtp_u32_v((int8x8_t)__a, 18); }
-__ai uint32x4_t vcvtpq_u32_f32(float32x4_t __a) {
-  return (uint32x4_t)__builtin_neon_vcvtpq_u32_v((int8x16_t)__a, 50); }
+#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(__aarch64__)
-__ai uint64x1_t vcvtp_u64_f64(float64x1_t __a) {
-  return (uint64x1_t)__builtin_neon_vcvtp_u64_v((int8x8_t)__a, 19); }
-__ai uint64x2_t vcvtpq_u64_f64(float64x2_t __a) {
-  return (uint64x2_t)__builtin_neon_vcvtpq_u64_v((int8x16_t)__a, 51); }
+#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
 
-#if defined(__aarch64__)
-#define vcvt_n_s64_f64(a, __b) __extension__ ({ \
-  float64x1_t __a = (a); \
-  (int64x1_t)__builtin_neon_vcvt_n_s64_v((int8x8_t)__a, __b, 3); })
-#define vcvtq_n_s64_f64(a, __b) __extension__ ({ \
-  float64x2_t __a = (a); \
-  (int64x2_t)__builtin_neon_vcvtq_n_s64_v((int8x16_t)__a, __b, 35); })
-
-#define vcvt_n_u64_f64(a, __b) __extension__ ({ \
-  float64x1_t __a = (a); \
-  (uint64x1_t)__builtin_neon_vcvt_n_u64_v((int8x8_t)__a, __b, 19); })
-#define vcvtq_n_u64_f64(a, __b) __extension__ ({ \
-  float64x2_t __a = (a); \
-  (uint64x2_t)__builtin_neon_vcvtq_n_u64_v((int8x16_t)__a, __b, 51); })
-
-__ai float32x2_t vdiv_f32(float32x2_t __a, float32x2_t __b) {
-  return __a / __b; }
-__ai float64x1_t vdiv_f64(float64x1_t __a, float64x1_t __b) {
-  return __a / __b; }
-__ai float32x4_t vdivq_f32(float32x4_t __a, float32x4_t __b) {
-  return __a / __b; }
-__ai float64x2_t vdivq_f64(float64x2_t __a, float64x2_t __b) {
-  return __a / __b; }
-
-__ai float32x2_t vmaxnm_f32(float32x2_t __a, float32x2_t __b) {
-  return (float32x2_t)__builtin_neon_vmaxnm_v((int8x8_t)__a, (int8x8_t)__b, 9); }
-__ai float64x1_t vmaxnm_f64(float64x1_t __a, float64x1_t __b) {
-  return (float64x1_t)__builtin_neon_vmaxnm_v((int8x8_t)__a, (int8x8_t)__b, 10); }
-__ai float32x4_t vmaxnmq_f32(float32x4_t __a, float32x4_t __b) {
-  return (float32x4_t)__builtin_neon_vmaxnmq_v((int8x16_t)__a, (int8x16_t)__b, 41); }
-__ai float64x2_t vmaxnmq_f64(float64x2_t __a, float64x2_t __b) {
-  return (float64x2_t)__builtin_neon_vmaxnmq_v((int8x16_t)__a, (int8x16_t)__b, 42); }
-
-__ai float32x2_t vpmaxnm_f32(float32x2_t __a, float32x2_t __b) {
-  return (float32x2_t)__builtin_neon_vpmaxnm_v((int8x8_t)__a, (int8x8_t)__b, 9); }
-__ai float32x4_t vpmaxnmq_f32(float32x4_t __a, float32x4_t __b) {
-  return (float32x4_t)__builtin_neon_vpmaxnmq_v((int8x16_t)__a, (int8x16_t)__b, 41); }
-__ai float64x2_t vpmaxnmq_f64(float64x2_t __a, float64x2_t __b) {
-  return (float64x2_t)__builtin_neon_vpmaxnmq_v((int8x16_t)__a, (int8x16_t)__b, 42); }
-
-__ai float32_t vmaxnmv_f32(float32x2_t __a) {
-  return (float32_t)__builtin_neon_vmaxnmv_f32(__a); }
-__ai float32_t vmaxnmvq_f32(float32x4_t __a) {
-  return (float32_t)__builtin_neon_vmaxnmvq_f32(__a); }
-__ai float64_t vmaxnmvq_f64(float64x2_t __a) {
-  return (float64_t)__builtin_neon_vmaxnmvq_f64(__a); }
-
-__ai float32x2_t vminnm_f32(float32x2_t __a, float32x2_t __b) {
-  return (float32x2_t)__builtin_neon_vminnm_v((int8x8_t)__a, (int8x8_t)__b, 9); }
-__ai float64x1_t vminnm_f64(float64x1_t __a, float64x1_t __b) {
-  return (float64x1_t)__builtin_neon_vminnm_v((int8x8_t)__a, (int8x8_t)__b, 10); }
-__ai float32x4_t vminnmq_f32(float32x4_t __a, float32x4_t __b) {
-  return (float32x4_t)__builtin_neon_vminnmq_v((int8x16_t)__a, (int8x16_t)__b, 41); }
-__ai float64x2_t vminnmq_f64(float64x2_t __a, float64x2_t __b) {
-  return (float64x2_t)__builtin_neon_vminnmq_v((int8x16_t)__a, (int8x16_t)__b, 42); }
-
-__ai float32x2_t vpminnm_f32(float32x2_t __a, float32x2_t __b) {
-  return (float32x2_t)__builtin_neon_vpminnm_v((int8x8_t)__a, (int8x8_t)__b, 9); }
-__ai float32x4_t vpminnmq_f32(float32x4_t __a, float32x4_t __b) {
-  return (float32x4_t)__builtin_neon_vpminnmq_v((int8x16_t)__a, (int8x16_t)__b, 41); }
-__ai float64x2_t vpminnmq_f64(float64x2_t __a, float64x2_t __b) {
-  return (float64x2_t)__builtin_neon_vpminnmq_v((int8x16_t)__a, (int8x16_t)__b, 42); }
-
-__ai float32_t vminnmv_f32(float32x2_t __a) {
-  return (float32_t)__builtin_neon_vminnmv_f32(__a); }
-__ai float32_t vminnmvq_f32(float32x4_t __a) {
-  return (float32_t)__builtin_neon_vminnmvq_f32(__a); }
-__ai float64_t vminnmvq_f64(float64x2_t __a) {
-  return (float64_t)__builtin_neon_vminnmvq_f64(__a); }
-
-__ai float64x1_t vfma_f64(float64x1_t __a, float64x1_t __b, float64x1_t __c) {
-  return (float64x1_t)__builtin_neon_vfma_v((int8x8_t)__a, (int8x8_t)__b, (int8x8_t)__c, 10); }
-__ai float64x2_t vfmaq_f64(float64x2_t __a, float64x2_t __b, float64x2_t __c) {
-  return (float64x2_t)__builtin_neon_vfmaq_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 42); }
-
-__ai float32x2_t vfma_n_f32(float32x2_t __a, float32x2_t __b, float32_t __c) {
-  return vfma_f32(__a, __b, (float32x2_t){ __c, __c }); }
-__ai float32x4_t vfmaq_n_f32(float32x4_t __a, float32x4_t __b, float32_t __c) {
-  return vfmaq_f32(__a, __b, (float32x4_t){ __c, __c, __c, __c }); }
-__ai float64x2_t vfmaq_n_f64(float64x2_t __a, float64x2_t __b, float64_t __c) {
-  return vfmaq_f64(__a, __b, (float64x2_t){ __c, __c }); }
-
-__ai float32x2_t vfms_f32(float32x2_t __a, float32x2_t __b, float32x2_t __c) {
-  return (float32x2_t)__builtin_neon_vfms_v((int8x8_t)__a, (int8x8_t)__b, (int8x8_t)__c, 9); }
-__ai float64x1_t vfms_f64(float64x1_t __a, float64x1_t __b, float64x1_t __c) {
-  return (float64x1_t)__builtin_neon_vfms_v((int8x8_t)__a, (int8x8_t)__b, (int8x8_t)__c, 10); }
-__ai float32x4_t vfmsq_f32(float32x4_t __a, float32x4_t __b, float32x4_t __c) {
-  return (float32x4_t)__builtin_neon_vfmsq_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 41); }
-__ai float64x2_t vfmsq_f64(float64x2_t __a, float64x2_t __b, float64x2_t __c) {
-  return (float64x2_t)__builtin_neon_vfmsq_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 42); }
-
-__ai float32x2_t vfms_n_f32(float32x2_t __a, float32x2_t __b, float32_t __c) {
-  return vfms_f32(__a, __b, (float32x2_t){ __c, __c }); }
-__ai float32x4_t vfmsq_n_f32(float32x4_t __a, float32x4_t __b, float32_t __c) {
-  return vfmsq_f32(__a, __b, (float32x4_t){ __c, __c, __c, __c }); }
-__ai float64x2_t vfmsq_n_f64(float64x2_t __a, float64x2_t __b, float64_t __c) {
-  return vfmsq_f64(__a, __b, (float64x2_t){ __c, __c }); }
-
-__ai float64x1_t vrecpe_f64(float64x1_t __a) {
-  return (float64x1_t)__builtin_neon_vrecpe_v((int8x8_t)__a, 10); }
-__ai float64x2_t vrecpeq_f64(float64x2_t __a) {
-  return (float64x2_t)__builtin_neon_vrecpeq_v((int8x16_t)__a, 42); }
-
-__ai float64x1_t vrecps_f64(float64x1_t __a, float64x1_t __b) {
-  return (float64x1_t)__builtin_neon_vrecps_v((int8x8_t)__a, (int8x8_t)__b, 10); }
-__ai float64x2_t vrecpsq_f64(float64x2_t __a, float64x2_t __b) {
-  return (float64x2_t)__builtin_neon_vrecpsq_v((int8x16_t)__a, (int8x16_t)__b, 42); }
-
-__ai float32x2_t vrnda_f32(float32x2_t __a) {
-  return (float32x2_t)__builtin_neon_vrnda_v((int8x8_t)__a, 9); }
-__ai float64x1_t vrnda_f64(float64x1_t __a) {
-  return (float64x1_t)__builtin_neon_vrnda_v((int8x8_t)__a, 10); }
-__ai float32x4_t vrndaq_f32(float32x4_t __a) {
-  return (float32x4_t)__builtin_neon_vrndaq_v((int8x16_t)__a, 41); }
-__ai float64x2_t vrndaq_f64(float64x2_t __a) {
-  return (float64x2_t)__builtin_neon_vrndaq_v((int8x16_t)__a, 42); }
-
-__ai float32x2_t vrndi_f32(float32x2_t __a) {
-  return (float32x2_t)__builtin_neon_vrndi_v((int8x8_t)__a, 9); }
-__ai float64x1_t vrndi_f64(float64x1_t __a) {
-  return (float64x1_t)__builtin_neon_vrndi_v((int8x8_t)__a, 10); }
-__ai float32x4_t vrndiq_f32(float32x4_t __a) {
-  return (float32x4_t)__builtin_neon_vrndiq_v((int8x16_t)__a, 41); }
-__ai float64x2_t vrndiq_f64(float64x2_t __a) {
-  return (float64x2_t)__builtin_neon_vrndiq_v((int8x16_t)__a, 42); }
-
-__ai float32x2_t vrndm_f32(float32x2_t __a) {
-  return (float32x2_t)__builtin_neon_vrndm_v((int8x8_t)__a, 9); }
-__ai float64x1_t vrndm_f64(float64x1_t __a) {
-  return (float64x1_t)__builtin_neon_vrndm_v((int8x8_t)__a, 10); }
-__ai float32x4_t vrndmq_f32(float32x4_t __a) {
-  return (float32x4_t)__builtin_neon_vrndmq_v((int8x16_t)__a, 41); }
-__ai float64x2_t vrndmq_f64(float64x2_t __a) {
-  return (float64x2_t)__builtin_neon_vrndmq_v((int8x16_t)__a, 42); }
-
-__ai float32x2_t vrndn_f32(float32x2_t __a) {
-  return (float32x2_t)__builtin_neon_vrndn_v((int8x8_t)__a, 9); }
-__ai float64x1_t vrndn_f64(float64x1_t __a) {
-  return (float64x1_t)__builtin_neon_vrndn_v((int8x8_t)__a, 10); }
-__ai float32x4_t vrndnq_f32(float32x4_t __a) {
-  return (float32x4_t)__builtin_neon_vrndnq_v((int8x16_t)__a, 41); }
-__ai float64x2_t vrndnq_f64(float64x2_t __a) {
-  return (float64x2_t)__builtin_neon_vrndnq_v((int8x16_t)__a, 42); }
-
-__ai float32x2_t vrndp_f32(float32x2_t __a) {
-  return (float32x2_t)__builtin_neon_vrndp_v((int8x8_t)__a, 9); }
-__ai float64x1_t vrndp_f64(float64x1_t __a) {
-  return (float64x1_t)__builtin_neon_vrndp_v((int8x8_t)__a, 10); }
-__ai float32x4_t vrndpq_f32(float32x4_t __a) {
-  return (float32x4_t)__builtin_neon_vrndpq_v((int8x16_t)__a, 41); }
-__ai float64x2_t vrndpq_f64(float64x2_t __a) {
-  return (float64x2_t)__builtin_neon_vrndpq_v((int8x16_t)__a, 42); }
-
-__ai float32x2_t vrndx_f32(float32x2_t __a) {
-  return (float32x2_t)__builtin_neon_vrndx_v((int8x8_t)__a, 9); }
-__ai float64x1_t vrndx_f64(float64x1_t __a) {
-  return (float64x1_t)__builtin_neon_vrndx_v((int8x8_t)__a, 10); }
-__ai float32x4_t vrndxq_f32(float32x4_t __a) {
-  return (float32x4_t)__builtin_neon_vrndxq_v((int8x16_t)__a, 41); }
-__ai float64x2_t vrndxq_f64(float64x2_t __a) {
-  return (float64x2_t)__builtin_neon_vrndxq_v((int8x16_t)__a, 42); }
-
-__ai float32x2_t vrnd_f32(float32x2_t __a) {
-  return (float32x2_t)__builtin_neon_vrnd_v((int8x8_t)__a, 9); }
-__ai float64x1_t vrnd_f64(float64x1_t __a) {
-  return (float64x1_t)__builtin_neon_vrnd_v((int8x8_t)__a, 10); }
-__ai float32x4_t vrndq_f32(float32x4_t __a) {
-  return (float32x4_t)__builtin_neon_vrndq_v((int8x16_t)__a, 41); }
-__ai float64x2_t vrndq_f64(float64x2_t __a) {
-  return (float64x2_t)__builtin_neon_vrndq_v((int8x16_t)__a, 42); }
-
-__ai float64x1_t vrsqrte_f64(float64x1_t __a) {
-  return (float64x1_t)__builtin_neon_vrsqrte_v((int8x8_t)__a, 10); }
-__ai float64x2_t vrsqrteq_f64(float64x2_t __a) {
-  return (float64x2_t)__builtin_neon_vrsqrteq_v((int8x16_t)__a, 42); }
-
-__ai float64x1_t vrsqrts_f64(float64x1_t __a, float64x1_t __b) {
-  return (float64x1_t)__builtin_neon_vrsqrts_v((int8x8_t)__a, (int8x8_t)__b, 10); }
-__ai float64x2_t vrsqrtsq_f64(float64x2_t __a, float64x2_t __b) {
-  return (float64x2_t)__builtin_neon_vrsqrtsq_v((int8x16_t)__a, (int8x16_t)__b, 42); }
-
-__ai float32x2_t vsqrt_f32(float32x2_t __a) {
-  return (float32x2_t)__builtin_neon_vsqrt_v((int8x8_t)__a, 9); }
-__ai float64x1_t vsqrt_f64(float64x1_t __a) {
-  return (float64x1_t)__builtin_neon_vsqrt_v((int8x8_t)__a, 10); }
-__ai float32x4_t vsqrtq_f32(float32x4_t __a) {
-  return (float32x4_t)__builtin_neon_vsqrtq_v((int8x16_t)__a, 41); }
-__ai float64x2_t vsqrtq_f64(float64x2_t __a) {
-  return (float64x2_t)__builtin_neon_vsqrtq_v((int8x16_t)__a, 42); }
-
-#define vget_lane_f64(a, __b) __extension__ ({ \
-  float64x1_t __a = (a); \
-  (float64_t)__builtin_neon_vget_lane_f64(__a, __b); })
-#define vgetq_lane_f64(a, __b) __extension__ ({ \
-  float64x2_t __a = (a); \
-  (float64_t)__builtin_neon_vgetq_lane_f64(__a, __b); })
-#define vget_lane_p64(a, __b) __extension__ ({ \
-  poly64x1_t __a = (a); \
-  (poly64_t)__builtin_neon_vget_lane_i64((int64x1_t)__a, __b); })
-#define vgetq_lane_p64(a, __b) __extension__ ({ \
-  poly64x2_t __a = (a); \
-  (poly64_t)__builtin_neon_vgetq_lane_i64((int64x2_t)__a, __b); })
-
-#define vld1_f64(__a) __extension__ ({ \
-  (float64x1_t)__builtin_neon_vld1_v(__a, 10); })
-#define vld1q_f64(__a) __extension__ ({ \
-  (float64x2_t)__builtin_neon_vld1q_v(__a, 42); })
-#define vld1_p64(__a) __extension__ ({ \
-  (poly64x1_t)__builtin_neon_vld1_v(__a, 6); })
-#define vld1q_p64(__a) __extension__ ({ \
-  (poly64x2_t)__builtin_neon_vld1q_v(__a, 38); })
-
-#define vld1_dup_f64(__a) __extension__ ({ \
-  (float64x1_t)__builtin_neon_vld1_dup_v(__a, 10); })
-#define vld1q_dup_f64(__a) __extension__ ({ \
-  (float64x2_t)__builtin_neon_vld1q_dup_v(__a, 42); })
-#define vld1_dup_p64(__a) __extension__ ({ \
-  (poly64x1_t)__builtin_neon_vld1_dup_v(__a, 6); })
-#define vld1q_dup_p64(__a) __extension__ ({ \
-  (poly64x2_t)__builtin_neon_vld1q_dup_v(__a, 38); })
-
-#define vld1_lane_f64(__a, b, __c) __extension__ ({ \
-  float64x1_t __b = (b); \
-  (float64x1_t)__builtin_neon_vld1_lane_v(__a, (int8x8_t)__b, __c, 10); })
-#define vld1q_lane_f64(__a, b, __c) __extension__ ({ \
-  float64x2_t __b = (b); \
-  (float64x2_t)__builtin_neon_vld1q_lane_v(__a, (int8x16_t)__b, __c, 42); })
-#define vld1_lane_p64(__a, b, __c) __extension__ ({ \
-  poly64x1_t __b = (b); \
-  (poly64x1_t)__builtin_neon_vld1_lane_v(__a, (int8x8_t)__b, __c, 6); })
-#define vld1q_lane_p64(__a, b, __c) __extension__ ({ \
-  poly64x2_t __b = (b); \
-  (poly64x2_t)__builtin_neon_vld1q_lane_v(__a, (int8x16_t)__b, __c, 38); })
-
-#define vld1q_u8_x2(__a) __extension__ ({ \
-  uint8x16x2_t r; __builtin_neon_vld1q_x2_v(&r, __a, 48); r; })
-#define vld1q_u16_x2(__a) __extension__ ({ \
-  uint16x8x2_t r; __builtin_neon_vld1q_x2_v(&r, __a, 49); r; })
-#define vld1q_u32_x2(__a) __extension__ ({ \
-  uint32x4x2_t r; __builtin_neon_vld1q_x2_v(&r, __a, 50); r; })
-#define vld1q_s8_x2(__a) __extension__ ({ \
-  int8x16x2_t r; __builtin_neon_vld1q_x2_v(&r, __a, 32); r; })
-#define vld1q_s16_x2(__a) __extension__ ({ \
-  int16x8x2_t r; __builtin_neon_vld1q_x2_v(&r, __a, 33); r; })
-#define vld1q_s32_x2(__a) __extension__ ({ \
-  int32x4x2_t r; __builtin_neon_vld1q_x2_v(&r, __a, 34); r; })
-#define vld1q_f16_x2(__a) __extension__ ({ \
-  float16x8x2_t r; __builtin_neon_vld1q_x2_v(&r, __a, 40); r; })
-#define vld1q_f32_x2(__a) __extension__ ({ \
-  float32x4x2_t r; __builtin_neon_vld1q_x2_v(&r, __a, 41); r; })
-#define vld1q_p8_x2(__a) __extension__ ({ \
-  poly8x16x2_t r; __builtin_neon_vld1q_x2_v(&r, __a, 36); r; })
-#define vld1q_p16_x2(__a) __extension__ ({ \
-  poly16x8x2_t r; __builtin_neon_vld1q_x2_v(&r, __a, 37); r; })
-#define vld1_u8_x2(__a) __extension__ ({ \
-  uint8x8x2_t r; __builtin_neon_vld1_x2_v(&r, __a, 16); r; })
-#define vld1_u16_x2(__a) __extension__ ({ \
-  uint16x4x2_t r; __builtin_neon_vld1_x2_v(&r, __a, 17); r; })
-#define vld1_u32_x2(__a) __extension__ ({ \
-  uint32x2x2_t r; __builtin_neon_vld1_x2_v(&r, __a, 18); r; })
-#define vld1_u64_x2(__a) __extension__ ({ \
-  uint64x1x2_t r; __builtin_neon_vld1_x2_v(&r, __a, 19); r; })
-#define vld1_s8_x2(__a) __extension__ ({ \
-  int8x8x2_t r; __builtin_neon_vld1_x2_v(&r, __a, 0); r; })
-#define vld1_s16_x2(__a) __extension__ ({ \
-  int16x4x2_t r; __builtin_neon_vld1_x2_v(&r, __a, 1); r; })
-#define vld1_s32_x2(__a) __extension__ ({ \
-  int32x2x2_t r; __builtin_neon_vld1_x2_v(&r, __a, 2); r; })
-#define vld1_s64_x2(__a) __extension__ ({ \
-  int64x1x2_t r; __builtin_neon_vld1_x2_v(&r, __a, 3); r; })
-#define vld1_f16_x2(__a) __extension__ ({ \
-  float16x4x2_t r; __builtin_neon_vld1_x2_v(&r, __a, 8); r; })
-#define vld1_f32_x2(__a) __extension__ ({ \
-  float32x2x2_t r; __builtin_neon_vld1_x2_v(&r, __a, 9); r; })
-#define vld1_p8_x2(__a) __extension__ ({ \
-  poly8x8x2_t r; __builtin_neon_vld1_x2_v(&r, __a, 4); r; })
-#define vld1_p16_x2(__a) __extension__ ({ \
-  poly16x4x2_t r; __builtin_neon_vld1_x2_v(&r, __a, 5); r; })
-#define vld1q_u64_x2(__a) __extension__ ({ \
-  uint64x2x2_t r; __builtin_neon_vld1q_x2_v(&r, __a, 51); r; })
-#define vld1q_s64_x2(__a) __extension__ ({ \
-  int64x2x2_t r; __builtin_neon_vld1q_x2_v(&r, __a, 35); r; })
-#define vld1_f64_x2(__a) __extension__ ({ \
-  float64x1x2_t r; __builtin_neon_vld1_x2_v(&r, __a, 10); r; })
-#define vld1q_f64_x2(__a) __extension__ ({ \
-  float64x2x2_t r; __builtin_neon_vld1q_x2_v(&r, __a, 42); r; })
-#define vld1_p64_x2(__a) __extension__ ({ \
-  poly64x1x2_t r; __builtin_neon_vld1_x2_v(&r, __a, 6); r; })
-#define vld1q_p64_x2(__a) __extension__ ({ \
-  poly64x2x2_t r; __builtin_neon_vld1q_x2_v(&r, __a, 38); r; })
-
-#define vld2q_u64(__a) __extension__ ({ \
-  uint64x2x2_t r; __builtin_neon_vld2q_v(&r, __a, 51); r; })
-#define vld2q_s64(__a) __extension__ ({ \
-  int64x2x2_t r; __builtin_neon_vld2q_v(&r, __a, 35); r; })
-#define vld2_f64(__a) __extension__ ({ \
-  float64x1x2_t r; __builtin_neon_vld2_v(&r, __a, 10); r; })
-#define vld2q_f64(__a) __extension__ ({ \
-  float64x2x2_t r; __builtin_neon_vld2q_v(&r, __a, 42); r; })
-#define vld2_p64(__a) __extension__ ({ \
-  poly64x1x2_t r; __builtin_neon_vld2_v(&r, __a, 6); r; })
-#define vld2q_p64(__a) __extension__ ({ \
-  poly64x2x2_t r; __builtin_neon_vld2q_v(&r, __a, 38); r; })
-
-#define vld2q_dup_u8(__a) __extension__ ({ \
-  uint8x16x2_t r; __builtin_neon_vld2q_dup_v(&r, __a, 48); r; })
-#define vld2q_dup_u16(__a) __extension__ ({ \
-  uint16x8x2_t r; __builtin_neon_vld2q_dup_v(&r, __a, 49); r; })
-#define vld2q_dup_u32(__a) __extension__ ({ \
-  uint32x4x2_t r; __builtin_neon_vld2q_dup_v(&r, __a, 50); r; })
-#define vld2q_dup_u64(__a) __extension__ ({ \
-  uint64x2x2_t r; __builtin_neon_vld2q_dup_v(&r, __a, 51); r; })
-#define vld2q_dup_s8(__a) __extension__ ({ \
-  int8x16x2_t r; __builtin_neon_vld2q_dup_v(&r, __a, 32); r; })
-#define vld2q_dup_s16(__a) __extension__ ({ \
-  int16x8x2_t r; __builtin_neon_vld2q_dup_v(&r, __a, 33); r; })
-#define vld2q_dup_s32(__a) __extension__ ({ \
-  int32x4x2_t r; __builtin_neon_vld2q_dup_v(&r, __a, 34); r; })
-#define vld2q_dup_s64(__a) __extension__ ({ \
-  int64x2x2_t r; __builtin_neon_vld2q_dup_v(&r, __a, 35); r; })
-#define vld2q_dup_f16(__a) __extension__ ({ \
-  float16x8x2_t r; __builtin_neon_vld2q_dup_v(&r, __a, 40); r; })
-#define vld2q_dup_f32(__a) __extension__ ({ \
-  float32x4x2_t r; __builtin_neon_vld2q_dup_v(&r, __a, 41); r; })
-#define vld2q_dup_f64(__a) __extension__ ({ \
-  float64x2x2_t r; __builtin_neon_vld2q_dup_v(&r, __a, 42); r; })
-#define vld2q_dup_p8(__a) __extension__ ({ \
-  poly8x16x2_t r; __builtin_neon_vld2q_dup_v(&r, __a, 36); r; })
-#define vld2q_dup_p16(__a) __extension__ ({ \
-  poly16x8x2_t r; __builtin_neon_vld2q_dup_v(&r, __a, 37); r; })
-#define vld2q_dup_p64(__a) __extension__ ({ \
-  poly64x2x2_t r; __builtin_neon_vld2q_dup_v(&r, __a, 38); r; })
-#define vld2_dup_f64(__a) __extension__ ({ \
-  float64x1x2_t r; __builtin_neon_vld2_dup_v(&r, __a, 10); r; })
-#define vld2_dup_p64(__a) __extension__ ({ \
-  poly64x1x2_t r; __builtin_neon_vld2_dup_v(&r, __a, 6); r; })
-
-#define vld2_lane_s64(__a, b, __c) __extension__ ({ \
-  int64x1x2_t __b = (b); \
-  int64x1x2_t r; __builtin_neon_vld2_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 3); r; })
-#define vld2_lane_u64(__a, b, __c) __extension__ ({ \
-  uint64x1x2_t __b = (b); \
-  uint64x1x2_t r; __builtin_neon_vld2_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 19); r; })
-#define vld2q_lane_s8(__a, b, __c) __extension__ ({ \
-  int8x16x2_t __b = (b); \
-  int8x16x2_t r; __builtin_neon_vld2q_lane_v(&r, __a, __b.val[0], __b.val[1], __c, 32); r; })
-#define vld2q_lane_u8(__a, b, __c) __extension__ ({ \
-  uint8x16x2_t __b = (b); \
-  uint8x16x2_t r; __builtin_neon_vld2q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 48); r; })
-#define vld2q_lane_p8(__a, b, __c) __extension__ ({ \
-  poly8x16x2_t __b = (b); \
-  poly8x16x2_t r; __builtin_neon_vld2q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 36); r; })
-#define vld2q_lane_s64(__a, b, __c) __extension__ ({ \
-  int64x2x2_t __b = (b); \
-  int64x2x2_t r; __builtin_neon_vld2q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 35); r; })
-#define vld2q_lane_u64(__a, b, __c) __extension__ ({ \
-  uint64x2x2_t __b = (b); \
-  uint64x2x2_t r; __builtin_neon_vld2q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 51); r; })
-#define vld2_lane_f64(__a, b, __c) __extension__ ({ \
-  float64x1x2_t __b = (b); \
-  float64x1x2_t r; __builtin_neon_vld2_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 10); r; })
-#define vld2q_lane_f64(__a, b, __c) __extension__ ({ \
-  float64x2x2_t __b = (b); \
-  float64x2x2_t r; __builtin_neon_vld2q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 42); r; })
-#define vld2_lane_p64(__a, b, __c) __extension__ ({ \
-  poly64x1x2_t __b = (b); \
-  poly64x1x2_t r; __builtin_neon_vld2_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 6); r; })
-#define vld2q_lane_p64(__a, b, __c) __extension__ ({ \
-  poly64x2x2_t __b = (b); \
-  poly64x2x2_t r; __builtin_neon_vld2q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 38); r; })
-
-#define vld3q_u64(__a) __extension__ ({ \
-  uint64x2x3_t r; __builtin_neon_vld3q_v(&r, __a, 51); r; })
-#define vld3q_s64(__a) __extension__ ({ \
-  int64x2x3_t r; __builtin_neon_vld3q_v(&r, __a, 35); r; })
-#define vld3_f64(__a) __extension__ ({ \
-  float64x1x3_t r; __builtin_neon_vld3_v(&r, __a, 10); r; })
-#define vld3q_f64(__a) __extension__ ({ \
-  float64x2x3_t r; __builtin_neon_vld3q_v(&r, __a, 42); r; })
-#define vld3_p64(__a) __extension__ ({ \
-  poly64x1x3_t r; __builtin_neon_vld3_v(&r, __a, 6); r; })
-#define vld3q_p64(__a) __extension__ ({ \
-  poly64x2x3_t r; __builtin_neon_vld3q_v(&r, __a, 38); r; })
-
-#define vld3q_dup_u8(__a) __extension__ ({ \
-  uint8x16x3_t r; __builtin_neon_vld3q_dup_v(&r, __a, 48); r; })
-#define vld3q_dup_u16(__a) __extension__ ({ \
-  uint16x8x3_t r; __builtin_neon_vld3q_dup_v(&r, __a, 49); r; })
-#define vld3q_dup_u32(__a) __extension__ ({ \
-  uint32x4x3_t r; __builtin_neon_vld3q_dup_v(&r, __a, 50); r; })
-#define vld3q_dup_u64(__a) __extension__ ({ \
-  uint64x2x3_t r; __builtin_neon_vld3q_dup_v(&r, __a, 51); r; })
-#define vld3q_dup_s8(__a) __extension__ ({ \
-  int8x16x3_t r; __builtin_neon_vld3q_dup_v(&r, __a, 32); r; })
-#define vld3q_dup_s16(__a) __extension__ ({ \
-  int16x8x3_t r; __builtin_neon_vld3q_dup_v(&r, __a, 33); r; })
-#define vld3q_dup_s32(__a) __extension__ ({ \
-  int32x4x3_t r; __builtin_neon_vld3q_dup_v(&r, __a, 34); r; })
-#define vld3q_dup_s64(__a) __extension__ ({ \
-  int64x2x3_t r; __builtin_neon_vld3q_dup_v(&r, __a, 35); r; })
-#define vld3q_dup_f16(__a) __extension__ ({ \
-  float16x8x3_t r; __builtin_neon_vld3q_dup_v(&r, __a, 40); r; })
-#define vld3q_dup_f32(__a) __extension__ ({ \
-  float32x4x3_t r; __builtin_neon_vld3q_dup_v(&r, __a, 41); r; })
-#define vld3q_dup_f64(__a) __extension__ ({ \
-  float64x2x3_t r; __builtin_neon_vld3q_dup_v(&r, __a, 42); r; })
-#define vld3q_dup_p8(__a) __extension__ ({ \
-  poly8x16x3_t r; __builtin_neon_vld3q_dup_v(&r, __a, 36); r; })
-#define vld3q_dup_p16(__a) __extension__ ({ \
-  poly16x8x3_t r; __builtin_neon_vld3q_dup_v(&r, __a, 37); r; })
-#define vld3q_dup_p64(__a) __extension__ ({ \
-  poly64x2x3_t r; __builtin_neon_vld3q_dup_v(&r, __a, 38); r; })
-#define vld3_dup_f64(__a) __extension__ ({ \
-  float64x1x3_t r; __builtin_neon_vld3_dup_v(&r, __a, 10); r; })
-#define vld3_dup_p64(__a) __extension__ ({ \
-  poly64x1x3_t r; __builtin_neon_vld3_dup_v(&r, __a, 6); r; })
-
-#define vld3_lane_s64(__a, b, __c) __extension__ ({ \
-  int64x1x3_t __b = (b); \
-  int64x1x3_t r; __builtin_neon_vld3_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 3); r; })
-#define vld3_lane_u64(__a, b, __c) __extension__ ({ \
-  uint64x1x3_t __b = (b); \
-  uint64x1x3_t r; __builtin_neon_vld3_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 19); r; })
-#define vld3q_lane_s8(__a, b, __c) __extension__ ({ \
-  int8x16x3_t __b = (b); \
-  int8x16x3_t r; __builtin_neon_vld3q_lane_v(&r, __a, __b.val[0], __b.val[1], __b.val[2], __c, 32); r; })
-#define vld3q_lane_u8(__a, b, __c) __extension__ ({ \
-  uint8x16x3_t __b = (b); \
-  uint8x16x3_t r; __builtin_neon_vld3q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 48); r; })
-#define vld3q_lane_p8(__a, b, __c) __extension__ ({ \
-  poly8x16x3_t __b = (b); \
-  poly8x16x3_t r; __builtin_neon_vld3q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 36); r; })
-#define vld3q_lane_s64(__a, b, __c) __extension__ ({ \
-  int64x2x3_t __b = (b); \
-  int64x2x3_t r; __builtin_neon_vld3q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 35); r; })
-#define vld3q_lane_u64(__a, b, __c) __extension__ ({ \
-  uint64x2x3_t __b = (b); \
-  uint64x2x3_t r; __builtin_neon_vld3q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 51); r; })
-#define vld3_lane_f64(__a, b, __c) __extension__ ({ \
-  float64x1x3_t __b = (b); \
-  float64x1x3_t r; __builtin_neon_vld3_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 10); r; })
-#define vld3q_lane_f64(__a, b, __c) __extension__ ({ \
-  float64x2x3_t __b = (b); \
-  float64x2x3_t r; __builtin_neon_vld3q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 42); r; })
-#define vld3_lane_p64(__a, b, __c) __extension__ ({ \
-  poly64x1x3_t __b = (b); \
-  poly64x1x3_t r; __builtin_neon_vld3_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 6); r; })
-#define vld3q_lane_p64(__a, b, __c) __extension__ ({ \
-  poly64x2x3_t __b = (b); \
-  poly64x2x3_t r; __builtin_neon_vld3q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 38); r; })
-
-#define vld1q_u8_x3(__a) __extension__ ({ \
-  uint8x16x3_t r; __builtin_neon_vld1q_x3_v(&r, __a, 48); r; })
-#define vld1q_u16_x3(__a) __extension__ ({ \
-  uint16x8x3_t r; __builtin_neon_vld1q_x3_v(&r, __a, 49); r; })
-#define vld1q_u32_x3(__a) __extension__ ({ \
-  uint32x4x3_t r; __builtin_neon_vld1q_x3_v(&r, __a, 50); r; })
-#define vld1q_s8_x3(__a) __extension__ ({ \
-  int8x16x3_t r; __builtin_neon_vld1q_x3_v(&r, __a, 32); r; })
-#define vld1q_s16_x3(__a) __extension__ ({ \
-  int16x8x3_t r; __builtin_neon_vld1q_x3_v(&r, __a, 33); r; })
-#define vld1q_s32_x3(__a) __extension__ ({ \
-  int32x4x3_t r; __builtin_neon_vld1q_x3_v(&r, __a, 34); r; })
-#define vld1q_f16_x3(__a) __extension__ ({ \
-  float16x8x3_t r; __builtin_neon_vld1q_x3_v(&r, __a, 40); r; })
-#define vld1q_f32_x3(__a) __extension__ ({ \
-  float32x4x3_t r; __builtin_neon_vld1q_x3_v(&r, __a, 41); r; })
-#define vld1q_p8_x3(__a) __extension__ ({ \
-  poly8x16x3_t r; __builtin_neon_vld1q_x3_v(&r, __a, 36); r; })
-#define vld1q_p16_x3(__a) __extension__ ({ \
-  poly16x8x3_t r; __builtin_neon_vld1q_x3_v(&r, __a, 37); r; })
-#define vld1_u8_x3(__a) __extension__ ({ \
-  uint8x8x3_t r; __builtin_neon_vld1_x3_v(&r, __a, 16); r; })
-#define vld1_u16_x3(__a) __extension__ ({ \
-  uint16x4x3_t r; __builtin_neon_vld1_x3_v(&r, __a, 17); r; })
-#define vld1_u32_x3(__a) __extension__ ({ \
-  uint32x2x3_t r; __builtin_neon_vld1_x3_v(&r, __a, 18); r; })
-#define vld1_u64_x3(__a) __extension__ ({ \
-  uint64x1x3_t r; __builtin_neon_vld1_x3_v(&r, __a, 19); r; })
-#define vld1_s8_x3(__a) __extension__ ({ \
-  int8x8x3_t r; __builtin_neon_vld1_x3_v(&r, __a, 0); r; })
-#define vld1_s16_x3(__a) __extension__ ({ \
-  int16x4x3_t r; __builtin_neon_vld1_x3_v(&r, __a, 1); r; })
-#define vld1_s32_x3(__a) __extension__ ({ \
-  int32x2x3_t r; __builtin_neon_vld1_x3_v(&r, __a, 2); r; })
-#define vld1_s64_x3(__a) __extension__ ({ \
-  int64x1x3_t r; __builtin_neon_vld1_x3_v(&r, __a, 3); r; })
-#define vld1_f16_x3(__a) __extension__ ({ \
-  float16x4x3_t r; __builtin_neon_vld1_x3_v(&r, __a, 8); r; })
-#define vld1_f32_x3(__a) __extension__ ({ \
-  float32x2x3_t r; __builtin_neon_vld1_x3_v(&r, __a, 9); r; })
-#define vld1_p8_x3(__a) __extension__ ({ \
-  poly8x8x3_t r; __builtin_neon_vld1_x3_v(&r, __a, 4); r; })
-#define vld1_p16_x3(__a) __extension__ ({ \
-  poly16x4x3_t r; __builtin_neon_vld1_x3_v(&r, __a, 5); r; })
-#define vld1q_u64_x3(__a) __extension__ ({ \
-  uint64x2x3_t r; __builtin_neon_vld1q_x3_v(&r, __a, 51); r; })
-#define vld1q_s64_x3(__a) __extension__ ({ \
-  int64x2x3_t r; __builtin_neon_vld1q_x3_v(&r, __a, 35); r; })
-#define vld1_f64_x3(__a) __extension__ ({ \
-  float64x1x3_t r; __builtin_neon_vld1_x3_v(&r, __a, 10); r; })
-#define vld1q_f64_x3(__a) __extension__ ({ \
-  float64x2x3_t r; __builtin_neon_vld1q_x3_v(&r, __a, 42); r; })
-#define vld1_p64_x3(__a) __extension__ ({ \
-  poly64x1x3_t r; __builtin_neon_vld1_x3_v(&r, __a, 6); r; })
-#define vld1q_p64_x3(__a) __extension__ ({ \
-  poly64x2x3_t r; __builtin_neon_vld1q_x3_v(&r, __a, 38); r; })
-
-#define vld4q_u64(__a) __extension__ ({ \
-  uint64x2x4_t r; __builtin_neon_vld4q_v(&r, __a, 51); r; })
-#define vld4q_s64(__a) __extension__ ({ \
-  int64x2x4_t r; __builtin_neon_vld4q_v(&r, __a, 35); r; })
-#define vld4_f64(__a) __extension__ ({ \
-  float64x1x4_t r; __builtin_neon_vld4_v(&r, __a, 10); r; })
-#define vld4q_f64(__a) __extension__ ({ \
-  float64x2x4_t r; __builtin_neon_vld4q_v(&r, __a, 42); r; })
-#define vld4_p64(__a) __extension__ ({ \
-  poly64x1x4_t r; __builtin_neon_vld4_v(&r, __a, 6); r; })
-#define vld4q_p64(__a) __extension__ ({ \
-  poly64x2x4_t r; __builtin_neon_vld4q_v(&r, __a, 38); r; })
-
-#define vld4q_dup_u8(__a) __extension__ ({ \
-  uint8x16x4_t r; __builtin_neon_vld4q_dup_v(&r, __a, 48); r; })
-#define vld4q_dup_u16(__a) __extension__ ({ \
-  uint16x8x4_t r; __builtin_neon_vld4q_dup_v(&r, __a, 49); r; })
-#define vld4q_dup_u32(__a) __extension__ ({ \
-  uint32x4x4_t r; __builtin_neon_vld4q_dup_v(&r, __a, 50); r; })
-#define vld4q_dup_u64(__a) __extension__ ({ \
-  uint64x2x4_t r; __builtin_neon_vld4q_dup_v(&r, __a, 51); r; })
-#define vld4q_dup_s8(__a) __extension__ ({ \
-  int8x16x4_t r; __builtin_neon_vld4q_dup_v(&r, __a, 32); r; })
-#define vld4q_dup_s16(__a) __extension__ ({ \
-  int16x8x4_t r; __builtin_neon_vld4q_dup_v(&r, __a, 33); r; })
-#define vld4q_dup_s32(__a) __extension__ ({ \
-  int32x4x4_t r; __builtin_neon_vld4q_dup_v(&r, __a, 34); r; })
-#define vld4q_dup_s64(__a) __extension__ ({ \
-  int64x2x4_t r; __builtin_neon_vld4q_dup_v(&r, __a, 35); r; })
-#define vld4q_dup_f16(__a) __extension__ ({ \
-  float16x8x4_t r; __builtin_neon_vld4q_dup_v(&r, __a, 40); r; })
-#define vld4q_dup_f32(__a) __extension__ ({ \
-  float32x4x4_t r; __builtin_neon_vld4q_dup_v(&r, __a, 41); r; })
-#define vld4q_dup_f64(__a) __extension__ ({ \
-  float64x2x4_t r; __builtin_neon_vld4q_dup_v(&r, __a, 42); r; })
-#define vld4q_dup_p8(__a) __extension__ ({ \
-  poly8x16x4_t r; __builtin_neon_vld4q_dup_v(&r, __a, 36); r; })
-#define vld4q_dup_p16(__a) __extension__ ({ \
-  poly16x8x4_t r; __builtin_neon_vld4q_dup_v(&r, __a, 37); r; })
-#define vld4q_dup_p64(__a) __extension__ ({ \
-  poly64x2x4_t r; __builtin_neon_vld4q_dup_v(&r, __a, 38); r; })
-#define vld4_dup_f64(__a) __extension__ ({ \
-  float64x1x4_t r; __builtin_neon_vld4_dup_v(&r, __a, 10); r; })
-#define vld4_dup_p64(__a) __extension__ ({ \
-  poly64x1x4_t r; __builtin_neon_vld4_dup_v(&r, __a, 6); r; })
-
-#define vld4_lane_s64(__a, b, __c) __extension__ ({ \
-  int64x1x4_t __b = (b); \
-  int64x1x4_t r; __builtin_neon_vld4_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 3); r; })
-#define vld4_lane_u64(__a, b, __c) __extension__ ({ \
-  uint64x1x4_t __b = (b); \
-  uint64x1x4_t r; __builtin_neon_vld4_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 19); r; })
-#define vld4q_lane_s8(__a, b, __c) __extension__ ({ \
-  int8x16x4_t __b = (b); \
-  int8x16x4_t r; __builtin_neon_vld4q_lane_v(&r, __a, __b.val[0], __b.val[1], __b.val[2], __b.val[3], __c, 32); r; })
-#define vld4q_lane_u8(__a, b, __c) __extension__ ({ \
-  uint8x16x4_t __b = (b); \
-  uint8x16x4_t r; __builtin_neon_vld4q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 48); r; })
-#define vld4q_lane_p8(__a, b, __c) __extension__ ({ \
-  poly8x16x4_t __b = (b); \
-  poly8x16x4_t r; __builtin_neon_vld4q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 36); r; })
-#define vld4q_lane_s64(__a, b, __c) __extension__ ({ \
-  int64x2x4_t __b = (b); \
-  int64x2x4_t r; __builtin_neon_vld4q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 35); r; })
-#define vld4q_lane_u64(__a, b, __c) __extension__ ({ \
-  uint64x2x4_t __b = (b); \
-  uint64x2x4_t r; __builtin_neon_vld4q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 51); r; })
-#define vld4_lane_f64(__a, b, __c) __extension__ ({ \
-  float64x1x4_t __b = (b); \
-  float64x1x4_t r; __builtin_neon_vld4_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 10); r; })
-#define vld4q_lane_f64(__a, b, __c) __extension__ ({ \
-  float64x2x4_t __b = (b); \
-  float64x2x4_t r; __builtin_neon_vld4q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 42); r; })
-#define vld4_lane_p64(__a, b, __c) __extension__ ({ \
-  poly64x1x4_t __b = (b); \
-  poly64x1x4_t r; __builtin_neon_vld4_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 6); r; })
-#define vld4q_lane_p64(__a, b, __c) __extension__ ({ \
-  poly64x2x4_t __b = (b); \
-  poly64x2x4_t r; __builtin_neon_vld4q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 38); r; })
-
-#define vld1q_u8_x4(__a) __extension__ ({ \
-  uint8x16x4_t r; __builtin_neon_vld1q_x4_v(&r, __a, 48); r; })
-#define vld1q_u16_x4(__a) __extension__ ({ \
-  uint16x8x4_t r; __builtin_neon_vld1q_x4_v(&r, __a, 49); r; })
-#define vld1q_u32_x4(__a) __extension__ ({ \
-  uint32x4x4_t r; __builtin_neon_vld1q_x4_v(&r, __a, 50); r; })
-#define vld1q_s8_x4(__a) __extension__ ({ \
-  int8x16x4_t r; __builtin_neon_vld1q_x4_v(&r, __a, 32); r; })
-#define vld1q_s16_x4(__a) __extension__ ({ \
-  int16x8x4_t r; __builtin_neon_vld1q_x4_v(&r, __a, 33); r; })
-#define vld1q_s32_x4(__a) __extension__ ({ \
-  int32x4x4_t r; __builtin_neon_vld1q_x4_v(&r, __a, 34); r; })
-#define vld1q_f16_x4(__a) __extension__ ({ \
-  float16x8x4_t r; __builtin_neon_vld1q_x4_v(&r, __a, 40); r; })
-#define vld1q_f32_x4(__a) __extension__ ({ \
-  float32x4x4_t r; __builtin_neon_vld1q_x4_v(&r, __a, 41); r; })
-#define vld1q_p8_x4(__a) __extension__ ({ \
-  poly8x16x4_t r; __builtin_neon_vld1q_x4_v(&r, __a, 36); r; })
-#define vld1q_p16_x4(__a) __extension__ ({ \
-  poly16x8x4_t r; __builtin_neon_vld1q_x4_v(&r, __a, 37); r; })
-#define vld1_u8_x4(__a) __extension__ ({ \
-  uint8x8x4_t r; __builtin_neon_vld1_x4_v(&r, __a, 16); r; })
-#define vld1_u16_x4(__a) __extension__ ({ \
-  uint16x4x4_t r; __builtin_neon_vld1_x4_v(&r, __a, 17); r; })
-#define vld1_u32_x4(__a) __extension__ ({ \
-  uint32x2x4_t r; __builtin_neon_vld1_x4_v(&r, __a, 18); r; })
-#define vld1_u64_x4(__a) __extension__ ({ \
-  uint64x1x4_t r; __builtin_neon_vld1_x4_v(&r, __a, 19); r; })
-#define vld1_s8_x4(__a) __extension__ ({ \
-  int8x8x4_t r; __builtin_neon_vld1_x4_v(&r, __a, 0); r; })
-#define vld1_s16_x4(__a) __extension__ ({ \
-  int16x4x4_t r; __builtin_neon_vld1_x4_v(&r, __a, 1); r; })
-#define vld1_s32_x4(__a) __extension__ ({ \
-  int32x2x4_t r; __builtin_neon_vld1_x4_v(&r, __a, 2); r; })
-#define vld1_s64_x4(__a) __extension__ ({ \
-  int64x1x4_t r; __builtin_neon_vld1_x4_v(&r, __a, 3); r; })
-#define vld1_f16_x4(__a) __extension__ ({ \
-  float16x4x4_t r; __builtin_neon_vld1_x4_v(&r, __a, 8); r; })
-#define vld1_f32_x4(__a) __extension__ ({ \
-  float32x2x4_t r; __builtin_neon_vld1_x4_v(&r, __a, 9); r; })
-#define vld1_p8_x4(__a) __extension__ ({ \
-  poly8x8x4_t r; __builtin_neon_vld1_x4_v(&r, __a, 4); r; })
-#define vld1_p16_x4(__a) __extension__ ({ \
-  poly16x4x4_t r; __builtin_neon_vld1_x4_v(&r, __a, 5); r; })
-#define vld1q_u64_x4(__a) __extension__ ({ \
-  uint64x2x4_t r; __builtin_neon_vld1q_x4_v(&r, __a, 51); r; })
-#define vld1q_s64_x4(__a) __extension__ ({ \
-  int64x2x4_t r; __builtin_neon_vld1q_x4_v(&r, __a, 35); r; })
-#define vld1_f64_x4(__a) __extension__ ({ \
-  float64x1x4_t r; __builtin_neon_vld1_x4_v(&r, __a, 10); r; })
-#define vld1q_f64_x4(__a) __extension__ ({ \
-  float64x2x4_t r; __builtin_neon_vld1q_x4_v(&r, __a, 42); r; })
-#define vld1_p64_x4(__a) __extension__ ({ \
-  poly64x1x4_t r; __builtin_neon_vld1_x4_v(&r, __a, 6); r; })
-#define vld1q_p64_x4(__a) __extension__ ({ \
-  poly64x2x4_t r; __builtin_neon_vld1q_x4_v(&r, __a, 38); r; })
-
-__ai float64x1_t vmax_f64(float64x1_t __a, float64x1_t __b) {
-  return (float64x1_t)__builtin_neon_vmax_v((int8x8_t)__a, (int8x8_t)__b, 10); }
-__ai float64x2_t vmaxq_f64(float64x2_t __a, float64x2_t __b) {
-  return (float64x2_t)__builtin_neon_vmaxq_v((int8x16_t)__a, (int8x16_t)__b, 42); }
-
-__ai int8x16_t vpmaxq_s8(int8x16_t __a, int8x16_t __b) {
-  return (int8x16_t)__builtin_neon_vpmaxq_v(__a, __b, 32); }
-__ai int16x8_t vpmaxq_s16(int16x8_t __a, int16x8_t __b) {
-  return (int16x8_t)__builtin_neon_vpmaxq_v((int8x16_t)__a, (int8x16_t)__b, 33); }
-__ai int32x4_t vpmaxq_s32(int32x4_t __a, int32x4_t __b) {
-  return (int32x4_t)__builtin_neon_vpmaxq_v((int8x16_t)__a, (int8x16_t)__b, 34); }
-__ai uint8x16_t vpmaxq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vpmaxq_v((int8x16_t)__a, (int8x16_t)__b, 48); }
-__ai uint16x8_t vpmaxq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint16x8_t)__builtin_neon_vpmaxq_v((int8x16_t)__a, (int8x16_t)__b, 49); }
-__ai uint32x4_t vpmaxq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vpmaxq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-__ai float32x4_t vpmaxq_f32(float32x4_t __a, float32x4_t __b) {
-  return (float32x4_t)__builtin_neon_vpmaxq_v((int8x16_t)__a, (int8x16_t)__b, 41); }
-__ai float64x2_t vpmaxq_f64(float64x2_t __a, float64x2_t __b) {
-  return (float64x2_t)__builtin_neon_vpmaxq_v((int8x16_t)__a, (int8x16_t)__b, 42); }
-
-__ai float64x1_t vmin_f64(float64x1_t __a, float64x1_t __b) {
-  return (float64x1_t)__builtin_neon_vmin_v((int8x8_t)__a, (int8x8_t)__b, 10); }
-__ai float64x2_t vminq_f64(float64x2_t __a, float64x2_t __b) {
-  return (float64x2_t)__builtin_neon_vminq_v((int8x16_t)__a, (int8x16_t)__b, 42); }
-
-__ai int8x16_t vpminq_s8(int8x16_t __a, int8x16_t __b) {
-  return (int8x16_t)__builtin_neon_vpminq_v(__a, __b, 32); }
-__ai int16x8_t vpminq_s16(int16x8_t __a, int16x8_t __b) {
-  return (int16x8_t)__builtin_neon_vpminq_v((int8x16_t)__a, (int8x16_t)__b, 33); }
-__ai int32x4_t vpminq_s32(int32x4_t __a, int32x4_t __b) {
-  return (int32x4_t)__builtin_neon_vpminq_v((int8x16_t)__a, (int8x16_t)__b, 34); }
-__ai uint8x16_t vpminq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vpminq_v((int8x16_t)__a, (int8x16_t)__b, 48); }
-__ai uint16x8_t vpminq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint16x8_t)__builtin_neon_vpminq_v((int8x16_t)__a, (int8x16_t)__b, 49); }
-__ai uint32x4_t vpminq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vpminq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-__ai float32x4_t vpminq_f32(float32x4_t __a, float32x4_t __b) {
-  return (float32x4_t)__builtin_neon_vpminq_v((int8x16_t)__a, (int8x16_t)__b, 41); }
-__ai float64x2_t vpminq_f64(float64x2_t __a, float64x2_t __b) {
-  return (float64x2_t)__builtin_neon_vpminq_v((int8x16_t)__a, (int8x16_t)__b, 42); }
-
-__ai float64x1_t vmla_f64(float64x1_t __a, float64x1_t __b, float64x1_t __c) {
-  return __a + (__b * __c); }
-__ai float64x2_t vmlaq_f64(float64x2_t __a, float64x2_t __b, float64x2_t __c) {
-  return __a + (__b * __c); }
-
-__ai float64x2_t vmlaq_n_f64(float64x2_t __a, float64x2_t __b, float64_t __c) {
-  return __a + (__b * (float64x2_t){ __c, __c }); }
-
-__ai float64x1_t vmls_f64(float64x1_t __a, float64x1_t __b, float64x1_t __c) {
-  return __a - (__b * __c); }
-__ai float64x2_t vmlsq_f64(float64x2_t __a, float64x2_t __b, float64x2_t __c) {
-  return __a - (__b * __c); }
-
-__ai float64x2_t vmlsq_n_f64(float64x2_t __a, float64x2_t __b, float64_t __c) {
-  return __a - (__b * (float64x2_t){ __c, __c }); }
-
-__ai float64x1_t vmov_n_f64(float64_t __a) {
-  return (float64x1_t){ __a }; }
-__ai float64x2_t vmovq_n_f64(float64_t __a) {
-  return (float64x2_t){ __a, __a }; }
-
-__ai float64x1_t vmul_f64(float64x1_t __a, float64x1_t __b) {
-  return __a * __b; }
-__ai float64x2_t vmulq_f64(float64x2_t __a, float64x2_t __b) {
-  return __a * __b; }
-
-__ai float32x2_t vmulx_f32(float32x2_t __a, float32x2_t __b) {
-  return (float32x2_t)__builtin_neon_vmulx_v((int8x8_t)__a, (int8x8_t)__b, 9); }
-__ai float64x1_t vmulx_f64(float64x1_t __a, float64x1_t __b) {
-  return (float64x1_t)__builtin_neon_vmulx_v((int8x8_t)__a, (int8x8_t)__b, 10); }
-__ai float32x4_t vmulxq_f32(float32x4_t __a, float32x4_t __b) {
-  return (float32x4_t)__builtin_neon_vmulxq_v((int8x16_t)__a, (int8x16_t)__b, 41); }
-__ai float64x2_t vmulxq_f64(float64x2_t __a, float64x2_t __b) {
-  return (float64x2_t)__builtin_neon_vmulxq_v((int8x16_t)__a, (int8x16_t)__b, 42); }
-
-__ai float64x1_t vneg_f64(float64x1_t __a) {
-  return  -__a; }
-__ai int64x1_t vneg_s64(int64x1_t __a) {
-  return  -__a; }
-__ai float64x2_t vnegq_f64(float64x2_t __a) {
-  return  -__a; }
-__ai int64x2_t vnegq_s64(int64x2_t __a) {
-  return  -__a; }
-
-__ai int64x1_t vqabs_s64(int64x1_t __a) {
-  return (int64x1_t)__builtin_neon_vqabs_v((int8x8_t)__a, 3); }
-__ai int64x2_t vqabsq_s64(int64x2_t __a) {
-  return (int64x2_t)__builtin_neon_vqabsq_v((int8x16_t)__a, 35); }
-
-__ai int64x1_t vqneg_s64(int64x1_t __a) {
-  return (int64x1_t)__builtin_neon_vqneg_v((int8x8_t)__a, 3); }
-__ai int64x2_t vqnegq_s64(int64x2_t __a) {
-  return (int64x2_t)__builtin_neon_vqnegq_v((int8x16_t)__a, 35); }
-
-#define vqrshrn_high_n_s16(a, b, __c) __extension__ ({ \
-  int8x8_t __a = (a); int16x8_t __b = (b); \
-  (int8x16_t)vcombine_s16(__a, vqrshrn_n_s16(__b, __c)); })
-#define vqrshrn_high_n_s32(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int32x4_t __b = (b); \
-  (int16x8_t)vcombine_s32(__a, vqrshrn_n_s32(__b, __c)); })
-#define vqrshrn_high_n_s64(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int64x2_t __b = (b); \
-  (int32x4_t)vcombine_s64(__a, vqrshrn_n_s64(__b, __c)); })
-#define vqrshrn_high_n_u16(a, b, __c) __extension__ ({ \
-  uint8x8_t __a = (a); uint16x8_t __b = (b); \
-  (uint8x16_t)vcombine_u16(__a, vqrshrn_n_u16(__b, __c)); })
-#define vqrshrn_high_n_u32(a, b, __c) __extension__ ({ \
-  uint16x4_t __a = (a); uint32x4_t __b = (b); \
-  (uint16x8_t)vcombine_u32(__a, vqrshrn_n_u32(__b, __c)); })
-#define vqrshrn_high_n_u64(a, b, __c) __extension__ ({ \
-  uint32x2_t __a = (a); uint64x2_t __b = (b); \
-  (uint32x4_t)vcombine_u64(__a, vqrshrn_n_u64(__b, __c)); })
-
-#define vqrshrun_high_n_s16(a, b, __c) __extension__ ({ \
-  int8x8_t __a = (a); int16x8_t __b = (b); \
-  (int8x16_t)vcombine_s16(__a, vqrshrun_n_s16(__b, __c)); })
-#define vqrshrun_high_n_s32(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int32x4_t __b = (b); \
-  (int16x8_t)vcombine_s32(__a, vqrshrun_n_s32(__b, __c)); })
-#define vqrshrun_high_n_s64(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int64x2_t __b = (b); \
-  (int32x4_t)vcombine_s64(__a, vqrshrun_n_s64(__b, __c)); })
-
-#define vqshrn_high_n_s16(a, b, __c) __extension__ ({ \
-  int8x8_t __a = (a); int16x8_t __b = (b); \
-  (int8x16_t)vcombine_s16(__a, vqshrn_n_s16(__b, __c)); })
-#define vqshrn_high_n_s32(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int32x4_t __b = (b); \
-  (int16x8_t)vcombine_s32(__a, vqshrn_n_s32(__b, __c)); })
-#define vqshrn_high_n_s64(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int64x2_t __b = (b); \
-  (int32x4_t)vcombine_s64(__a, vqshrn_n_s64(__b, __c)); })
-#define vqshrn_high_n_u16(a, b, __c) __extension__ ({ \
-  uint8x8_t __a = (a); uint16x8_t __b = (b); \
-  (uint8x16_t)vcombine_u16(__a, vqshrn_n_u16(__b, __c)); })
-#define vqshrn_high_n_u32(a, b, __c) __extension__ ({ \
-  uint16x4_t __a = (a); uint32x4_t __b = (b); \
-  (uint16x8_t)vcombine_u32(__a, vqshrn_n_u32(__b, __c)); })
-#define vqshrn_high_n_u64(a, b, __c) __extension__ ({ \
-  uint32x2_t __a = (a); uint64x2_t __b = (b); \
-  (uint32x4_t)vcombine_u64(__a, vqshrn_n_u64(__b, __c)); })
-
-#define vqshrun_high_n_s16(a, b, __c) __extension__ ({ \
-  int8x8_t __a = (a); int16x8_t __b = (b); \
-  (int8x16_t)vcombine_s16(__a, vqshrun_n_s16(__b, __c)); })
-#define vqshrun_high_n_s32(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int32x4_t __b = (b); \
-  (int16x8_t)vcombine_s32(__a, vqshrun_n_s32(__b, __c)); })
-#define vqshrun_high_n_s64(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int64x2_t __b = (b); \
-  (int32x4_t)vcombine_s64(__a, vqshrun_n_s64(__b, __c)); })
-
-__ai int8x16_t vqmovn_high_s16(int8x8_t __a, int16x8_t __b) {
-  int8x8_t __a1 = vqmovn_s16(__b);
-  return __builtin_shufflevector(__a, __a1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); }
-__ai int16x8_t vqmovn_high_s32(int16x4_t __a, int32x4_t __b) {
-  int16x4_t __a1 = vqmovn_s32(__b);
-  return __builtin_shufflevector(__a, __a1, 0, 1, 2, 3, 4, 5, 6, 7); }
-__ai int32x4_t vqmovn_high_s64(int32x2_t __a, int64x2_t __b) {
-  int32x2_t __a1 = vqmovn_s64(__b);
-  return __builtin_shufflevector(__a, __a1, 0, 1, 2, 3); }
-__ai uint8x16_t vqmovn_high_u16(uint8x8_t __a, uint16x8_t __b) {
-  uint8x8_t __a1 = vqmovn_u16(__b);
-  return __builtin_shufflevector(__a, __a1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); }
-__ai uint16x8_t vqmovn_high_u32(uint16x4_t __a, uint32x4_t __b) {
-  uint16x4_t __a1 = vqmovn_u32(__b);
-  return __builtin_shufflevector(__a, __a1, 0, 1, 2, 3, 4, 5, 6, 7); }
-__ai uint32x4_t vqmovn_high_u64(uint32x2_t __a, uint64x2_t __b) {
-  uint32x2_t __a1 = vqmovn_u64(__b);
-  return __builtin_shufflevector(__a, __a1, 0, 1, 2, 3); }
-
-__ai int8x8_t vrbit_s8(int8x8_t __a) {
-  return (int8x8_t)__builtin_neon_vrbit_v(__a, 0); }
-__ai uint8x8_t vrbit_u8(uint8x8_t __a) {
-  return (uint8x8_t)__builtin_neon_vrbit_v((int8x8_t)__a, 16); }
-__ai poly8x8_t vrbit_p8(poly8x8_t __a) {
-  return (poly8x8_t)__builtin_neon_vrbit_v((int8x8_t)__a, 4); }
-__ai int8x16_t vrbitq_s8(int8x16_t __a) {
-  return (int8x16_t)__builtin_neon_vrbitq_v(__a, 32); }
-__ai uint8x16_t vrbitq_u8(uint8x16_t __a) {
-  return (uint8x16_t)__builtin_neon_vrbitq_v((int8x16_t)__a, 48); }
-__ai poly8x16_t vrbitq_p8(poly8x16_t __a) {
-  return (poly8x16_t)__builtin_neon_vrbitq_v((int8x16_t)__a, 36); }
-
-#define vrshrn_high_n_s16(a, b, __c) __extension__ ({ \
-  int8x8_t __a = (a); int16x8_t __b = (b); \
-  (int8x16_t)vcombine_s16(__a, vrshrn_n_s16(__b, __c)); })
-#define vrshrn_high_n_s32(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int32x4_t __b = (b); \
-  (int16x8_t)vcombine_s32(__a, vrshrn_n_s32(__b, __c)); })
-#define vrshrn_high_n_s64(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int64x2_t __b = (b); \
-  (int32x4_t)vcombine_s64(__a, vrshrn_n_s64(__b, __c)); })
-#define vrshrn_high_n_u16(a, b, __c) __extension__ ({ \
-  uint8x8_t __a = (a); uint16x8_t __b = (b); \
-  (uint8x16_t)vcombine_u16(__a, vrshrn_n_u16(__b, __c)); })
-#define vrshrn_high_n_u32(a, b, __c) __extension__ ({ \
-  uint16x4_t __a = (a); uint32x4_t __b = (b); \
-  (uint16x8_t)vcombine_u32(__a, vrshrn_n_u32(__b, __c)); })
-#define vrshrn_high_n_u64(a, b, __c) __extension__ ({ \
-  uint32x2_t __a = (a); uint64x2_t __b = (b); \
-  (uint32x4_t)vcombine_u64(__a, vrshrn_n_u64(__b, __c)); })
-
-__ai float32_t vabds_f32(float32_t __a, float32_t __b) {
-  return (float32_t)__builtin_neon_vabds_f32(__a, __b); }
-__ai float64_t vabdd_f64(float64_t __a, float64_t __b) {
-  return (float64_t)__builtin_neon_vabdd_f64(__a, __b); }
-
-__ai int64_t vabsd_s64(int64_t __a) {
-  return (int64_t)__builtin_neon_vabsd_s64(__a); }
-
-__ai int64_t vaddd_s64(int64_t __a, int64_t __b) {
-  return (int64_t)__builtin_neon_vaddd_s64(__a, __b); }
-__ai uint64_t vaddd_u64(uint64_t __a, uint64_t __b) {
-  return (uint64_t)__builtin_neon_vaddd_u64(__a, __b); }
-
-__ai float32_t vpadds_f32(float32x2_t __a) {
-  return (float32_t)__builtin_neon_vpadds_f32(__a); }
-__ai int64_t vpaddd_s64(int64x2_t __a) {
-  return (int64_t)__builtin_neon_vpaddd_s64(__a); }
-__ai float64_t vpaddd_f64(float64x2_t __a) {
-  return (float64_t)__builtin_neon_vpaddd_f64(__a); }
-__ai uint64_t vpaddd_u64(uint64x2_t __a) {
-  return (uint64_t)__builtin_neon_vpaddd_u64((int64x2_t)__a); }
-
-__ai int64_t vceqd_s64(int64_t __a, int64_t __b) {
-  return (int64_t)__builtin_neon_vceqd_s64(__a, __b); }
-__ai uint64_t vceqd_u64(uint64_t __a, uint64_t __b) {
-  return (uint64_t)__builtin_neon_vceqd_u64(__a, __b); }
-
-__ai int64_t vceqzd_s64(int64_t __a) {
-  return (int64_t)__builtin_neon_vceqzd_s64(__a); }
-__ai uint64_t vceqzd_u64(uint64_t __a) {
-  return (uint64_t)__builtin_neon_vceqzd_u64(__a); }
-
-__ai int64_t vcged_s64(int64_t __a, int64_t __b) {
-  return (int64_t)__builtin_neon_vcged_s64(__a, __b); }
-
-__ai int64_t vcgezd_s64(int64_t __a) {
-  return (int64_t)__builtin_neon_vcgezd_s64(__a); }
-
-__ai int64_t vcgtd_s64(int64_t __a, int64_t __b) {
-  return (int64_t)__builtin_neon_vcgtd_s64(__a, __b); }
-
-__ai int64_t vcgtzd_s64(int64_t __a) {
-  return (int64_t)__builtin_neon_vcgtzd_s64(__a); }
-
-__ai uint64_t vcgtd_u64(uint64_t __a, uint64_t __b) {
-  return (uint64_t)__builtin_neon_vcgtd_u64(__a, __b); }
-
-__ai uint64_t vcged_u64(uint64_t __a, uint64_t __b) {
-  return (uint64_t)__builtin_neon_vcged_u64(__a, __b); }
-
-__ai int64_t vcled_s64(int64_t __a, int64_t __b) {
-  return (int64_t)__builtin_neon_vcled_s64(__a, __b); }
-__ai uint64_t vcled_u64(uint64_t __a, uint64_t __b) {
-  return (uint64_t)__builtin_neon_vcled_u64(__a, __b); }
-
-__ai int64_t vclezd_s64(int64_t __a) {
-  return (int64_t)__builtin_neon_vclezd_s64(__a); }
-
-__ai int64_t vcltd_s64(int64_t __a, int64_t __b) {
-  return (int64_t)__builtin_neon_vcltd_s64(__a, __b); }
-__ai uint64_t vcltd_u64(uint64_t __a, uint64_t __b) {
-  return (uint64_t)__builtin_neon_vcltd_u64(__a, __b); }
-
-__ai int64_t vcltzd_s64(int64_t __a) {
-  return (int64_t)__builtin_neon_vcltzd_s64(__a); }
-
-__ai int64_t vtstd_s64(int64_t __a, int64_t __b) {
-  return (int64_t)__builtin_neon_vtstd_s64(__a, __b); }
-__ai uint64_t vtstd_u64(uint64_t __a, uint64_t __b) {
-  return (uint64_t)__builtin_neon_vtstd_u64(__a, __b); }
-
-__ai uint32_t vcages_f32(float32_t __a, float32_t __b) {
-  return (uint32_t)__builtin_neon_vcages_f32(__a, __b); }
-__ai uint64_t vcaged_f64(float64_t __a, float64_t __b) {
-  return (uint64_t)__builtin_neon_vcaged_f64(__a, __b); }
-
-__ai uint32_t vcagts_f32(float32_t __a, float32_t __b) {
-  return (uint32_t)__builtin_neon_vcagts_f32(__a, __b); }
-__ai uint64_t vcagtd_f64(float64_t __a, float64_t __b) {
-  return (uint64_t)__builtin_neon_vcagtd_f64(__a, __b); }
-
-__ai uint32_t vcales_f32(float32_t __a, float32_t __b) {
-  return (uint32_t)__builtin_neon_vcales_f32(__a, __b); }
-__ai uint64_t vcaled_f64(float64_t __a, float64_t __b) {
-  return (uint64_t)__builtin_neon_vcaled_f64(__a, __b); }
-
-__ai uint32_t vcalts_f32(float32_t __a, float32_t __b) {
-  return (uint32_t)__builtin_neon_vcalts_f32(__a, __b); }
-__ai uint64_t vcaltd_f64(float64_t __a, float64_t __b) {
-  return (uint64_t)__builtin_neon_vcaltd_f64(__a, __b); }
-
-__ai uint32_t vceqs_f32(float32_t __a, float32_t __b) {
-  return (uint32_t)__builtin_neon_vceqs_f32(__a, __b); }
-__ai uint64_t vceqd_f64(float64_t __a, float64_t __b) {
-  return (uint64_t)__builtin_neon_vceqd_f64(__a, __b); }
+#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
 
-__ai uint32_t vceqzs_f32(float32_t __a) {
-  return (uint32_t)__builtin_neon_vceqzs_f32(__a); }
-__ai uint64_t vceqzd_f64(float64_t __a) {
-  return (uint64_t)__builtin_neon_vceqzd_f64(__a); }
+#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
 
-__ai uint32_t vcges_f32(float32_t __a, float32_t __b) {
-  return (uint32_t)__builtin_neon_vcges_f32(__a, __b); }
-__ai uint64_t vcged_f64(float64_t __a, float64_t __b) {
-  return (uint64_t)__builtin_neon_vcged_f64(__a, __b); }
+#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
 
-__ai uint32_t vcgezs_f32(float32_t __a) {
-  return (uint32_t)__builtin_neon_vcgezs_f32(__a); }
-__ai uint64_t vcgezd_f64(float64_t __a) {
-  return (uint64_t)__builtin_neon_vcgezd_f64(__a); }
+#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
 
-__ai uint32_t vcgts_f32(float32_t __a, float32_t __b) {
-  return (uint32_t)__builtin_neon_vcgts_f32(__a, __b); }
-__ai uint64_t vcgtd_f64(float64_t __a, float64_t __b) {
-  return (uint64_t)__builtin_neon_vcgtd_f64(__a, __b); }
+#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
 
-__ai uint32_t vcgtzs_f32(float32_t __a) {
-  return (uint32_t)__builtin_neon_vcgtzs_f32(__a); }
-__ai uint64_t vcgtzd_f64(float64_t __a) {
-  return (uint64_t)__builtin_neon_vcgtzd_f64(__a); }
+#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
 
-__ai uint32_t vcles_f32(float32_t __a, float32_t __b) {
-  return (uint32_t)__builtin_neon_vcles_f32(__a, __b); }
-__ai uint64_t vcled_f64(float64_t __a, float64_t __b) {
-  return (uint64_t)__builtin_neon_vcled_f64(__a, __b); }
+#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
 
-__ai uint32_t vclezs_f32(float32_t __a) {
-  return (uint32_t)__builtin_neon_vclezs_f32(__a); }
-__ai uint64_t vclezd_f64(float64_t __a) {
-  return (uint64_t)__builtin_neon_vclezd_f64(__a); }
+#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
 
-__ai uint32_t vclts_f32(float32_t __a, float32_t __b) {
-  return (uint32_t)__builtin_neon_vclts_f32(__a, __b); }
-__ai uint64_t vcltd_f64(float64_t __a, float64_t __b) {
-  return (uint64_t)__builtin_neon_vcltd_f64(__a, __b); }
+#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
 
-__ai uint32_t vcltzs_f32(float32_t __a) {
-  return (uint32_t)__builtin_neon_vcltzs_f32(__a); }
-__ai uint64_t vcltzd_f64(float64_t __a) {
-  return (uint64_t)__builtin_neon_vcltzd_f64(__a); }
+#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
 
-__ai int64_t vcvtad_s64_f64(float64_t __a) {
-  return (int64_t)__builtin_neon_vcvtad_s64_f64(__a); }
+#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
 
-__ai int32_t vcvtas_s32_f32(float32_t __a) {
-  return (int32_t)__builtin_neon_vcvtas_s32_f32(__a); }
+#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
 
-__ai uint64_t vcvtad_u64_f64(float64_t __a) {
-  return (uint64_t)__builtin_neon_vcvtad_u64_f64(__a); }
+#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
 
-__ai uint32_t vcvtas_u32_f32(float32_t __a) {
-  return (uint32_t)__builtin_neon_vcvtas_u32_f32(__a); }
-
-__ai int64_t vcvtmd_s64_f64(float64_t __a) {
-  return (int64_t)__builtin_neon_vcvtmd_s64_f64(__a); }
-
-__ai int32_t vcvtms_s32_f32(float32_t __a) {
-  return (int32_t)__builtin_neon_vcvtms_s32_f32(__a); }
-
-__ai uint64_t vcvtmd_u64_f64(float64_t __a) {
-  return (uint64_t)__builtin_neon_vcvtmd_u64_f64(__a); }
-
-__ai uint32_t vcvtms_u32_f32(float32_t __a) {
-  return (uint32_t)__builtin_neon_vcvtms_u32_f32(__a); }
-
-__ai int64_t vcvtnd_s64_f64(float64_t __a) {
-  return (int64_t)__builtin_neon_vcvtnd_s64_f64(__a); }
-
-__ai int32_t vcvtns_s32_f32(float32_t __a) {
-  return (int32_t)__builtin_neon_vcvtns_s32_f32(__a); }
-
-__ai uint64_t vcvtnd_u64_f64(float64_t __a) {
-  return (uint64_t)__builtin_neon_vcvtnd_u64_f64(__a); }
-
-__ai uint32_t vcvtns_u32_f32(float32_t __a) {
-  return (uint32_t)__builtin_neon_vcvtns_u32_f32(__a); }
-
-__ai int64_t vcvtpd_s64_f64(float64_t __a) {
-  return (int64_t)__builtin_neon_vcvtpd_s64_f64(__a); }
-
-__ai int32_t vcvtps_s32_f32(float32_t __a) {
-  return (int32_t)__builtin_neon_vcvtps_s32_f32(__a); }
-
-__ai uint64_t vcvtpd_u64_f64(float64_t __a) {
-  return (uint64_t)__builtin_neon_vcvtpd_u64_f64(__a); }
-
-__ai uint32_t vcvtps_u32_f32(float32_t __a) {
-  return (uint32_t)__builtin_neon_vcvtps_u32_f32(__a); }
-
-__ai float32_t vcvtxd_f32_f64(float64_t __a) {
-  return (float32_t)__builtin_neon_vcvtxd_f32_f64(__a); }
-
-__ai int64_t vcvtd_s64_f64(float64_t __a) {
-  return (int64_t)__builtin_neon_vcvtd_s64_f64(__a); }
-
-__ai int32_t vcvts_s32_f32(float32_t __a) {
-  return (int32_t)__builtin_neon_vcvts_s32_f32(__a); }
-
-#define vcvts_n_s32_f32(a, __b) __extension__ ({ \
-  float32_t __a = (a); \
-  (int32_t)__builtin_neon_vcvts_n_s32_f32(__a, __b); })
-
-#define vcvtd_n_s64_f64(a, __b) __extension__ ({ \
-  float64_t __a = (a); \
-  (int64_t)__builtin_neon_vcvtd_n_s64_f64(__a, __b); })
-
-__ai uint64_t vcvtd_u64_f64(float64_t __a) {
-  return (uint64_t)__builtin_neon_vcvtd_u64_f64(__a); }
-
-__ai uint32_t vcvts_u32_f32(float32_t __a) {
-  return (uint32_t)__builtin_neon_vcvts_u32_f32(__a); }
-
-#define vcvts_n_u32_f32(a, __b) __extension__ ({ \
-  float32_t __a = (a); \
-  (uint32_t)__builtin_neon_vcvts_n_u32_f32(__a, __b); })
-
-#define vcvtd_n_u64_f64(a, __b) __extension__ ({ \
-  float64_t __a = (a); \
-  (uint64_t)__builtin_neon_vcvtd_n_u64_f64(__a, __b); })
-
-__ai float32_t vpmaxnms_f32(float32x2_t __a) {
-  return (float32_t)__builtin_neon_vpmaxnms_f32(__a); }
-__ai float64_t vpmaxnmqd_f64(float64x2_t __a) {
-  return (float64_t)__builtin_neon_vpmaxnmqd_f64(__a); }
-
-__ai float32_t vpmaxs_f32(float32x2_t __a) {
-  return (float32_t)__builtin_neon_vpmaxs_f32(__a); }
-__ai float64_t vpmaxqd_f64(float64x2_t __a) {
-  return (float64_t)__builtin_neon_vpmaxqd_f64(__a); }
-
-__ai float32_t vpminnms_f32(float32x2_t __a) {
-  return (float32_t)__builtin_neon_vpminnms_f32(__a); }
-__ai float64_t vpminnmqd_f64(float64x2_t __a) {
-  return (float64_t)__builtin_neon_vpminnmqd_f64(__a); }
-
-__ai float32_t vpmins_f32(float32x2_t __a) {
-  return (float32_t)__builtin_neon_vpmins_f32(__a); }
-__ai float64_t vpminqd_f64(float64x2_t __a) {
-  return (float64_t)__builtin_neon_vpminqd_f64(__a); }
-
-#define vfmas_lane_f32(a, b, c, __d) __extension__ ({ \
-  float32_t __a = (a); float32_t __b = (b); float32x2_t __c = (c); \
-  (float32_t)__builtin_neon_vfmas_lane_f32(__a, __b, __c, __d); })
-#define vfmad_lane_f64(a, b, c, __d) __extension__ ({ \
-  float64_t __a = (a); float64_t __b = (b); float64x1_t __c = (c); \
-  (float64_t)__builtin_neon_vfmad_lane_f64(__a, __b, __c, __d); })
-
-#define vfmas_laneq_f32(a, b, c, __d) __extension__ ({ \
-  float32_t __a = (a); float32_t __b = (b); float32x4_t __c = (c); \
-  (float32_t)__builtin_neon_vfmas_laneq_f32(__a, __b, __c, __d); })
-#define vfmad_laneq_f64(a, b, c, __d) __extension__ ({ \
-  float64_t __a = (a); float64_t __b = (b); float64x2_t __c = (c); \
-  (float64_t)__builtin_neon_vfmad_laneq_f64(__a, __b, __c, __d); })
-
-#define vfmss_lane_f32(a, b, c, __d) __extension__ ({ \
-  float32_t __a = (a); float32_t __b = (b); float32x2_t __c = (c); \
-  float32_t __a1 = __a; \
-  float32_t __b1 = __b; \
-  float32x2_t __c1 = __c; \
-  vfmas_lane_f32(__a1, __b1, -__c1, __d); })
-#define vfmsd_lane_f64(a, b, c, __d) __extension__ ({ \
-  float64_t __a = (a); float64_t __b = (b); float64x1_t __c = (c); \
-  float64_t __a1 = __a; \
-  float64_t __b1 = __b; \
-  float64x1_t __c1 = __c; \
-  vfmad_lane_f64(__a1, __b1, -__c1, __d); })
-
-#define vfmss_laneq_f32(a, b, c, __d) __extension__ ({ \
-  float32_t __a = (a); float32_t __b = (b); float32x4_t __c = (c); \
-  float32_t __a1 = __a; \
-  float32_t __b1 = __b; \
-  float32x4_t __c1 = __c; \
-  vfmas_laneq_f32(__a1, __b1, -__c1, __d); })
-#define vfmsd_laneq_f64(a, b, c, __d) __extension__ ({ \
-  float64_t __a = (a); float64_t __b = (b); float64x2_t __c = (c); \
-  float64_t __a1 = __a; \
-  float64_t __b1 = __b; \
-  float64x2_t __c1 = __c; \
-  vfmad_laneq_f64(__a1, __b1, -__c1, __d); })
-
-__ai float32_t vmulxs_f32(float32_t __a, float32_t __b) {
-  return (float32_t)__builtin_neon_vmulxs_f32(__a, __b); }
-__ai float64_t vmulxd_f64(float64_t __a, float64_t __b) {
-  return (float64_t)__builtin_neon_vmulxd_f64(__a, __b); }
-
-#define vmulxs_lane_f32(a, b, __c) __extension__ ({ \
-  float32_t __a = (a); float32x2_t __b = (b); \
-  float32_t __a1 = __a; \
-  float32x2_t __b1 = __b; \
-  float32_t __d1 = vget_lane_f32(__b1, __c);\
-  vmulxs_f32(__a1, __d1); })
-#define vmulxd_lane_f64(a, b, __c) __extension__ ({ \
-  float64_t __a = (a); float64x1_t __b = (b); \
-  float64_t __a1 = __a; \
-  float64x1_t __b1 = __b; \
-  float64_t __d1 = vget_lane_f64(__b1, __c);\
-  vmulxd_f64(__a1, __d1); })
-
-#define vmulxs_laneq_f32(a, b, __c) __extension__ ({ \
-  float32_t __a = (a); float32x4_t __b = (b); \
-  float32_t __a1 = __a; \
-  float32x4_t __b1 = __b; \
-  float32_t __d1 = vgetq_lane_f32(__b1, __c);\
-  vmulxs_f32(__a1, __d1); })
-#define vmulxd_laneq_f64(a, b, __c) __extension__ ({ \
-  float64_t __a = (a); float64x2_t __b = (b); \
-  float64_t __a1 = __a; \
-  float64x2_t __b1 = __b; \
-  float64_t __d1 = vgetq_lane_f64(__b1, __c);\
-  vmulxd_f64(__a1, __d1); })
-
-#define vmuls_lane_f32(a, b, __c) __extension__ ({ \
-  float32_t __a = (a); float32x2_t __b = (b); \
-  float32_t __d1 = vget_lane_f32(__b, __c);\
-  __a * __d1; })
-#define vmuld_lane_f64(a, b, __c) __extension__ ({ \
-  float64_t __a = (a); float64x1_t __b = (b); \
-  float64_t __d1 = vget_lane_f64(__b, __c);\
-  __a * __d1; })
-
-#define vmuls_laneq_f32(a, b, __c) __extension__ ({ \
-  float32_t __a = (a); float32x4_t __b = (b); \
-  float32_t __a1 = __a; \
-  float32x4_t __b1 = __b; \
-  float32_t __d1 = vgetq_lane_f32(__b1, __c);\
-  __a1 * __d1; })
-#define vmuld_laneq_f64(a, b, __c) __extension__ ({ \
-  float64_t __a = (a); float64x2_t __b = (b); \
-  float64_t __a1 = __a; \
-  float64x2_t __b1 = __b; \
-  float64_t __d1 = vgetq_lane_f64(__b1, __c);\
-  __a1 * __d1; })
-
-__ai float32_t vrecpes_f32(float32_t __a) {
-  return (float32_t)__builtin_neon_vrecpes_f32(__a); }
-__ai float64_t vrecped_f64(float64_t __a) {
-  return (float64_t)__builtin_neon_vrecped_f64(__a); }
-
-__ai float32_t vrecpss_f32(float32_t __a, float32_t __b) {
-  return (float32_t)__builtin_neon_vrecpss_f32(__a, __b); }
-__ai float64_t vrecpsd_f64(float64_t __a, float64_t __b) {
-  return (float64_t)__builtin_neon_vrecpsd_f64(__a, __b); }
-
-__ai float32_t vrecpxs_f32(float32_t __a) {
-  return (float32_t)__builtin_neon_vrecpxs_f32(__a); }
-__ai float64_t vrecpxd_f64(float64_t __a) {
-  return (float64_t)__builtin_neon_vrecpxd_f64(__a); }
-
-__ai float32_t vrsqrtes_f32(float32_t __a) {
-  return (float32_t)__builtin_neon_vrsqrtes_f32(__a); }
-__ai float64_t vrsqrted_f64(float64_t __a) {
-  return (float64_t)__builtin_neon_vrsqrted_f64(__a); }
-
-__ai float32_t vrsqrtss_f32(float32_t __a, float32_t __b) {
-  return (float32_t)__builtin_neon_vrsqrtss_f32(__a, __b); }
-__ai float64_t vrsqrtsd_f64(float64_t __a, float64_t __b) {
-  return (float64_t)__builtin_neon_vrsqrtsd_f64(__a, __b); }
-
-#define vget_lane_f16(a, __b) __extension__ ({ \
-  float16x4_t __a = (a); \
-  float16x4_t __a1 = __a; \
-  int16x4_t __a2 = *(int16x4_t *)(&__a1);\
-  int16_t __a3 = vget_lane_s16(__a2, __b);\
-  float16_t __a4 = *(float16_t *)(&__a3);\
-  __a4; })
-#define vgetq_lane_f16(a, __b) __extension__ ({ \
-  float16x8_t __a = (a); \
-  float16x8_t __a1 = __a; \
-  int16x8_t __a2 = *(int16x8_t *)(&__a1);\
-  int16_t __a3 = vgetq_lane_s16(__a2, __b);\
-  float16_t __a4 = *(float16_t *)(&__a3);\
-  __a4; })
-
-__ai int64_t vnegd_s64(int64_t __a) {
-  return (int64_t)__builtin_neon_vnegd_s64(__a); }
-
-__ai int8_t vqaddb_s8(int8_t __a, int8_t __b) {
-  return (int8_t)__builtin_neon_vqaddb_s8(__a, __b); }
-__ai int16_t vqaddh_s16(int16_t __a, int16_t __b) {
-  return (int16_t)__builtin_neon_vqaddh_s16(__a, __b); }
-__ai int32_t vqadds_s32(int32_t __a, int32_t __b) {
-  return (int32_t)__builtin_neon_vqadds_s32(__a, __b); }
-__ai int64_t vqaddd_s64(int64_t __a, int64_t __b) {
-  return (int64_t)__builtin_neon_vqaddd_s64(__a, __b); }
-__ai uint8_t vqaddb_u8(uint8_t __a, uint8_t __b) {
-  return (uint8_t)__builtin_neon_vqaddb_u8(__a, __b); }
-__ai uint16_t vqaddh_u16(uint16_t __a, uint16_t __b) {
-  return (uint16_t)__builtin_neon_vqaddh_u16(__a, __b); }
-__ai uint32_t vqadds_u32(uint32_t __a, uint32_t __b) {
-  return (uint32_t)__builtin_neon_vqadds_u32(__a, __b); }
-__ai uint64_t vqaddd_u64(uint64_t __a, uint64_t __b) {
-  return (uint64_t)__builtin_neon_vqaddd_u64(__a, __b); }
-
-__ai int8_t vqrshlb_s8(int8_t __a, int8_t __b) {
-  return (int8_t)__builtin_neon_vqrshlb_s8(__a, __b); }
-__ai int16_t vqrshlh_s16(int16_t __a, int16_t __b) {
-  return (int16_t)__builtin_neon_vqrshlh_s16(__a, __b); }
-__ai int32_t vqrshls_s32(int32_t __a, int32_t __b) {
-  return (int32_t)__builtin_neon_vqrshls_s32(__a, __b); }
-__ai int64_t vqrshld_s64(int64_t __a, int64_t __b) {
-  return (int64_t)__builtin_neon_vqrshld_s64(__a, __b); }
-__ai uint8_t vqrshlb_u8(uint8_t __a, uint8_t __b) {
-  return (uint8_t)__builtin_neon_vqrshlb_u8(__a, __b); }
-__ai uint16_t vqrshlh_u16(uint16_t __a, uint16_t __b) {
-  return (uint16_t)__builtin_neon_vqrshlh_u16(__a, __b); }
-__ai uint32_t vqrshls_u32(uint32_t __a, uint32_t __b) {
-  return (uint32_t)__builtin_neon_vqrshls_u32(__a, __b); }
-__ai uint64_t vqrshld_u64(uint64_t __a, uint64_t __b) {
-  return (uint64_t)__builtin_neon_vqrshld_u64(__a, __b); }
-
-__ai int8_t vqshlb_s8(int8_t __a, int8_t __b) {
-  return (int8_t)__builtin_neon_vqshlb_s8(__a, __b); }
-__ai int16_t vqshlh_s16(int16_t __a, int16_t __b) {
-  return (int16_t)__builtin_neon_vqshlh_s16(__a, __b); }
-__ai int32_t vqshls_s32(int32_t __a, int32_t __b) {
-  return (int32_t)__builtin_neon_vqshls_s32(__a, __b); }
-__ai int64_t vqshld_s64(int64_t __a, int64_t __b) {
-  return (int64_t)__builtin_neon_vqshld_s64(__a, __b); }
-__ai uint8_t vqshlb_u8(uint8_t __a, uint8_t __b) {
-  return (uint8_t)__builtin_neon_vqshlb_u8(__a, __b); }
-__ai uint16_t vqshlh_u16(uint16_t __a, uint16_t __b) {
-  return (uint16_t)__builtin_neon_vqshlh_u16(__a, __b); }
-__ai uint32_t vqshls_u32(uint32_t __a, uint32_t __b) {
-  return (uint32_t)__builtin_neon_vqshls_u32(__a, __b); }
-__ai uint64_t vqshld_u64(uint64_t __a, uint64_t __b) {
-  return (uint64_t)__builtin_neon_vqshld_u64(__a, __b); }
-
-__ai int8_t vqsubb_s8(int8_t __a, int8_t __b) {
-  return (int8_t)__builtin_neon_vqsubb_s8(__a, __b); }
-__ai int16_t vqsubh_s16(int16_t __a, int16_t __b) {
-  return (int16_t)__builtin_neon_vqsubh_s16(__a, __b); }
-__ai int32_t vqsubs_s32(int32_t __a, int32_t __b) {
-  return (int32_t)__builtin_neon_vqsubs_s32(__a, __b); }
-__ai int64_t vqsubd_s64(int64_t __a, int64_t __b) {
-  return (int64_t)__builtin_neon_vqsubd_s64(__a, __b); }
-__ai uint8_t vqsubb_u8(uint8_t __a, uint8_t __b) {
-  return (uint8_t)__builtin_neon_vqsubb_u8(__a, __b); }
-__ai uint16_t vqsubh_u16(uint16_t __a, uint16_t __b) {
-  return (uint16_t)__builtin_neon_vqsubh_u16(__a, __b); }
-__ai uint32_t vqsubs_u32(uint32_t __a, uint32_t __b) {
-  return (uint32_t)__builtin_neon_vqsubs_u32(__a, __b); }
-__ai uint64_t vqsubd_u64(uint64_t __a, uint64_t __b) {
-  return (uint64_t)__builtin_neon_vqsubd_u64(__a, __b); }
-
-__ai int64_t vrshld_s64(int64_t __a, int64_t __b) {
-  return (int64_t)__builtin_neon_vrshld_s64(__a, __b); }
-__ai uint64_t vrshld_u64(uint64_t __a, uint64_t __b) {
-  return (uint64_t)__builtin_neon_vrshld_u64(__a, __b); }
-
-__ai float64_t vcvtd_f64_s64(int64_t __a) {
-  return (float64_t)__builtin_neon_vcvtd_f64_s64(__a); }
-
-__ai float32_t vcvts_f32_s32(int32_t __a) {
-  return (float32_t)__builtin_neon_vcvts_f32_s32(__a); }
-
-#define vcvts_n_f32_s32(a, __b) __extension__ ({ \
-  int32_t __a = (a); \
-  (float32_t)__builtin_neon_vcvts_n_f32_s32(__a, __b); })
-#define vcvts_n_f32_u32(a, __b) __extension__ ({ \
-  uint32_t __a = (a); \
-  (float32_t)__builtin_neon_vcvts_n_f32_u32(__a, __b); })
-
-#define vcvtd_n_f64_s64(a, __b) __extension__ ({ \
-  int64_t __a = (a); \
-  (float64_t)__builtin_neon_vcvtd_n_f64_s64(__a, __b); })
-#define vcvtd_n_f64_u64(a, __b) __extension__ ({ \
-  uint64_t __a = (a); \
-  (float64_t)__builtin_neon_vcvtd_n_f64_u64(__a, __b); })
-
-#define vset_lane_f16(a, b, __c) __extension__ ({ \
-  float16_t __a = (a); float16x4_t __b = (b); \
-  float16_t __a1 = __a;\
-  int16_t __a2 = *(int16_t *)(&__a1);\
-  int16x4_t __b2 = *(int16x4_t *)(&__b);\
-  int16x4_t __b3 = vset_lane_s16(__a2, __b2, __c);\
-  float16x4_t __b4 = *(float16x4_t *)(&__b3);\
-__b4; })
-#define vsetq_lane_f16(a, b, __c) __extension__ ({ \
-  float16_t __a = (a); float16x8_t __b = (b); \
-  float16_t __a1 = __a;\
-  int16_t __a2 = *(int16_t *)(&__a1);\
-  int16x8_t __b2 = *(int16x8_t *)(&__b);\
-  int16x8_t __b3 = vsetq_lane_s16(__a2, __b2, __c);\
-  float16x8_t __b4 = *(float16x8_t *)(&__b3);\
-__b4; })
-
-__ai int64_t vshld_s64(int64_t __a, int64_t __b) {
-  return (int64_t)__builtin_neon_vshld_s64(__a, __b); }
-__ai uint64_t vshld_u64(uint64_t __a, uint64_t __b) {
-  return (uint64_t)__builtin_neon_vshld_u64(__a, __b); }
-
-#define vshld_n_s64(a, __b) __extension__ ({ \
-  int64_t __a = (a); \
-  (int64_t)__builtin_neon_vshld_n_s64(__a, __b); })
-#define vshld_n_u64(a, __b) __extension__ ({ \
-  uint64_t __a = (a); \
-  (uint64_t)__builtin_neon_vshld_n_u64(__a, __b); })
-
-#define vslid_n_s64(a, b, __c) __extension__ ({ \
-  int64_t __a = (a); int64_t __b = (b); \
-  (int64_t)__builtin_neon_vslid_n_s64(__a, __b, __c); })
-#define vslid_n_u64(a, b, __c) __extension__ ({ \
-  uint64_t __a = (a); uint64_t __b = (b); \
-  (uint64_t)__builtin_neon_vslid_n_u64(__a, __b, __c); })
-
-__ai int8_t vqabsb_s8(int8_t __a) {
-  return (int8_t)__builtin_neon_vqabsb_s8(__a); }
-__ai int16_t vqabsh_s16(int16_t __a) {
-  return (int16_t)__builtin_neon_vqabsh_s16(__a); }
-__ai int32_t vqabss_s32(int32_t __a) {
-  return (int32_t)__builtin_neon_vqabss_s32(__a); }
-__ai int64_t vqabsd_s64(int64_t __a) {
-  return (int64_t)__builtin_neon_vqabsd_s64(__a); }
-
-__ai int32_t vqdmlalh_s16(int32_t __a, int16_t __b, int16_t __c) {
-  return (int32_t)__builtin_neon_vqdmlalh_s16(__a, __b, __c); }
-__ai int64_t vqdmlals_s32(int64_t __a, int32_t __b, int32_t __c) {
-  return (int64_t)__builtin_neon_vqdmlals_s32(__a, __b, __c); }
-
-#define vqdmlalh_lane_s16(a, b, c, __d) __extension__ ({ \
-  int32_t __a = (a); int16_t __b = (b); int16x4_t __c = (c); \
-  (int32_t)__builtin_neon_vqdmlalh_lane_s16(__a, __b, __c, __d); })
-#define vqdmlals_lane_s32(a, b, c, __d) __extension__ ({ \
-  int64_t __a = (a); int32_t __b = (b); int32x2_t __c = (c); \
-  (int64_t)__builtin_neon_vqdmlals_lane_s32(__a, __b, __c, __d); })
-
-#define vqdmlalh_laneq_s16(a, b, c, __d) __extension__ ({ \
-  int32_t __a = (a); int16_t __b = (b); int16x8_t __c = (c); \
-  (int32_t)__builtin_neon_vqdmlalh_laneq_s16(__a, __b, __c, __d); })
-#define vqdmlals_laneq_s32(a, b, c, __d) __extension__ ({ \
-  int64_t __a = (a); int32_t __b = (b); int32x4_t __c = (c); \
-  (int64_t)__builtin_neon_vqdmlals_laneq_s32(__a, __b, __c, __d); })
-
-__ai int32_t vqdmlslh_s16(int32_t __a, int16_t __b, int16_t __c) {
-  return (int32_t)__builtin_neon_vqdmlslh_s16(__a, __b, __c); }
-__ai int64_t vqdmlsls_s32(int64_t __a, int32_t __b, int32_t __c) {
-  return (int64_t)__builtin_neon_vqdmlsls_s32(__a, __b, __c); }
-
-#define vqdmlslh_lane_s16(a, b, c, __d) __extension__ ({ \
-  int32_t __a = (a); int16_t __b = (b); int16x4_t __c = (c); \
-  (int32_t)__builtin_neon_vqdmlslh_lane_s16(__a, __b, __c, __d); })
-#define vqdmlsls_lane_s32(a, b, c, __d) __extension__ ({ \
-  int64_t __a = (a); int32_t __b = (b); int32x2_t __c = (c); \
-  (int64_t)__builtin_neon_vqdmlsls_lane_s32(__a, __b, __c, __d); })
-
-#define vqdmlslh_laneq_s16(a, b, c, __d) __extension__ ({ \
-  int32_t __a = (a); int16_t __b = (b); int16x8_t __c = (c); \
-  (int32_t)__builtin_neon_vqdmlslh_laneq_s16(__a, __b, __c, __d); })
-#define vqdmlsls_laneq_s32(a, b, c, __d) __extension__ ({ \
-  int64_t __a = (a); int32_t __b = (b); int32x4_t __c = (c); \
-  (int64_t)__builtin_neon_vqdmlsls_laneq_s32(__a, __b, __c, __d); })
-
-__ai int16_t vqdmulhh_s16(int16_t __a, int16_t __b) {
-  return (int16_t)__builtin_neon_vqdmulhh_s16(__a, __b); }
-__ai int32_t vqdmulhs_s32(int32_t __a, int32_t __b) {
-  return (int32_t)__builtin_neon_vqdmulhs_s32(__a, __b); }
-
-#define vqdmulhh_lane_s16(a, b, __c) __extension__ ({ \
-  int16_t __a = (a); int16x4_t __b = (b); \
-  int16_t __a1 = __a; \
-  int16x4_t __b1 = __b; \
-  vqdmulhh_s16(__a1, vget_lane_s16(__b1, __c)); })
-#define vqdmulhs_lane_s32(a, b, __c) __extension__ ({ \
-  int32_t __a = (a); int32x2_t __b = (b); \
-  int32_t __a1 = __a; \
-  int32x2_t __b1 = __b; \
-  vqdmulhs_s32(__a1, vget_lane_s32(__b1, __c)); })
-
-#define vqdmulhh_laneq_s16(a, b, __c) __extension__ ({ \
-  int16_t __a = (a); int16x8_t __b = (b); \
-  int16_t __a1 = __a; \
-  int16x8_t __b1 = __b; \
-  vqdmulhh_s16(__a1, vgetq_lane_s16(__b1, __c)); })
-#define vqdmulhs_laneq_s32(a, b, __c) __extension__ ({ \
-  int32_t __a = (a); int32x4_t __b = (b); \
-  int32_t __a1 = __a; \
-  int32x4_t __b1 = __b; \
-  vqdmulhs_s32(__a1, vgetq_lane_s32(__b1, __c)); })
-
-__ai int32_t vqdmullh_s16(int16_t __a, int16_t __b) {
-  return (int32_t)__builtin_neon_vqdmullh_s16(__a, __b); }
-__ai int64_t vqdmulls_s32(int32_t __a, int32_t __b) {
-  return (int64_t)__builtin_neon_vqdmulls_s32(__a, __b); }
-
-#define vqdmullh_lane_s16(a, b, __c) __extension__ ({ \
-  int16_t __a = (a); int16x4_t __b = (b); \
-  int16_t __a1 = __a; \
-  int16x4_t __b1 = __b; \
-  vqdmullh_s16(__a1, vget_lane_s16(__b1, __c)); })
-#define vqdmulls_lane_s32(a, b, __c) __extension__ ({ \
-  int32_t __a = (a); int32x2_t __b = (b); \
-  int32_t __a1 = __a; \
-  int32x2_t __b1 = __b; \
-  vqdmulls_s32(__a1, vget_lane_s32(__b1, __c)); })
-
-#define vqdmullh_laneq_s16(a, b, __c) __extension__ ({ \
-  int16_t __a = (a); int16x8_t __b = (b); \
-  int16_t __a1 = __a; \
-  int16x8_t __b1 = __b; \
-  vqdmullh_s16(__a1, vgetq_lane_s16(__b1, __c)); })
-#define vqdmulls_laneq_s32(a, b, __c) __extension__ ({ \
-  int32_t __a = (a); int32x4_t __b = (b); \
-  int32_t __a1 = __a; \
-  int32x4_t __b1 = __b; \
-  vqdmulls_s32(__a1, vgetq_lane_s32(__b1, __c)); })
-
-__ai int8_t vqnegb_s8(int8_t __a) {
-  return (int8_t)__builtin_neon_vqnegb_s8(__a); }
-__ai int16_t vqnegh_s16(int16_t __a) {
-  return (int16_t)__builtin_neon_vqnegh_s16(__a); }
-__ai int32_t vqnegs_s32(int32_t __a) {
-  return (int32_t)__builtin_neon_vqnegs_s32(__a); }
-__ai int64_t vqnegd_s64(int64_t __a) {
-  return (int64_t)__builtin_neon_vqnegd_s64(__a); }
-
-__ai int16_t vqrdmulhh_s16(int16_t __a, int16_t __b) {
-  return (int16_t)__builtin_neon_vqrdmulhh_s16(__a, __b); }
-__ai int32_t vqrdmulhs_s32(int32_t __a, int32_t __b) {
-  return (int32_t)__builtin_neon_vqrdmulhs_s32(__a, __b); }
-
-#define vqrdmulhh_lane_s16(a, b, __c) __extension__ ({ \
-  int16_t __a = (a); int16x4_t __b = (b); \
-  int16_t __a1 = __a; \
-  int16x4_t __b1 = __b; \
-  vqrdmulhh_s16(__a1, vget_lane_s16(__b1, __c)); })
-#define vqrdmulhs_lane_s32(a, b, __c) __extension__ ({ \
-  int32_t __a = (a); int32x2_t __b = (b); \
-  int32_t __a1 = __a; \
-  int32x2_t __b1 = __b; \
-  vqrdmulhs_s32(__a1, vget_lane_s32(__b1, __c)); })
-
-#define vqrdmulhh_laneq_s16(a, b, __c) __extension__ ({ \
-  int16_t __a = (a); int16x8_t __b = (b); \
-  int16_t __a1 = __a; \
-  int16x8_t __b1 = __b; \
-  vqrdmulhh_s16(__a1, vgetq_lane_s16(__b1, __c)); })
-#define vqrdmulhs_laneq_s32(a, b, __c) __extension__ ({ \
-  int32_t __a = (a); int32x4_t __b = (b); \
-  int32_t __a1 = __a; \
-  int32x4_t __b1 = __b; \
-  vqrdmulhs_s32(__a1, vgetq_lane_s32(__b1, __c)); })
-
-#define vqrshrnh_n_s16(a, __b) __extension__ ({ \
-  int16_t __a = (a); \
-  (int8_t)__builtin_neon_vqrshrnh_n_s16(__a, __b); })
-#define vqrshrns_n_s32(a, __b) __extension__ ({ \
-  int32_t __a = (a); \
-  (int16_t)__builtin_neon_vqrshrns_n_s32(__a, __b); })
-#define vqrshrnd_n_s64(a, __b) __extension__ ({ \
-  int64_t __a = (a); \
-  (int32_t)__builtin_neon_vqrshrnd_n_s64(__a, __b); })
-#define vqrshrnh_n_u16(a, __b) __extension__ ({ \
-  uint16_t __a = (a); \
-  (uint8_t)__builtin_neon_vqrshrnh_n_u16(__a, __b); })
-#define vqrshrns_n_u32(a, __b) __extension__ ({ \
-  uint32_t __a = (a); \
-  (uint16_t)__builtin_neon_vqrshrns_n_u32(__a, __b); })
-#define vqrshrnd_n_u64(a, __b) __extension__ ({ \
-  uint64_t __a = (a); \
-  (uint32_t)__builtin_neon_vqrshrnd_n_u64(__a, __b); })
-
-#define vqrshrunh_n_s16(a, __b) __extension__ ({ \
-  int16_t __a = (a); \
-  (int8_t)__builtin_neon_vqrshrunh_n_s16(__a, __b); })
-#define vqrshruns_n_s32(a, __b) __extension__ ({ \
-  int32_t __a = (a); \
-  (int16_t)__builtin_neon_vqrshruns_n_s32(__a, __b); })
-#define vqrshrund_n_s64(a, __b) __extension__ ({ \
-  int64_t __a = (a); \
-  (int32_t)__builtin_neon_vqrshrund_n_s64(__a, __b); })
-
-#define vqshlub_n_s8(a, __b) __extension__ ({ \
-  int8_t __a = (a); \
-  (int8_t)__builtin_neon_vqshlub_n_s8(__a, __b); })
-#define vqshluh_n_s16(a, __b) __extension__ ({ \
-  int16_t __a = (a); \
-  (int16_t)__builtin_neon_vqshluh_n_s16(__a, __b); })
-#define vqshlus_n_s32(a, __b) __extension__ ({ \
-  int32_t __a = (a); \
-  (int32_t)__builtin_neon_vqshlus_n_s32(__a, __b); })
-#define vqshlud_n_s64(a, __b) __extension__ ({ \
-  int64_t __a = (a); \
-  (int64_t)__builtin_neon_vqshlud_n_s64(__a, __b); })
-
-#define vqshlb_n_s8(a, __b) __extension__ ({ \
-  int8_t __a = (a); \
-  (int8_t)__builtin_neon_vqshlb_n_s8(__a, __b); })
-#define vqshlh_n_s16(a, __b) __extension__ ({ \
-  int16_t __a = (a); \
-  (int16_t)__builtin_neon_vqshlh_n_s16(__a, __b); })
-#define vqshls_n_s32(a, __b) __extension__ ({ \
-  int32_t __a = (a); \
-  (int32_t)__builtin_neon_vqshls_n_s32(__a, __b); })
-#define vqshld_n_s64(a, __b) __extension__ ({ \
-  int64_t __a = (a); \
-  (int64_t)__builtin_neon_vqshld_n_s64(__a, __b); })
-#define vqshlb_n_u8(a, __b) __extension__ ({ \
-  uint8_t __a = (a); \
-  (uint8_t)__builtin_neon_vqshlb_n_u8(__a, __b); })
-#define vqshlh_n_u16(a, __b) __extension__ ({ \
-  uint16_t __a = (a); \
-  (uint16_t)__builtin_neon_vqshlh_n_u16(__a, __b); })
-#define vqshls_n_u32(a, __b) __extension__ ({ \
-  uint32_t __a = (a); \
-  (uint32_t)__builtin_neon_vqshls_n_u32(__a, __b); })
-#define vqshld_n_u64(a, __b) __extension__ ({ \
-  uint64_t __a = (a); \
-  (uint64_t)__builtin_neon_vqshld_n_u64(__a, __b); })
-
-#define vqshrnh_n_s16(a, __b) __extension__ ({ \
-  int16_t __a = (a); \
-  (int8_t)__builtin_neon_vqshrnh_n_s16(__a, __b); })
-#define vqshrns_n_s32(a, __b) __extension__ ({ \
-  int32_t __a = (a); \
-  (int16_t)__builtin_neon_vqshrns_n_s32(__a, __b); })
-#define vqshrnd_n_s64(a, __b) __extension__ ({ \
-  int64_t __a = (a); \
-  (int32_t)__builtin_neon_vqshrnd_n_s64(__a, __b); })
-#define vqshrnh_n_u16(a, __b) __extension__ ({ \
-  uint16_t __a = (a); \
-  (uint8_t)__builtin_neon_vqshrnh_n_u16(__a, __b); })
-#define vqshrns_n_u32(a, __b) __extension__ ({ \
-  uint32_t __a = (a); \
-  (uint16_t)__builtin_neon_vqshrns_n_u32(__a, __b); })
-#define vqshrnd_n_u64(a, __b) __extension__ ({ \
-  uint64_t __a = (a); \
-  (uint32_t)__builtin_neon_vqshrnd_n_u64(__a, __b); })
-
-#define vqshrunh_n_s16(a, __b) __extension__ ({ \
-  int16_t __a = (a); \
-  (int8_t)__builtin_neon_vqshrunh_n_s16(__a, __b); })
-#define vqshruns_n_s32(a, __b) __extension__ ({ \
-  int32_t __a = (a); \
-  (int16_t)__builtin_neon_vqshruns_n_s32(__a, __b); })
-#define vqshrund_n_s64(a, __b) __extension__ ({ \
-  int64_t __a = (a); \
-  (int32_t)__builtin_neon_vqshrund_n_s64(__a, __b); })
-
-__ai int8_t vqmovnh_s16(int16_t __a) {
-  return (int8_t)__builtin_neon_vqmovnh_s16(__a); }
-__ai int16_t vqmovns_s32(int32_t __a) {
-  return (int16_t)__builtin_neon_vqmovns_s32(__a); }
-__ai int32_t vqmovnd_s64(int64_t __a) {
-  return (int32_t)__builtin_neon_vqmovnd_s64(__a); }
-
-__ai int8_t vqmovunh_s16(int16_t __a) {
-  return (int8_t)__builtin_neon_vqmovunh_s16(__a); }
-__ai int16_t vqmovuns_s32(int32_t __a) {
-  return (int16_t)__builtin_neon_vqmovuns_s32(__a); }
-__ai int32_t vqmovund_s64(int64_t __a) {
-  return (int32_t)__builtin_neon_vqmovund_s64(__a); }
-
-#define vsrid_n_s64(a, b, __c) __extension__ ({ \
-  int64_t __a = (a); int64_t __b = (b); \
-  (int64_t)__builtin_neon_vsrid_n_s64(__a, __b, __c); })
-#define vsrid_n_u64(a, b, __c) __extension__ ({ \
-  uint64_t __a = (a); uint64_t __b = (b); \
-  (uint64_t)__builtin_neon_vsrid_n_u64(__a, __b, __c); })
-
-#define vrshrd_n_s64(a, __b) __extension__ ({ \
-  int64_t __a = (a); \
-  (int64_t)__builtin_neon_vrshrd_n_s64(__a, __b); })
-#define vrshrd_n_u64(a, __b) __extension__ ({ \
-  uint64_t __a = (a); \
-  (uint64_t)__builtin_neon_vrshrd_n_u64(__a, __b); })
-
-#define vrsrad_n_s64(a, b, __c) __extension__ ({ \
-  int64_t __a = (a); int64_t __b = (b); \
-  (int64_t)__builtin_neon_vrsrad_n_s64(__a, __b, __c); })
-#define vrsrad_n_u64(a, b, __c) __extension__ ({ \
-  uint64_t __a = (a); uint64_t __b = (b); \
-  (uint64_t)__builtin_neon_vrsrad_n_u64(__a, __b, __c); })
-
-#define vshrd_n_s64(a, __b) __extension__ ({ \
-  int64_t __a = (a); \
-  (int64_t)__builtin_neon_vshrd_n_s64(__a, __b); })
-#define vshrd_n_u64(a, __b) __extension__ ({ \
-  uint64_t __a = (a); \
-  (uint64_t)__builtin_neon_vshrd_n_u64(__a, __b); })
-
-#define vsrad_n_s64(a, b, __c) __extension__ ({ \
-  int64_t __a = (a); int64_t __b = (b); \
-  (int64_t)__builtin_neon_vsrad_n_s64(__a, __b, __c); })
-#define vsrad_n_u64(a, b, __c) __extension__ ({ \
-  uint64_t __a = (a); uint64_t __b = (b); \
-  (uint64_t)__builtin_neon_vsrad_n_u64(__a, __b, __c); })
-
-__ai int64_t vsubd_s64(int64_t __a, int64_t __b) {
-  return (int64_t)__builtin_neon_vsubd_s64(__a, __b); }
-__ai uint64_t vsubd_u64(uint64_t __a, uint64_t __b) {
-  return (uint64_t)__builtin_neon_vsubd_u64(__a, __b); }
-
-__ai int8_t vuqaddb_s8(int8_t __a, int8_t __b) {
-  return (int8_t)__builtin_neon_vuqaddb_s8(__a, __b); }
-__ai int16_t vuqaddh_s16(int16_t __a, int16_t __b) {
-  return (int16_t)__builtin_neon_vuqaddh_s16(__a, __b); }
-__ai int32_t vuqadds_s32(int32_t __a, int32_t __b) {
-  return (int32_t)__builtin_neon_vuqadds_s32(__a, __b); }
-__ai int64_t vuqaddd_s64(int64_t __a, int64_t __b) {
-  return (int64_t)__builtin_neon_vuqaddd_s64(__a, __b); }
-
-__ai float64_t vcvtd_f64_u64(uint64_t __a) {
-  return (float64_t)__builtin_neon_vcvtd_f64_u64(__a); }
-
-__ai float32_t vcvts_f32_u32(uint32_t __a) {
-  return (float32_t)__builtin_neon_vcvts_f32_u32(__a); }
-
-__ai uint8_t vqmovnh_u16(uint16_t __a) {
-  return (uint8_t)__builtin_neon_vqmovnh_u16(__a); }
-__ai uint16_t vqmovns_u32(uint32_t __a) {
-  return (uint16_t)__builtin_neon_vqmovns_u32(__a); }
-__ai uint32_t vqmovnd_u64(uint64_t __a) {
-  return (uint32_t)__builtin_neon_vqmovnd_u64(__a); }
-
-__ai uint8_t vsqaddb_u8(uint8_t __a, uint8_t __b) {
-  return (uint8_t)__builtin_neon_vsqaddb_u8(__a, __b); }
-__ai uint16_t vsqaddh_u16(uint16_t __a, uint16_t __b) {
-  return (uint16_t)__builtin_neon_vsqaddh_u16(__a, __b); }
-__ai uint32_t vsqadds_u32(uint32_t __a, uint32_t __b) {
-  return (uint32_t)__builtin_neon_vsqadds_u32(__a, __b); }
-__ai uint64_t vsqaddd_u64(uint64_t __a, uint64_t __b) {
-  return (uint64_t)__builtin_neon_vsqaddd_u64(__a, __b); }
-
-#define vdupb_lane_s8(a, __b) __extension__ ({ \
-  int8x8_t __a = (a); \
-  (int8_t)__builtin_neon_vdupb_lane_i8(__a, __b); })
-#define vduph_lane_s16(a, __b) __extension__ ({ \
-  int16x4_t __a = (a); \
-  (int16_t)__builtin_neon_vduph_lane_i16(__a, __b); })
-#define vdups_lane_s32(a, __b) __extension__ ({ \
-  int32x2_t __a = (a); \
-  (int32_t)__builtin_neon_vdups_lane_i32(__a, __b); })
-#define vdupd_lane_s64(a, __b) __extension__ ({ \
-  int64x1_t __a = (a); \
-  (int64_t)__builtin_neon_vdupd_lane_i64(__a, __b); })
-#define vdups_lane_f32(a, __b) __extension__ ({ \
-  float32x2_t __a = (a); \
-  (float32_t)__builtin_neon_vdups_lane_f32(__a, __b); })
-#define vdupd_lane_f64(a, __b) __extension__ ({ \
-  float64x1_t __a = (a); \
-  (float64_t)__builtin_neon_vdupd_lane_f64(__a, __b); })
-#define vdupb_lane_u8(a, __b) __extension__ ({ \
-  uint8x8_t __a = (a); \
-  (uint8_t)__builtin_neon_vdupb_lane_i8((int8x8_t)__a, __b); })
-#define vduph_lane_u16(a, __b) __extension__ ({ \
-  uint16x4_t __a = (a); \
-  (uint16_t)__builtin_neon_vduph_lane_i16((int16x4_t)__a, __b); })
-#define vdups_lane_u32(a, __b) __extension__ ({ \
-  uint32x2_t __a = (a); \
-  (uint32_t)__builtin_neon_vdups_lane_i32((int32x2_t)__a, __b); })
-#define vdupd_lane_u64(a, __b) __extension__ ({ \
-  uint64x1_t __a = (a); \
-  (uint64_t)__builtin_neon_vdupd_lane_i64((int64x1_t)__a, __b); })
-#define vdupb_lane_p8(a, __b) __extension__ ({ \
-  poly8x8_t __a = (a); \
-  (poly8_t)__builtin_neon_vdupb_lane_i8((int8x8_t)__a, __b); })
-#define vduph_lane_p16(a, __b) __extension__ ({ \
-  poly16x4_t __a = (a); \
-  (poly16_t)__builtin_neon_vduph_lane_i16((int16x4_t)__a, __b); })
-
-#define vdupb_laneq_s8(a, __b) __extension__ ({ \
-  int8x16_t __a = (a); \
-  (int8_t)__builtin_neon_vdupb_laneq_i8(__a, __b); })
-#define vduph_laneq_s16(a, __b) __extension__ ({ \
-  int16x8_t __a = (a); \
-  (int16_t)__builtin_neon_vduph_laneq_i16(__a, __b); })
-#define vdups_laneq_s32(a, __b) __extension__ ({ \
-  int32x4_t __a = (a); \
-  (int32_t)__builtin_neon_vdups_laneq_i32(__a, __b); })
-#define vdupd_laneq_s64(a, __b) __extension__ ({ \
-  int64x2_t __a = (a); \
-  (int64_t)__builtin_neon_vdupd_laneq_i64(__a, __b); })
-#define vdups_laneq_f32(a, __b) __extension__ ({ \
-  float32x4_t __a = (a); \
-  (float32_t)__builtin_neon_vdups_laneq_f32(__a, __b); })
-#define vdupd_laneq_f64(a, __b) __extension__ ({ \
-  float64x2_t __a = (a); \
-  (float64_t)__builtin_neon_vdupd_laneq_f64(__a, __b); })
-#define vdupb_laneq_u8(a, __b) __extension__ ({ \
-  uint8x16_t __a = (a); \
-  (uint8_t)__builtin_neon_vdupb_laneq_i8((int8x16_t)__a, __b); })
-#define vduph_laneq_u16(a, __b) __extension__ ({ \
-  uint16x8_t __a = (a); \
-  (uint16_t)__builtin_neon_vduph_laneq_i16((int16x8_t)__a, __b); })
-#define vdups_laneq_u32(a, __b) __extension__ ({ \
-  uint32x4_t __a = (a); \
-  (uint32_t)__builtin_neon_vdups_laneq_i32((int32x4_t)__a, __b); })
-#define vdupd_laneq_u64(a, __b) __extension__ ({ \
-  uint64x2_t __a = (a); \
-  (uint64_t)__builtin_neon_vdupd_laneq_i64((int64x2_t)__a, __b); })
-#define vdupb_laneq_p8(a, __b) __extension__ ({ \
-  poly8x16_t __a = (a); \
-  (poly8_t)__builtin_neon_vdupb_laneq_i8((int8x16_t)__a, __b); })
-#define vduph_laneq_p16(a, __b) __extension__ ({ \
-  poly16x8_t __a = (a); \
-  (poly16_t)__builtin_neon_vduph_laneq_i16((int16x8_t)__a, __b); })
-
-#define vmulx_lane_f64(a, b, __c) __extension__ ({ \
-  float64x1_t __a = (a); float64x1_t __b = (b); \
-  float64x1_t __a1 = __a; \
-  float64x1_t __b1 = __b; \
-  float64_t __d1 = vget_lane_f64(__a1, 0);\
-  float64_t __e1 = vget_lane_f64(__b1, __c);\
-  float64_t __f1 = vmulxd_f64(__d1, __e1);\
-  float64x1_t __g1;\
-  vset_lane_f64(__f1, __g1, __c); })
-
-#define vmulx_laneq_f64(a, b, __c) __extension__ ({ \
-  float64x1_t __a = (a); float64x2_t __b = (b); \
-  float64x1_t __a1 = __a; \
-  float64x2_t __b1 = __b; \
-  float64_t __d1 = vget_lane_f64(__a1, 0);\
-  float64_t __e1 = vgetq_lane_f64(__b1, __c);\
-  float64_t __f1 = vmulxd_f64(__d1, __e1);\
-  float64x1_t __g1;\
-  vset_lane_f64(__f1, __g1, 0); })
-
-#define vmul_lane_f64(a, b, __c) __extension__ ({ \
-  float64x1_t __a = (a); float64x1_t __b = (b); \
-  (float64x1_t)__builtin_neon_vmul_lane_v((int8x8_t)__a, (int8x8_t)__b, __c, 10); })
-
-#define vmul_laneq_f64(a, b, __c) __extension__ ({ \
-  float64x1_t __a = (a); float64x2_t __b = (b); \
-  (float64x1_t)__builtin_neon_vmul_laneq_v((int8x8_t)__a, (int8x16_t)__b, __c, 10); })
-
-__ai float64x1_t vmul_n_f64(float64x1_t __a, float64_t __b) {
-  return (float64x1_t)__builtin_neon_vmul_n_f64(__a, __b); }
-
-#define vset_lane_f64(a, b, __c) __extension__ ({ \
-  float64_t __a = (a); float64x1_t __b = (b); \
-  (float64x1_t)__builtin_neon_vset_lane_f64(__a, __b, __c); })
-#define vsetq_lane_f64(a, b, __c) __extension__ ({ \
-  float64_t __a = (a); float64x2_t __b = (b); \
-  (float64x2_t)__builtin_neon_vsetq_lane_f64(__a, __b, __c); })
-#define vset_lane_p64(a, b, __c) __extension__ ({ \
-  poly64_t __a = (a); poly64x1_t __b = (b); \
-  (poly64x1_t)__builtin_neon_vset_lane_i64(__a, (int64x1_t)__b, __c); })
-#define vsetq_lane_p64(a, b, __c) __extension__ ({ \
-  poly64_t __a = (a); poly64x2_t __b = (b); \
-  (poly64x2_t)__builtin_neon_vsetq_lane_i64(__a, (int64x2_t)__b, __c); })
+#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_FEATURE_CRYPTO
-__ai uint32x4_t vsha1cq_u32(uint32x4_t __a, uint32_t __b, uint32x4_t __c) {
-  return (uint32x4_t)__builtin_neon_vsha1cq_u32((int32x4_t)__a, __b, (int32x4_t)__c); }
+#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
 
-__ai uint32_t vsha1h_u32(uint32_t __a) {
-  return (uint32_t)__builtin_neon_vsha1h_u32(__a); }
+#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
 
-__ai uint32x4_t vsha1mq_u32(uint32x4_t __a, uint32_t __b, uint32x4_t __c) {
-  return (uint32x4_t)__builtin_neon_vsha1mq_u32((int32x4_t)__a, __b, (int32x4_t)__c); }
+#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
 
-__ai uint32x4_t vsha1pq_u32(uint32x4_t __a, uint32_t __b, uint32x4_t __c) {
-  return (uint32x4_t)__builtin_neon_vsha1pq_u32((int32x4_t)__a, __b, (int32x4_t)__c); }
+#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
 
-__ai uint32x4_t vsha1su0q_u32(uint32x4_t __a, uint32x4_t __b, uint32x4_t __c) {
-  return (uint32x4_t)__builtin_neon_vsha1su0q_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 50); }
+#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
 
-__ai uint32x4_t vsha1su1q_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vsha1su1q_v((int8x16_t)__a, (int8x16_t)__b, 50); }
+#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
 
-__ai uint32x4_t vsha256hq_u32(uint32x4_t __a, uint32x4_t __b, uint32x4_t __c) {
-  return (uint32x4_t)__builtin_neon_vsha256hq_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 50); }
+#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
 
-__ai uint32x4_t vsha256h2q_u32(uint32x4_t __a, uint32x4_t __b, uint32x4_t __c) {
-  return (uint32x4_t)__builtin_neon_vsha256h2q_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 50); }
+#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
 
-__ai uint32x4_t vsha256su0q_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vsha256su0q_v((int8x16_t)__a, (int8x16_t)__b, 50); }
+#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
 
-__ai uint32x4_t vsha256su1q_u32(uint32x4_t __a, uint32x4_t __b, uint32x4_t __c) {
-  return (uint32x4_t)__builtin_neon_vsha256su1q_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 50); }
+#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
 
-#if defined(__aarch64__)
-#define vshll_high_n_s8(a, __b) __extension__ ({ \
-  int8x16_t __a = (a); \
-  int8x8_t __a1 = vget_high_s8(__a); \
-  (int16x8_t)vshll_n_s8(__a1, __b); })
-#define vshll_high_n_s16(a, __b) __extension__ ({ \
-  int16x8_t __a = (a); \
-  int16x4_t __a1 = vget_high_s16(__a); \
-  (int32x4_t)vshll_n_s16(__a1, __b); })
-#define vshll_high_n_s32(a, __b) __extension__ ({ \
-  int32x4_t __a = (a); \
-  int32x2_t __a1 = vget_high_s32(__a); \
-  (int64x2_t)vshll_n_s32(__a1, __b); })
-#define vshll_high_n_u8(a, __b) __extension__ ({ \
-  uint8x16_t __a = (a); \
-  uint8x8_t __a1 = vget_high_u8(__a); \
-  (uint16x8_t)vshll_n_u8(__a1, __b); })
-#define vshll_high_n_u16(a, __b) __extension__ ({ \
-  uint16x8_t __a = (a); \
-  uint16x4_t __a1 = vget_high_u16(__a); \
-  (uint32x4_t)vshll_n_u16(__a1, __b); })
-#define vshll_high_n_u32(a, __b) __extension__ ({ \
-  uint32x4_t __a = (a); \
-  uint32x2_t __a1 = vget_high_u32(__a); \
-  (uint64x2_t)vshll_n_u32(__a1, __b); })
-
-#define vshrn_high_n_s16(a, b, __c) __extension__ ({ \
-  int8x8_t __a = (a); int16x8_t __b = (b); \
-  (int8x16_t)vcombine_s16(__a, vshrn_n_s16(__b, __c)); })
-#define vshrn_high_n_s32(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int32x4_t __b = (b); \
-  (int16x8_t)vcombine_s32(__a, vshrn_n_s32(__b, __c)); })
-#define vshrn_high_n_s64(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int64x2_t __b = (b); \
-  (int32x4_t)vcombine_s64(__a, vshrn_n_s64(__b, __c)); })
-#define vshrn_high_n_u16(a, b, __c) __extension__ ({ \
-  uint8x8_t __a = (a); uint16x8_t __b = (b); \
-  (uint8x16_t)vcombine_u16(__a, vshrn_n_u16(__b, __c)); })
-#define vshrn_high_n_u32(a, b, __c) __extension__ ({ \
-  uint16x4_t __a = (a); uint32x4_t __b = (b); \
-  (uint16x8_t)vcombine_u32(__a, vshrn_n_u32(__b, __c)); })
-#define vshrn_high_n_u64(a, b, __c) __extension__ ({ \
-  uint32x2_t __a = (a); uint64x2_t __b = (b); \
-  (uint32x4_t)vcombine_u64(__a, vshrn_n_u64(__b, __c)); })
-
-#define vsli_n_p64(a, b, __c) __extension__ ({ \
-  poly64x1_t __a = (a); poly64x1_t __b = (b); \
-  (poly64x1_t)__builtin_neon_vsli_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 6); })
-#define vsliq_n_p64(a, b, __c) __extension__ ({ \
-  poly64x2_t __a = (a); poly64x2_t __b = (b); \
-  (poly64x2_t)__builtin_neon_vsliq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 38); })
-
-__ai int8x16_t vqmovun_high_s16(int8x8_t __a, int16x8_t __b) {
-  int8x8_t __a1 = vqmovun_s16(__b);
-  return __builtin_shufflevector(__a, __a1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); }
-__ai int16x8_t vqmovun_high_s32(int16x4_t __a, int32x4_t __b) {
-  int16x4_t __a1 = vqmovun_s32(__b);
-  return __builtin_shufflevector(__a, __a1, 0, 1, 2, 3, 4, 5, 6, 7); }
-__ai int32x4_t vqmovun_high_s64(int32x2_t __a, int64x2_t __b) {
-  int32x2_t __a1 = vqmovun_s64(__b);
-  return __builtin_shufflevector(__a, __a1, 0, 1, 2, 3); }
-
-#define vsri_n_p64(a, b, __c) __extension__ ({ \
-  poly64x1_t __a = (a); poly64x1_t __b = (b); \
-  (poly64x1_t)__builtin_neon_vsri_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 6); })
-#define vsriq_n_p64(a, b, __c) __extension__ ({ \
-  poly64x2_t __a = (a); poly64x2_t __b = (b); \
-  (poly64x2_t)__builtin_neon_vsriq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 38); })
-
-#define vst1_f64(__a, b) __extension__ ({ \
-  float64x1_t __b = (b); \
-  __builtin_neon_vst1_v(__a, (int8x8_t)__b, 10); })
-#define vst1q_f64(__a, b) __extension__ ({ \
-  float64x2_t __b = (b); \
-  __builtin_neon_vst1q_v(__a, (int8x16_t)__b, 42); })
-#define vst1_p64(__a, b) __extension__ ({ \
-  poly64x1_t __b = (b); \
-  __builtin_neon_vst1_v(__a, (int8x8_t)__b, 6); })
-#define vst1q_p64(__a, b) __extension__ ({ \
-  poly64x2_t __b = (b); \
-  __builtin_neon_vst1q_v(__a, (int8x16_t)__b, 38); })
-
-#define vst1_lane_f64(__a, b, __c) __extension__ ({ \
-  float64x1_t __b = (b); \
-  __builtin_neon_vst1_lane_v(__a, (int8x8_t)__b, __c, 10); })
-#define vst1q_lane_f64(__a, b, __c) __extension__ ({ \
-  float64x2_t __b = (b); \
-  __builtin_neon_vst1q_lane_v(__a, (int8x16_t)__b, __c, 42); })
-#define vst1_lane_p64(__a, b, __c) __extension__ ({ \
-  poly64x1_t __b = (b); \
-  __builtin_neon_vst1_lane_v(__a, (int8x8_t)__b, __c, 6); })
-#define vst1q_lane_p64(__a, b, __c) __extension__ ({ \
-  poly64x2_t __b = (b); \
-  __builtin_neon_vst1q_lane_v(__a, (int8x16_t)__b, __c, 38); })
-
-#define vst1q_u8_x2(__a, b) __extension__ ({ \
-  uint8x16x2_t __b = (b); \
-  __builtin_neon_vst1q_x2_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 48); })
-#define vst1q_u16_x2(__a, b) __extension__ ({ \
-  uint16x8x2_t __b = (b); \
-  __builtin_neon_vst1q_x2_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 49); })
-#define vst1q_u32_x2(__a, b) __extension__ ({ \
-  uint32x4x2_t __b = (b); \
-  __builtin_neon_vst1q_x2_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 50); })
-#define vst1q_s8_x2(__a, b) __extension__ ({ \
-  int8x16x2_t __b = (b); \
-  __builtin_neon_vst1q_x2_v(__a, __b.val[0], __b.val[1], 32); })
-#define vst1q_s16_x2(__a, b) __extension__ ({ \
-  int16x8x2_t __b = (b); \
-  __builtin_neon_vst1q_x2_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 33); })
-#define vst1q_s32_x2(__a, b) __extension__ ({ \
-  int32x4x2_t __b = (b); \
-  __builtin_neon_vst1q_x2_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 34); })
-#define vst1q_f16_x2(__a, b) __extension__ ({ \
-  float16x8x2_t __b = (b); \
-  __builtin_neon_vst1q_x2_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 40); })
-#define vst1q_f32_x2(__a, b) __extension__ ({ \
-  float32x4x2_t __b = (b); \
-  __builtin_neon_vst1q_x2_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 41); })
-#define vst1q_p8_x2(__a, b) __extension__ ({ \
-  poly8x16x2_t __b = (b); \
-  __builtin_neon_vst1q_x2_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 36); })
-#define vst1q_p16_x2(__a, b) __extension__ ({ \
-  poly16x8x2_t __b = (b); \
-  __builtin_neon_vst1q_x2_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 37); })
-#define vst1_u8_x2(__a, b) __extension__ ({ \
-  uint8x8x2_t __b = (b); \
-  __builtin_neon_vst1_x2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 16); })
-#define vst1_u16_x2(__a, b) __extension__ ({ \
-  uint16x4x2_t __b = (b); \
-  __builtin_neon_vst1_x2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 17); })
-#define vst1_u32_x2(__a, b) __extension__ ({ \
-  uint32x2x2_t __b = (b); \
-  __builtin_neon_vst1_x2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 18); })
-#define vst1_u64_x2(__a, b) __extension__ ({ \
-  uint64x1x2_t __b = (b); \
-  __builtin_neon_vst1_x2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 19); })
-#define vst1_s8_x2(__a, b) __extension__ ({ \
-  int8x8x2_t __b = (b); \
-  __builtin_neon_vst1_x2_v(__a, __b.val[0], __b.val[1], 0); })
-#define vst1_s16_x2(__a, b) __extension__ ({ \
-  int16x4x2_t __b = (b); \
-  __builtin_neon_vst1_x2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 1); })
-#define vst1_s32_x2(__a, b) __extension__ ({ \
-  int32x2x2_t __b = (b); \
-  __builtin_neon_vst1_x2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 2); })
-#define vst1_s64_x2(__a, b) __extension__ ({ \
-  int64x1x2_t __b = (b); \
-  __builtin_neon_vst1_x2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 3); })
-#define vst1_f16_x2(__a, b) __extension__ ({ \
-  float16x4x2_t __b = (b); \
-  __builtin_neon_vst1_x2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 8); })
-#define vst1_f32_x2(__a, b) __extension__ ({ \
-  float32x2x2_t __b = (b); \
-  __builtin_neon_vst1_x2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 9); })
-#define vst1_p8_x2(__a, b) __extension__ ({ \
-  poly8x8x2_t __b = (b); \
-  __builtin_neon_vst1_x2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 4); })
-#define vst1_p16_x2(__a, b) __extension__ ({ \
-  poly16x4x2_t __b = (b); \
-  __builtin_neon_vst1_x2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 5); })
-#define vst1q_u64_x2(__a, b) __extension__ ({ \
-  uint64x2x2_t __b = (b); \
-  __builtin_neon_vst1q_x2_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 51); })
-#define vst1q_s64_x2(__a, b) __extension__ ({ \
-  int64x2x2_t __b = (b); \
-  __builtin_neon_vst1q_x2_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 35); })
-#define vst1_f64_x2(__a, b) __extension__ ({ \
-  float64x1x2_t __b = (b); \
-  __builtin_neon_vst1_x2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 10); })
-#define vst1q_f64_x2(__a, b) __extension__ ({ \
-  float64x2x2_t __b = (b); \
-  __builtin_neon_vst1q_x2_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 42); })
-#define vst1_p64_x2(__a, b) __extension__ ({ \
-  poly64x1x2_t __b = (b); \
-  __builtin_neon_vst1_x2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 6); })
-#define vst1q_p64_x2(__a, b) __extension__ ({ \
-  poly64x2x2_t __b = (b); \
-  __builtin_neon_vst1q_x2_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 38); })
-
-#define vst1q_u8_x3(__a, b) __extension__ ({ \
-  uint8x16x3_t __b = (b); \
-  __builtin_neon_vst1q_x3_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 48); })
-#define vst1q_u16_x3(__a, b) __extension__ ({ \
-  uint16x8x3_t __b = (b); \
-  __builtin_neon_vst1q_x3_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 49); })
-#define vst1q_u32_x3(__a, b) __extension__ ({ \
-  uint32x4x3_t __b = (b); \
-  __builtin_neon_vst1q_x3_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 50); })
-#define vst1q_s8_x3(__a, b) __extension__ ({ \
-  int8x16x3_t __b = (b); \
-  __builtin_neon_vst1q_x3_v(__a, __b.val[0], __b.val[1], __b.val[2], 32); })
-#define vst1q_s16_x3(__a, b) __extension__ ({ \
-  int16x8x3_t __b = (b); \
-  __builtin_neon_vst1q_x3_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 33); })
-#define vst1q_s32_x3(__a, b) __extension__ ({ \
-  int32x4x3_t __b = (b); \
-  __builtin_neon_vst1q_x3_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 34); })
-#define vst1q_f16_x3(__a, b) __extension__ ({ \
-  float16x8x3_t __b = (b); \
-  __builtin_neon_vst1q_x3_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 40); })
-#define vst1q_f32_x3(__a, b) __extension__ ({ \
-  float32x4x3_t __b = (b); \
-  __builtin_neon_vst1q_x3_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 41); })
-#define vst1q_p8_x3(__a, b) __extension__ ({ \
-  poly8x16x3_t __b = (b); \
-  __builtin_neon_vst1q_x3_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 36); })
-#define vst1q_p16_x3(__a, b) __extension__ ({ \
-  poly16x8x3_t __b = (b); \
-  __builtin_neon_vst1q_x3_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 37); })
-#define vst1_u8_x3(__a, b) __extension__ ({ \
-  uint8x8x3_t __b = (b); \
-  __builtin_neon_vst1_x3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 16); })
-#define vst1_u16_x3(__a, b) __extension__ ({ \
-  uint16x4x3_t __b = (b); \
-  __builtin_neon_vst1_x3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 17); })
-#define vst1_u32_x3(__a, b) __extension__ ({ \
-  uint32x2x3_t __b = (b); \
-  __builtin_neon_vst1_x3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 18); })
-#define vst1_u64_x3(__a, b) __extension__ ({ \
-  uint64x1x3_t __b = (b); \
-  __builtin_neon_vst1_x3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 19); })
-#define vst1_s8_x3(__a, b) __extension__ ({ \
-  int8x8x3_t __b = (b); \
-  __builtin_neon_vst1_x3_v(__a, __b.val[0], __b.val[1], __b.val[2], 0); })
-#define vst1_s16_x3(__a, b) __extension__ ({ \
-  int16x4x3_t __b = (b); \
-  __builtin_neon_vst1_x3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 1); })
-#define vst1_s32_x3(__a, b) __extension__ ({ \
-  int32x2x3_t __b = (b); \
-  __builtin_neon_vst1_x3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 2); })
-#define vst1_s64_x3(__a, b) __extension__ ({ \
-  int64x1x3_t __b = (b); \
-  __builtin_neon_vst1_x3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 3); })
-#define vst1_f16_x3(__a, b) __extension__ ({ \
-  float16x4x3_t __b = (b); \
-  __builtin_neon_vst1_x3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 8); })
-#define vst1_f32_x3(__a, b) __extension__ ({ \
-  float32x2x3_t __b = (b); \
-  __builtin_neon_vst1_x3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 9); })
-#define vst1_p8_x3(__a, b) __extension__ ({ \
-  poly8x8x3_t __b = (b); \
-  __builtin_neon_vst1_x3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 4); })
-#define vst1_p16_x3(__a, b) __extension__ ({ \
-  poly16x4x3_t __b = (b); \
-  __builtin_neon_vst1_x3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 5); })
-#define vst1q_u64_x3(__a, b) __extension__ ({ \
-  uint64x2x3_t __b = (b); \
-  __builtin_neon_vst1q_x3_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 51); })
-#define vst1q_s64_x3(__a, b) __extension__ ({ \
-  int64x2x3_t __b = (b); \
-  __builtin_neon_vst1q_x3_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 35); })
-#define vst1_f64_x3(__a, b) __extension__ ({ \
-  float64x1x3_t __b = (b); \
-  __builtin_neon_vst1_x3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 10); })
-#define vst1q_f64_x3(__a, b) __extension__ ({ \
-  float64x2x3_t __b = (b); \
-  __builtin_neon_vst1q_x3_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 42); })
-#define vst1_p64_x3(__a, b) __extension__ ({ \
-  poly64x1x3_t __b = (b); \
-  __builtin_neon_vst1_x3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 6); })
-#define vst1q_p64_x3(__a, b) __extension__ ({ \
-  poly64x2x3_t __b = (b); \
-  __builtin_neon_vst1q_x3_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 38); })
-
-#define vst1q_u8_x4(__a, b) __extension__ ({ \
-  uint8x16x4_t __b = (b); \
-  __builtin_neon_vst1q_x4_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 48); })
-#define vst1q_u16_x4(__a, b) __extension__ ({ \
-  uint16x8x4_t __b = (b); \
-  __builtin_neon_vst1q_x4_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 49); })
-#define vst1q_u32_x4(__a, b) __extension__ ({ \
-  uint32x4x4_t __b = (b); \
-  __builtin_neon_vst1q_x4_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 50); })
-#define vst1q_s8_x4(__a, b) __extension__ ({ \
-  int8x16x4_t __b = (b); \
-  __builtin_neon_vst1q_x4_v(__a, __b.val[0], __b.val[1], __b.val[2], __b.val[3], 32); })
-#define vst1q_s16_x4(__a, b) __extension__ ({ \
-  int16x8x4_t __b = (b); \
-  __builtin_neon_vst1q_x4_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 33); })
-#define vst1q_s32_x4(__a, b) __extension__ ({ \
-  int32x4x4_t __b = (b); \
-  __builtin_neon_vst1q_x4_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 34); })
-#define vst1q_f16_x4(__a, b) __extension__ ({ \
-  float16x8x4_t __b = (b); \
-  __builtin_neon_vst1q_x4_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 40); })
-#define vst1q_f32_x4(__a, b) __extension__ ({ \
-  float32x4x4_t __b = (b); \
-  __builtin_neon_vst1q_x4_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 41); })
-#define vst1q_p8_x4(__a, b) __extension__ ({ \
-  poly8x16x4_t __b = (b); \
-  __builtin_neon_vst1q_x4_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 36); })
-#define vst1q_p16_x4(__a, b) __extension__ ({ \
-  poly16x8x4_t __b = (b); \
-  __builtin_neon_vst1q_x4_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 37); })
-#define vst1_u8_x4(__a, b) __extension__ ({ \
-  uint8x8x4_t __b = (b); \
-  __builtin_neon_vst1_x4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 16); })
-#define vst1_u16_x4(__a, b) __extension__ ({ \
-  uint16x4x4_t __b = (b); \
-  __builtin_neon_vst1_x4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 17); })
-#define vst1_u32_x4(__a, b) __extension__ ({ \
-  uint32x2x4_t __b = (b); \
-  __builtin_neon_vst1_x4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 18); })
-#define vst1_u64_x4(__a, b) __extension__ ({ \
-  uint64x1x4_t __b = (b); \
-  __builtin_neon_vst1_x4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 19); })
-#define vst1_s8_x4(__a, b) __extension__ ({ \
-  int8x8x4_t __b = (b); \
-  __builtin_neon_vst1_x4_v(__a, __b.val[0], __b.val[1], __b.val[2], __b.val[3], 0); })
-#define vst1_s16_x4(__a, b) __extension__ ({ \
-  int16x4x4_t __b = (b); \
-  __builtin_neon_vst1_x4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 1); })
-#define vst1_s32_x4(__a, b) __extension__ ({ \
-  int32x2x4_t __b = (b); \
-  __builtin_neon_vst1_x4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 2); })
-#define vst1_s64_x4(__a, b) __extension__ ({ \
-  int64x1x4_t __b = (b); \
-  __builtin_neon_vst1_x4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 3); })
-#define vst1_f16_x4(__a, b) __extension__ ({ \
-  float16x4x4_t __b = (b); \
-  __builtin_neon_vst1_x4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 8); })
-#define vst1_f32_x4(__a, b) __extension__ ({ \
-  float32x2x4_t __b = (b); \
-  __builtin_neon_vst1_x4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 9); })
-#define vst1_p8_x4(__a, b) __extension__ ({ \
-  poly8x8x4_t __b = (b); \
-  __builtin_neon_vst1_x4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 4); })
-#define vst1_p16_x4(__a, b) __extension__ ({ \
-  poly16x4x4_t __b = (b); \
-  __builtin_neon_vst1_x4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 5); })
-#define vst1q_u64_x4(__a, b) __extension__ ({ \
-  uint64x2x4_t __b = (b); \
-  __builtin_neon_vst1q_x4_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 51); })
-#define vst1q_s64_x4(__a, b) __extension__ ({ \
-  int64x2x4_t __b = (b); \
-  __builtin_neon_vst1q_x4_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 35); })
-#define vst1_f64_x4(__a, b) __extension__ ({ \
-  float64x1x4_t __b = (b); \
-  __builtin_neon_vst1_x4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 10); })
-#define vst1q_f64_x4(__a, b) __extension__ ({ \
-  float64x2x4_t __b = (b); \
-  __builtin_neon_vst1q_x4_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 42); })
-#define vst1_p64_x4(__a, b) __extension__ ({ \
-  poly64x1x4_t __b = (b); \
-  __builtin_neon_vst1_x4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 6); })
-#define vst1q_p64_x4(__a, b) __extension__ ({ \
-  poly64x2x4_t __b = (b); \
-  __builtin_neon_vst1q_x4_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 38); })
-
-#define vst2q_u64(__a, b) __extension__ ({ \
-  uint64x2x2_t __b = (b); \
-  __builtin_neon_vst2q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 51); })
-#define vst2q_s64(__a, b) __extension__ ({ \
-  int64x2x2_t __b = (b); \
-  __builtin_neon_vst2q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 35); })
-#define vst2_f64(__a, b) __extension__ ({ \
-  float64x1x2_t __b = (b); \
-  __builtin_neon_vst2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 10); })
-#define vst2q_f64(__a, b) __extension__ ({ \
-  float64x2x2_t __b = (b); \
-  __builtin_neon_vst2q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 42); })
-#define vst2_p64(__a, b) __extension__ ({ \
-  poly64x1x2_t __b = (b); \
-  __builtin_neon_vst2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 6); })
-#define vst2q_p64(__a, b) __extension__ ({ \
-  poly64x2x2_t __b = (b); \
-  __builtin_neon_vst2q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 38); })
-
-#define vst2_lane_s64(__a, b, __c) __extension__ ({ \
-  int64x1x2_t __b = (b); \
-  __builtin_neon_vst2_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 3); })
-#define vst2_lane_u64(__a, b, __c) __extension__ ({ \
-  uint64x1x2_t __b = (b); \
-  __builtin_neon_vst2_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 19); })
-#define vst2q_lane_s8(__a, b, __c) __extension__ ({ \
-  int8x16x2_t __b = (b); \
-  __builtin_neon_vst2q_lane_v(__a, __b.val[0], __b.val[1], __c, 32); })
-#define vst2q_lane_u8(__a, b, __c) __extension__ ({ \
-  uint8x16x2_t __b = (b); \
-  __builtin_neon_vst2q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 48); })
-#define vst2q_lane_p8(__a, b, __c) __extension__ ({ \
-  poly8x16x2_t __b = (b); \
-  __builtin_neon_vst2q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 36); })
-#define vst2q_lane_s64(__a, b, __c) __extension__ ({ \
-  int64x2x2_t __b = (b); \
-  __builtin_neon_vst2q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 35); })
-#define vst2q_lane_u64(__a, b, __c) __extension__ ({ \
-  uint64x2x2_t __b = (b); \
-  __builtin_neon_vst2q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 51); })
-#define vst2_lane_f64(__a, b, __c) __extension__ ({ \
-  float64x1x2_t __b = (b); \
-  __builtin_neon_vst2_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 10); })
-#define vst2q_lane_f64(__a, b, __c) __extension__ ({ \
-  float64x2x2_t __b = (b); \
-  __builtin_neon_vst2q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 42); })
-#define vst2_lane_p64(__a, b, __c) __extension__ ({ \
-  poly64x1x2_t __b = (b); \
-  __builtin_neon_vst2_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 6); })
-#define vst2q_lane_p64(__a, b, __c) __extension__ ({ \
-  poly64x2x2_t __b = (b); \
-  __builtin_neon_vst2q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 38); })
-
-#define vst3q_u64(__a, b) __extension__ ({ \
-  uint64x2x3_t __b = (b); \
-  __builtin_neon_vst3q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 51); })
-#define vst3q_s64(__a, b) __extension__ ({ \
-  int64x2x3_t __b = (b); \
-  __builtin_neon_vst3q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 35); })
-#define vst3_f64(__a, b) __extension__ ({ \
-  float64x1x3_t __b = (b); \
-  __builtin_neon_vst3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 10); })
-#define vst3q_f64(__a, b) __extension__ ({ \
-  float64x2x3_t __b = (b); \
-  __builtin_neon_vst3q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 42); })
-#define vst3_p64(__a, b) __extension__ ({ \
-  poly64x1x3_t __b = (b); \
-  __builtin_neon_vst3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 6); })
-#define vst3q_p64(__a, b) __extension__ ({ \
-  poly64x2x3_t __b = (b); \
-  __builtin_neon_vst3q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 38); })
-
-#define vst3_lane_s64(__a, b, __c) __extension__ ({ \
-  int64x1x3_t __b = (b); \
-  __builtin_neon_vst3_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 3); })
-#define vst3_lane_u64(__a, b, __c) __extension__ ({ \
-  uint64x1x3_t __b = (b); \
-  __builtin_neon_vst3_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 19); })
-#define vst3q_lane_s8(__a, b, __c) __extension__ ({ \
-  int8x16x3_t __b = (b); \
-  __builtin_neon_vst3q_lane_v(__a, __b.val[0], __b.val[1], __b.val[2], __c, 32); })
-#define vst3q_lane_u8(__a, b, __c) __extension__ ({ \
-  uint8x16x3_t __b = (b); \
-  __builtin_neon_vst3q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 48); })
-#define vst3q_lane_p8(__a, b, __c) __extension__ ({ \
-  poly8x16x3_t __b = (b); \
-  __builtin_neon_vst3q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 36); })
-#define vst3q_lane_s64(__a, b, __c) __extension__ ({ \
-  int64x2x3_t __b = (b); \
-  __builtin_neon_vst3q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 35); })
-#define vst3q_lane_u64(__a, b, __c) __extension__ ({ \
-  uint64x2x3_t __b = (b); \
-  __builtin_neon_vst3q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 51); })
-#define vst3_lane_f64(__a, b, __c) __extension__ ({ \
-  float64x1x3_t __b = (b); \
-  __builtin_neon_vst3_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 10); })
-#define vst3q_lane_f64(__a, b, __c) __extension__ ({ \
-  float64x2x3_t __b = (b); \
-  __builtin_neon_vst3q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 42); })
-#define vst3_lane_p64(__a, b, __c) __extension__ ({ \
-  poly64x1x3_t __b = (b); \
-  __builtin_neon_vst3_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 6); })
-#define vst3q_lane_p64(__a, b, __c) __extension__ ({ \
-  poly64x2x3_t __b = (b); \
-  __builtin_neon_vst3q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 38); })
-
-#define vst4q_u64(__a, b) __extension__ ({ \
-  uint64x2x4_t __b = (b); \
-  __builtin_neon_vst4q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 51); })
-#define vst4q_s64(__a, b) __extension__ ({ \
-  int64x2x4_t __b = (b); \
-  __builtin_neon_vst4q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 35); })
-#define vst4_f64(__a, b) __extension__ ({ \
-  float64x1x4_t __b = (b); \
-  __builtin_neon_vst4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 10); })
-#define vst4q_f64(__a, b) __extension__ ({ \
-  float64x2x4_t __b = (b); \
-  __builtin_neon_vst4q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 42); })
-#define vst4_p64(__a, b) __extension__ ({ \
-  poly64x1x4_t __b = (b); \
-  __builtin_neon_vst4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 6); })
-#define vst4q_p64(__a, b) __extension__ ({ \
-  poly64x2x4_t __b = (b); \
-  __builtin_neon_vst4q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 38); })
-
-#define vst4_lane_s64(__a, b, __c) __extension__ ({ \
-  int64x1x4_t __b = (b); \
-  __builtin_neon_vst4_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 3); })
-#define vst4_lane_u64(__a, b, __c) __extension__ ({ \
-  uint64x1x4_t __b = (b); \
-  __builtin_neon_vst4_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 19); })
-#define vst4q_lane_s8(__a, b, __c) __extension__ ({ \
-  int8x16x4_t __b = (b); \
-  __builtin_neon_vst4q_lane_v(__a, __b.val[0], __b.val[1], __b.val[2], __b.val[3], __c, 32); })
-#define vst4q_lane_u8(__a, b, __c) __extension__ ({ \
-  uint8x16x4_t __b = (b); \
-  __builtin_neon_vst4q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 48); })
-#define vst4q_lane_p8(__a, b, __c) __extension__ ({ \
-  poly8x16x4_t __b = (b); \
-  __builtin_neon_vst4q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 36); })
-#define vst4q_lane_s64(__a, b, __c) __extension__ ({ \
-  int64x2x4_t __b = (b); \
-  __builtin_neon_vst4q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 35); })
-#define vst4q_lane_u64(__a, b, __c) __extension__ ({ \
-  uint64x2x4_t __b = (b); \
-  __builtin_neon_vst4q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 51); })
-#define vst4_lane_f64(__a, b, __c) __extension__ ({ \
-  float64x1x4_t __b = (b); \
-  __builtin_neon_vst4_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 10); })
-#define vst4q_lane_f64(__a, b, __c) __extension__ ({ \
-  float64x2x4_t __b = (b); \
-  __builtin_neon_vst4q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 42); })
-#define vst4_lane_p64(__a, b, __c) __extension__ ({ \
-  poly64x1x4_t __b = (b); \
-  __builtin_neon_vst4_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 6); })
-#define vst4q_lane_p64(__a, b, __c) __extension__ ({ \
-  poly64x2x4_t __b = (b); \
-  __builtin_neon_vst4q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 38); })
-
-__ai float64x1_t vsub_f64(float64x1_t __a, float64x1_t __b) {
-  return __a - __b; }
-__ai float64x2_t vsubq_f64(float64x2_t __a, float64x2_t __b) {
-  return __a - __b; }
-
-__ai int8x8_t vuqadd_s8(int8x8_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vuqadd_v(__a, __b, 0); }
-__ai int16x4_t vuqadd_s16(int16x4_t __a, int16x4_t __b) {
-  return (int16x4_t)__builtin_neon_vuqadd_v((int8x8_t)__a, (int8x8_t)__b, 1); }
-__ai int32x2_t vuqadd_s32(int32x2_t __a, int32x2_t __b) {
-  return (int32x2_t)__builtin_neon_vuqadd_v((int8x8_t)__a, (int8x8_t)__b, 2); }
-__ai int64x1_t vuqadd_s64(int64x1_t __a, int64x1_t __b) {
-  return (int64x1_t)__builtin_neon_vuqadd_v((int8x8_t)__a, (int8x8_t)__b, 3); }
-__ai int8x16_t vuqaddq_s8(int8x16_t __a, int8x16_t __b) {
-  return (int8x16_t)__builtin_neon_vuqaddq_v(__a, __b, 32); }
-__ai int16x8_t vuqaddq_s16(int16x8_t __a, int16x8_t __b) {
-  return (int16x8_t)__builtin_neon_vuqaddq_v((int8x16_t)__a, (int8x16_t)__b, 33); }
-__ai int32x4_t vuqaddq_s32(int32x4_t __a, int32x4_t __b) {
-  return (int32x4_t)__builtin_neon_vuqaddq_v((int8x16_t)__a, (int8x16_t)__b, 34); }
-__ai int64x2_t vuqaddq_s64(int64x2_t __a, int64x2_t __b) {
-  return (int64x2_t)__builtin_neon_vuqaddq_v((int8x16_t)__a, (int8x16_t)__b, 35); }
-
-__ai uint8x8_t vsqadd_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vsqadd_v((int8x8_t)__a, (int8x8_t)__b, 16); }
-__ai uint16x4_t vsqadd_u16(uint16x4_t __a, uint16x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vsqadd_v((int8x8_t)__a, (int8x8_t)__b, 17); }
-__ai uint32x2_t vsqadd_u32(uint32x2_t __a, uint32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vsqadd_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai uint64x1_t vsqadd_u64(uint64x1_t __a, uint64x1_t __b) {
-  return (uint64x1_t)__builtin_neon_vsqadd_v((int8x8_t)__a, (int8x8_t)__b, 19); }
-__ai uint8x16_t vsqaddq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vsqaddq_v((int8x16_t)__a, (int8x16_t)__b, 48); }
-__ai uint16x8_t vsqaddq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint16x8_t)__builtin_neon_vsqaddq_v((int8x16_t)__a, (int8x16_t)__b, 49); }
-__ai uint32x4_t vsqaddq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vsqaddq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-__ai uint64x2_t vsqaddq_u64(uint64x2_t __a, uint64x2_t __b) {
-  return (uint64x2_t)__builtin_neon_vsqaddq_v((int8x16_t)__a, (int8x16_t)__b, 51); }
-
-#endif
-
-__ai int8x8_t vaba_s8(int8x8_t __a, int8x8_t __b, int8x8_t __c) {
-  return __a + vabd_s8(__b, __c); }
-__ai int16x4_t vaba_s16(int16x4_t __a, int16x4_t __b, int16x4_t __c) {
-  return __a + vabd_s16(__b, __c); }
-__ai int32x2_t vaba_s32(int32x2_t __a, int32x2_t __b, int32x2_t __c) {
-  return __a + vabd_s32(__b, __c); }
-__ai uint8x8_t vaba_u8(uint8x8_t __a, uint8x8_t __b, uint8x8_t __c) {
-  return __a + vabd_u8(__b, __c); }
-__ai uint16x4_t vaba_u16(uint16x4_t __a, uint16x4_t __b, uint16x4_t __c) {
-  return __a + vabd_u16(__b, __c); }
-__ai uint32x2_t vaba_u32(uint32x2_t __a, uint32x2_t __b, uint32x2_t __c) {
-  return __a + vabd_u32(__b, __c); }
-__ai int8x16_t vabaq_s8(int8x16_t __a, int8x16_t __b, int8x16_t __c) {
-  return __a + vabdq_s8(__b, __c); }
-__ai int16x8_t vabaq_s16(int16x8_t __a, int16x8_t __b, int16x8_t __c) {
-  return __a + vabdq_s16(__b, __c); }
-__ai int32x4_t vabaq_s32(int32x4_t __a, int32x4_t __b, int32x4_t __c) {
-  return __a + vabdq_s32(__b, __c); }
-__ai uint8x16_t vabaq_u8(uint8x16_t __a, uint8x16_t __b, uint8x16_t __c) {
-  return __a + vabdq_u8(__b, __c); }
-__ai uint16x8_t vabaq_u16(uint16x8_t __a, uint16x8_t __b, uint16x8_t __c) {
-  return __a + vabdq_u16(__b, __c); }
-__ai uint32x4_t vabaq_u32(uint32x4_t __a, uint32x4_t __b, uint32x4_t __c) {
-  return __a + vabdq_u32(__b, __c); }
-
-__ai int16x8_t vabal_s8(int16x8_t __a, int8x8_t __b, int8x8_t __c) {
-  return __a + vabdl_s8(__b, __c); }
-__ai int32x4_t vabal_s16(int32x4_t __a, int16x4_t __b, int16x4_t __c) {
-  return __a + vabdl_s16(__b, __c); }
-__ai int64x2_t vabal_s32(int64x2_t __a, int32x2_t __b, int32x2_t __c) {
-  return __a + vabdl_s32(__b, __c); }
-__ai uint16x8_t vabal_u8(uint16x8_t __a, uint8x8_t __b, uint8x8_t __c) {
-  return __a + vabdl_u8(__b, __c); }
-__ai uint32x4_t vabal_u16(uint32x4_t __a, uint16x4_t __b, uint16x4_t __c) {
-  return __a + vabdl_u16(__b, __c); }
-__ai uint64x2_t vabal_u32(uint64x2_t __a, uint32x2_t __b, uint32x2_t __c) {
-  return __a + vabdl_u32(__b, __c); }
+#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
 
-#if defined(__aarch64__)
-__ai int16x8_t vabal_high_s8(int16x8_t __a, int8x16_t __b, int8x16_t __c) {
-  return vabal_s8(__a, vget_high_s8(__b), vget_high_s8(__c)); }
-__ai int32x4_t vabal_high_s16(int32x4_t __a, int16x8_t __b, int16x8_t __c) {
-  return vabal_s16(__a, vget_high_s16(__b), vget_high_s16(__c)); }
-__ai int64x2_t vabal_high_s32(int64x2_t __a, int32x4_t __b, int32x4_t __c) {
-  return vabal_s32(__a, vget_high_s32(__b), vget_high_s32(__c)); }
-__ai uint16x8_t vabal_high_u8(uint16x8_t __a, uint8x16_t __b, uint8x16_t __c) {
-  return vabal_u8(__a, vget_high_u8(__b), vget_high_u8(__c)); }
-__ai uint32x4_t vabal_high_u16(uint32x4_t __a, uint16x8_t __b, uint16x8_t __c) {
-  return vabal_u16(__a, vget_high_u16(__b), vget_high_u16(__c)); }
-__ai uint64x2_t vabal_high_u32(uint64x2_t __a, uint32x4_t __b, uint32x4_t __c) {
-  return vabal_u32(__a, vget_high_u32(__b), vget_high_u32(__c)); }
-
-__ai int16x8_t vabdl_high_s8(int8x16_t __a, int8x16_t __b) {
-  return vabdl_s8(vget_high_s8(__a), vget_high_s8(__b)); }
-__ai int32x4_t vabdl_high_s16(int16x8_t __a, int16x8_t __b) {
-  return vabdl_s16(vget_high_s16(__a), vget_high_s16(__b)); }
-__ai int64x2_t vabdl_high_s32(int32x4_t __a, int32x4_t __b) {
-  return vabdl_s32(vget_high_s32(__a), vget_high_s32(__b)); }
-__ai uint16x8_t vabdl_high_u8(uint8x16_t __a, uint8x16_t __b) {
-  return vabdl_u8(vget_high_u8(__a), vget_high_u8(__b)); }
-__ai uint32x4_t vabdl_high_u16(uint16x8_t __a, uint16x8_t __b) {
-  return vabdl_u16(vget_high_u16(__a), vget_high_u16(__b)); }
-__ai uint64x2_t vabdl_high_u32(uint32x4_t __a, uint32x4_t __b) {
-  return vabdl_u32(vget_high_u32(__a), vget_high_u32(__b)); }
-
-#endif
-
-__ai int8x8_t vabs_s8(int8x8_t __a) {
-  return (int8x8_t)__builtin_neon_vabs_v(__a, 0); }
-__ai int16x4_t vabs_s16(int16x4_t __a) {
-  return (int16x4_t)__builtin_neon_vabs_v((int8x8_t)__a, 1); }
-__ai int32x2_t vabs_s32(int32x2_t __a) {
-  return (int32x2_t)__builtin_neon_vabs_v((int8x8_t)__a, 2); }
-__ai float32x2_t vabs_f32(float32x2_t __a) {
-  return (float32x2_t)__builtin_neon_vabs_v((int8x8_t)__a, 9); }
-__ai int8x16_t vabsq_s8(int8x16_t __a) {
-  return (int8x16_t)__builtin_neon_vabsq_v(__a, 32); }
-__ai int16x8_t vabsq_s16(int16x8_t __a) {
-  return (int16x8_t)__builtin_neon_vabsq_v((int8x16_t)__a, 33); }
-__ai int32x4_t vabsq_s32(int32x4_t __a) {
-  return (int32x4_t)__builtin_neon_vabsq_v((int8x16_t)__a, 34); }
-__ai float32x4_t vabsq_f32(float32x4_t __a) {
-  return (float32x4_t)__builtin_neon_vabsq_v((int8x16_t)__a, 41); }
-
-__ai int8x8_t vadd_s8(int8x8_t __a, int8x8_t __b) {
-  return __a + __b; }
-__ai int16x4_t vadd_s16(int16x4_t __a, int16x4_t __b) {
-  return __a + __b; }
-__ai int32x2_t vadd_s32(int32x2_t __a, int32x2_t __b) {
-  return __a + __b; }
-__ai int64x1_t vadd_s64(int64x1_t __a, int64x1_t __b) {
-  return __a + __b; }
-__ai float32x2_t vadd_f32(float32x2_t __a, float32x2_t __b) {
-  return __a + __b; }
-__ai uint8x8_t vadd_u8(uint8x8_t __a, uint8x8_t __b) {
-  return __a + __b; }
-__ai uint16x4_t vadd_u16(uint16x4_t __a, uint16x4_t __b) {
-  return __a + __b; }
-__ai uint32x2_t vadd_u32(uint32x2_t __a, uint32x2_t __b) {
-  return __a + __b; }
-__ai uint64x1_t vadd_u64(uint64x1_t __a, uint64x1_t __b) {
-  return __a + __b; }
-__ai int8x16_t vaddq_s8(int8x16_t __a, int8x16_t __b) {
-  return __a + __b; }
-__ai int16x8_t vaddq_s16(int16x8_t __a, int16x8_t __b) {
-  return __a + __b; }
-__ai int32x4_t vaddq_s32(int32x4_t __a, int32x4_t __b) {
-  return __a + __b; }
-__ai int64x2_t vaddq_s64(int64x2_t __a, int64x2_t __b) {
-  return __a + __b; }
-__ai float32x4_t vaddq_f32(float32x4_t __a, float32x4_t __b) {
-  return __a + __b; }
-__ai uint8x16_t vaddq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return __a + __b; }
-__ai uint16x8_t vaddq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return __a + __b; }
-__ai uint32x4_t vaddq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return __a + __b; }
-__ai uint64x2_t vaddq_u64(uint64x2_t __a, uint64x2_t __b) {
-  return __a + __b; }
-
-__ai int8x8_t vaddhn_s16(int16x8_t __a, int16x8_t __b) {
-  return (int8x8_t)__builtin_neon_vaddhn_v((int8x16_t)__a, (int8x16_t)__b, 0); }
-__ai int16x4_t vaddhn_s32(int32x4_t __a, int32x4_t __b) {
-  return (int16x4_t)__builtin_neon_vaddhn_v((int8x16_t)__a, (int8x16_t)__b, 1); }
-__ai int32x2_t vaddhn_s64(int64x2_t __a, int64x2_t __b) {
-  return (int32x2_t)__builtin_neon_vaddhn_v((int8x16_t)__a, (int8x16_t)__b, 2); }
-__ai uint8x8_t vaddhn_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vaddhn_v((int8x16_t)__a, (int8x16_t)__b, 16); }
-__ai uint16x4_t vaddhn_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vaddhn_v((int8x16_t)__a, (int8x16_t)__b, 17); }
-__ai uint32x2_t vaddhn_u64(uint64x2_t __a, uint64x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vaddhn_v((int8x16_t)__a, (int8x16_t)__b, 18); }
+#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
 
-#if defined(__aarch64__)
-__ai int8x16_t vaddhn_high_s16(int8x8_t __a, int16x8_t __b, int16x8_t __c) {
-  return vcombine_s8(__a, vaddhn_s16(__b, __c)); }
-__ai int16x8_t vaddhn_high_s32(int16x4_t __a, int32x4_t __b, int32x4_t __c) {
-  return vcombine_s16(__a, vaddhn_s32(__b, __c)); }
-__ai int32x4_t vaddhn_high_s64(int32x2_t __a, int64x2_t __b, int64x2_t __c) {
-  return vcombine_s32(__a, vaddhn_s64(__b, __c)); }
-__ai uint8x16_t vaddhn_high_u16(uint8x8_t __a, uint16x8_t __b, uint16x8_t __c) {
-  return vcombine_u8(__a, vaddhn_u16(__b, __c)); }
-__ai uint16x8_t vaddhn_high_u32(uint16x4_t __a, uint32x4_t __b, uint32x4_t __c) {
-  return vcombine_u16(__a, vaddhn_u32(__b, __c)); }
-__ai uint32x4_t vaddhn_high_u64(uint32x2_t __a, uint64x2_t __b, uint64x2_t __c) {
-  return vcombine_u32(__a, vaddhn_u64(__b, __c)); }
-
-#endif
-
-__ai int16x8_t vaddl_s8(int8x8_t __a, int8x8_t __b) {
-  return vmovl_s8(__a) + vmovl_s8(__b); }
-__ai int32x4_t vaddl_s16(int16x4_t __a, int16x4_t __b) {
-  return vmovl_s16(__a) + vmovl_s16(__b); }
-__ai int64x2_t vaddl_s32(int32x2_t __a, int32x2_t __b) {
-  return vmovl_s32(__a) + vmovl_s32(__b); }
-__ai uint16x8_t vaddl_u8(uint8x8_t __a, uint8x8_t __b) {
-  return vmovl_u8(__a) + vmovl_u8(__b); }
-__ai uint32x4_t vaddl_u16(uint16x4_t __a, uint16x4_t __b) {
-  return vmovl_u16(__a) + vmovl_u16(__b); }
-__ai uint64x2_t vaddl_u32(uint32x2_t __a, uint32x2_t __b) {
-  return vmovl_u32(__a) + vmovl_u32(__b); }
+#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__)
-__ai int16_t vaddlv_s8(int8x8_t __a) {
-  return (int16_t)__builtin_neon_vaddlv_s8(__a); }
-__ai int32_t vaddlv_s16(int16x4_t __a) {
-  return (int32_t)__builtin_neon_vaddlv_s16(__a); }
-__ai int64_t vaddlv_s32(int32x2_t __a) {
-  return (int64_t)__builtin_neon_vaddlv_s32(__a); }
-__ai uint16_t vaddlv_u8(uint8x8_t __a) {
-  return (uint16_t)__builtin_neon_vaddlv_u8((int8x8_t)__a); }
-__ai uint32_t vaddlv_u16(uint16x4_t __a) {
-  return (uint32_t)__builtin_neon_vaddlv_u16((int16x4_t)__a); }
-__ai uint64_t vaddlv_u32(uint32x2_t __a) {
-  return (uint64_t)__builtin_neon_vaddlv_u32((int32x2_t)__a); }
-__ai int16_t vaddlvq_s8(int8x16_t __a) {
-  return (int16_t)__builtin_neon_vaddlvq_s8(__a); }
-__ai int32_t vaddlvq_s16(int16x8_t __a) {
-  return (int32_t)__builtin_neon_vaddlvq_s16(__a); }
-__ai int64_t vaddlvq_s32(int32x4_t __a) {
-  return (int64_t)__builtin_neon_vaddlvq_s32(__a); }
-__ai uint16_t vaddlvq_u8(uint8x16_t __a) {
-  return (uint16_t)__builtin_neon_vaddlvq_u8((int8x16_t)__a); }
-__ai uint32_t vaddlvq_u16(uint16x8_t __a) {
-  return (uint32_t)__builtin_neon_vaddlvq_u16((int16x8_t)__a); }
-__ai uint64_t vaddlvq_u32(uint32x4_t __a) {
-  return (uint64_t)__builtin_neon_vaddlvq_u32((int32x4_t)__a); }
-
-__ai int16x8_t vaddl_high_s8(int8x16_t __a, int8x16_t __b) {
-  return vmovl_high_s8(__a) + vmovl_high_s8(__b); }
-__ai int32x4_t vaddl_high_s16(int16x8_t __a, int16x8_t __b) {
-  return vmovl_high_s16(__a) + vmovl_high_s16(__b); }
-__ai int64x2_t vaddl_high_s32(int32x4_t __a, int32x4_t __b) {
-  return vmovl_high_s32(__a) + vmovl_high_s32(__b); }
-__ai uint16x8_t vaddl_high_u8(uint8x16_t __a, uint8x16_t __b) {
-  return vmovl_high_u8(__a) + vmovl_high_u8(__b); }
-__ai uint32x4_t vaddl_high_u16(uint16x8_t __a, uint16x8_t __b) {
-  return vmovl_high_u16(__a) + vmovl_high_u16(__b); }
-__ai uint64x2_t vaddl_high_u32(uint32x4_t __a, uint32x4_t __b) {
-  return vmovl_high_u32(__a) + vmovl_high_u32(__b); }
-
-__ai int8_t vaddv_s8(int8x8_t __a) {
-  return (int8_t)__builtin_neon_vaddv_s8(__a); }
-__ai int16_t vaddv_s16(int16x4_t __a) {
-  return (int16_t)__builtin_neon_vaddv_s16(__a); }
-__ai int32_t vaddv_s32(int32x2_t __a) {
-  return (int32_t)__builtin_neon_vaddv_s32(__a); }
-__ai float32_t vaddv_f32(float32x2_t __a) {
-  return (float32_t)__builtin_neon_vaddv_f32(__a); }
-__ai uint8_t vaddv_u8(uint8x8_t __a) {
-  return (uint8_t)__builtin_neon_vaddv_u8((int8x8_t)__a); }
-__ai uint16_t vaddv_u16(uint16x4_t __a) {
-  return (uint16_t)__builtin_neon_vaddv_u16((int16x4_t)__a); }
-__ai uint32_t vaddv_u32(uint32x2_t __a) {
-  return (uint32_t)__builtin_neon_vaddv_u32((int32x2_t)__a); }
-__ai int8_t vaddvq_s8(int8x16_t __a) {
-  return (int8_t)__builtin_neon_vaddvq_s8(__a); }
-__ai int16_t vaddvq_s16(int16x8_t __a) {
-  return (int16_t)__builtin_neon_vaddvq_s16(__a); }
-__ai int32_t vaddvq_s32(int32x4_t __a) {
-  return (int32_t)__builtin_neon_vaddvq_s32(__a); }
-__ai uint8_t vaddvq_u8(uint8x16_t __a) {
-  return (uint8_t)__builtin_neon_vaddvq_u8((int8x16_t)__a); }
-__ai uint16_t vaddvq_u16(uint16x8_t __a) {
-  return (uint16_t)__builtin_neon_vaddvq_u16((int16x8_t)__a); }
-__ai uint32_t vaddvq_u32(uint32x4_t __a) {
-  return (uint32_t)__builtin_neon_vaddvq_u32((int32x4_t)__a); }
-__ai float32_t vaddvq_f32(float32x4_t __a) {
-  return (float32_t)__builtin_neon_vaddvq_f32(__a); }
-__ai float64_t vaddvq_f64(float64x2_t __a) {
-  return (float64_t)__builtin_neon_vaddvq_f64(__a); }
-__ai int64_t vaddvq_s64(int64x2_t __a) {
-  return (int64_t)__builtin_neon_vaddvq_s64(__a); }
-__ai uint64_t vaddvq_u64(uint64x2_t __a) {
-  return (uint64_t)__builtin_neon_vaddvq_u64((int64x2_t)__a); }
-
-#endif
-
-__ai int16x8_t vaddw_s8(int16x8_t __a, int8x8_t __b) {
-  return __a + vmovl_s8(__b); }
-__ai int32x4_t vaddw_s16(int32x4_t __a, int16x4_t __b) {
-  return __a + vmovl_s16(__b); }
-__ai int64x2_t vaddw_s32(int64x2_t __a, int32x2_t __b) {
-  return __a + vmovl_s32(__b); }
-__ai uint16x8_t vaddw_u8(uint16x8_t __a, uint8x8_t __b) {
-  return __a + vmovl_u8(__b); }
-__ai uint32x4_t vaddw_u16(uint32x4_t __a, uint16x4_t __b) {
-  return __a + vmovl_u16(__b); }
-__ai uint64x2_t vaddw_u32(uint64x2_t __a, uint32x2_t __b) {
-  return __a + vmovl_u32(__b); }
+#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
 
-#if defined(__aarch64__)
-__ai int16x8_t vaddw_high_s8(int16x8_t __a, int8x16_t __b) {
-  return __a + vmovl_high_s8(__b); }
-__ai int32x4_t vaddw_high_s16(int32x4_t __a, int16x8_t __b) {
-  return __a + vmovl_high_s16(__b); }
-__ai int64x2_t vaddw_high_s32(int64x2_t __a, int32x4_t __b) {
-  return __a + vmovl_high_s32(__b); }
-__ai uint16x8_t vaddw_high_u8(uint16x8_t __a, uint8x16_t __b) {
-  return __a + vmovl_high_u8(__b); }
-__ai uint32x4_t vaddw_high_u16(uint32x4_t __a, uint16x8_t __b) {
-  return __a + vmovl_high_u16(__b); }
-__ai uint64x2_t vaddw_high_u32(uint64x2_t __a, uint32x4_t __b) {
-  return __a + vmovl_high_u32(__b); }
-
-#endif
-
-__ai int8x8_t vand_s8(int8x8_t __a, int8x8_t __b) {
-  return __a & __b; }
-__ai int16x4_t vand_s16(int16x4_t __a, int16x4_t __b) {
-  return __a & __b; }
-__ai int32x2_t vand_s32(int32x2_t __a, int32x2_t __b) {
-  return __a & __b; }
-__ai int64x1_t vand_s64(int64x1_t __a, int64x1_t __b) {
-  return __a & __b; }
-__ai uint8x8_t vand_u8(uint8x8_t __a, uint8x8_t __b) {
-  return __a & __b; }
-__ai uint16x4_t vand_u16(uint16x4_t __a, uint16x4_t __b) {
-  return __a & __b; }
-__ai uint32x2_t vand_u32(uint32x2_t __a, uint32x2_t __b) {
-  return __a & __b; }
-__ai uint64x1_t vand_u64(uint64x1_t __a, uint64x1_t __b) {
-  return __a & __b; }
-__ai int8x16_t vandq_s8(int8x16_t __a, int8x16_t __b) {
-  return __a & __b; }
-__ai int16x8_t vandq_s16(int16x8_t __a, int16x8_t __b) {
-  return __a & __b; }
-__ai int32x4_t vandq_s32(int32x4_t __a, int32x4_t __b) {
-  return __a & __b; }
-__ai int64x2_t vandq_s64(int64x2_t __a, int64x2_t __b) {
-  return __a & __b; }
-__ai uint8x16_t vandq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return __a & __b; }
-__ai uint16x8_t vandq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return __a & __b; }
-__ai uint32x4_t vandq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return __a & __b; }
-__ai uint64x2_t vandq_u64(uint64x2_t __a, uint64x2_t __b) {
-  return __a & __b; }
-
-__ai int8x8_t vbic_s8(int8x8_t __a, int8x8_t __b) {
-  return __a & ~__b; }
-__ai int16x4_t vbic_s16(int16x4_t __a, int16x4_t __b) {
-  return __a & ~__b; }
-__ai int32x2_t vbic_s32(int32x2_t __a, int32x2_t __b) {
-  return __a & ~__b; }
-__ai int64x1_t vbic_s64(int64x1_t __a, int64x1_t __b) {
-  return __a & ~__b; }
-__ai uint8x8_t vbic_u8(uint8x8_t __a, uint8x8_t __b) {
-  return __a & ~__b; }
-__ai uint16x4_t vbic_u16(uint16x4_t __a, uint16x4_t __b) {
-  return __a & ~__b; }
-__ai uint32x2_t vbic_u32(uint32x2_t __a, uint32x2_t __b) {
-  return __a & ~__b; }
-__ai uint64x1_t vbic_u64(uint64x1_t __a, uint64x1_t __b) {
-  return __a & ~__b; }
-__ai int8x16_t vbicq_s8(int8x16_t __a, int8x16_t __b) {
-  return __a & ~__b; }
-__ai int16x8_t vbicq_s16(int16x8_t __a, int16x8_t __b) {
-  return __a & ~__b; }
-__ai int32x4_t vbicq_s32(int32x4_t __a, int32x4_t __b) {
-  return __a & ~__b; }
-__ai int64x2_t vbicq_s64(int64x2_t __a, int64x2_t __b) {
-  return __a & ~__b; }
-__ai uint8x16_t vbicq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return __a & ~__b; }
-__ai uint16x8_t vbicq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return __a & ~__b; }
-__ai uint32x4_t vbicq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return __a & ~__b; }
-__ai uint64x2_t vbicq_u64(uint64x2_t __a, uint64x2_t __b) {
-  return __a & ~__b; }
-
-__ai int8x8_t vbsl_s8(uint8x8_t __a, int8x8_t __b, int8x8_t __c) {
-  return (int8x8_t)__builtin_neon_vbsl_v((int8x8_t)__a, __b, __c, 0); }
-__ai int16x4_t vbsl_s16(uint16x4_t __a, int16x4_t __b, int16x4_t __c) {
-  return (int16x4_t)__builtin_neon_vbsl_v((int8x8_t)__a, (int8x8_t)__b, (int8x8_t)__c, 1); }
-__ai int32x2_t vbsl_s32(uint32x2_t __a, int32x2_t __b, int32x2_t __c) {
-  return (int32x2_t)__builtin_neon_vbsl_v((int8x8_t)__a, (int8x8_t)__b, (int8x8_t)__c, 2); }
-__ai int64x1_t vbsl_s64(uint64x1_t __a, int64x1_t __b, int64x1_t __c) {
-  return (int64x1_t)__builtin_neon_vbsl_v((int8x8_t)__a, (int8x8_t)__b, (int8x8_t)__c, 3); }
-__ai uint8x8_t vbsl_u8(uint8x8_t __a, uint8x8_t __b, uint8x8_t __c) {
-  return (uint8x8_t)__builtin_neon_vbsl_v((int8x8_t)__a, (int8x8_t)__b, (int8x8_t)__c, 16); }
-__ai uint16x4_t vbsl_u16(uint16x4_t __a, uint16x4_t __b, uint16x4_t __c) {
-  return (uint16x4_t)__builtin_neon_vbsl_v((int8x8_t)__a, (int8x8_t)__b, (int8x8_t)__c, 17); }
-__ai uint32x2_t vbsl_u32(uint32x2_t __a, uint32x2_t __b, uint32x2_t __c) {
-  return (uint32x2_t)__builtin_neon_vbsl_v((int8x8_t)__a, (int8x8_t)__b, (int8x8_t)__c, 18); }
-__ai uint64x1_t vbsl_u64(uint64x1_t __a, uint64x1_t __b, uint64x1_t __c) {
-  return (uint64x1_t)__builtin_neon_vbsl_v((int8x8_t)__a, (int8x8_t)__b, (int8x8_t)__c, 19); }
-__ai float32x2_t vbsl_f32(uint32x2_t __a, float32x2_t __b, float32x2_t __c) {
-  return (float32x2_t)__builtin_neon_vbsl_v((int8x8_t)__a, (int8x8_t)__b, (int8x8_t)__c, 9); }
-__ai poly8x8_t vbsl_p8(uint8x8_t __a, poly8x8_t __b, poly8x8_t __c) {
-  return (poly8x8_t)__builtin_neon_vbsl_v((int8x8_t)__a, (int8x8_t)__b, (int8x8_t)__c, 4); }
-__ai poly16x4_t vbsl_p16(uint16x4_t __a, poly16x4_t __b, poly16x4_t __c) {
-  return (poly16x4_t)__builtin_neon_vbsl_v((int8x8_t)__a, (int8x8_t)__b, (int8x8_t)__c, 5); }
-__ai int8x16_t vbslq_s8(uint8x16_t __a, int8x16_t __b, int8x16_t __c) {
-  return (int8x16_t)__builtin_neon_vbslq_v((int8x16_t)__a, __b, __c, 32); }
-__ai int16x8_t vbslq_s16(uint16x8_t __a, int16x8_t __b, int16x8_t __c) {
-  return (int16x8_t)__builtin_neon_vbslq_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 33); }
-__ai int32x4_t vbslq_s32(uint32x4_t __a, int32x4_t __b, int32x4_t __c) {
-  return (int32x4_t)__builtin_neon_vbslq_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 34); }
-__ai int64x2_t vbslq_s64(uint64x2_t __a, int64x2_t __b, int64x2_t __c) {
-  return (int64x2_t)__builtin_neon_vbslq_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 35); }
-__ai uint8x16_t vbslq_u8(uint8x16_t __a, uint8x16_t __b, uint8x16_t __c) {
-  return (uint8x16_t)__builtin_neon_vbslq_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 48); }
-__ai uint16x8_t vbslq_u16(uint16x8_t __a, uint16x8_t __b, uint16x8_t __c) {
-  return (uint16x8_t)__builtin_neon_vbslq_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 49); }
-__ai uint32x4_t vbslq_u32(uint32x4_t __a, uint32x4_t __b, uint32x4_t __c) {
-  return (uint32x4_t)__builtin_neon_vbslq_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 50); }
-__ai uint64x2_t vbslq_u64(uint64x2_t __a, uint64x2_t __b, uint64x2_t __c) {
-  return (uint64x2_t)__builtin_neon_vbslq_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 51); }
-__ai float32x4_t vbslq_f32(uint32x4_t __a, float32x4_t __b, float32x4_t __c) {
-  return (float32x4_t)__builtin_neon_vbslq_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 41); }
-__ai poly8x16_t vbslq_p8(uint8x16_t __a, poly8x16_t __b, poly8x16_t __c) {
-  return (poly8x16_t)__builtin_neon_vbslq_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 36); }
-__ai poly16x8_t vbslq_p16(uint16x8_t __a, poly16x8_t __b, poly16x8_t __c) {
-  return (poly16x8_t)__builtin_neon_vbslq_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 37); }
-
-__ai uint32x2_t vcage_f32(float32x2_t __a, float32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vcage_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai uint32x4_t vcageq_f32(float32x4_t __a, float32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vcageq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-
-__ai uint32x2_t vcagt_f32(float32x2_t __a, float32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vcagt_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai uint32x4_t vcagtq_f32(float32x4_t __a, float32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vcagtq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-
-__ai uint32x2_t vcale_f32(float32x2_t __a, float32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vcale_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai uint32x4_t vcaleq_f32(float32x4_t __a, float32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vcaleq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-
-__ai uint32x2_t vcalt_f32(float32x2_t __a, float32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vcalt_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai uint32x4_t vcaltq_f32(float32x4_t __a, float32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vcaltq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-
-__ai uint8x8_t vceq_s8(int8x8_t __a, int8x8_t __b) {
-  return (uint8x8_t)(__a == __b); }
-__ai uint16x4_t vceq_s16(int16x4_t __a, int16x4_t __b) {
-  return (uint16x4_t)(__a == __b); }
-__ai uint32x2_t vceq_s32(int32x2_t __a, int32x2_t __b) {
-  return (uint32x2_t)(__a == __b); }
-__ai uint32x2_t vceq_f32(float32x2_t __a, float32x2_t __b) {
-  return (uint32x2_t)(__a == __b); }
-__ai uint8x8_t vceq_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint8x8_t)(__a == __b); }
-__ai uint16x4_t vceq_u16(uint16x4_t __a, uint16x4_t __b) {
-  return (uint16x4_t)(__a == __b); }
-__ai uint32x2_t vceq_u32(uint32x2_t __a, uint32x2_t __b) {
-  return (uint32x2_t)(__a == __b); }
-__ai uint8x8_t vceq_p8(poly8x8_t __a, poly8x8_t __b) {
-  return (uint8x8_t)(__a == __b); }
-__ai uint8x16_t vceqq_s8(int8x16_t __a, int8x16_t __b) {
-  return (uint8x16_t)(__a == __b); }
-__ai uint16x8_t vceqq_s16(int16x8_t __a, int16x8_t __b) {
-  return (uint16x8_t)(__a == __b); }
-__ai uint32x4_t vceqq_s32(int32x4_t __a, int32x4_t __b) {
-  return (uint32x4_t)(__a == __b); }
-__ai uint32x4_t vceqq_f32(float32x4_t __a, float32x4_t __b) {
-  return (uint32x4_t)(__a == __b); }
-__ai uint8x16_t vceqq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)(__a == __b); }
-__ai uint16x8_t vceqq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint16x8_t)(__a == __b); }
-__ai uint32x4_t vceqq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint32x4_t)(__a == __b); }
-__ai uint8x16_t vceqq_p8(poly8x16_t __a, poly8x16_t __b) {
-  return (uint8x16_t)(__a == __b); }
-
-__ai uint8x8_t vcge_s8(int8x8_t __a, int8x8_t __b) {
-  return (uint8x8_t)(__a >= __b); }
-__ai uint16x4_t vcge_s16(int16x4_t __a, int16x4_t __b) {
-  return (uint16x4_t)(__a >= __b); }
-__ai uint32x2_t vcge_s32(int32x2_t __a, int32x2_t __b) {
-  return (uint32x2_t)(__a >= __b); }
-__ai uint32x2_t vcge_f32(float32x2_t __a, float32x2_t __b) {
-  return (uint32x2_t)(__a >= __b); }
-__ai uint8x8_t vcge_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint8x8_t)(__a >= __b); }
-__ai uint16x4_t vcge_u16(uint16x4_t __a, uint16x4_t __b) {
-  return (uint16x4_t)(__a >= __b); }
-__ai uint32x2_t vcge_u32(uint32x2_t __a, uint32x2_t __b) {
-  return (uint32x2_t)(__a >= __b); }
-__ai uint8x16_t vcgeq_s8(int8x16_t __a, int8x16_t __b) {
-  return (uint8x16_t)(__a >= __b); }
-__ai uint16x8_t vcgeq_s16(int16x8_t __a, int16x8_t __b) {
-  return (uint16x8_t)(__a >= __b); }
-__ai uint32x4_t vcgeq_s32(int32x4_t __a, int32x4_t __b) {
-  return (uint32x4_t)(__a >= __b); }
-__ai uint32x4_t vcgeq_f32(float32x4_t __a, float32x4_t __b) {
-  return (uint32x4_t)(__a >= __b); }
-__ai uint8x16_t vcgeq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)(__a >= __b); }
-__ai uint16x8_t vcgeq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint16x8_t)(__a >= __b); }
-__ai uint32x4_t vcgeq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint32x4_t)(__a >= __b); }
-
-__ai uint8x8_t vcgt_s8(int8x8_t __a, int8x8_t __b) {
-  return (uint8x8_t)(__a > __b); }
-__ai uint16x4_t vcgt_s16(int16x4_t __a, int16x4_t __b) {
-  return (uint16x4_t)(__a > __b); }
-__ai uint32x2_t vcgt_s32(int32x2_t __a, int32x2_t __b) {
-  return (uint32x2_t)(__a > __b); }
-__ai uint32x2_t vcgt_f32(float32x2_t __a, float32x2_t __b) {
-  return (uint32x2_t)(__a > __b); }
-__ai uint8x8_t vcgt_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint8x8_t)(__a > __b); }
-__ai uint16x4_t vcgt_u16(uint16x4_t __a, uint16x4_t __b) {
-  return (uint16x4_t)(__a > __b); }
-__ai uint32x2_t vcgt_u32(uint32x2_t __a, uint32x2_t __b) {
-  return (uint32x2_t)(__a > __b); }
-__ai uint8x16_t vcgtq_s8(int8x16_t __a, int8x16_t __b) {
-  return (uint8x16_t)(__a > __b); }
-__ai uint16x8_t vcgtq_s16(int16x8_t __a, int16x8_t __b) {
-  return (uint16x8_t)(__a > __b); }
-__ai uint32x4_t vcgtq_s32(int32x4_t __a, int32x4_t __b) {
-  return (uint32x4_t)(__a > __b); }
-__ai uint32x4_t vcgtq_f32(float32x4_t __a, float32x4_t __b) {
-  return (uint32x4_t)(__a > __b); }
-__ai uint8x16_t vcgtq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)(__a > __b); }
-__ai uint16x8_t vcgtq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint16x8_t)(__a > __b); }
-__ai uint32x4_t vcgtq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint32x4_t)(__a > __b); }
-
-__ai uint8x8_t vcle_s8(int8x8_t __a, int8x8_t __b) {
-  return (uint8x8_t)(__a <= __b); }
-__ai uint16x4_t vcle_s16(int16x4_t __a, int16x4_t __b) {
-  return (uint16x4_t)(__a <= __b); }
-__ai uint32x2_t vcle_s32(int32x2_t __a, int32x2_t __b) {
-  return (uint32x2_t)(__a <= __b); }
-__ai uint32x2_t vcle_f32(float32x2_t __a, float32x2_t __b) {
-  return (uint32x2_t)(__a <= __b); }
-__ai uint8x8_t vcle_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint8x8_t)(__a <= __b); }
-__ai uint16x4_t vcle_u16(uint16x4_t __a, uint16x4_t __b) {
-  return (uint16x4_t)(__a <= __b); }
-__ai uint32x2_t vcle_u32(uint32x2_t __a, uint32x2_t __b) {
-  return (uint32x2_t)(__a <= __b); }
-__ai uint8x16_t vcleq_s8(int8x16_t __a, int8x16_t __b) {
-  return (uint8x16_t)(__a <= __b); }
-__ai uint16x8_t vcleq_s16(int16x8_t __a, int16x8_t __b) {
-  return (uint16x8_t)(__a <= __b); }
-__ai uint32x4_t vcleq_s32(int32x4_t __a, int32x4_t __b) {
-  return (uint32x4_t)(__a <= __b); }
-__ai uint32x4_t vcleq_f32(float32x4_t __a, float32x4_t __b) {
-  return (uint32x4_t)(__a <= __b); }
-__ai uint8x16_t vcleq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)(__a <= __b); }
-__ai uint16x8_t vcleq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint16x8_t)(__a <= __b); }
-__ai uint32x4_t vcleq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint32x4_t)(__a <= __b); }
-
-__ai int8x8_t vcls_s8(int8x8_t __a) {
-  return (int8x8_t)__builtin_neon_vcls_v(__a, 0); }
-__ai int16x4_t vcls_s16(int16x4_t __a) {
-  return (int16x4_t)__builtin_neon_vcls_v((int8x8_t)__a, 1); }
-__ai int32x2_t vcls_s32(int32x2_t __a) {
-  return (int32x2_t)__builtin_neon_vcls_v((int8x8_t)__a, 2); }
-__ai int8x16_t vclsq_s8(int8x16_t __a) {
-  return (int8x16_t)__builtin_neon_vclsq_v(__a, 32); }
-__ai int16x8_t vclsq_s16(int16x8_t __a) {
-  return (int16x8_t)__builtin_neon_vclsq_v((int8x16_t)__a, 33); }
-__ai int32x4_t vclsq_s32(int32x4_t __a) {
-  return (int32x4_t)__builtin_neon_vclsq_v((int8x16_t)__a, 34); }
-
-__ai uint8x8_t vclt_s8(int8x8_t __a, int8x8_t __b) {
-  return (uint8x8_t)(__a < __b); }
-__ai uint16x4_t vclt_s16(int16x4_t __a, int16x4_t __b) {
-  return (uint16x4_t)(__a < __b); }
-__ai uint32x2_t vclt_s32(int32x2_t __a, int32x2_t __b) {
-  return (uint32x2_t)(__a < __b); }
-__ai uint32x2_t vclt_f32(float32x2_t __a, float32x2_t __b) {
-  return (uint32x2_t)(__a < __b); }
-__ai uint8x8_t vclt_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint8x8_t)(__a < __b); }
-__ai uint16x4_t vclt_u16(uint16x4_t __a, uint16x4_t __b) {
-  return (uint16x4_t)(__a < __b); }
-__ai uint32x2_t vclt_u32(uint32x2_t __a, uint32x2_t __b) {
-  return (uint32x2_t)(__a < __b); }
-__ai uint8x16_t vcltq_s8(int8x16_t __a, int8x16_t __b) {
-  return (uint8x16_t)(__a < __b); }
-__ai uint16x8_t vcltq_s16(int16x8_t __a, int16x8_t __b) {
-  return (uint16x8_t)(__a < __b); }
-__ai uint32x4_t vcltq_s32(int32x4_t __a, int32x4_t __b) {
-  return (uint32x4_t)(__a < __b); }
-__ai uint32x4_t vcltq_f32(float32x4_t __a, float32x4_t __b) {
-  return (uint32x4_t)(__a < __b); }
-__ai uint8x16_t vcltq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)(__a < __b); }
-__ai uint16x8_t vcltq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint16x8_t)(__a < __b); }
-__ai uint32x4_t vcltq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint32x4_t)(__a < __b); }
-
-__ai int8x8_t vclz_s8(int8x8_t __a) {
-  return (int8x8_t)__builtin_neon_vclz_v(__a, 0); }
-__ai int16x4_t vclz_s16(int16x4_t __a) {
-  return (int16x4_t)__builtin_neon_vclz_v((int8x8_t)__a, 1); }
-__ai int32x2_t vclz_s32(int32x2_t __a) {
-  return (int32x2_t)__builtin_neon_vclz_v((int8x8_t)__a, 2); }
-__ai uint8x8_t vclz_u8(uint8x8_t __a) {
-  return (uint8x8_t)__builtin_neon_vclz_v((int8x8_t)__a, 16); }
-__ai uint16x4_t vclz_u16(uint16x4_t __a) {
-  return (uint16x4_t)__builtin_neon_vclz_v((int8x8_t)__a, 17); }
-__ai uint32x2_t vclz_u32(uint32x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vclz_v((int8x8_t)__a, 18); }
-__ai int8x16_t vclzq_s8(int8x16_t __a) {
-  return (int8x16_t)__builtin_neon_vclzq_v(__a, 32); }
-__ai int16x8_t vclzq_s16(int16x8_t __a) {
-  return (int16x8_t)__builtin_neon_vclzq_v((int8x16_t)__a, 33); }
-__ai int32x4_t vclzq_s32(int32x4_t __a) {
-  return (int32x4_t)__builtin_neon_vclzq_v((int8x16_t)__a, 34); }
-__ai uint8x16_t vclzq_u8(uint8x16_t __a) {
-  return (uint8x16_t)__builtin_neon_vclzq_v((int8x16_t)__a, 48); }
-__ai uint16x8_t vclzq_u16(uint16x8_t __a) {
-  return (uint16x8_t)__builtin_neon_vclzq_v((int8x16_t)__a, 49); }
-__ai uint32x4_t vclzq_u32(uint32x4_t __a) {
-  return (uint32x4_t)__builtin_neon_vclzq_v((int8x16_t)__a, 50); }
-
-__ai uint8x8_t vcnt_u8(uint8x8_t __a) {
-  return (uint8x8_t)__builtin_neon_vcnt_v((int8x8_t)__a, 16); }
-__ai int8x8_t vcnt_s8(int8x8_t __a) {
-  return (int8x8_t)__builtin_neon_vcnt_v(__a, 0); }
-__ai poly8x8_t vcnt_p8(poly8x8_t __a) {
-  return (poly8x8_t)__builtin_neon_vcnt_v((int8x8_t)__a, 4); }
-__ai uint8x16_t vcntq_u8(uint8x16_t __a) {
-  return (uint8x16_t)__builtin_neon_vcntq_v((int8x16_t)__a, 48); }
-__ai int8x16_t vcntq_s8(int8x16_t __a) {
-  return (int8x16_t)__builtin_neon_vcntq_v(__a, 32); }
-__ai poly8x16_t vcntq_p8(poly8x16_t __a) {
-  return (poly8x16_t)__builtin_neon_vcntq_v((int8x16_t)__a, 36); }
-
-__ai int8x8_t vcreate_s8(uint64_t __a) {
-  return (int8x8_t)__a; }
-__ai int16x4_t vcreate_s16(uint64_t __a) {
-  return (int16x4_t)__a; }
-__ai int32x2_t vcreate_s32(uint64_t __a) {
-  return (int32x2_t)__a; }
-__ai float16x4_t vcreate_f16(uint64_t __a) {
-  return (float16x4_t)__a; }
-__ai float32x2_t vcreate_f32(uint64_t __a) {
-  return (float32x2_t)__a; }
-__ai uint8x8_t vcreate_u8(uint64_t __a) {
-  return (uint8x8_t)__a; }
-__ai uint16x4_t vcreate_u16(uint64_t __a) {
-  return (uint16x4_t)__a; }
-__ai uint32x2_t vcreate_u32(uint64_t __a) {
-  return (uint32x2_t)__a; }
-__ai uint64x1_t vcreate_u64(uint64_t __a) {
-  return (uint64x1_t)__a; }
-__ai poly8x8_t vcreate_p8(uint64_t __a) {
-  return (poly8x8_t)__a; }
-__ai poly16x4_t vcreate_p16(uint64_t __a) {
-  return (poly16x4_t)__a; }
-__ai int64x1_t vcreate_s64(uint64_t __a) {
-  return (int64x1_t)__a; }
+#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
 
-#if defined(__aarch64__)
-__ai float32x2_t vcvtx_f32_f64(float64x2_t __a) {
-  return (float32x2_t)__builtin_neon_vcvtx_f32_v((int8x16_t)__a, 42); }
+#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
 
-__ai float32x4_t vcvtx_high_f32_f64(float32x2_t __a, float64x2_t __b) {
-  float32x2_t __a1 = vcvtx_f32_f64(__b);
-  return __builtin_shufflevector(__a, __a1, 0, 1, 2, 3); }
+#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
 
-__ai float16x4_t vcvt_f16_f32(float32x4_t __a) {
-  return (float16x4_t)__builtin_neon_vcvt_f16_v((int8x16_t)__a, 8); }
+#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
 
-__ai float32x2_t vcvt_f32_s32(int32x2_t __a) {
-  return (float32x2_t)__builtin_neon_vcvt_f32_v((int8x8_t)__a, 2); }
-__ai float32x2_t vcvt_f32_u32(uint32x2_t __a) {
-  return (float32x2_t)__builtin_neon_vcvt_f32_v((int8x8_t)__a, 18); }
-__ai float32x4_t vcvtq_f32_s32(int32x4_t __a) {
-  return (float32x4_t)__builtin_neon_vcvtq_f32_v((int8x16_t)__a, 34); }
-__ai float32x4_t vcvtq_f32_u32(uint32x4_t __a) {
-  return (float32x4_t)__builtin_neon_vcvtq_f32_v((int8x16_t)__a, 50); }
+#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
 
-__ai float32x4_t vcvt_f32_f16(float16x4_t __a) {
-  return (float32x4_t)__builtin_neon_vcvt_f32_f16((int8x8_t)__a, 8); }
+#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
 
-#if defined(__aarch64__)
-__ai float32x2_t vcvt_f32_f64(float64x2_t __a) {
-  return (float32x2_t)__builtin_neon_vcvt_f32_f64((int8x16_t)__a, 9); }
-
-__ai float64x1_t vcvt_f64_s64(int64x1_t __a) {
-  return (float64x1_t)__builtin_neon_vcvt_f64_v((int8x8_t)__a, 3); }
-__ai float64x1_t vcvt_f64_u64(uint64x1_t __a) {
-  return (float64x1_t)__builtin_neon_vcvt_f64_v((int8x8_t)__a, 19); }
-__ai float64x2_t vcvtq_f64_s64(int64x2_t __a) {
-  return (float64x2_t)__builtin_neon_vcvtq_f64_v((int8x16_t)__a, 35); }
-__ai float64x2_t vcvtq_f64_u64(uint64x2_t __a) {
-  return (float64x2_t)__builtin_neon_vcvtq_f64_v((int8x16_t)__a, 51); }
-
-__ai float64x2_t vcvt_f64_f32(float32x2_t __a) {
-  return (float64x2_t)__builtin_neon_vcvt_f64_f32((int8x8_t)__a, 42); }
-
-__ai float16x8_t vcvt_high_f16_f32(float16x4_t __a, float32x4_t __b) {
-  float16x4_t __a1 = vcvt_f16_f32(__b);
-  return __builtin_shufflevector(__a, __a1, 0, 1, 2, 3, 4, 5, 6, 7); }
-
-__ai float32x4_t vcvt_high_f32_f16(float16x8_t __a) {
-  float16x4_t __a1 = vget_high_f16(__a);
-  return vcvt_f32_f16(__a1); }
-
-__ai float32x4_t vcvt_high_f32_f64(float32x2_t __a, float64x2_t __b) {
-  float32x2_t __a1 = vcvt_f32_f64(__b);
-  return __builtin_shufflevector(__a, __a1, 0, 1, 2, 3); }
-
-__ai float64x2_t vcvt_high_f64_f32(float32x4_t __a) {
-  float32x2_t __a1 = vget_high_f32(__a);
-  return vcvt_f64_f32(__a1); }
-
-#endif
-
-#define vcvt_n_f32_s32(a, __b) __extension__ ({ \
-  int32x2_t __a = (a); \
-  (float32x2_t)__builtin_neon_vcvt_n_f32_v((int8x8_t)__a, __b, 2); })
-#define vcvt_n_f32_u32(a, __b) __extension__ ({ \
-  uint32x2_t __a = (a); \
-  (float32x2_t)__builtin_neon_vcvt_n_f32_v((int8x8_t)__a, __b, 18); })
-#define vcvtq_n_f32_s32(a, __b) __extension__ ({ \
-  int32x4_t __a = (a); \
-  (float32x4_t)__builtin_neon_vcvtq_n_f32_v((int8x16_t)__a, __b, 34); })
-#define vcvtq_n_f32_u32(a, __b) __extension__ ({ \
-  uint32x4_t __a = (a); \
-  (float32x4_t)__builtin_neon_vcvtq_n_f32_v((int8x16_t)__a, __b, 50); })
-
-#define vcvt_n_s32_f32(a, __b) __extension__ ({ \
-  float32x2_t __a = (a); \
-  (int32x2_t)__builtin_neon_vcvt_n_s32_v((int8x8_t)__a, __b, 2); })
-#define vcvtq_n_s32_f32(a, __b) __extension__ ({ \
-  float32x4_t __a = (a); \
-  (int32x4_t)__builtin_neon_vcvtq_n_s32_v((int8x16_t)__a, __b, 34); })
-
-#define vcvt_n_u32_f32(a, __b) __extension__ ({ \
-  float32x2_t __a = (a); \
-  (uint32x2_t)__builtin_neon_vcvt_n_u32_v((int8x8_t)__a, __b, 18); })
-#define vcvtq_n_u32_f32(a, __b) __extension__ ({ \
-  float32x4_t __a = (a); \
-  (uint32x4_t)__builtin_neon_vcvtq_n_u32_v((int8x16_t)__a, __b, 50); })
-
-__ai int32x2_t vcvt_s32_f32(float32x2_t __a) {
-  return (int32x2_t)__builtin_neon_vcvt_s32_v((int8x8_t)__a, 2); }
-__ai int32x4_t vcvtq_s32_f32(float32x4_t __a) {
-  return (int32x4_t)__builtin_neon_vcvtq_s32_v((int8x16_t)__a, 34); }
+#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
 
-#if defined(__aarch64__)
-__ai int64x1_t vcvt_s64_f64(float64x1_t __a) {
-  return (int64x1_t)__builtin_neon_vcvt_s64_v((int8x8_t)__a, 3); }
-__ai int64x2_t vcvtq_s64_f64(float64x2_t __a) {
-  return (int64x2_t)__builtin_neon_vcvtq_s64_v((int8x16_t)__a, 35); }
+#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
 
-__ai uint32x2_t vcvt_u32_f32(float32x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vcvt_u32_v((int8x8_t)__a, 18); }
-__ai uint32x4_t vcvtq_u32_f32(float32x4_t __a) {
-  return (uint32x4_t)__builtin_neon_vcvtq_u32_v((int8x16_t)__a, 50); }
+#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
 
-#if defined(__aarch64__)
-__ai uint64x1_t vcvt_u64_f64(float64x1_t __a) {
-  return (uint64x1_t)__builtin_neon_vcvt_u64_v((int8x8_t)__a, 19); }
-__ai uint64x2_t vcvtq_u64_f64(float64x2_t __a) {
-  return (uint64x2_t)__builtin_neon_vcvtq_u64_v((int8x16_t)__a, 51); }
-
-#endif
-
-#define vdup_lane_u8(a, __b) __extension__ ({ \
-  uint8x8_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b, __b, __b, __b, __b); })
-#define vdup_lane_u16(a, __b) __extension__ ({ \
-  uint16x4_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b); })
-#define vdup_lane_u32(a, __b) __extension__ ({ \
-  uint32x2_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b); })
-#define vdup_lane_s8(a, __b) __extension__ ({ \
-  int8x8_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b, __b, __b, __b, __b); })
-#define vdup_lane_s16(a, __b) __extension__ ({ \
-  int16x4_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b); })
-#define vdup_lane_s32(a, __b) __extension__ ({ \
-  int32x2_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b); })
-#define vdup_lane_p8(a, __b) __extension__ ({ \
-  poly8x8_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b, __b, __b, __b, __b); })
-#define vdup_lane_p16(a, __b) __extension__ ({ \
-  poly16x4_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b); })
-#define vdup_lane_f32(a, __b) __extension__ ({ \
-  float32x2_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b); })
-#define vdupq_lane_u8(a, __b) __extension__ ({ \
-  uint8x8_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b); })
-#define vdupq_lane_u16(a, __b) __extension__ ({ \
-  uint16x4_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b, __b, __b, __b, __b); })
-#define vdupq_lane_u32(a, __b) __extension__ ({ \
-  uint32x2_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b); })
-#define vdupq_lane_s8(a, __b) __extension__ ({ \
-  int8x8_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b); })
-#define vdupq_lane_s16(a, __b) __extension__ ({ \
-  int16x4_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b, __b, __b, __b, __b); })
-#define vdupq_lane_s32(a, __b) __extension__ ({ \
-  int32x2_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b); })
-#define vdupq_lane_p8(a, __b) __extension__ ({ \
-  poly8x8_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b); })
-#define vdupq_lane_p16(a, __b) __extension__ ({ \
-  poly16x4_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b, __b, __b, __b, __b); })
-#define vdupq_lane_f32(a, __b) __extension__ ({ \
-  float32x2_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b); })
-#define vdup_lane_s64(a, __b) __extension__ ({ \
-  int64x1_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b); })
-#define vdup_lane_u64(a, __b) __extension__ ({ \
-  uint64x1_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b); })
-#define vdupq_lane_s64(a, __b) __extension__ ({ \
-  int64x1_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b); })
-#define vdupq_lane_u64(a, __b) __extension__ ({ \
-  uint64x1_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b); })
+#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
 
-#if defined(__aarch64__)
-#define vdup_lane_f16(a, __b) __extension__ ({ \
-  float16x4_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b); })
-#define vdup_lane_f64(a, __b) __extension__ ({ \
-  float64x1_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b); })
-#define vdupq_lane_f16(a, __b) __extension__ ({ \
-  float16x4_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b, __b, __b, __b, __b); })
-#define vdupq_lane_f64(a, __b) __extension__ ({ \
-  float64x1_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b); })
-#define vdup_lane_p64(a, __b) __extension__ ({ \
-  poly64x1_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b); })
-#define vdupq_lane_p64(a, __b) __extension__ ({ \
-  poly64x1_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b); })
-
-#define vdup_laneq_s8(a, __b) __extension__ ({ \
-  int8x16_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b, __b, __b, __b, __b); })
-#define vdup_laneq_s16(a, __b) __extension__ ({ \
-  int16x8_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b); })
-#define vdup_laneq_s32(a, __b) __extension__ ({ \
-  int32x4_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b); })
-#define vdup_laneq_s64(a, __b) __extension__ ({ \
-  int64x2_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b); })
-#define vdup_laneq_u8(a, __b) __extension__ ({ \
-  uint8x16_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b, __b, __b, __b, __b); })
-#define vdup_laneq_u16(a, __b) __extension__ ({ \
-  uint16x8_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b); })
-#define vdup_laneq_u32(a, __b) __extension__ ({ \
-  uint32x4_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b); })
-#define vdup_laneq_u64(a, __b) __extension__ ({ \
-  uint64x2_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b); })
-#define vdup_laneq_p8(a, __b) __extension__ ({ \
-  poly8x16_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b, __b, __b, __b, __b); })
-#define vdup_laneq_p16(a, __b) __extension__ ({ \
-  poly16x8_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b); })
-#define vdup_laneq_f16(a, __b) __extension__ ({ \
-  float16x8_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b); })
-#define vdup_laneq_f32(a, __b) __extension__ ({ \
-  float32x4_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b); })
-#define vdup_laneq_f64(a, __b) __extension__ ({ \
-  float64x2_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b); })
-#define vdupq_laneq_s8(a, __b) __extension__ ({ \
-  int8x16_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b); })
-#define vdupq_laneq_s16(a, __b) __extension__ ({ \
-  int16x8_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b, __b, __b, __b, __b); })
-#define vdupq_laneq_s32(a, __b) __extension__ ({ \
-  int32x4_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b); })
-#define vdupq_laneq_s64(a, __b) __extension__ ({ \
-  int64x2_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b); })
-#define vdupq_laneq_p8(a, __b) __extension__ ({ \
-  poly8x16_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b); })
-#define vdupq_laneq_p16(a, __b) __extension__ ({ \
-  poly16x8_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b, __b, __b, __b, __b); })
-#define vdupq_laneq_u8(a, __b) __extension__ ({ \
-  uint8x16_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b); })
-#define vdupq_laneq_u16(a, __b) __extension__ ({ \
-  uint16x8_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b, __b, __b, __b, __b); })
-#define vdupq_laneq_u32(a, __b) __extension__ ({ \
-  uint32x4_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b); })
-#define vdupq_laneq_u64(a, __b) __extension__ ({ \
-  uint64x2_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b); })
-#define vdupq_laneq_f16(a, __b) __extension__ ({ \
-  float16x8_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b, __b, __b, __b, __b); })
-#define vdupq_laneq_f32(a, __b) __extension__ ({ \
-  float32x4_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b, __b, __b); })
-#define vdupq_laneq_f64(a, __b) __extension__ ({ \
-  float64x2_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b); })
-#define vdup_laneq_p64(a, __b) __extension__ ({ \
-  poly64x2_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b); })
-#define vdupq_laneq_p64(a, __b) __extension__ ({ \
-  poly64x2_t __a = (a); \
-  __builtin_shufflevector(__a, __a, __b, __b); })
-
-#endif
-
-__ai uint8x8_t vdup_n_u8(uint8_t __a) {
-  return (uint8x8_t){ __a, __a, __a, __a, __a, __a, __a, __a }; }
-__ai uint16x4_t vdup_n_u16(uint16_t __a) {
-  return (uint16x4_t){ __a, __a, __a, __a }; }
-__ai uint32x2_t vdup_n_u32(uint32_t __a) {
-  return (uint32x2_t){ __a, __a }; }
-__ai int8x8_t vdup_n_s8(int8_t __a) {
-  return (int8x8_t){ __a, __a, __a, __a, __a, __a, __a, __a }; }
-__ai int16x4_t vdup_n_s16(int16_t __a) {
-  return (int16x4_t){ __a, __a, __a, __a }; }
-__ai int32x2_t vdup_n_s32(int32_t __a) {
-  return (int32x2_t){ __a, __a }; }
-__ai poly8x8_t vdup_n_p8(poly8_t __a) {
-  return (poly8x8_t){ __a, __a, __a, __a, __a, __a, __a, __a }; }
-__ai poly16x4_t vdup_n_p16(poly16_t __a) {
-  return (poly16x4_t){ __a, __a, __a, __a }; }
-#define vdup_n_f16(a) __extension__ ({ \
-  float16_t __a = (a); \
-  (float16x4_t){ __a, __a, __a, __a }; })
-__ai float32x2_t vdup_n_f32(float32_t __a) {
-  return (float32x2_t){ __a, __a }; }
-__ai uint8x16_t vdupq_n_u8(uint8_t __a) {
-  return (uint8x16_t){ __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a }; }
-__ai uint16x8_t vdupq_n_u16(uint16_t __a) {
-  return (uint16x8_t){ __a, __a, __a, __a, __a, __a, __a, __a }; }
-__ai uint32x4_t vdupq_n_u32(uint32_t __a) {
-  return (uint32x4_t){ __a, __a, __a, __a }; }
-__ai int8x16_t vdupq_n_s8(int8_t __a) {
-  return (int8x16_t){ __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a }; }
-__ai int16x8_t vdupq_n_s16(int16_t __a) {
-  return (int16x8_t){ __a, __a, __a, __a, __a, __a, __a, __a }; }
-__ai int32x4_t vdupq_n_s32(int32_t __a) {
-  return (int32x4_t){ __a, __a, __a, __a }; }
-__ai poly8x16_t vdupq_n_p8(poly8_t __a) {
-  return (poly8x16_t){ __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a }; }
-__ai poly16x8_t vdupq_n_p16(poly16_t __a) {
-  return (poly16x8_t){ __a, __a, __a, __a, __a, __a, __a, __a }; }
-#define vdupq_n_f16(a) __extension__ ({ \
-  float16_t __a = (a); \
-  (float16x8_t){ __a, __a, __a, __a, __a, __a, __a, __a }; })
-__ai float32x4_t vdupq_n_f32(float32_t __a) {
-  return (float32x4_t){ __a, __a, __a, __a }; }
-__ai int64x1_t vdup_n_s64(int64_t __a) {
-  return (int64x1_t){ __a }; }
-__ai uint64x1_t vdup_n_u64(uint64_t __a) {
-  return (uint64x1_t){ __a }; }
-__ai int64x2_t vdupq_n_s64(int64_t __a) {
-  return (int64x2_t){ __a, __a }; }
-__ai uint64x2_t vdupq_n_u64(uint64_t __a) {
-  return (uint64x2_t){ __a, __a }; }
-
-__ai int8x8_t veor_s8(int8x8_t __a, int8x8_t __b) {
-  return __a ^ __b; }
-__ai int16x4_t veor_s16(int16x4_t __a, int16x4_t __b) {
-  return __a ^ __b; }
-__ai int32x2_t veor_s32(int32x2_t __a, int32x2_t __b) {
-  return __a ^ __b; }
-__ai int64x1_t veor_s64(int64x1_t __a, int64x1_t __b) {
-  return __a ^ __b; }
-__ai uint8x8_t veor_u8(uint8x8_t __a, uint8x8_t __b) {
-  return __a ^ __b; }
-__ai uint16x4_t veor_u16(uint16x4_t __a, uint16x4_t __b) {
-  return __a ^ __b; }
-__ai uint32x2_t veor_u32(uint32x2_t __a, uint32x2_t __b) {
-  return __a ^ __b; }
-__ai uint64x1_t veor_u64(uint64x1_t __a, uint64x1_t __b) {
-  return __a ^ __b; }
-__ai int8x16_t veorq_s8(int8x16_t __a, int8x16_t __b) {
-  return __a ^ __b; }
-__ai int16x8_t veorq_s16(int16x8_t __a, int16x8_t __b) {
-  return __a ^ __b; }
-__ai int32x4_t veorq_s32(int32x4_t __a, int32x4_t __b) {
-  return __a ^ __b; }
-__ai int64x2_t veorq_s64(int64x2_t __a, int64x2_t __b) {
-  return __a ^ __b; }
-__ai uint8x16_t veorq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return __a ^ __b; }
-__ai uint16x8_t veorq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return __a ^ __b; }
-__ai uint32x4_t veorq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return __a ^ __b; }
-__ai uint64x2_t veorq_u64(uint64x2_t __a, uint64x2_t __b) {
-  return __a ^ __b; }
-
-#define vext_s8(a, b, __c) __extension__ ({ \
-  int8x8_t __a = (a); int8x8_t __b = (b); \
-  (int8x8_t)__builtin_neon_vext_v(__a, __b, __c, 0); })
-#define vext_u8(a, b, __c) __extension__ ({ \
-  uint8x8_t __a = (a); uint8x8_t __b = (b); \
-  (uint8x8_t)__builtin_neon_vext_v((int8x8_t)__a, (int8x8_t)__b, __c, 16); })
-#define vext_p8(a, b, __c) __extension__ ({ \
-  poly8x8_t __a = (a); poly8x8_t __b = (b); \
-  (poly8x8_t)__builtin_neon_vext_v((int8x8_t)__a, (int8x8_t)__b, __c, 4); })
-#define vext_s16(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int16x4_t __b = (b); \
-  (int16x4_t)__builtin_neon_vext_v((int8x8_t)__a, (int8x8_t)__b, __c, 1); })
-#define vext_u16(a, b, __c) __extension__ ({ \
-  uint16x4_t __a = (a); uint16x4_t __b = (b); \
-  (uint16x4_t)__builtin_neon_vext_v((int8x8_t)__a, (int8x8_t)__b, __c, 17); })
-#define vext_p16(a, b, __c) __extension__ ({ \
-  poly16x4_t __a = (a); poly16x4_t __b = (b); \
-  (poly16x4_t)__builtin_neon_vext_v((int8x8_t)__a, (int8x8_t)__b, __c, 5); })
-#define vext_s32(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int32x2_t __b = (b); \
-  (int32x2_t)__builtin_neon_vext_v((int8x8_t)__a, (int8x8_t)__b, __c, 2); })
-#define vext_u32(a, b, __c) __extension__ ({ \
-  uint32x2_t __a = (a); uint32x2_t __b = (b); \
-  (uint32x2_t)__builtin_neon_vext_v((int8x8_t)__a, (int8x8_t)__b, __c, 18); })
-#define vext_s64(a, b, __c) __extension__ ({ \
-  int64x1_t __a = (a); int64x1_t __b = (b); \
-  (int64x1_t)__builtin_neon_vext_v((int8x8_t)__a, (int8x8_t)__b, __c, 3); })
-#define vext_u64(a, b, __c) __extension__ ({ \
-  uint64x1_t __a = (a); uint64x1_t __b = (b); \
-  (uint64x1_t)__builtin_neon_vext_v((int8x8_t)__a, (int8x8_t)__b, __c, 19); })
-#define vext_f32(a, b, __c) __extension__ ({ \
-  float32x2_t __a = (a); float32x2_t __b = (b); \
-  (float32x2_t)__builtin_neon_vext_v((int8x8_t)__a, (int8x8_t)__b, __c, 9); })
-#define vextq_s8(a, b, __c) __extension__ ({ \
-  int8x16_t __a = (a); int8x16_t __b = (b); \
-  (int8x16_t)__builtin_neon_vextq_v(__a, __b, __c, 32); })
-#define vextq_u8(a, b, __c) __extension__ ({ \
-  uint8x16_t __a = (a); uint8x16_t __b = (b); \
-  (uint8x16_t)__builtin_neon_vextq_v((int8x16_t)__a, (int8x16_t)__b, __c, 48); })
-#define vextq_p8(a, b, __c) __extension__ ({ \
-  poly8x16_t __a = (a); poly8x16_t __b = (b); \
-  (poly8x16_t)__builtin_neon_vextq_v((int8x16_t)__a, (int8x16_t)__b, __c, 36); })
-#define vextq_s16(a, b, __c) __extension__ ({ \
-  int16x8_t __a = (a); int16x8_t __b = (b); \
-  (int16x8_t)__builtin_neon_vextq_v((int8x16_t)__a, (int8x16_t)__b, __c, 33); })
-#define vextq_u16(a, b, __c) __extension__ ({ \
-  uint16x8_t __a = (a); uint16x8_t __b = (b); \
-  (uint16x8_t)__builtin_neon_vextq_v((int8x16_t)__a, (int8x16_t)__b, __c, 49); })
-#define vextq_p16(a, b, __c) __extension__ ({ \
-  poly16x8_t __a = (a); poly16x8_t __b = (b); \
-  (poly16x8_t)__builtin_neon_vextq_v((int8x16_t)__a, (int8x16_t)__b, __c, 37); })
-#define vextq_s32(a, b, __c) __extension__ ({ \
-  int32x4_t __a = (a); int32x4_t __b = (b); \
-  (int32x4_t)__builtin_neon_vextq_v((int8x16_t)__a, (int8x16_t)__b, __c, 34); })
-#define vextq_u32(a, b, __c) __extension__ ({ \
-  uint32x4_t __a = (a); uint32x4_t __b = (b); \
-  (uint32x4_t)__builtin_neon_vextq_v((int8x16_t)__a, (int8x16_t)__b, __c, 50); })
-#define vextq_s64(a, b, __c) __extension__ ({ \
-  int64x2_t __a = (a); int64x2_t __b = (b); \
-  (int64x2_t)__builtin_neon_vextq_v((int8x16_t)__a, (int8x16_t)__b, __c, 35); })
-#define vextq_u64(a, b, __c) __extension__ ({ \
-  uint64x2_t __a = (a); uint64x2_t __b = (b); \
-  (uint64x2_t)__builtin_neon_vextq_v((int8x16_t)__a, (int8x16_t)__b, __c, 51); })
-#define vextq_f32(a, b, __c) __extension__ ({ \
-  float32x4_t __a = (a); float32x4_t __b = (b); \
-  (float32x4_t)__builtin_neon_vextq_v((int8x16_t)__a, (int8x16_t)__b, __c, 41); })
+#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
 
-#if defined(__aarch64__)
-#define vext_f64(a, b, __c) __extension__ ({ \
-  float64x1_t __a = (a); float64x1_t __b = (b); \
-  (float64x1_t)__builtin_neon_vext_v((int8x8_t)__a, (int8x8_t)__b, __c, 10); })
-#define vextq_f64(a, b, __c) __extension__ ({ \
-  float64x2_t __a = (a); float64x2_t __b = (b); \
-  (float64x2_t)__builtin_neon_vextq_v((int8x16_t)__a, (int8x16_t)__b, __c, 42); })
-#define vext_p64(a, b, __c) __extension__ ({ \
-  poly64x1_t __a = (a); poly64x1_t __b = (b); \
-  (poly64x1_t)__builtin_neon_vext_v((int8x8_t)__a, (int8x8_t)__b, __c, 6); })
-#define vextq_p64(a, b, __c) __extension__ ({ \
-  poly64x2_t __a = (a); poly64x2_t __b = (b); \
-  (poly64x2_t)__builtin_neon_vextq_v((int8x16_t)__a, (int8x16_t)__b, __c, 38); })
-
-#define vfma_lane_f32(a, b, c, __d) __extension__ ({ \
-  float32x2_t __a = (a); float32x2_t __b = (b); float32x2_t __c = (c); \
-  (float32x2_t)__builtin_neon_vfma_lane_v((int8x8_t)__a, (int8x8_t)__b, (int8x8_t)__c, __d, 9); })
-#define vfma_lane_f64(a, b, c, __d) __extension__ ({ \
-  float64x1_t __a = (a); float64x1_t __b = (b); float64x1_t __c = (c); \
-  (float64x1_t)__builtin_neon_vfma_lane_v((int8x8_t)__a, (int8x8_t)__b, (int8x8_t)__c, __d, 10); })
-#define vfmaq_lane_f32(a, b, c, __d) __extension__ ({ \
-  float32x4_t __a = (a); float32x4_t __b = (b); float32x2_t __c = (c); \
-  (float32x4_t)__builtin_neon_vfmaq_lane_v((int8x16_t)__a, (int8x16_t)__b, (int8x8_t)__c, __d, 41); })
-#define vfmaq_lane_f64(a, b, c, __d) __extension__ ({ \
-  float64x2_t __a = (a); float64x2_t __b = (b); float64x1_t __c = (c); \
-  (float64x2_t)__builtin_neon_vfmaq_lane_v((int8x16_t)__a, (int8x16_t)__b, (int8x8_t)__c, __d, 42); })
-
-#define vfma_laneq_f32(a, b, c, __d) __extension__ ({ \
-  float32x2_t __a = (a); float32x2_t __b = (b); float32x4_t __c = (c); \
-  (float32x2_t)__builtin_neon_vfma_laneq_v((int8x8_t)__a, (int8x8_t)__b, (int8x16_t)__c, __d, 9); })
-#define vfma_laneq_f64(a, b, c, __d) __extension__ ({ \
-  float64x1_t __a = (a); float64x1_t __b = (b); float64x2_t __c = (c); \
-  (float64x1_t)__builtin_neon_vfma_laneq_v((int8x8_t)__a, (int8x8_t)__b, (int8x16_t)__c, __d, 10); })
-#define vfmaq_laneq_f32(a, b, c, __d) __extension__ ({ \
-  float32x4_t __a = (a); float32x4_t __b = (b); float32x4_t __c = (c); \
-  (float32x4_t)__builtin_neon_vfmaq_laneq_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, __d, 41); })
-#define vfmaq_laneq_f64(a, b, c, __d) __extension__ ({ \
-  float64x2_t __a = (a); float64x2_t __b = (b); float64x2_t __c = (c); \
-  (float64x2_t)__builtin_neon_vfmaq_laneq_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, __d, 42); })
-
-#define vfms_lane_f32(a, b, c, __d) __extension__ ({ \
-  float32x2_t __a = (a); float32x2_t __b = (b); float32x2_t __c = (c); \
-  float32x2_t __a1 = __a; \
-  float32x2_t __b1 = __b; \
-  float32x2_t __c1 = __c; \
-  vfma_lane_f32(__a1, __b1, -__c1, __d); })
-#define vfms_lane_f64(a, b, c, __d) __extension__ ({ \
-  float64x1_t __a = (a); float64x1_t __b = (b); float64x1_t __c = (c); \
-  float64x1_t __a1 = __a; \
-  float64x1_t __b1 = __b; \
-  float64x1_t __c1 = __c; \
-  vfma_lane_f64(__a1, __b1, -__c1, __d); })
-#define vfmsq_lane_f32(a, b, c, __d) __extension__ ({ \
-  float32x4_t __a = (a); float32x4_t __b = (b); float32x2_t __c = (c); \
-  float32x4_t __a1 = __a; \
-  float32x4_t __b1 = __b; \
-  float32x2_t __c1 = __c; \
-  vfmaq_lane_f32(__a1, __b1, -__c1, __d); })
-#define vfmsq_lane_f64(a, b, c, __d) __extension__ ({ \
-  float64x2_t __a = (a); float64x2_t __b = (b); float64x1_t __c = (c); \
-  float64x2_t __a1 = __a; \
-  float64x2_t __b1 = __b; \
-  float64x1_t __c1 = __c; \
-  vfmaq_lane_f64(__a1, __b1, -__c1, __d); })
-
-#define vfms_laneq_f32(a, b, c, __d) __extension__ ({ \
-  float32x2_t __a = (a); float32x2_t __b = (b); float32x4_t __c = (c); \
-  float32x2_t __a1 = __a; \
-  float32x2_t __b1 = __b; \
-  float32x4_t __c1 = __c; \
-  vfma_laneq_f32(__a1, __b1, -__c1, __d); })
-#define vfms_laneq_f64(a, b, c, __d) __extension__ ({ \
-  float64x1_t __a = (a); float64x1_t __b = (b); float64x2_t __c = (c); \
-  float64x1_t __a1 = __a; \
-  float64x1_t __b1 = __b; \
-  float64x2_t __c1 = __c; \
-  vfma_laneq_f64(__a1, __b1, -__c1, __d); })
-#define vfmsq_laneq_f32(a, b, c, __d) __extension__ ({ \
-  float32x4_t __a = (a); float32x4_t __b = (b); float32x4_t __c = (c); \
-  float32x4_t __a1 = __a; \
-  float32x4_t __b1 = __b; \
-  float32x4_t __c1 = __c; \
-  vfmaq_laneq_f32(__a1, __b1, -__c1, __d); })
-#define vfmsq_laneq_f64(a, b, c, __d) __extension__ ({ \
-  float64x2_t __a = (a); float64x2_t __b = (b); float64x2_t __c = (c); \
-  float64x2_t __a1 = __a; \
-  float64x2_t __b1 = __b; \
-  float64x2_t __c1 = __c; \
-  vfmaq_laneq_f64(__a1, __b1, -__c1, __d); })
-
-__ai float64x1_t vget_high_f64(float64x2_t __a) {
-  return __builtin_shufflevector(__a, __a, 1); }
-__ai poly64x1_t vget_high_p64(poly64x2_t __a) {
-  return __builtin_shufflevector(__a, __a, 1); }
-
-#endif
-
-#define vget_lane_u8(a, __b) __extension__ ({ \
-  uint8x8_t __a = (a); \
-  (uint8_t)__builtin_neon_vget_lane_i8((int8x8_t)__a, __b); })
-#define vget_lane_u16(a, __b) __extension__ ({ \
-  uint16x4_t __a = (a); \
-  (uint16_t)__builtin_neon_vget_lane_i16((int16x4_t)__a, __b); })
-#define vget_lane_u32(a, __b) __extension__ ({ \
-  uint32x2_t __a = (a); \
-  (uint32_t)__builtin_neon_vget_lane_i32((int32x2_t)__a, __b); })
-#define vget_lane_s8(a, __b) __extension__ ({ \
-  int8x8_t __a = (a); \
-  (int8_t)__builtin_neon_vget_lane_i8(__a, __b); })
-#define vget_lane_s16(a, __b) __extension__ ({ \
-  int16x4_t __a = (a); \
-  (int16_t)__builtin_neon_vget_lane_i16(__a, __b); })
-#define vget_lane_s32(a, __b) __extension__ ({ \
-  int32x2_t __a = (a); \
-  (int32_t)__builtin_neon_vget_lane_i32(__a, __b); })
-#define vget_lane_p8(a, __b) __extension__ ({ \
-  poly8x8_t __a = (a); \
-  (poly8_t)__builtin_neon_vget_lane_i8((int8x8_t)__a, __b); })
-#define vget_lane_p16(a, __b) __extension__ ({ \
-  poly16x4_t __a = (a); \
-  (poly16_t)__builtin_neon_vget_lane_i16((int16x4_t)__a, __b); })
-#define vget_lane_f32(a, __b) __extension__ ({ \
-  float32x2_t __a = (a); \
-  (float32_t)__builtin_neon_vget_lane_f32(__a, __b); })
-#define vgetq_lane_u8(a, __b) __extension__ ({ \
-  uint8x16_t __a = (a); \
-  (uint8_t)__builtin_neon_vgetq_lane_i8((int8x16_t)__a, __b); })
-#define vgetq_lane_u16(a, __b) __extension__ ({ \
-  uint16x8_t __a = (a); \
-  (uint16_t)__builtin_neon_vgetq_lane_i16((int16x8_t)__a, __b); })
-#define vgetq_lane_u32(a, __b) __extension__ ({ \
-  uint32x4_t __a = (a); \
-  (uint32_t)__builtin_neon_vgetq_lane_i32((int32x4_t)__a, __b); })
-#define vgetq_lane_s8(a, __b) __extension__ ({ \
-  int8x16_t __a = (a); \
-  (int8_t)__builtin_neon_vgetq_lane_i8(__a, __b); })
-#define vgetq_lane_s16(a, __b) __extension__ ({ \
-  int16x8_t __a = (a); \
-  (int16_t)__builtin_neon_vgetq_lane_i16(__a, __b); })
-#define vgetq_lane_s32(a, __b) __extension__ ({ \
-  int32x4_t __a = (a); \
-  (int32_t)__builtin_neon_vgetq_lane_i32(__a, __b); })
-#define vgetq_lane_p8(a, __b) __extension__ ({ \
-  poly8x16_t __a = (a); \
-  (poly8_t)__builtin_neon_vgetq_lane_i8((int8x16_t)__a, __b); })
-#define vgetq_lane_p16(a, __b) __extension__ ({ \
-  poly16x8_t __a = (a); \
-  (poly16_t)__builtin_neon_vgetq_lane_i16((int16x8_t)__a, __b); })
-#define vgetq_lane_f32(a, __b) __extension__ ({ \
-  float32x4_t __a = (a); \
-  (float32_t)__builtin_neon_vgetq_lane_f32(__a, __b); })
-#define vget_lane_s64(a, __b) __extension__ ({ \
-  int64x1_t __a = (a); \
-  (int64_t)__builtin_neon_vget_lane_i64(__a, __b); })
-#define vget_lane_u64(a, __b) __extension__ ({ \
-  uint64x1_t __a = (a); \
-  (uint64_t)__builtin_neon_vget_lane_i64((int64x1_t)__a, __b); })
-#define vgetq_lane_s64(a, __b) __extension__ ({ \
-  int64x2_t __a = (a); \
-  (int64_t)__builtin_neon_vgetq_lane_i64(__a, __b); })
-#define vgetq_lane_u64(a, __b) __extension__ ({ \
-  uint64x2_t __a = (a); \
-  (uint64_t)__builtin_neon_vgetq_lane_i64((int64x2_t)__a, __b); })
-
-__ai int8x8_t vget_low_s8(int8x16_t __a) {
-  return __builtin_shufflevector(__a, __a, 0, 1, 2, 3, 4, 5, 6, 7); }
-__ai int16x4_t vget_low_s16(int16x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 0, 1, 2, 3); }
-__ai int32x2_t vget_low_s32(int32x4_t __a) {
-  return __builtin_shufflevector(__a, __a, 0, 1); }
-__ai int64x1_t vget_low_s64(int64x2_t __a) {
-  return __builtin_shufflevector(__a, __a, 0); }
-__ai float16x4_t vget_low_f16(float16x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 0, 1, 2, 3); }
-__ai float32x2_t vget_low_f32(float32x4_t __a) {
-  return __builtin_shufflevector(__a, __a, 0, 1); }
-__ai uint8x8_t vget_low_u8(uint8x16_t __a) {
-  return __builtin_shufflevector(__a, __a, 0, 1, 2, 3, 4, 5, 6, 7); }
-__ai uint16x4_t vget_low_u16(uint16x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 0, 1, 2, 3); }
-__ai uint32x2_t vget_low_u32(uint32x4_t __a) {
-  return __builtin_shufflevector(__a, __a, 0, 1); }
-__ai uint64x1_t vget_low_u64(uint64x2_t __a) {
-  return __builtin_shufflevector(__a, __a, 0); }
-__ai poly8x8_t vget_low_p8(poly8x16_t __a) {
-  return __builtin_shufflevector(__a, __a, 0, 1, 2, 3, 4, 5, 6, 7); }
-__ai poly16x4_t vget_low_p16(poly16x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 0, 1, 2, 3); }
+#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
 
-#if defined(__aarch64__)
-__ai float64x1_t vget_low_f64(float64x2_t __a) {
-  return __builtin_shufflevector(__a, __a, 0); }
-__ai poly64x1_t vget_low_p64(poly64x2_t __a) {
-  return __builtin_shufflevector(__a, __a, 0); }
-
-#endif
-
-__ai int8x8_t vhadd_s8(int8x8_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vhadd_v(__a, __b, 0); }
-__ai int16x4_t vhadd_s16(int16x4_t __a, int16x4_t __b) {
-  return (int16x4_t)__builtin_neon_vhadd_v((int8x8_t)__a, (int8x8_t)__b, 1); }
-__ai int32x2_t vhadd_s32(int32x2_t __a, int32x2_t __b) {
-  return (int32x2_t)__builtin_neon_vhadd_v((int8x8_t)__a, (int8x8_t)__b, 2); }
-__ai uint8x8_t vhadd_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vhadd_v((int8x8_t)__a, (int8x8_t)__b, 16); }
-__ai uint16x4_t vhadd_u16(uint16x4_t __a, uint16x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vhadd_v((int8x8_t)__a, (int8x8_t)__b, 17); }
-__ai uint32x2_t vhadd_u32(uint32x2_t __a, uint32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vhadd_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai int8x16_t vhaddq_s8(int8x16_t __a, int8x16_t __b) {
-  return (int8x16_t)__builtin_neon_vhaddq_v(__a, __b, 32); }
-__ai int16x8_t vhaddq_s16(int16x8_t __a, int16x8_t __b) {
-  return (int16x8_t)__builtin_neon_vhaddq_v((int8x16_t)__a, (int8x16_t)__b, 33); }
-__ai int32x4_t vhaddq_s32(int32x4_t __a, int32x4_t __b) {
-  return (int32x4_t)__builtin_neon_vhaddq_v((int8x16_t)__a, (int8x16_t)__b, 34); }
-__ai uint8x16_t vhaddq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vhaddq_v((int8x16_t)__a, (int8x16_t)__b, 48); }
-__ai uint16x8_t vhaddq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint16x8_t)__builtin_neon_vhaddq_v((int8x16_t)__a, (int8x16_t)__b, 49); }
-__ai uint32x4_t vhaddq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vhaddq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-
-__ai int8x8_t vhsub_s8(int8x8_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vhsub_v(__a, __b, 0); }
-__ai int16x4_t vhsub_s16(int16x4_t __a, int16x4_t __b) {
-  return (int16x4_t)__builtin_neon_vhsub_v((int8x8_t)__a, (int8x8_t)__b, 1); }
-__ai int32x2_t vhsub_s32(int32x2_t __a, int32x2_t __b) {
-  return (int32x2_t)__builtin_neon_vhsub_v((int8x8_t)__a, (int8x8_t)__b, 2); }
-__ai uint8x8_t vhsub_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vhsub_v((int8x8_t)__a, (int8x8_t)__b, 16); }
-__ai uint16x4_t vhsub_u16(uint16x4_t __a, uint16x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vhsub_v((int8x8_t)__a, (int8x8_t)__b, 17); }
-__ai uint32x2_t vhsub_u32(uint32x2_t __a, uint32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vhsub_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai int8x16_t vhsubq_s8(int8x16_t __a, int8x16_t __b) {
-  return (int8x16_t)__builtin_neon_vhsubq_v(__a, __b, 32); }
-__ai int16x8_t vhsubq_s16(int16x8_t __a, int16x8_t __b) {
-  return (int16x8_t)__builtin_neon_vhsubq_v((int8x16_t)__a, (int8x16_t)__b, 33); }
-__ai int32x4_t vhsubq_s32(int32x4_t __a, int32x4_t __b) {
-  return (int32x4_t)__builtin_neon_vhsubq_v((int8x16_t)__a, (int8x16_t)__b, 34); }
-__ai uint8x16_t vhsubq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vhsubq_v((int8x16_t)__a, (int8x16_t)__b, 48); }
-__ai uint16x8_t vhsubq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint16x8_t)__builtin_neon_vhsubq_v((int8x16_t)__a, (int8x16_t)__b, 49); }
-__ai uint32x4_t vhsubq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vhsubq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-
-#define vld1q_u8(__a) __extension__ ({ \
-  (uint8x16_t)__builtin_neon_vld1q_v(__a, 48); })
-#define vld1q_u16(__a) __extension__ ({ \
-  (uint16x8_t)__builtin_neon_vld1q_v(__a, 49); })
-#define vld1q_u32(__a) __extension__ ({ \
-  (uint32x4_t)__builtin_neon_vld1q_v(__a, 50); })
-#define vld1q_u64(__a) __extension__ ({ \
-  (uint64x2_t)__builtin_neon_vld1q_v(__a, 51); })
-#define vld1q_s8(__a) __extension__ ({ \
-  (int8x16_t)__builtin_neon_vld1q_v(__a, 32); })
-#define vld1q_s16(__a) __extension__ ({ \
-  (int16x8_t)__builtin_neon_vld1q_v(__a, 33); })
-#define vld1q_s32(__a) __extension__ ({ \
-  (int32x4_t)__builtin_neon_vld1q_v(__a, 34); })
-#define vld1q_s64(__a) __extension__ ({ \
-  (int64x2_t)__builtin_neon_vld1q_v(__a, 35); })
-#define vld1q_f16(__a) __extension__ ({ \
-  (float16x8_t)__builtin_neon_vld1q_v(__a, 40); })
-#define vld1q_f32(__a) __extension__ ({ \
-  (float32x4_t)__builtin_neon_vld1q_v(__a, 41); })
-#define vld1q_p8(__a) __extension__ ({ \
-  (poly8x16_t)__builtin_neon_vld1q_v(__a, 36); })
-#define vld1q_p16(__a) __extension__ ({ \
-  (poly16x8_t)__builtin_neon_vld1q_v(__a, 37); })
-#define vld1_u8(__a) __extension__ ({ \
-  (uint8x8_t)__builtin_neon_vld1_v(__a, 16); })
-#define vld1_u16(__a) __extension__ ({ \
-  (uint16x4_t)__builtin_neon_vld1_v(__a, 17); })
-#define vld1_u32(__a) __extension__ ({ \
-  (uint32x2_t)__builtin_neon_vld1_v(__a, 18); })
-#define vld1_u64(__a) __extension__ ({ \
-  (uint64x1_t)__builtin_neon_vld1_v(__a, 19); })
-#define vld1_s8(__a) __extension__ ({ \
-  (int8x8_t)__builtin_neon_vld1_v(__a, 0); })
-#define vld1_s16(__a) __extension__ ({ \
-  (int16x4_t)__builtin_neon_vld1_v(__a, 1); })
-#define vld1_s32(__a) __extension__ ({ \
-  (int32x2_t)__builtin_neon_vld1_v(__a, 2); })
-#define vld1_s64(__a) __extension__ ({ \
-  (int64x1_t)__builtin_neon_vld1_v(__a, 3); })
-#define vld1_f16(__a) __extension__ ({ \
-  (float16x4_t)__builtin_neon_vld1_v(__a, 8); })
-#define vld1_f32(__a) __extension__ ({ \
-  (float32x2_t)__builtin_neon_vld1_v(__a, 9); })
-#define vld1_p8(__a) __extension__ ({ \
-  (poly8x8_t)__builtin_neon_vld1_v(__a, 4); })
-#define vld1_p16(__a) __extension__ ({ \
-  (poly16x4_t)__builtin_neon_vld1_v(__a, 5); })
-
-#define vld1q_dup_u8(__a) __extension__ ({ \
-  (uint8x16_t)__builtin_neon_vld1q_dup_v(__a, 48); })
-#define vld1q_dup_u16(__a) __extension__ ({ \
-  (uint16x8_t)__builtin_neon_vld1q_dup_v(__a, 49); })
-#define vld1q_dup_u32(__a) __extension__ ({ \
-  (uint32x4_t)__builtin_neon_vld1q_dup_v(__a, 50); })
-#define vld1q_dup_u64(__a) __extension__ ({ \
-  (uint64x2_t)__builtin_neon_vld1q_dup_v(__a, 51); })
-#define vld1q_dup_s8(__a) __extension__ ({ \
-  (int8x16_t)__builtin_neon_vld1q_dup_v(__a, 32); })
-#define vld1q_dup_s16(__a) __extension__ ({ \
-  (int16x8_t)__builtin_neon_vld1q_dup_v(__a, 33); })
-#define vld1q_dup_s32(__a) __extension__ ({ \
-  (int32x4_t)__builtin_neon_vld1q_dup_v(__a, 34); })
-#define vld1q_dup_s64(__a) __extension__ ({ \
-  (int64x2_t)__builtin_neon_vld1q_dup_v(__a, 35); })
-#define vld1q_dup_f16(__a) __extension__ ({ \
-  (float16x8_t)__builtin_neon_vld1q_dup_v(__a, 40); })
-#define vld1q_dup_f32(__a) __extension__ ({ \
-  (float32x4_t)__builtin_neon_vld1q_dup_v(__a, 41); })
-#define vld1q_dup_p8(__a) __extension__ ({ \
-  (poly8x16_t)__builtin_neon_vld1q_dup_v(__a, 36); })
-#define vld1q_dup_p16(__a) __extension__ ({ \
-  (poly16x8_t)__builtin_neon_vld1q_dup_v(__a, 37); })
-#define vld1_dup_u8(__a) __extension__ ({ \
-  (uint8x8_t)__builtin_neon_vld1_dup_v(__a, 16); })
-#define vld1_dup_u16(__a) __extension__ ({ \
-  (uint16x4_t)__builtin_neon_vld1_dup_v(__a, 17); })
-#define vld1_dup_u32(__a) __extension__ ({ \
-  (uint32x2_t)__builtin_neon_vld1_dup_v(__a, 18); })
-#define vld1_dup_u64(__a) __extension__ ({ \
-  (uint64x1_t)__builtin_neon_vld1_dup_v(__a, 19); })
-#define vld1_dup_s8(__a) __extension__ ({ \
-  (int8x8_t)__builtin_neon_vld1_dup_v(__a, 0); })
-#define vld1_dup_s16(__a) __extension__ ({ \
-  (int16x4_t)__builtin_neon_vld1_dup_v(__a, 1); })
-#define vld1_dup_s32(__a) __extension__ ({ \
-  (int32x2_t)__builtin_neon_vld1_dup_v(__a, 2); })
-#define vld1_dup_s64(__a) __extension__ ({ \
-  (int64x1_t)__builtin_neon_vld1_dup_v(__a, 3); })
-#define vld1_dup_f16(__a) __extension__ ({ \
-  (float16x4_t)__builtin_neon_vld1_dup_v(__a, 8); })
-#define vld1_dup_f32(__a) __extension__ ({ \
-  (float32x2_t)__builtin_neon_vld1_dup_v(__a, 9); })
-#define vld1_dup_p8(__a) __extension__ ({ \
-  (poly8x8_t)__builtin_neon_vld1_dup_v(__a, 4); })
-#define vld1_dup_p16(__a) __extension__ ({ \
-  (poly16x4_t)__builtin_neon_vld1_dup_v(__a, 5); })
-
-#define vld1q_lane_u8(__a, b, __c) __extension__ ({ \
-  uint8x16_t __b = (b); \
-  (uint8x16_t)__builtin_neon_vld1q_lane_v(__a, (int8x16_t)__b, __c, 48); })
-#define vld1q_lane_u16(__a, b, __c) __extension__ ({ \
-  uint16x8_t __b = (b); \
-  (uint16x8_t)__builtin_neon_vld1q_lane_v(__a, (int8x16_t)__b, __c, 49); })
-#define vld1q_lane_u32(__a, b, __c) __extension__ ({ \
-  uint32x4_t __b = (b); \
-  (uint32x4_t)__builtin_neon_vld1q_lane_v(__a, (int8x16_t)__b, __c, 50); })
-#define vld1q_lane_u64(__a, b, __c) __extension__ ({ \
-  uint64x2_t __b = (b); \
-  (uint64x2_t)__builtin_neon_vld1q_lane_v(__a, (int8x16_t)__b, __c, 51); })
-#define vld1q_lane_s8(__a, b, __c) __extension__ ({ \
-  int8x16_t __b = (b); \
-  (int8x16_t)__builtin_neon_vld1q_lane_v(__a, __b, __c, 32); })
-#define vld1q_lane_s16(__a, b, __c) __extension__ ({ \
-  int16x8_t __b = (b); \
-  (int16x8_t)__builtin_neon_vld1q_lane_v(__a, (int8x16_t)__b, __c, 33); })
-#define vld1q_lane_s32(__a, b, __c) __extension__ ({ \
-  int32x4_t __b = (b); \
-  (int32x4_t)__builtin_neon_vld1q_lane_v(__a, (int8x16_t)__b, __c, 34); })
-#define vld1q_lane_s64(__a, b, __c) __extension__ ({ \
-  int64x2_t __b = (b); \
-  (int64x2_t)__builtin_neon_vld1q_lane_v(__a, (int8x16_t)__b, __c, 35); })
-#define vld1q_lane_f16(__a, b, __c) __extension__ ({ \
-  float16x8_t __b = (b); \
-  (float16x8_t)__builtin_neon_vld1q_lane_v(__a, (int8x16_t)__b, __c, 40); })
-#define vld1q_lane_f32(__a, b, __c) __extension__ ({ \
-  float32x4_t __b = (b); \
-  (float32x4_t)__builtin_neon_vld1q_lane_v(__a, (int8x16_t)__b, __c, 41); })
-#define vld1q_lane_p8(__a, b, __c) __extension__ ({ \
-  poly8x16_t __b = (b); \
-  (poly8x16_t)__builtin_neon_vld1q_lane_v(__a, (int8x16_t)__b, __c, 36); })
-#define vld1q_lane_p16(__a, b, __c) __extension__ ({ \
-  poly16x8_t __b = (b); \
-  (poly16x8_t)__builtin_neon_vld1q_lane_v(__a, (int8x16_t)__b, __c, 37); })
-#define vld1_lane_u8(__a, b, __c) __extension__ ({ \
-  uint8x8_t __b = (b); \
-  (uint8x8_t)__builtin_neon_vld1_lane_v(__a, (int8x8_t)__b, __c, 16); })
-#define vld1_lane_u16(__a, b, __c) __extension__ ({ \
-  uint16x4_t __b = (b); \
-  (uint16x4_t)__builtin_neon_vld1_lane_v(__a, (int8x8_t)__b, __c, 17); })
-#define vld1_lane_u32(__a, b, __c) __extension__ ({ \
-  uint32x2_t __b = (b); \
-  (uint32x2_t)__builtin_neon_vld1_lane_v(__a, (int8x8_t)__b, __c, 18); })
-#define vld1_lane_u64(__a, b, __c) __extension__ ({ \
-  uint64x1_t __b = (b); \
-  (uint64x1_t)__builtin_neon_vld1_lane_v(__a, (int8x8_t)__b, __c, 19); })
-#define vld1_lane_s8(__a, b, __c) __extension__ ({ \
-  int8x8_t __b = (b); \
-  (int8x8_t)__builtin_neon_vld1_lane_v(__a, __b, __c, 0); })
-#define vld1_lane_s16(__a, b, __c) __extension__ ({ \
-  int16x4_t __b = (b); \
-  (int16x4_t)__builtin_neon_vld1_lane_v(__a, (int8x8_t)__b, __c, 1); })
-#define vld1_lane_s32(__a, b, __c) __extension__ ({ \
-  int32x2_t __b = (b); \
-  (int32x2_t)__builtin_neon_vld1_lane_v(__a, (int8x8_t)__b, __c, 2); })
-#define vld1_lane_s64(__a, b, __c) __extension__ ({ \
-  int64x1_t __b = (b); \
-  (int64x1_t)__builtin_neon_vld1_lane_v(__a, (int8x8_t)__b, __c, 3); })
-#define vld1_lane_f16(__a, b, __c) __extension__ ({ \
-  float16x4_t __b = (b); \
-  (float16x4_t)__builtin_neon_vld1_lane_v(__a, (int8x8_t)__b, __c, 8); })
-#define vld1_lane_f32(__a, b, __c) __extension__ ({ \
-  float32x2_t __b = (b); \
-  (float32x2_t)__builtin_neon_vld1_lane_v(__a, (int8x8_t)__b, __c, 9); })
-#define vld1_lane_p8(__a, b, __c) __extension__ ({ \
-  poly8x8_t __b = (b); \
-  (poly8x8_t)__builtin_neon_vld1_lane_v(__a, (int8x8_t)__b, __c, 4); })
-#define vld1_lane_p16(__a, b, __c) __extension__ ({ \
-  poly16x4_t __b = (b); \
-  (poly16x4_t)__builtin_neon_vld1_lane_v(__a, (int8x8_t)__b, __c, 5); })
-
-#define vld2q_u8(__a) __extension__ ({ \
-  uint8x16x2_t r; __builtin_neon_vld2q_v(&r, __a, 48); r; })
-#define vld2q_u16(__a) __extension__ ({ \
-  uint16x8x2_t r; __builtin_neon_vld2q_v(&r, __a, 49); r; })
-#define vld2q_u32(__a) __extension__ ({ \
-  uint32x4x2_t r; __builtin_neon_vld2q_v(&r, __a, 50); r; })
-#define vld2q_s8(__a) __extension__ ({ \
-  int8x16x2_t r; __builtin_neon_vld2q_v(&r, __a, 32); r; })
-#define vld2q_s16(__a) __extension__ ({ \
-  int16x8x2_t r; __builtin_neon_vld2q_v(&r, __a, 33); r; })
-#define vld2q_s32(__a) __extension__ ({ \
-  int32x4x2_t r; __builtin_neon_vld2q_v(&r, __a, 34); r; })
-#define vld2q_f16(__a) __extension__ ({ \
-  float16x8x2_t r; __builtin_neon_vld2q_v(&r, __a, 40); r; })
-#define vld2q_f32(__a) __extension__ ({ \
-  float32x4x2_t r; __builtin_neon_vld2q_v(&r, __a, 41); r; })
-#define vld2q_p8(__a) __extension__ ({ \
-  poly8x16x2_t r; __builtin_neon_vld2q_v(&r, __a, 36); r; })
-#define vld2q_p16(__a) __extension__ ({ \
-  poly16x8x2_t r; __builtin_neon_vld2q_v(&r, __a, 37); r; })
-#define vld2_u8(__a) __extension__ ({ \
-  uint8x8x2_t r; __builtin_neon_vld2_v(&r, __a, 16); r; })
-#define vld2_u16(__a) __extension__ ({ \
-  uint16x4x2_t r; __builtin_neon_vld2_v(&r, __a, 17); r; })
-#define vld2_u32(__a) __extension__ ({ \
-  uint32x2x2_t r; __builtin_neon_vld2_v(&r, __a, 18); r; })
-#define vld2_u64(__a) __extension__ ({ \
-  uint64x1x2_t r; __builtin_neon_vld2_v(&r, __a, 19); r; })
-#define vld2_s8(__a) __extension__ ({ \
-  int8x8x2_t r; __builtin_neon_vld2_v(&r, __a, 0); r; })
-#define vld2_s16(__a) __extension__ ({ \
-  int16x4x2_t r; __builtin_neon_vld2_v(&r, __a, 1); r; })
-#define vld2_s32(__a) __extension__ ({ \
-  int32x2x2_t r; __builtin_neon_vld2_v(&r, __a, 2); r; })
-#define vld2_s64(__a) __extension__ ({ \
-  int64x1x2_t r; __builtin_neon_vld2_v(&r, __a, 3); r; })
-#define vld2_f16(__a) __extension__ ({ \
-  float16x4x2_t r; __builtin_neon_vld2_v(&r, __a, 8); r; })
-#define vld2_f32(__a) __extension__ ({ \
-  float32x2x2_t r; __builtin_neon_vld2_v(&r, __a, 9); r; })
-#define vld2_p8(__a) __extension__ ({ \
-  poly8x8x2_t r; __builtin_neon_vld2_v(&r, __a, 4); r; })
-#define vld2_p16(__a) __extension__ ({ \
-  poly16x4x2_t r; __builtin_neon_vld2_v(&r, __a, 5); r; })
-
-#define vld2_dup_u8(__a) __extension__ ({ \
-  uint8x8x2_t r; __builtin_neon_vld2_dup_v(&r, __a, 16); r; })
-#define vld2_dup_u16(__a) __extension__ ({ \
-  uint16x4x2_t r; __builtin_neon_vld2_dup_v(&r, __a, 17); r; })
-#define vld2_dup_u32(__a) __extension__ ({ \
-  uint32x2x2_t r; __builtin_neon_vld2_dup_v(&r, __a, 18); r; })
-#define vld2_dup_u64(__a) __extension__ ({ \
-  uint64x1x2_t r; __builtin_neon_vld2_dup_v(&r, __a, 19); r; })
-#define vld2_dup_s8(__a) __extension__ ({ \
-  int8x8x2_t r; __builtin_neon_vld2_dup_v(&r, __a, 0); r; })
-#define vld2_dup_s16(__a) __extension__ ({ \
-  int16x4x2_t r; __builtin_neon_vld2_dup_v(&r, __a, 1); r; })
-#define vld2_dup_s32(__a) __extension__ ({ \
-  int32x2x2_t r; __builtin_neon_vld2_dup_v(&r, __a, 2); r; })
-#define vld2_dup_s64(__a) __extension__ ({ \
-  int64x1x2_t r; __builtin_neon_vld2_dup_v(&r, __a, 3); r; })
-#define vld2_dup_f16(__a) __extension__ ({ \
-  float16x4x2_t r; __builtin_neon_vld2_dup_v(&r, __a, 8); r; })
-#define vld2_dup_f32(__a) __extension__ ({ \
-  float32x2x2_t r; __builtin_neon_vld2_dup_v(&r, __a, 9); r; })
-#define vld2_dup_p8(__a) __extension__ ({ \
-  poly8x8x2_t r; __builtin_neon_vld2_dup_v(&r, __a, 4); r; })
-#define vld2_dup_p16(__a) __extension__ ({ \
-  poly16x4x2_t r; __builtin_neon_vld2_dup_v(&r, __a, 5); r; })
-
-#define vld2q_lane_u16(__a, b, __c) __extension__ ({ \
-  uint16x8x2_t __b = (b); \
-  uint16x8x2_t r; __builtin_neon_vld2q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 49); r; })
-#define vld2q_lane_u32(__a, b, __c) __extension__ ({ \
-  uint32x4x2_t __b = (b); \
-  uint32x4x2_t r; __builtin_neon_vld2q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 50); r; })
-#define vld2q_lane_s16(__a, b, __c) __extension__ ({ \
-  int16x8x2_t __b = (b); \
-  int16x8x2_t r; __builtin_neon_vld2q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 33); r; })
-#define vld2q_lane_s32(__a, b, __c) __extension__ ({ \
-  int32x4x2_t __b = (b); \
-  int32x4x2_t r; __builtin_neon_vld2q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 34); r; })
-#define vld2q_lane_f16(__a, b, __c) __extension__ ({ \
-  float16x8x2_t __b = (b); \
-  float16x8x2_t r; __builtin_neon_vld2q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 40); r; })
-#define vld2q_lane_f32(__a, b, __c) __extension__ ({ \
-  float32x4x2_t __b = (b); \
-  float32x4x2_t r; __builtin_neon_vld2q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 41); r; })
-#define vld2q_lane_p16(__a, b, __c) __extension__ ({ \
-  poly16x8x2_t __b = (b); \
-  poly16x8x2_t r; __builtin_neon_vld2q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 37); r; })
-#define vld2_lane_u8(__a, b, __c) __extension__ ({ \
-  uint8x8x2_t __b = (b); \
-  uint8x8x2_t r; __builtin_neon_vld2_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 16); r; })
-#define vld2_lane_u16(__a, b, __c) __extension__ ({ \
-  uint16x4x2_t __b = (b); \
-  uint16x4x2_t r; __builtin_neon_vld2_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 17); r; })
-#define vld2_lane_u32(__a, b, __c) __extension__ ({ \
-  uint32x2x2_t __b = (b); \
-  uint32x2x2_t r; __builtin_neon_vld2_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 18); r; })
-#define vld2_lane_s8(__a, b, __c) __extension__ ({ \
-  int8x8x2_t __b = (b); \
-  int8x8x2_t r; __builtin_neon_vld2_lane_v(&r, __a, __b.val[0], __b.val[1], __c, 0); r; })
-#define vld2_lane_s16(__a, b, __c) __extension__ ({ \
-  int16x4x2_t __b = (b); \
-  int16x4x2_t r; __builtin_neon_vld2_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 1); r; })
-#define vld2_lane_s32(__a, b, __c) __extension__ ({ \
-  int32x2x2_t __b = (b); \
-  int32x2x2_t r; __builtin_neon_vld2_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 2); r; })
-#define vld2_lane_f16(__a, b, __c) __extension__ ({ \
-  float16x4x2_t __b = (b); \
-  float16x4x2_t r; __builtin_neon_vld2_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 8); r; })
-#define vld2_lane_f32(__a, b, __c) __extension__ ({ \
-  float32x2x2_t __b = (b); \
-  float32x2x2_t r; __builtin_neon_vld2_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 9); r; })
-#define vld2_lane_p8(__a, b, __c) __extension__ ({ \
-  poly8x8x2_t __b = (b); \
-  poly8x8x2_t r; __builtin_neon_vld2_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 4); r; })
-#define vld2_lane_p16(__a, b, __c) __extension__ ({ \
-  poly16x4x2_t __b = (b); \
-  poly16x4x2_t r; __builtin_neon_vld2_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 5); r; })
-
-#define vld3q_u8(__a) __extension__ ({ \
-  uint8x16x3_t r; __builtin_neon_vld3q_v(&r, __a, 48); r; })
-#define vld3q_u16(__a) __extension__ ({ \
-  uint16x8x3_t r; __builtin_neon_vld3q_v(&r, __a, 49); r; })
-#define vld3q_u32(__a) __extension__ ({ \
-  uint32x4x3_t r; __builtin_neon_vld3q_v(&r, __a, 50); r; })
-#define vld3q_s8(__a) __extension__ ({ \
-  int8x16x3_t r; __builtin_neon_vld3q_v(&r, __a, 32); r; })
-#define vld3q_s16(__a) __extension__ ({ \
-  int16x8x3_t r; __builtin_neon_vld3q_v(&r, __a, 33); r; })
-#define vld3q_s32(__a) __extension__ ({ \
-  int32x4x3_t r; __builtin_neon_vld3q_v(&r, __a, 34); r; })
-#define vld3q_f16(__a) __extension__ ({ \
-  float16x8x3_t r; __builtin_neon_vld3q_v(&r, __a, 40); r; })
-#define vld3q_f32(__a) __extension__ ({ \
-  float32x4x3_t r; __builtin_neon_vld3q_v(&r, __a, 41); r; })
-#define vld3q_p8(__a) __extension__ ({ \
-  poly8x16x3_t r; __builtin_neon_vld3q_v(&r, __a, 36); r; })
-#define vld3q_p16(__a) __extension__ ({ \
-  poly16x8x3_t r; __builtin_neon_vld3q_v(&r, __a, 37); r; })
-#define vld3_u8(__a) __extension__ ({ \
-  uint8x8x3_t r; __builtin_neon_vld3_v(&r, __a, 16); r; })
-#define vld3_u16(__a) __extension__ ({ \
-  uint16x4x3_t r; __builtin_neon_vld3_v(&r, __a, 17); r; })
-#define vld3_u32(__a) __extension__ ({ \
-  uint32x2x3_t r; __builtin_neon_vld3_v(&r, __a, 18); r; })
-#define vld3_u64(__a) __extension__ ({ \
-  uint64x1x3_t r; __builtin_neon_vld3_v(&r, __a, 19); r; })
-#define vld3_s8(__a) __extension__ ({ \
-  int8x8x3_t r; __builtin_neon_vld3_v(&r, __a, 0); r; })
-#define vld3_s16(__a) __extension__ ({ \
-  int16x4x3_t r; __builtin_neon_vld3_v(&r, __a, 1); r; })
-#define vld3_s32(__a) __extension__ ({ \
-  int32x2x3_t r; __builtin_neon_vld3_v(&r, __a, 2); r; })
-#define vld3_s64(__a) __extension__ ({ \
-  int64x1x3_t r; __builtin_neon_vld3_v(&r, __a, 3); r; })
-#define vld3_f16(__a) __extension__ ({ \
-  float16x4x3_t r; __builtin_neon_vld3_v(&r, __a, 8); r; })
-#define vld3_f32(__a) __extension__ ({ \
-  float32x2x3_t r; __builtin_neon_vld3_v(&r, __a, 9); r; })
-#define vld3_p8(__a) __extension__ ({ \
-  poly8x8x3_t r; __builtin_neon_vld3_v(&r, __a, 4); r; })
-#define vld3_p16(__a) __extension__ ({ \
-  poly16x4x3_t r; __builtin_neon_vld3_v(&r, __a, 5); r; })
-
-#define vld3_dup_u8(__a) __extension__ ({ \
-  uint8x8x3_t r; __builtin_neon_vld3_dup_v(&r, __a, 16); r; })
-#define vld3_dup_u16(__a) __extension__ ({ \
-  uint16x4x3_t r; __builtin_neon_vld3_dup_v(&r, __a, 17); r; })
-#define vld3_dup_u32(__a) __extension__ ({ \
-  uint32x2x3_t r; __builtin_neon_vld3_dup_v(&r, __a, 18); r; })
-#define vld3_dup_u64(__a) __extension__ ({ \
-  uint64x1x3_t r; __builtin_neon_vld3_dup_v(&r, __a, 19); r; })
-#define vld3_dup_s8(__a) __extension__ ({ \
-  int8x8x3_t r; __builtin_neon_vld3_dup_v(&r, __a, 0); r; })
-#define vld3_dup_s16(__a) __extension__ ({ \
-  int16x4x3_t r; __builtin_neon_vld3_dup_v(&r, __a, 1); r; })
-#define vld3_dup_s32(__a) __extension__ ({ \
-  int32x2x3_t r; __builtin_neon_vld3_dup_v(&r, __a, 2); r; })
-#define vld3_dup_s64(__a) __extension__ ({ \
-  int64x1x3_t r; __builtin_neon_vld3_dup_v(&r, __a, 3); r; })
-#define vld3_dup_f16(__a) __extension__ ({ \
-  float16x4x3_t r; __builtin_neon_vld3_dup_v(&r, __a, 8); r; })
-#define vld3_dup_f32(__a) __extension__ ({ \
-  float32x2x3_t r; __builtin_neon_vld3_dup_v(&r, __a, 9); r; })
-#define vld3_dup_p8(__a) __extension__ ({ \
-  poly8x8x3_t r; __builtin_neon_vld3_dup_v(&r, __a, 4); r; })
-#define vld3_dup_p16(__a) __extension__ ({ \
-  poly16x4x3_t r; __builtin_neon_vld3_dup_v(&r, __a, 5); r; })
-
-#define vld3q_lane_u16(__a, b, __c) __extension__ ({ \
-  uint16x8x3_t __b = (b); \
-  uint16x8x3_t r; __builtin_neon_vld3q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 49); r; })
-#define vld3q_lane_u32(__a, b, __c) __extension__ ({ \
-  uint32x4x3_t __b = (b); \
-  uint32x4x3_t r; __builtin_neon_vld3q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 50); r; })
-#define vld3q_lane_s16(__a, b, __c) __extension__ ({ \
-  int16x8x3_t __b = (b); \
-  int16x8x3_t r; __builtin_neon_vld3q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 33); r; })
-#define vld3q_lane_s32(__a, b, __c) __extension__ ({ \
-  int32x4x3_t __b = (b); \
-  int32x4x3_t r; __builtin_neon_vld3q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 34); r; })
-#define vld3q_lane_f16(__a, b, __c) __extension__ ({ \
-  float16x8x3_t __b = (b); \
-  float16x8x3_t r; __builtin_neon_vld3q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 40); r; })
-#define vld3q_lane_f32(__a, b, __c) __extension__ ({ \
-  float32x4x3_t __b = (b); \
-  float32x4x3_t r; __builtin_neon_vld3q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 41); r; })
-#define vld3q_lane_p16(__a, b, __c) __extension__ ({ \
-  poly16x8x3_t __b = (b); \
-  poly16x8x3_t r; __builtin_neon_vld3q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 37); r; })
-#define vld3_lane_u8(__a, b, __c) __extension__ ({ \
-  uint8x8x3_t __b = (b); \
-  uint8x8x3_t r; __builtin_neon_vld3_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 16); r; })
-#define vld3_lane_u16(__a, b, __c) __extension__ ({ \
-  uint16x4x3_t __b = (b); \
-  uint16x4x3_t r; __builtin_neon_vld3_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 17); r; })
-#define vld3_lane_u32(__a, b, __c) __extension__ ({ \
-  uint32x2x3_t __b = (b); \
-  uint32x2x3_t r; __builtin_neon_vld3_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 18); r; })
-#define vld3_lane_s8(__a, b, __c) __extension__ ({ \
-  int8x8x3_t __b = (b); \
-  int8x8x3_t r; __builtin_neon_vld3_lane_v(&r, __a, __b.val[0], __b.val[1], __b.val[2], __c, 0); r; })
-#define vld3_lane_s16(__a, b, __c) __extension__ ({ \
-  int16x4x3_t __b = (b); \
-  int16x4x3_t r; __builtin_neon_vld3_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 1); r; })
-#define vld3_lane_s32(__a, b, __c) __extension__ ({ \
-  int32x2x3_t __b = (b); \
-  int32x2x3_t r; __builtin_neon_vld3_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 2); r; })
-#define vld3_lane_f16(__a, b, __c) __extension__ ({ \
-  float16x4x3_t __b = (b); \
-  float16x4x3_t r; __builtin_neon_vld3_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 8); r; })
-#define vld3_lane_f32(__a, b, __c) __extension__ ({ \
-  float32x2x3_t __b = (b); \
-  float32x2x3_t r; __builtin_neon_vld3_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 9); r; })
-#define vld3_lane_p8(__a, b, __c) __extension__ ({ \
-  poly8x8x3_t __b = (b); \
-  poly8x8x3_t r; __builtin_neon_vld3_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 4); r; })
-#define vld3_lane_p16(__a, b, __c) __extension__ ({ \
-  poly16x4x3_t __b = (b); \
-  poly16x4x3_t r; __builtin_neon_vld3_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 5); r; })
-
-#define vld4q_u8(__a) __extension__ ({ \
-  uint8x16x4_t r; __builtin_neon_vld4q_v(&r, __a, 48); r; })
-#define vld4q_u16(__a) __extension__ ({ \
-  uint16x8x4_t r; __builtin_neon_vld4q_v(&r, __a, 49); r; })
-#define vld4q_u32(__a) __extension__ ({ \
-  uint32x4x4_t r; __builtin_neon_vld4q_v(&r, __a, 50); r; })
-#define vld4q_s8(__a) __extension__ ({ \
-  int8x16x4_t r; __builtin_neon_vld4q_v(&r, __a, 32); r; })
-#define vld4q_s16(__a) __extension__ ({ \
-  int16x8x4_t r; __builtin_neon_vld4q_v(&r, __a, 33); r; })
-#define vld4q_s32(__a) __extension__ ({ \
-  int32x4x4_t r; __builtin_neon_vld4q_v(&r, __a, 34); r; })
-#define vld4q_f16(__a) __extension__ ({ \
-  float16x8x4_t r; __builtin_neon_vld4q_v(&r, __a, 40); r; })
-#define vld4q_f32(__a) __extension__ ({ \
-  float32x4x4_t r; __builtin_neon_vld4q_v(&r, __a, 41); r; })
-#define vld4q_p8(__a) __extension__ ({ \
-  poly8x16x4_t r; __builtin_neon_vld4q_v(&r, __a, 36); r; })
-#define vld4q_p16(__a) __extension__ ({ \
-  poly16x8x4_t r; __builtin_neon_vld4q_v(&r, __a, 37); r; })
-#define vld4_u8(__a) __extension__ ({ \
-  uint8x8x4_t r; __builtin_neon_vld4_v(&r, __a, 16); r; })
-#define vld4_u16(__a) __extension__ ({ \
-  uint16x4x4_t r; __builtin_neon_vld4_v(&r, __a, 17); r; })
-#define vld4_u32(__a) __extension__ ({ \
-  uint32x2x4_t r; __builtin_neon_vld4_v(&r, __a, 18); r; })
-#define vld4_u64(__a) __extension__ ({ \
-  uint64x1x4_t r; __builtin_neon_vld4_v(&r, __a, 19); r; })
-#define vld4_s8(__a) __extension__ ({ \
-  int8x8x4_t r; __builtin_neon_vld4_v(&r, __a, 0); r; })
-#define vld4_s16(__a) __extension__ ({ \
-  int16x4x4_t r; __builtin_neon_vld4_v(&r, __a, 1); r; })
-#define vld4_s32(__a) __extension__ ({ \
-  int32x2x4_t r; __builtin_neon_vld4_v(&r, __a, 2); r; })
-#define vld4_s64(__a) __extension__ ({ \
-  int64x1x4_t r; __builtin_neon_vld4_v(&r, __a, 3); r; })
-#define vld4_f16(__a) __extension__ ({ \
-  float16x4x4_t r; __builtin_neon_vld4_v(&r, __a, 8); r; })
-#define vld4_f32(__a) __extension__ ({ \
-  float32x2x4_t r; __builtin_neon_vld4_v(&r, __a, 9); r; })
-#define vld4_p8(__a) __extension__ ({ \
-  poly8x8x4_t r; __builtin_neon_vld4_v(&r, __a, 4); r; })
-#define vld4_p16(__a) __extension__ ({ \
-  poly16x4x4_t r; __builtin_neon_vld4_v(&r, __a, 5); r; })
-
-#define vld4_dup_u8(__a) __extension__ ({ \
-  uint8x8x4_t r; __builtin_neon_vld4_dup_v(&r, __a, 16); r; })
-#define vld4_dup_u16(__a) __extension__ ({ \
-  uint16x4x4_t r; __builtin_neon_vld4_dup_v(&r, __a, 17); r; })
-#define vld4_dup_u32(__a) __extension__ ({ \
-  uint32x2x4_t r; __builtin_neon_vld4_dup_v(&r, __a, 18); r; })
-#define vld4_dup_u64(__a) __extension__ ({ \
-  uint64x1x4_t r; __builtin_neon_vld4_dup_v(&r, __a, 19); r; })
-#define vld4_dup_s8(__a) __extension__ ({ \
-  int8x8x4_t r; __builtin_neon_vld4_dup_v(&r, __a, 0); r; })
-#define vld4_dup_s16(__a) __extension__ ({ \
-  int16x4x4_t r; __builtin_neon_vld4_dup_v(&r, __a, 1); r; })
-#define vld4_dup_s32(__a) __extension__ ({ \
-  int32x2x4_t r; __builtin_neon_vld4_dup_v(&r, __a, 2); r; })
-#define vld4_dup_s64(__a) __extension__ ({ \
-  int64x1x4_t r; __builtin_neon_vld4_dup_v(&r, __a, 3); r; })
-#define vld4_dup_f16(__a) __extension__ ({ \
-  float16x4x4_t r; __builtin_neon_vld4_dup_v(&r, __a, 8); r; })
-#define vld4_dup_f32(__a) __extension__ ({ \
-  float32x2x4_t r; __builtin_neon_vld4_dup_v(&r, __a, 9); r; })
-#define vld4_dup_p8(__a) __extension__ ({ \
-  poly8x8x4_t r; __builtin_neon_vld4_dup_v(&r, __a, 4); r; })
-#define vld4_dup_p16(__a) __extension__ ({ \
-  poly16x4x4_t r; __builtin_neon_vld4_dup_v(&r, __a, 5); r; })
-
-#define vld4q_lane_u16(__a, b, __c) __extension__ ({ \
-  uint16x8x4_t __b = (b); \
-  uint16x8x4_t r; __builtin_neon_vld4q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 49); r; })
-#define vld4q_lane_u32(__a, b, __c) __extension__ ({ \
-  uint32x4x4_t __b = (b); \
-  uint32x4x4_t r; __builtin_neon_vld4q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 50); r; })
-#define vld4q_lane_s16(__a, b, __c) __extension__ ({ \
-  int16x8x4_t __b = (b); \
-  int16x8x4_t r; __builtin_neon_vld4q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 33); r; })
-#define vld4q_lane_s32(__a, b, __c) __extension__ ({ \
-  int32x4x4_t __b = (b); \
-  int32x4x4_t r; __builtin_neon_vld4q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 34); r; })
-#define vld4q_lane_f16(__a, b, __c) __extension__ ({ \
-  float16x8x4_t __b = (b); \
-  float16x8x4_t r; __builtin_neon_vld4q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 40); r; })
-#define vld4q_lane_f32(__a, b, __c) __extension__ ({ \
-  float32x4x4_t __b = (b); \
-  float32x4x4_t r; __builtin_neon_vld4q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 41); r; })
-#define vld4q_lane_p16(__a, b, __c) __extension__ ({ \
-  poly16x8x4_t __b = (b); \
-  poly16x8x4_t r; __builtin_neon_vld4q_lane_v(&r, __a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 37); r; })
-#define vld4_lane_u8(__a, b, __c) __extension__ ({ \
-  uint8x8x4_t __b = (b); \
-  uint8x8x4_t r; __builtin_neon_vld4_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 16); r; })
-#define vld4_lane_u16(__a, b, __c) __extension__ ({ \
-  uint16x4x4_t __b = (b); \
-  uint16x4x4_t r; __builtin_neon_vld4_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 17); r; })
-#define vld4_lane_u32(__a, b, __c) __extension__ ({ \
-  uint32x2x4_t __b = (b); \
-  uint32x2x4_t r; __builtin_neon_vld4_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 18); r; })
-#define vld4_lane_s8(__a, b, __c) __extension__ ({ \
-  int8x8x4_t __b = (b); \
-  int8x8x4_t r; __builtin_neon_vld4_lane_v(&r, __a, __b.val[0], __b.val[1], __b.val[2], __b.val[3], __c, 0); r; })
-#define vld4_lane_s16(__a, b, __c) __extension__ ({ \
-  int16x4x4_t __b = (b); \
-  int16x4x4_t r; __builtin_neon_vld4_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 1); r; })
-#define vld4_lane_s32(__a, b, __c) __extension__ ({ \
-  int32x2x4_t __b = (b); \
-  int32x2x4_t r; __builtin_neon_vld4_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 2); r; })
-#define vld4_lane_f16(__a, b, __c) __extension__ ({ \
-  float16x4x4_t __b = (b); \
-  float16x4x4_t r; __builtin_neon_vld4_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 8); r; })
-#define vld4_lane_f32(__a, b, __c) __extension__ ({ \
-  float32x2x4_t __b = (b); \
-  float32x2x4_t r; __builtin_neon_vld4_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 9); r; })
-#define vld4_lane_p8(__a, b, __c) __extension__ ({ \
-  poly8x8x4_t __b = (b); \
-  poly8x8x4_t r; __builtin_neon_vld4_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 4); r; })
-#define vld4_lane_p16(__a, b, __c) __extension__ ({ \
-  poly16x4x4_t __b = (b); \
-  poly16x4x4_t r; __builtin_neon_vld4_lane_v(&r, __a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 5); r; })
+#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
 
-#if defined(__aarch64__)
-#define vldrq_p128(__a) __extension__ ({ \
-  (poly128_t)__builtin_neon_vldrq_p128(__a); })
-
-#endif
-
-__ai int8x8_t vmax_s8(int8x8_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vmax_v(__a, __b, 0); }
-__ai int16x4_t vmax_s16(int16x4_t __a, int16x4_t __b) {
-  return (int16x4_t)__builtin_neon_vmax_v((int8x8_t)__a, (int8x8_t)__b, 1); }
-__ai int32x2_t vmax_s32(int32x2_t __a, int32x2_t __b) {
-  return (int32x2_t)__builtin_neon_vmax_v((int8x8_t)__a, (int8x8_t)__b, 2); }
-__ai uint8x8_t vmax_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vmax_v((int8x8_t)__a, (int8x8_t)__b, 16); }
-__ai uint16x4_t vmax_u16(uint16x4_t __a, uint16x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vmax_v((int8x8_t)__a, (int8x8_t)__b, 17); }
-__ai uint32x2_t vmax_u32(uint32x2_t __a, uint32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vmax_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai float32x2_t vmax_f32(float32x2_t __a, float32x2_t __b) {
-  return (float32x2_t)__builtin_neon_vmax_v((int8x8_t)__a, (int8x8_t)__b, 9); }
-__ai int8x16_t vmaxq_s8(int8x16_t __a, int8x16_t __b) {
-  return (int8x16_t)__builtin_neon_vmaxq_v(__a, __b, 32); }
-__ai int16x8_t vmaxq_s16(int16x8_t __a, int16x8_t __b) {
-  return (int16x8_t)__builtin_neon_vmaxq_v((int8x16_t)__a, (int8x16_t)__b, 33); }
-__ai int32x4_t vmaxq_s32(int32x4_t __a, int32x4_t __b) {
-  return (int32x4_t)__builtin_neon_vmaxq_v((int8x16_t)__a, (int8x16_t)__b, 34); }
-__ai uint8x16_t vmaxq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vmaxq_v((int8x16_t)__a, (int8x16_t)__b, 48); }
-__ai uint16x8_t vmaxq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint16x8_t)__builtin_neon_vmaxq_v((int8x16_t)__a, (int8x16_t)__b, 49); }
-__ai uint32x4_t vmaxq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vmaxq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-__ai float32x4_t vmaxq_f32(float32x4_t __a, float32x4_t __b) {
-  return (float32x4_t)__builtin_neon_vmaxq_v((int8x16_t)__a, (int8x16_t)__b, 41); }
+#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
 
-#if defined(__aarch64__)
-__ai int8_t vmaxv_s8(int8x8_t __a) {
-  return (int8_t)__builtin_neon_vmaxv_s8(__a); }
-__ai int16_t vmaxv_s16(int16x4_t __a) {
-  return (int16_t)__builtin_neon_vmaxv_s16(__a); }
-__ai int32_t vmaxv_s32(int32x2_t __a) {
-  return (int32_t)__builtin_neon_vmaxv_s32(__a); }
-__ai float32_t vmaxv_f32(float32x2_t __a) {
-  return (float32_t)__builtin_neon_vmaxv_f32(__a); }
-__ai uint8_t vmaxv_u8(uint8x8_t __a) {
-  return (uint8_t)__builtin_neon_vmaxv_u8((int8x8_t)__a); }
-__ai uint16_t vmaxv_u16(uint16x4_t __a) {
-  return (uint16_t)__builtin_neon_vmaxv_u16((int16x4_t)__a); }
-__ai uint32_t vmaxv_u32(uint32x2_t __a) {
-  return (uint32_t)__builtin_neon_vmaxv_u32((int32x2_t)__a); }
-__ai int8_t vmaxvq_s8(int8x16_t __a) {
-  return (int8_t)__builtin_neon_vmaxvq_s8(__a); }
-__ai int16_t vmaxvq_s16(int16x8_t __a) {
-  return (int16_t)__builtin_neon_vmaxvq_s16(__a); }
-__ai int32_t vmaxvq_s32(int32x4_t __a) {
-  return (int32_t)__builtin_neon_vmaxvq_s32(__a); }
-__ai uint8_t vmaxvq_u8(uint8x16_t __a) {
-  return (uint8_t)__builtin_neon_vmaxvq_u8((int8x16_t)__a); }
-__ai uint16_t vmaxvq_u16(uint16x8_t __a) {
-  return (uint16_t)__builtin_neon_vmaxvq_u16((int16x8_t)__a); }
-__ai uint32_t vmaxvq_u32(uint32x4_t __a) {
-  return (uint32_t)__builtin_neon_vmaxvq_u32((int32x4_t)__a); }
-__ai float32_t vmaxvq_f32(float32x4_t __a) {
-  return (float32_t)__builtin_neon_vmaxvq_f32(__a); }
-__ai float64_t vmaxvq_f64(float64x2_t __a) {
-  return (float64_t)__builtin_neon_vmaxvq_f64(__a); }
-
-#endif
-
-__ai int8x8_t vmin_s8(int8x8_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vmin_v(__a, __b, 0); }
-__ai int16x4_t vmin_s16(int16x4_t __a, int16x4_t __b) {
-  return (int16x4_t)__builtin_neon_vmin_v((int8x8_t)__a, (int8x8_t)__b, 1); }
-__ai int32x2_t vmin_s32(int32x2_t __a, int32x2_t __b) {
-  return (int32x2_t)__builtin_neon_vmin_v((int8x8_t)__a, (int8x8_t)__b, 2); }
-__ai uint8x8_t vmin_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vmin_v((int8x8_t)__a, (int8x8_t)__b, 16); }
-__ai uint16x4_t vmin_u16(uint16x4_t __a, uint16x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vmin_v((int8x8_t)__a, (int8x8_t)__b, 17); }
-__ai uint32x2_t vmin_u32(uint32x2_t __a, uint32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vmin_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai float32x2_t vmin_f32(float32x2_t __a, float32x2_t __b) {
-  return (float32x2_t)__builtin_neon_vmin_v((int8x8_t)__a, (int8x8_t)__b, 9); }
-__ai int8x16_t vminq_s8(int8x16_t __a, int8x16_t __b) {
-  return (int8x16_t)__builtin_neon_vminq_v(__a, __b, 32); }
-__ai int16x8_t vminq_s16(int16x8_t __a, int16x8_t __b) {
-  return (int16x8_t)__builtin_neon_vminq_v((int8x16_t)__a, (int8x16_t)__b, 33); }
-__ai int32x4_t vminq_s32(int32x4_t __a, int32x4_t __b) {
-  return (int32x4_t)__builtin_neon_vminq_v((int8x16_t)__a, (int8x16_t)__b, 34); }
-__ai uint8x16_t vminq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vminq_v((int8x16_t)__a, (int8x16_t)__b, 48); }
-__ai uint16x8_t vminq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint16x8_t)__builtin_neon_vminq_v((int8x16_t)__a, (int8x16_t)__b, 49); }
-__ai uint32x4_t vminq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vminq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-__ai float32x4_t vminq_f32(float32x4_t __a, float32x4_t __b) {
-  return (float32x4_t)__builtin_neon_vminq_v((int8x16_t)__a, (int8x16_t)__b, 41); }
+#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
 
-#if defined(__aarch64__)
-__ai int8_t vminv_s8(int8x8_t __a) {
-  return (int8_t)__builtin_neon_vminv_s8(__a); }
-__ai int16_t vminv_s16(int16x4_t __a) {
-  return (int16_t)__builtin_neon_vminv_s16(__a); }
-__ai int32_t vminv_s32(int32x2_t __a) {
-  return (int32_t)__builtin_neon_vminv_s32(__a); }
-__ai float32_t vminv_f32(float32x2_t __a) {
-  return (float32_t)__builtin_neon_vminv_f32(__a); }
-__ai uint8_t vminv_u8(uint8x8_t __a) {
-  return (uint8_t)__builtin_neon_vminv_u8((int8x8_t)__a); }
-__ai uint16_t vminv_u16(uint16x4_t __a) {
-  return (uint16_t)__builtin_neon_vminv_u16((int16x4_t)__a); }
-__ai uint32_t vminv_u32(uint32x2_t __a) {
-  return (uint32_t)__builtin_neon_vminv_u32((int32x2_t)__a); }
-__ai int8_t vminvq_s8(int8x16_t __a) {
-  return (int8_t)__builtin_neon_vminvq_s8(__a); }
-__ai int16_t vminvq_s16(int16x8_t __a) {
-  return (int16_t)__builtin_neon_vminvq_s16(__a); }
-__ai int32_t vminvq_s32(int32x4_t __a) {
-  return (int32_t)__builtin_neon_vminvq_s32(__a); }
-__ai uint8_t vminvq_u8(uint8x16_t __a) {
-  return (uint8_t)__builtin_neon_vminvq_u8((int8x16_t)__a); }
-__ai uint16_t vminvq_u16(uint16x8_t __a) {
-  return (uint16_t)__builtin_neon_vminvq_u16((int16x8_t)__a); }
-__ai uint32_t vminvq_u32(uint32x4_t __a) {
-  return (uint32_t)__builtin_neon_vminvq_u32((int32x4_t)__a); }
-__ai float32_t vminvq_f32(float32x4_t __a) {
-  return (float32_t)__builtin_neon_vminvq_f32(__a); }
-__ai float64_t vminvq_f64(float64x2_t __a) {
-  return (float64_t)__builtin_neon_vminvq_f64(__a); }
-
-#endif
-
-__ai int8x8_t vmla_s8(int8x8_t __a, int8x8_t __b, int8x8_t __c) {
-  return __a + (__b * __c); }
-__ai int16x4_t vmla_s16(int16x4_t __a, int16x4_t __b, int16x4_t __c) {
-  return __a + (__b * __c); }
-__ai int32x2_t vmla_s32(int32x2_t __a, int32x2_t __b, int32x2_t __c) {
-  return __a + (__b * __c); }
-__ai float32x2_t vmla_f32(float32x2_t __a, float32x2_t __b, float32x2_t __c) {
-  return __a + (__b * __c); }
-__ai uint8x8_t vmla_u8(uint8x8_t __a, uint8x8_t __b, uint8x8_t __c) {
-  return __a + (__b * __c); }
-__ai uint16x4_t vmla_u16(uint16x4_t __a, uint16x4_t __b, uint16x4_t __c) {
-  return __a + (__b * __c); }
-__ai uint32x2_t vmla_u32(uint32x2_t __a, uint32x2_t __b, uint32x2_t __c) {
-  return __a + (__b * __c); }
-__ai int8x16_t vmlaq_s8(int8x16_t __a, int8x16_t __b, int8x16_t __c) {
-  return __a + (__b * __c); }
-__ai int16x8_t vmlaq_s16(int16x8_t __a, int16x8_t __b, int16x8_t __c) {
-  return __a + (__b * __c); }
-__ai int32x4_t vmlaq_s32(int32x4_t __a, int32x4_t __b, int32x4_t __c) {
-  return __a + (__b * __c); }
-__ai float32x4_t vmlaq_f32(float32x4_t __a, float32x4_t __b, float32x4_t __c) {
-  return __a + (__b * __c); }
-__ai uint8x16_t vmlaq_u8(uint8x16_t __a, uint8x16_t __b, uint8x16_t __c) {
-  return __a + (__b * __c); }
-__ai uint16x8_t vmlaq_u16(uint16x8_t __a, uint16x8_t __b, uint16x8_t __c) {
-  return __a + (__b * __c); }
-__ai uint32x4_t vmlaq_u32(uint32x4_t __a, uint32x4_t __b, uint32x4_t __c) {
-  return __a + (__b * __c); }
-
-__ai int16x8_t vmlal_s8(int16x8_t __a, int8x8_t __b, int8x8_t __c) {
-  return __a + vmull_s8(__b, __c); }
-__ai int32x4_t vmlal_s16(int32x4_t __a, int16x4_t __b, int16x4_t __c) {
-  return __a + vmull_s16(__b, __c); }
-__ai int64x2_t vmlal_s32(int64x2_t __a, int32x2_t __b, int32x2_t __c) {
-  return __a + vmull_s32(__b, __c); }
-__ai uint16x8_t vmlal_u8(uint16x8_t __a, uint8x8_t __b, uint8x8_t __c) {
-  return __a + vmull_u8(__b, __c); }
-__ai uint32x4_t vmlal_u16(uint32x4_t __a, uint16x4_t __b, uint16x4_t __c) {
-  return __a + vmull_u16(__b, __c); }
-__ai uint64x2_t vmlal_u32(uint64x2_t __a, uint32x2_t __b, uint32x2_t __c) {
-  return __a + vmull_u32(__b, __c); }
+#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
 
-#if defined(__aarch64__)
-__ai int16x8_t vmlal_high_s8(int16x8_t __a, int8x16_t __b, int8x16_t __c) {
-  return vmlal_s8(__a, vget_high_s8(__b), vget_high_s8(__c)); }
-__ai int32x4_t vmlal_high_s16(int32x4_t __a, int16x8_t __b, int16x8_t __c) {
-  return vmlal_s16(__a, vget_high_s16(__b), vget_high_s16(__c)); }
-__ai int64x2_t vmlal_high_s32(int64x2_t __a, int32x4_t __b, int32x4_t __c) {
-  return vmlal_s32(__a, vget_high_s32(__b), vget_high_s32(__c)); }
-__ai uint16x8_t vmlal_high_u8(uint16x8_t __a, uint8x16_t __b, uint8x16_t __c) {
-  return vmlal_u8(__a, vget_high_u8(__b), vget_high_u8(__c)); }
-__ai uint32x4_t vmlal_high_u16(uint32x4_t __a, uint16x8_t __b, uint16x8_t __c) {
-  return vmlal_u16(__a, vget_high_u16(__b), vget_high_u16(__c)); }
-__ai uint64x2_t vmlal_high_u32(uint64x2_t __a, uint32x4_t __b, uint32x4_t __c) {
-  return vmlal_u32(__a, vget_high_u32(__b), vget_high_u32(__c)); }
-
-#define vmlal_high_lane_s16(a, b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int16x8_t __b = (b); int16x4_t __c = (c); \
-  __a + vmull_s16(vget_high_s16(__b), __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlal_high_lane_s32(a, b, c, __d) __extension__ ({ \
-  int64x2_t __a = (a); int32x4_t __b = (b); int32x2_t __c = (c); \
-  __a + vmull_s32(vget_high_s32(__b), __builtin_shufflevector(__c, __c, __d, __d)); })
-#define vmlal_high_lane_u16(a, b, c, __d) __extension__ ({ \
-  uint32x4_t __a = (a); uint16x8_t __b = (b); uint16x4_t __c = (c); \
-  __a + vmull_u16(vget_high_u16(__b), __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlal_high_lane_u32(a, b, c, __d) __extension__ ({ \
-  uint64x2_t __a = (a); uint32x4_t __b = (b); uint32x2_t __c = (c); \
-  __a + vmull_u32(vget_high_u32(__b), __builtin_shufflevector(__c, __c, __d, __d)); })
-
-#define vmlal_high_laneq_s16(a, b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int16x8_t __b = (b); int16x8_t __c = (c); \
-  __a + vmull_s16(vget_high_s16(__b), __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlal_high_laneq_s32(a, b, c, __d) __extension__ ({ \
-  int64x2_t __a = (a); int32x4_t __b = (b); int32x4_t __c = (c); \
-  __a + vmull_s32(vget_high_s32(__b), __builtin_shufflevector(__c, __c, __d, __d)); })
-#define vmlal_high_laneq_u16(a, b, c, __d) __extension__ ({ \
-  uint32x4_t __a = (a); uint16x8_t __b = (b); uint16x8_t __c = (c); \
-  __a + vmull_u16(vget_high_u16(__b), __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlal_high_laneq_u32(a, b, c, __d) __extension__ ({ \
-  uint64x2_t __a = (a); uint32x4_t __b = (b); uint32x4_t __c = (c); \
-  __a + vmull_u32(vget_high_u32(__b), __builtin_shufflevector(__c, __c, __d, __d)); })
-
-__ai int32x4_t vmlal_high_n_s16(int32x4_t __a, int16x8_t __b, int16_t __c) {
-  return vmlal_n_s16(__a, vget_high_s16(__b), __c); }
-__ai int64x2_t vmlal_high_n_s32(int64x2_t __a, int32x4_t __b, int32_t __c) {
-  return vmlal_n_s32(__a, vget_high_s32(__b), __c); }
-__ai uint32x4_t vmlal_high_n_u16(uint32x4_t __a, uint16x8_t __b, uint16_t __c) {
-  return vmlal_n_u16(__a, vget_high_u16(__b), __c); }
-__ai uint64x2_t vmlal_high_n_u32(uint64x2_t __a, uint32x4_t __b, uint32_t __c) {
-  return vmlal_n_u32(__a, vget_high_u32(__b), __c); }
-
-#endif
-
-#define vmlal_lane_s16(a, b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int16x4_t __b = (b); int16x4_t __c = (c); \
-  __a + vmull_s16(__b, __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlal_lane_s32(a, b, c, __d) __extension__ ({ \
-  int64x2_t __a = (a); int32x2_t __b = (b); int32x2_t __c = (c); \
-  __a + vmull_s32(__b, __builtin_shufflevector(__c, __c, __d, __d)); })
-#define vmlal_lane_u16(a, b, c, __d) __extension__ ({ \
-  uint32x4_t __a = (a); uint16x4_t __b = (b); uint16x4_t __c = (c); \
-  __a + vmull_u16(__b, __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlal_lane_u32(a, b, c, __d) __extension__ ({ \
-  uint64x2_t __a = (a); uint32x2_t __b = (b); uint32x2_t __c = (c); \
-  __a + vmull_u32(__b, __builtin_shufflevector(__c, __c, __d, __d)); })
+#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
 
-#if defined(__aarch64__)
-#define vmlal_laneq_s16(a, b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int16x4_t __b = (b); int16x8_t __c = (c); \
-  __a + vmull_s16(__b, __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlal_laneq_s32(a, b, c, __d) __extension__ ({ \
-  int64x2_t __a = (a); int32x2_t __b = (b); int32x4_t __c = (c); \
-  __a + vmull_s32(__b, __builtin_shufflevector(__c, __c, __d, __d)); })
-#define vmlal_laneq_u16(a, b, c, __d) __extension__ ({ \
-  uint32x4_t __a = (a); uint16x4_t __b = (b); uint16x8_t __c = (c); \
-  __a + vmull_u16(__b, __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlal_laneq_u32(a, b, c, __d) __extension__ ({ \
-  uint64x2_t __a = (a); uint32x2_t __b = (b); uint32x4_t __c = (c); \
-  __a + vmull_u32(__b, __builtin_shufflevector(__c, __c, __d, __d)); })
-
-#endif
-
-#define vmla_lane_s16(a, b, c, __d) __extension__ ({ \
-  int16x4_t __a = (a); int16x4_t __b = (b); int16x4_t __c = (c); \
-  __a + (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmla_lane_s32(a, b, c, __d) __extension__ ({ \
-  int32x2_t __a = (a); int32x2_t __b = (b); int32x2_t __c = (c); \
-  __a + (__b * __builtin_shufflevector(__c, __c, __d, __d)); })
-#define vmla_lane_u16(a, b, c, __d) __extension__ ({ \
-  uint16x4_t __a = (a); uint16x4_t __b = (b); uint16x4_t __c = (c); \
-  __a + (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmla_lane_u32(a, b, c, __d) __extension__ ({ \
-  uint32x2_t __a = (a); uint32x2_t __b = (b); uint32x2_t __c = (c); \
-  __a + (__b * __builtin_shufflevector(__c, __c, __d, __d)); })
-#define vmla_lane_f32(a, b, c, __d) __extension__ ({ \
-  float32x2_t __a = (a); float32x2_t __b = (b); float32x2_t __c = (c); \
-  __a + (__b * __builtin_shufflevector(__c, __c, __d, __d)); })
-#define vmlaq_lane_s16(a, b, c, __d) __extension__ ({ \
-  int16x8_t __a = (a); int16x8_t __b = (b); int16x4_t __c = (c); \
-  __a + (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d, __d, __d, __d, __d)); })
-#define vmlaq_lane_s32(a, b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int32x4_t __b = (b); int32x2_t __c = (c); \
-  __a + (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlaq_lane_u16(a, b, c, __d) __extension__ ({ \
-  uint16x8_t __a = (a); uint16x8_t __b = (b); uint16x4_t __c = (c); \
-  __a + (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d, __d, __d, __d, __d)); })
-#define vmlaq_lane_u32(a, b, c, __d) __extension__ ({ \
-  uint32x4_t __a = (a); uint32x4_t __b = (b); uint32x2_t __c = (c); \
-  __a + (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlaq_lane_f32(a, b, c, __d) __extension__ ({ \
-  float32x4_t __a = (a); float32x4_t __b = (b); float32x2_t __c = (c); \
-  __a + (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
+#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
 
-#if defined(__aarch64__)
-#define vmla_laneq_s16(a, b, c, __d) __extension__ ({ \
-  int16x4_t __a = (a); int16x4_t __b = (b); int16x8_t __c = (c); \
-  __a + (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmla_laneq_s32(a, b, c, __d) __extension__ ({ \
-  int32x2_t __a = (a); int32x2_t __b = (b); int32x4_t __c = (c); \
-  __a + (__b * __builtin_shufflevector(__c, __c, __d, __d)); })
-#define vmla_laneq_u16(a, b, c, __d) __extension__ ({ \
-  uint16x4_t __a = (a); uint16x4_t __b = (b); uint16x8_t __c = (c); \
-  __a + (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmla_laneq_u32(a, b, c, __d) __extension__ ({ \
-  uint32x2_t __a = (a); uint32x2_t __b = (b); uint32x4_t __c = (c); \
-  __a + (__b * __builtin_shufflevector(__c, __c, __d, __d)); })
-#define vmla_laneq_f32(a, b, c, __d) __extension__ ({ \
-  float32x2_t __a = (a); float32x2_t __b = (b); float32x4_t __c = (c); \
-  __a + (__b * __builtin_shufflevector(__c, __c, __d, __d)); })
-#define vmlaq_laneq_s16(a, b, c, __d) __extension__ ({ \
-  int16x8_t __a = (a); int16x8_t __b = (b); int16x8_t __c = (c); \
-  __a + (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d, __d, __d, __d, __d)); })
-#define vmlaq_laneq_s32(a, b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int32x4_t __b = (b); int32x4_t __c = (c); \
-  __a + (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlaq_laneq_u16(a, b, c, __d) __extension__ ({ \
-  uint16x8_t __a = (a); uint16x8_t __b = (b); uint16x8_t __c = (c); \
-  __a + (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d, __d, __d, __d, __d)); })
-#define vmlaq_laneq_u32(a, b, c, __d) __extension__ ({ \
-  uint32x4_t __a = (a); uint32x4_t __b = (b); uint32x4_t __c = (c); \
-  __a + (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlaq_laneq_f32(a, b, c, __d) __extension__ ({ \
-  float32x4_t __a = (a); float32x4_t __b = (b); float32x4_t __c = (c); \
-  __a + (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-
-#endif
-
-__ai int16x4_t vmla_n_s16(int16x4_t __a, int16x4_t __b, int16_t __c) {
-  return __a + (__b * (int16x4_t){ __c, __c, __c, __c }); }
-__ai int32x2_t vmla_n_s32(int32x2_t __a, int32x2_t __b, int32_t __c) {
-  return __a + (__b * (int32x2_t){ __c, __c }); }
-__ai uint16x4_t vmla_n_u16(uint16x4_t __a, uint16x4_t __b, uint16_t __c) {
-  return __a + (__b * (uint16x4_t){ __c, __c, __c, __c }); }
-__ai uint32x2_t vmla_n_u32(uint32x2_t __a, uint32x2_t __b, uint32_t __c) {
-  return __a + (__b * (uint32x2_t){ __c, __c }); }
-__ai float32x2_t vmla_n_f32(float32x2_t __a, float32x2_t __b, float32_t __c) {
-  return __a + (__b * (float32x2_t){ __c, __c }); }
-__ai int16x8_t vmlaq_n_s16(int16x8_t __a, int16x8_t __b, int16_t __c) {
-  return __a + (__b * (int16x8_t){ __c, __c, __c, __c, __c, __c, __c, __c }); }
-__ai int32x4_t vmlaq_n_s32(int32x4_t __a, int32x4_t __b, int32_t __c) {
-  return __a + (__b * (int32x4_t){ __c, __c, __c, __c }); }
-__ai uint16x8_t vmlaq_n_u16(uint16x8_t __a, uint16x8_t __b, uint16_t __c) {
-  return __a + (__b * (uint16x8_t){ __c, __c, __c, __c, __c, __c, __c, __c }); }
-__ai uint32x4_t vmlaq_n_u32(uint32x4_t __a, uint32x4_t __b, uint32_t __c) {
-  return __a + (__b * (uint32x4_t){ __c, __c, __c, __c }); }
-__ai float32x4_t vmlaq_n_f32(float32x4_t __a, float32x4_t __b, float32_t __c) {
-  return __a + (__b * (float32x4_t){ __c, __c, __c, __c }); }
-
-__ai int8x8_t vmls_s8(int8x8_t __a, int8x8_t __b, int8x8_t __c) {
-  return __a - (__b * __c); }
-__ai int16x4_t vmls_s16(int16x4_t __a, int16x4_t __b, int16x4_t __c) {
-  return __a - (__b * __c); }
-__ai int32x2_t vmls_s32(int32x2_t __a, int32x2_t __b, int32x2_t __c) {
-  return __a - (__b * __c); }
-__ai float32x2_t vmls_f32(float32x2_t __a, float32x2_t __b, float32x2_t __c) {
-  return __a - (__b * __c); }
-__ai uint8x8_t vmls_u8(uint8x8_t __a, uint8x8_t __b, uint8x8_t __c) {
-  return __a - (__b * __c); }
-__ai uint16x4_t vmls_u16(uint16x4_t __a, uint16x4_t __b, uint16x4_t __c) {
-  return __a - (__b * __c); }
-__ai uint32x2_t vmls_u32(uint32x2_t __a, uint32x2_t __b, uint32x2_t __c) {
-  return __a - (__b * __c); }
-__ai int8x16_t vmlsq_s8(int8x16_t __a, int8x16_t __b, int8x16_t __c) {
-  return __a - (__b * __c); }
-__ai int16x8_t vmlsq_s16(int16x8_t __a, int16x8_t __b, int16x8_t __c) {
-  return __a - (__b * __c); }
-__ai int32x4_t vmlsq_s32(int32x4_t __a, int32x4_t __b, int32x4_t __c) {
-  return __a - (__b * __c); }
-__ai float32x4_t vmlsq_f32(float32x4_t __a, float32x4_t __b, float32x4_t __c) {
-  return __a - (__b * __c); }
-__ai uint8x16_t vmlsq_u8(uint8x16_t __a, uint8x16_t __b, uint8x16_t __c) {
-  return __a - (__b * __c); }
-__ai uint16x8_t vmlsq_u16(uint16x8_t __a, uint16x8_t __b, uint16x8_t __c) {
-  return __a - (__b * __c); }
-__ai uint32x4_t vmlsq_u32(uint32x4_t __a, uint32x4_t __b, uint32x4_t __c) {
-  return __a - (__b * __c); }
-
-__ai int16x8_t vmlsl_s8(int16x8_t __a, int8x8_t __b, int8x8_t __c) {
-  return __a - vmull_s8(__b, __c); }
-__ai int32x4_t vmlsl_s16(int32x4_t __a, int16x4_t __b, int16x4_t __c) {
-  return __a - vmull_s16(__b, __c); }
-__ai int64x2_t vmlsl_s32(int64x2_t __a, int32x2_t __b, int32x2_t __c) {
-  return __a - vmull_s32(__b, __c); }
-__ai uint16x8_t vmlsl_u8(uint16x8_t __a, uint8x8_t __b, uint8x8_t __c) {
-  return __a - vmull_u8(__b, __c); }
-__ai uint32x4_t vmlsl_u16(uint32x4_t __a, uint16x4_t __b, uint16x4_t __c) {
-  return __a - vmull_u16(__b, __c); }
-__ai uint64x2_t vmlsl_u32(uint64x2_t __a, uint32x2_t __b, uint32x2_t __c) {
-  return __a - vmull_u32(__b, __c); }
+#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
 
-#if defined(__aarch64__)
-__ai int16x8_t vmlsl_high_s8(int16x8_t __a, int8x16_t __b, int8x16_t __c) {
-  return vmlsl_s8(__a, vget_high_s8(__b), vget_high_s8(__c)); }
-__ai int32x4_t vmlsl_high_s16(int32x4_t __a, int16x8_t __b, int16x8_t __c) {
-  return vmlsl_s16(__a, vget_high_s16(__b), vget_high_s16(__c)); }
-__ai int64x2_t vmlsl_high_s32(int64x2_t __a, int32x4_t __b, int32x4_t __c) {
-  return vmlsl_s32(__a, vget_high_s32(__b), vget_high_s32(__c)); }
-__ai uint16x8_t vmlsl_high_u8(uint16x8_t __a, uint8x16_t __b, uint8x16_t __c) {
-  return vmlsl_u8(__a, vget_high_u8(__b), vget_high_u8(__c)); }
-__ai uint32x4_t vmlsl_high_u16(uint32x4_t __a, uint16x8_t __b, uint16x8_t __c) {
-  return vmlsl_u16(__a, vget_high_u16(__b), vget_high_u16(__c)); }
-__ai uint64x2_t vmlsl_high_u32(uint64x2_t __a, uint32x4_t __b, uint32x4_t __c) {
-  return vmlsl_u32(__a, vget_high_u32(__b), vget_high_u32(__c)); }
-
-#define vmlsl_high_lane_s16(a, b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int16x8_t __b = (b); int16x4_t __c = (c); \
-  __a - vmull_s16(vget_high_s16(__b), __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlsl_high_lane_s32(a, b, c, __d) __extension__ ({ \
-  int64x2_t __a = (a); int32x4_t __b = (b); int32x2_t __c = (c); \
-  __a - vmull_s32(vget_high_s32(__b), __builtin_shufflevector(__c, __c, __d, __d)); })
-#define vmlsl_high_lane_u16(a, b, c, __d) __extension__ ({ \
-  uint32x4_t __a = (a); uint16x8_t __b = (b); uint16x4_t __c = (c); \
-  __a - vmull_u16(vget_high_u16(__b), __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlsl_high_lane_u32(a, b, c, __d) __extension__ ({ \
-  uint64x2_t __a = (a); uint32x4_t __b = (b); uint32x2_t __c = (c); \
-  __a - vmull_u32(vget_high_u32(__b), __builtin_shufflevector(__c, __c, __d, __d)); })
-
-#define vmlsl_high_laneq_s16(a, b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int16x8_t __b = (b); int16x8_t __c = (c); \
-  __a - vmull_s16(vget_high_s16(__b), __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlsl_high_laneq_s32(a, b, c, __d) __extension__ ({ \
-  int64x2_t __a = (a); int32x4_t __b = (b); int32x4_t __c = (c); \
-  __a - vmull_s32(vget_high_s32(__b), __builtin_shufflevector(__c, __c, __d, __d)); })
-#define vmlsl_high_laneq_u16(a, b, c, __d) __extension__ ({ \
-  uint32x4_t __a = (a); uint16x8_t __b = (b); uint16x8_t __c = (c); \
-  __a - vmull_u16(vget_high_u16(__b), __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlsl_high_laneq_u32(a, b, c, __d) __extension__ ({ \
-  uint64x2_t __a = (a); uint32x4_t __b = (b); uint32x4_t __c = (c); \
-  __a - vmull_u32(vget_high_u32(__b), __builtin_shufflevector(__c, __c, __d, __d)); })
-
-__ai int32x4_t vmlsl_high_n_s16(int32x4_t __a, int16x8_t __b, int16_t __c) {
-  return vmlsl_n_s16(__a, vget_high_s16(__b), __c); }
-__ai int64x2_t vmlsl_high_n_s32(int64x2_t __a, int32x4_t __b, int32_t __c) {
-  return vmlsl_n_s32(__a, vget_high_s32(__b), __c); }
-__ai uint32x4_t vmlsl_high_n_u16(uint32x4_t __a, uint16x8_t __b, uint16_t __c) {
-  return vmlsl_n_u16(__a, vget_high_u16(__b), __c); }
-__ai uint64x2_t vmlsl_high_n_u32(uint64x2_t __a, uint32x4_t __b, uint32_t __c) {
-  return vmlsl_n_u32(__a, vget_high_u32(__b), __c); }
-
-#endif
-
-#define vmlsl_lane_s16(a, b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int16x4_t __b = (b); int16x4_t __c = (c); \
-  __a - vmull_s16(__b, __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlsl_lane_s32(a, b, c, __d) __extension__ ({ \
-  int64x2_t __a = (a); int32x2_t __b = (b); int32x2_t __c = (c); \
-  __a - vmull_s32(__b, __builtin_shufflevector(__c, __c, __d, __d)); })
-#define vmlsl_lane_u16(a, b, c, __d) __extension__ ({ \
-  uint32x4_t __a = (a); uint16x4_t __b = (b); uint16x4_t __c = (c); \
-  __a - vmull_u16(__b, __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlsl_lane_u32(a, b, c, __d) __extension__ ({ \
-  uint64x2_t __a = (a); uint32x2_t __b = (b); uint32x2_t __c = (c); \
-  __a - vmull_u32(__b, __builtin_shufflevector(__c, __c, __d, __d)); })
+#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
 
-#if defined(__aarch64__)
-#define vmlsl_laneq_s16(a, b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int16x4_t __b = (b); int16x8_t __c = (c); \
-  __a - vmull_s16(__b, __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlsl_laneq_s32(a, b, c, __d) __extension__ ({ \
-  int64x2_t __a = (a); int32x2_t __b = (b); int32x4_t __c = (c); \
-  __a - vmull_s32(__b, __builtin_shufflevector(__c, __c, __d, __d)); })
-#define vmlsl_laneq_u16(a, b, c, __d) __extension__ ({ \
-  uint32x4_t __a = (a); uint16x4_t __b = (b); uint16x8_t __c = (c); \
-  __a - vmull_u16(__b, __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlsl_laneq_u32(a, b, c, __d) __extension__ ({ \
-  uint64x2_t __a = (a); uint32x2_t __b = (b); uint32x4_t __c = (c); \
-  __a - vmull_u32(__b, __builtin_shufflevector(__c, __c, __d, __d)); })
-
-#endif
-
-#define vmls_lane_s16(a, b, c, __d) __extension__ ({ \
-  int16x4_t __a = (a); int16x4_t __b = (b); int16x4_t __c = (c); \
-  __a - (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmls_lane_s32(a, b, c, __d) __extension__ ({ \
-  int32x2_t __a = (a); int32x2_t __b = (b); int32x2_t __c = (c); \
-  __a - (__b * __builtin_shufflevector(__c, __c, __d, __d)); })
-#define vmls_lane_u16(a, b, c, __d) __extension__ ({ \
-  uint16x4_t __a = (a); uint16x4_t __b = (b); uint16x4_t __c = (c); \
-  __a - (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmls_lane_u32(a, b, c, __d) __extension__ ({ \
-  uint32x2_t __a = (a); uint32x2_t __b = (b); uint32x2_t __c = (c); \
-  __a - (__b * __builtin_shufflevector(__c, __c, __d, __d)); })
-#define vmls_lane_f32(a, b, c, __d) __extension__ ({ \
-  float32x2_t __a = (a); float32x2_t __b = (b); float32x2_t __c = (c); \
-  __a - (__b * __builtin_shufflevector(__c, __c, __d, __d)); })
-#define vmlsq_lane_s16(a, b, c, __d) __extension__ ({ \
-  int16x8_t __a = (a); int16x8_t __b = (b); int16x4_t __c = (c); \
-  __a - (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d, __d, __d, __d, __d)); })
-#define vmlsq_lane_s32(a, b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int32x4_t __b = (b); int32x2_t __c = (c); \
-  __a - (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlsq_lane_u16(a, b, c, __d) __extension__ ({ \
-  uint16x8_t __a = (a); uint16x8_t __b = (b); uint16x4_t __c = (c); \
-  __a - (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d, __d, __d, __d, __d)); })
-#define vmlsq_lane_u32(a, b, c, __d) __extension__ ({ \
-  uint32x4_t __a = (a); uint32x4_t __b = (b); uint32x2_t __c = (c); \
-  __a - (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlsq_lane_f32(a, b, c, __d) __extension__ ({ \
-  float32x4_t __a = (a); float32x4_t __b = (b); float32x2_t __c = (c); \
-  __a - (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
+#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
 
-#if defined(__aarch64__)
-#define vmls_laneq_s16(a, b, c, __d) __extension__ ({ \
-  int16x4_t __a = (a); int16x4_t __b = (b); int16x8_t __c = (c); \
-  __a - (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmls_laneq_s32(a, b, c, __d) __extension__ ({ \
-  int32x2_t __a = (a); int32x2_t __b = (b); int32x4_t __c = (c); \
-  __a - (__b * __builtin_shufflevector(__c, __c, __d, __d)); })
-#define vmls_laneq_u16(a, b, c, __d) __extension__ ({ \
-  uint16x4_t __a = (a); uint16x4_t __b = (b); uint16x8_t __c = (c); \
-  __a - (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmls_laneq_u32(a, b, c, __d) __extension__ ({ \
-  uint32x2_t __a = (a); uint32x2_t __b = (b); uint32x4_t __c = (c); \
-  __a - (__b * __builtin_shufflevector(__c, __c, __d, __d)); })
-#define vmls_laneq_f32(a, b, c, __d) __extension__ ({ \
-  float32x2_t __a = (a); float32x2_t __b = (b); float32x4_t __c = (c); \
-  __a - (__b * __builtin_shufflevector(__c, __c, __d, __d)); })
-#define vmlsq_laneq_s16(a, b, c, __d) __extension__ ({ \
-  int16x8_t __a = (a); int16x8_t __b = (b); int16x8_t __c = (c); \
-  __a - (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d, __d, __d, __d, __d)); })
-#define vmlsq_laneq_s32(a, b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int32x4_t __b = (b); int32x4_t __c = (c); \
-  __a - (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlsq_laneq_u16(a, b, c, __d) __extension__ ({ \
-  uint16x8_t __a = (a); uint16x8_t __b = (b); uint16x8_t __c = (c); \
-  __a - (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d, __d, __d, __d, __d)); })
-#define vmlsq_laneq_u32(a, b, c, __d) __extension__ ({ \
-  uint32x4_t __a = (a); uint32x4_t __b = (b); uint32x4_t __c = (c); \
-  __a - (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vmlsq_laneq_f32(a, b, c, __d) __extension__ ({ \
-  float32x4_t __a = (a); float32x4_t __b = (b); float32x4_t __c = (c); \
-  __a - (__b * __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-
-#endif
-
-__ai int16x4_t vmls_n_s16(int16x4_t __a, int16x4_t __b, int16_t __c) {
-  return __a - (__b * (int16x4_t){ __c, __c, __c, __c }); }
-__ai int32x2_t vmls_n_s32(int32x2_t __a, int32x2_t __b, int32_t __c) {
-  return __a - (__b * (int32x2_t){ __c, __c }); }
-__ai uint16x4_t vmls_n_u16(uint16x4_t __a, uint16x4_t __b, uint16_t __c) {
-  return __a - (__b * (uint16x4_t){ __c, __c, __c, __c }); }
-__ai uint32x2_t vmls_n_u32(uint32x2_t __a, uint32x2_t __b, uint32_t __c) {
-  return __a - (__b * (uint32x2_t){ __c, __c }); }
-__ai float32x2_t vmls_n_f32(float32x2_t __a, float32x2_t __b, float32_t __c) {
-  return __a - (__b * (float32x2_t){ __c, __c }); }
-__ai int16x8_t vmlsq_n_s16(int16x8_t __a, int16x8_t __b, int16_t __c) {
-  return __a - (__b * (int16x8_t){ __c, __c, __c, __c, __c, __c, __c, __c }); }
-__ai int32x4_t vmlsq_n_s32(int32x4_t __a, int32x4_t __b, int32_t __c) {
-  return __a - (__b * (int32x4_t){ __c, __c, __c, __c }); }
-__ai uint16x8_t vmlsq_n_u16(uint16x8_t __a, uint16x8_t __b, uint16_t __c) {
-  return __a - (__b * (uint16x8_t){ __c, __c, __c, __c, __c, __c, __c, __c }); }
-__ai uint32x4_t vmlsq_n_u32(uint32x4_t __a, uint32x4_t __b, uint32_t __c) {
-  return __a - (__b * (uint32x4_t){ __c, __c, __c, __c }); }
-__ai float32x4_t vmlsq_n_f32(float32x4_t __a, float32x4_t __b, float32_t __c) {
-  return __a - (__b * (float32x4_t){ __c, __c, __c, __c }); }
-
-__ai int8x8_t vmovn_s16(int16x8_t __a) {
-  return (int8x8_t)__builtin_neon_vmovn_v((int8x16_t)__a, 0); }
-__ai int16x4_t vmovn_s32(int32x4_t __a) {
-  return (int16x4_t)__builtin_neon_vmovn_v((int8x16_t)__a, 1); }
-__ai int32x2_t vmovn_s64(int64x2_t __a) {
-  return (int32x2_t)__builtin_neon_vmovn_v((int8x16_t)__a, 2); }
-__ai uint8x8_t vmovn_u16(uint16x8_t __a) {
-  return (uint8x8_t)__builtin_neon_vmovn_v((int8x16_t)__a, 16); }
-__ai uint16x4_t vmovn_u32(uint32x4_t __a) {
-  return (uint16x4_t)__builtin_neon_vmovn_v((int8x16_t)__a, 17); }
-__ai uint32x2_t vmovn_u64(uint64x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vmovn_v((int8x16_t)__a, 18); }
-
-__ai uint8x8_t vmov_n_u8(uint8_t __a) {
-  return (uint8x8_t){ __a, __a, __a, __a, __a, __a, __a, __a }; }
-__ai uint16x4_t vmov_n_u16(uint16_t __a) {
-  return (uint16x4_t){ __a, __a, __a, __a }; }
-__ai uint32x2_t vmov_n_u32(uint32_t __a) {
-  return (uint32x2_t){ __a, __a }; }
-__ai int8x8_t vmov_n_s8(int8_t __a) {
-  return (int8x8_t){ __a, __a, __a, __a, __a, __a, __a, __a }; }
-__ai int16x4_t vmov_n_s16(int16_t __a) {
-  return (int16x4_t){ __a, __a, __a, __a }; }
-__ai int32x2_t vmov_n_s32(int32_t __a) {
-  return (int32x2_t){ __a, __a }; }
-__ai poly8x8_t vmov_n_p8(poly8_t __a) {
-  return (poly8x8_t){ __a, __a, __a, __a, __a, __a, __a, __a }; }
-__ai poly16x4_t vmov_n_p16(poly16_t __a) {
-  return (poly16x4_t){ __a, __a, __a, __a }; }
-#define vmov_n_f16(a) __extension__ ({ \
-  float16_t __a = (a); \
-  (float16x4_t){ __a, __a, __a, __a }; })
-__ai float32x2_t vmov_n_f32(float32_t __a) {
-  return (float32x2_t){ __a, __a }; }
-__ai uint8x16_t vmovq_n_u8(uint8_t __a) {
-  return (uint8x16_t){ __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a }; }
-__ai uint16x8_t vmovq_n_u16(uint16_t __a) {
-  return (uint16x8_t){ __a, __a, __a, __a, __a, __a, __a, __a }; }
-__ai uint32x4_t vmovq_n_u32(uint32_t __a) {
-  return (uint32x4_t){ __a, __a, __a, __a }; }
-__ai int8x16_t vmovq_n_s8(int8_t __a) {
-  return (int8x16_t){ __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a }; }
-__ai int16x8_t vmovq_n_s16(int16_t __a) {
-  return (int16x8_t){ __a, __a, __a, __a, __a, __a, __a, __a }; }
-__ai int32x4_t vmovq_n_s32(int32_t __a) {
-  return (int32x4_t){ __a, __a, __a, __a }; }
-__ai poly8x16_t vmovq_n_p8(poly8_t __a) {
-  return (poly8x16_t){ __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a, __a }; }
-__ai poly16x8_t vmovq_n_p16(poly16_t __a) {
-  return (poly16x8_t){ __a, __a, __a, __a, __a, __a, __a, __a }; }
-#define vmovq_n_f16(a) __extension__ ({ \
-  float16_t __a = (a); \
-  (float16x8_t){ __a, __a, __a, __a, __a, __a, __a, __a }; })
-__ai float32x4_t vmovq_n_f32(float32_t __a) {
-  return (float32x4_t){ __a, __a, __a, __a }; }
-__ai int64x1_t vmov_n_s64(int64_t __a) {
-  return (int64x1_t){ __a }; }
-__ai uint64x1_t vmov_n_u64(uint64_t __a) {
-  return (uint64x1_t){ __a }; }
-__ai int64x2_t vmovq_n_s64(int64_t __a) {
-  return (int64x2_t){ __a, __a }; }
-__ai uint64x2_t vmovq_n_u64(uint64_t __a) {
-  return (uint64x2_t){ __a, __a }; }
-
-__ai int8x8_t vmul_s8(int8x8_t __a, int8x8_t __b) {
-  return __a * __b; }
-__ai int16x4_t vmul_s16(int16x4_t __a, int16x4_t __b) {
-  return __a * __b; }
-__ai int32x2_t vmul_s32(int32x2_t __a, int32x2_t __b) {
-  return __a * __b; }
-__ai float32x2_t vmul_f32(float32x2_t __a, float32x2_t __b) {
-  return __a * __b; }
-__ai uint8x8_t vmul_u8(uint8x8_t __a, uint8x8_t __b) {
-  return __a * __b; }
-__ai uint16x4_t vmul_u16(uint16x4_t __a, uint16x4_t __b) {
-  return __a * __b; }
-__ai uint32x2_t vmul_u32(uint32x2_t __a, uint32x2_t __b) {
-  return __a * __b; }
-__ai int8x16_t vmulq_s8(int8x16_t __a, int8x16_t __b) {
-  return __a * __b; }
-__ai int16x8_t vmulq_s16(int16x8_t __a, int16x8_t __b) {
-  return __a * __b; }
-__ai int32x4_t vmulq_s32(int32x4_t __a, int32x4_t __b) {
-  return __a * __b; }
-__ai float32x4_t vmulq_f32(float32x4_t __a, float32x4_t __b) {
-  return __a * __b; }
-__ai uint8x16_t vmulq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return __a * __b; }
-__ai uint16x8_t vmulq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return __a * __b; }
-__ai uint32x4_t vmulq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return __a * __b; }
+#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
 
-#if defined(__aarch64__)
-__ai int16x8_t vmull_high_s8(int8x16_t __a, int8x16_t __b) {
-  return vmull_s8(vget_high_s8(__a), vget_high_s8(__b)); }
-__ai int32x4_t vmull_high_s16(int16x8_t __a, int16x8_t __b) {
-  return vmull_s16(vget_high_s16(__a), vget_high_s16(__b)); }
-__ai int64x2_t vmull_high_s32(int32x4_t __a, int32x4_t __b) {
-  return vmull_s32(vget_high_s32(__a), vget_high_s32(__b)); }
-__ai uint16x8_t vmull_high_u8(uint8x16_t __a, uint8x16_t __b) {
-  return vmull_u8(vget_high_u8(__a), vget_high_u8(__b)); }
-__ai uint32x4_t vmull_high_u16(uint16x8_t __a, uint16x8_t __b) {
-  return vmull_u16(vget_high_u16(__a), vget_high_u16(__b)); }
-__ai uint64x2_t vmull_high_u32(uint32x4_t __a, uint32x4_t __b) {
-  return vmull_u32(vget_high_u32(__a), vget_high_u32(__b)); }
-__ai poly16x8_t vmull_high_p8(poly8x16_t __a, poly8x16_t __b) {
-  return vmull_p8(vget_high_p8(__a), vget_high_p8(__b)); }
-
-#define vmull_high_lane_s16(a, b, __c) __extension__ ({ \
-  int16x8_t __a = (a); int16x4_t __b = (b); \
-  vmull_s16(vget_high_s16(__a), __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
-#define vmull_high_lane_s32(a, b, __c) __extension__ ({ \
-  int32x4_t __a = (a); int32x2_t __b = (b); \
-  vmull_s32(vget_high_s32(__a), __builtin_shufflevector(__b, __b, __c, __c)); })
-#define vmull_high_lane_u16(a, b, __c) __extension__ ({ \
-  uint16x8_t __a = (a); uint16x4_t __b = (b); \
-  vmull_u16(vget_high_u16(__a), __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
-#define vmull_high_lane_u32(a, b, __c) __extension__ ({ \
-  uint32x4_t __a = (a); uint32x2_t __b = (b); \
-  vmull_u32(vget_high_u32(__a), __builtin_shufflevector(__b, __b, __c, __c)); })
-
-#define vmull_high_laneq_s16(a, b, __c) __extension__ ({ \
-  int16x8_t __a = (a); int16x8_t __b = (b); \
-  vmull_s16(vget_high_s16(__a), __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
-#define vmull_high_laneq_s32(a, b, __c) __extension__ ({ \
-  int32x4_t __a = (a); int32x4_t __b = (b); \
-  vmull_s32(vget_high_s32(__a), __builtin_shufflevector(__b, __b, __c, __c)); })
-#define vmull_high_laneq_u16(a, b, __c) __extension__ ({ \
-  uint16x8_t __a = (a); uint16x8_t __b = (b); \
-  vmull_u16(vget_high_u16(__a), __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
-#define vmull_high_laneq_u32(a, b, __c) __extension__ ({ \
-  uint32x4_t __a = (a); uint32x4_t __b = (b); \
-  vmull_u32(vget_high_u32(__a), __builtin_shufflevector(__b, __b, __c, __c)); })
-
-__ai int32x4_t vmull_high_n_s16(int16x8_t __a, int16_t __b) {
-  return vmull_n_s16(vget_high_s16(__a), __b); }
-__ai int64x2_t vmull_high_n_s32(int32x4_t __a, int32_t __b) {
-  return vmull_n_s32(vget_high_s32(__a), __b); }
-__ai uint32x4_t vmull_high_n_u16(uint16x8_t __a, uint16_t __b) {
-  return vmull_n_u16(vget_high_u16(__a), __b); }
-__ai uint64x2_t vmull_high_n_u32(uint32x4_t __a, uint32_t __b) {
-  return vmull_n_u32(vget_high_u32(__a), __b); }
-
-__ai poly128_t vmull_high_p64(poly64x2_t __a, poly64x2_t __b) {
-  return vmull_p64((poly64_t)vget_high_p64(__a), (poly64_t)vget_high_p64(__b)); }
-
-#endif
-
-#define vmull_lane_s16(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int16x4_t __b = (b); \
-  vmull_s16(__a, __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
-#define vmull_lane_s32(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int32x2_t __b = (b); \
-  vmull_s32(__a, __builtin_shufflevector(__b, __b, __c, __c)); })
-#define vmull_lane_u16(a, b, __c) __extension__ ({ \
-  uint16x4_t __a = (a); uint16x4_t __b = (b); \
-  vmull_u16(__a, __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
-#define vmull_lane_u32(a, b, __c) __extension__ ({ \
-  uint32x2_t __a = (a); uint32x2_t __b = (b); \
-  vmull_u32(__a, __builtin_shufflevector(__b, __b, __c, __c)); })
+#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
 
-#if defined(__aarch64__)
-#define vmull_laneq_s16(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int16x8_t __b = (b); \
-  vmull_s16(__a, __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
-#define vmull_laneq_s32(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int32x4_t __b = (b); \
-  vmull_s32(__a, __builtin_shufflevector(__b, __b, __c, __c)); })
-#define vmull_laneq_u16(a, b, __c) __extension__ ({ \
-  uint16x4_t __a = (a); uint16x8_t __b = (b); \
-  vmull_u16(__a, __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
-#define vmull_laneq_u32(a, b, __c) __extension__ ({ \
-  uint32x2_t __a = (a); uint32x4_t __b = (b); \
-  vmull_u32(__a, __builtin_shufflevector(__b, __b, __c, __c)); })
-
-#endif
-
-__ai poly8x8_t vmul_p8(poly8x8_t __a, poly8x8_t __b) {
-  return (poly8x8_t)__builtin_neon_vmul_v((int8x8_t)__a, (int8x8_t)__b, 4); }
-__ai poly8x16_t vmulq_p8(poly8x16_t __a, poly8x16_t __b) {
-  return (poly8x16_t)__builtin_neon_vmulq_v((int8x16_t)__a, (int8x16_t)__b, 36); }
+#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
 
-#if defined(__aarch64__)
-#define vmulx_lane_f32(a, b, __c) __extension__ ({ \
-  float32x2_t __a = (a); float32x2_t __b = (b); \
-  vmulx_f32(__a, __builtin_shufflevector(__b, __b, __c, __c)); })
-#define vmulxq_lane_f32(a, b, __c) __extension__ ({ \
-  float32x4_t __a = (a); float32x2_t __b = (b); \
-  vmulxq_f32(__a, __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
-#define vmulxq_lane_f64(a, b, __c) __extension__ ({ \
-  float64x2_t __a = (a); float64x1_t __b = (b); \
-  vmulxq_f64(__a, __builtin_shufflevector(__b, __b, __c, __c)); })
-
-#define vmulx_laneq_f32(a, b, __c) __extension__ ({ \
-  float32x2_t __a = (a); float32x4_t __b = (b); \
-  vmulx_f32(__a, __builtin_shufflevector(__b, __b, __c, __c)); })
-#define vmulxq_laneq_f32(a, b, __c) __extension__ ({ \
-  float32x4_t __a = (a); float32x4_t __b = (b); \
-  vmulxq_f32(__a, __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
-#define vmulxq_laneq_f64(a, b, __c) __extension__ ({ \
-  float64x2_t __a = (a); float64x2_t __b = (b); \
-  vmulxq_f64(__a, __builtin_shufflevector(__b, __b, __c, __c)); })
-
-#endif
-
-#define vmul_lane_s16(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int16x4_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c, __c, __c); })
-#define vmul_lane_s32(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int32x2_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c); })
-#define vmul_lane_f32(a, b, __c) __extension__ ({ \
-  float32x2_t __a = (a); float32x2_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c); })
-#define vmul_lane_u16(a, b, __c) __extension__ ({ \
-  uint16x4_t __a = (a); uint16x4_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c, __c, __c); })
-#define vmul_lane_u32(a, b, __c) __extension__ ({ \
-  uint32x2_t __a = (a); uint32x2_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c); })
-#define vmulq_lane_s16(a, b, __c) __extension__ ({ \
-  int16x8_t __a = (a); int16x4_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c, __c, __c, __c, __c, __c, __c); })
-#define vmulq_lane_s32(a, b, __c) __extension__ ({ \
-  int32x4_t __a = (a); int32x2_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c, __c, __c); })
-#define vmulq_lane_f32(a, b, __c) __extension__ ({ \
-  float32x4_t __a = (a); float32x2_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c, __c, __c); })
-#define vmulq_lane_u16(a, b, __c) __extension__ ({ \
-  uint16x8_t __a = (a); uint16x4_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c, __c, __c, __c, __c, __c, __c); })
-#define vmulq_lane_u32(a, b, __c) __extension__ ({ \
-  uint32x4_t __a = (a); uint32x2_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c, __c, __c); })
+#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
 
-#if defined(__aarch64__)
-#define vmul_laneq_s16(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int16x8_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c, __c, __c); })
-#define vmul_laneq_s32(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int32x4_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c); })
-#define vmul_laneq_f32(a, b, __c) __extension__ ({ \
-  float32x2_t __a = (a); float32x4_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c); })
-#define vmul_laneq_u16(a, b, __c) __extension__ ({ \
-  uint16x4_t __a = (a); uint16x8_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c, __c, __c); })
-#define vmul_laneq_u32(a, b, __c) __extension__ ({ \
-  uint32x2_t __a = (a); uint32x4_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c); })
-#define vmulq_laneq_s16(a, b, __c) __extension__ ({ \
-  int16x8_t __a = (a); int16x8_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c, __c, __c, __c, __c, __c, __c); })
-#define vmulq_laneq_s32(a, b, __c) __extension__ ({ \
-  int32x4_t __a = (a); int32x4_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c, __c, __c); })
-#define vmulq_laneq_u16(a, b, __c) __extension__ ({ \
-  uint16x8_t __a = (a); uint16x8_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c, __c, __c, __c, __c, __c, __c); })
-#define vmulq_laneq_u32(a, b, __c) __extension__ ({ \
-  uint32x4_t __a = (a); uint32x4_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c, __c, __c); })
-#define vmulq_laneq_f32(a, b, __c) __extension__ ({ \
-  float32x4_t __a = (a); float32x4_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c, __c, __c); })
-#define vmulq_laneq_f64(a, b, __c) __extension__ ({ \
-  float64x2_t __a = (a); float64x2_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c); })
-
-#define vmulq_lane_f64(a, b, __c) __extension__ ({ \
-  float64x2_t __a = (a); float64x1_t __b = (b); \
-  __a * __builtin_shufflevector(__b, __b, __c, __c); })
-
-#endif
-
-__ai int16x4_t vmul_n_s16(int16x4_t __a, int16_t __b) {
-  return __a * (int16x4_t){ __b, __b, __b, __b }; }
-__ai int32x2_t vmul_n_s32(int32x2_t __a, int32_t __b) {
-  return __a * (int32x2_t){ __b, __b }; }
-__ai float32x2_t vmul_n_f32(float32x2_t __a, float32_t __b) {
-  return __a * (float32x2_t){ __b, __b }; }
-__ai uint16x4_t vmul_n_u16(uint16x4_t __a, uint16_t __b) {
-  return __a * (uint16x4_t){ __b, __b, __b, __b }; }
-__ai uint32x2_t vmul_n_u32(uint32x2_t __a, uint32_t __b) {
-  return __a * (uint32x2_t){ __b, __b }; }
-__ai int16x8_t vmulq_n_s16(int16x8_t __a, int16_t __b) {
-  return __a * (int16x8_t){ __b, __b, __b, __b, __b, __b, __b, __b }; }
-__ai int32x4_t vmulq_n_s32(int32x4_t __a, int32_t __b) {
-  return __a * (int32x4_t){ __b, __b, __b, __b }; }
-__ai float32x4_t vmulq_n_f32(float32x4_t __a, float32_t __b) {
-  return __a * (float32x4_t){ __b, __b, __b, __b }; }
-__ai uint16x8_t vmulq_n_u16(uint16x8_t __a, uint16_t __b) {
-  return __a * (uint16x8_t){ __b, __b, __b, __b, __b, __b, __b, __b }; }
-__ai uint32x4_t vmulq_n_u32(uint32x4_t __a, uint32_t __b) {
-  return __a * (uint32x4_t){ __b, __b, __b, __b }; }
+#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
 
-#if defined(__aarch64__)
-__ai float64x2_t vmulq_n_f64(float64x2_t __a, float64_t __b) {
-  return __a * (float64x2_t){ __b, __b }; }
-
-#endif
-
-__ai int8x8_t vmvn_s8(int8x8_t __a) {
-  return  ~__a; }
-__ai int16x4_t vmvn_s16(int16x4_t __a) {
-  return  ~__a; }
-__ai int32x2_t vmvn_s32(int32x2_t __a) {
-  return  ~__a; }
-__ai uint8x8_t vmvn_u8(uint8x8_t __a) {
-  return  ~__a; }
-__ai uint16x4_t vmvn_u16(uint16x4_t __a) {
-  return  ~__a; }
-__ai uint32x2_t vmvn_u32(uint32x2_t __a) {
-  return  ~__a; }
-__ai poly8x8_t vmvn_p8(poly8x8_t __a) {
-  return  ~__a; }
-__ai int8x16_t vmvnq_s8(int8x16_t __a) {
-  return  ~__a; }
-__ai int16x8_t vmvnq_s16(int16x8_t __a) {
-  return  ~__a; }
-__ai int32x4_t vmvnq_s32(int32x4_t __a) {
-  return  ~__a; }
-__ai uint8x16_t vmvnq_u8(uint8x16_t __a) {
-  return  ~__a; }
-__ai uint16x8_t vmvnq_u16(uint16x8_t __a) {
-  return  ~__a; }
-__ai uint32x4_t vmvnq_u32(uint32x4_t __a) {
-  return  ~__a; }
-__ai poly8x16_t vmvnq_p8(poly8x16_t __a) {
-  return  ~__a; }
-
-__ai int8x8_t vneg_s8(int8x8_t __a) {
-  return  -__a; }
-__ai int16x4_t vneg_s16(int16x4_t __a) {
-  return  -__a; }
-__ai int32x2_t vneg_s32(int32x2_t __a) {
-  return  -__a; }
-__ai float32x2_t vneg_f32(float32x2_t __a) {
-  return  -__a; }
-__ai int8x16_t vnegq_s8(int8x16_t __a) {
-  return  -__a; }
-__ai int16x8_t vnegq_s16(int16x8_t __a) {
-  return  -__a; }
-__ai int32x4_t vnegq_s32(int32x4_t __a) {
-  return  -__a; }
-__ai float32x4_t vnegq_f32(float32x4_t __a) {
-  return  -__a; }
-
-__ai int8x8_t vorn_s8(int8x8_t __a, int8x8_t __b) {
-  return __a | ~__b; }
-__ai int16x4_t vorn_s16(int16x4_t __a, int16x4_t __b) {
-  return __a | ~__b; }
-__ai int32x2_t vorn_s32(int32x2_t __a, int32x2_t __b) {
-  return __a | ~__b; }
-__ai int64x1_t vorn_s64(int64x1_t __a, int64x1_t __b) {
-  return __a | ~__b; }
-__ai uint8x8_t vorn_u8(uint8x8_t __a, uint8x8_t __b) {
-  return __a | ~__b; }
-__ai uint16x4_t vorn_u16(uint16x4_t __a, uint16x4_t __b) {
-  return __a | ~__b; }
-__ai uint32x2_t vorn_u32(uint32x2_t __a, uint32x2_t __b) {
-  return __a | ~__b; }
-__ai uint64x1_t vorn_u64(uint64x1_t __a, uint64x1_t __b) {
-  return __a | ~__b; }
-__ai int8x16_t vornq_s8(int8x16_t __a, int8x16_t __b) {
-  return __a | ~__b; }
-__ai int16x8_t vornq_s16(int16x8_t __a, int16x8_t __b) {
-  return __a | ~__b; }
-__ai int32x4_t vornq_s32(int32x4_t __a, int32x4_t __b) {
-  return __a | ~__b; }
-__ai int64x2_t vornq_s64(int64x2_t __a, int64x2_t __b) {
-  return __a | ~__b; }
-__ai uint8x16_t vornq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return __a | ~__b; }
-__ai uint16x8_t vornq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return __a | ~__b; }
-__ai uint32x4_t vornq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return __a | ~__b; }
-__ai uint64x2_t vornq_u64(uint64x2_t __a, uint64x2_t __b) {
-  return __a | ~__b; }
-
-__ai int8x8_t vorr_s8(int8x8_t __a, int8x8_t __b) {
-  return __a | __b; }
-__ai int16x4_t vorr_s16(int16x4_t __a, int16x4_t __b) {
-  return __a | __b; }
-__ai int32x2_t vorr_s32(int32x2_t __a, int32x2_t __b) {
-  return __a | __b; }
-__ai int64x1_t vorr_s64(int64x1_t __a, int64x1_t __b) {
-  return __a | __b; }
-__ai uint8x8_t vorr_u8(uint8x8_t __a, uint8x8_t __b) {
-  return __a | __b; }
-__ai uint16x4_t vorr_u16(uint16x4_t __a, uint16x4_t __b) {
-  return __a | __b; }
-__ai uint32x2_t vorr_u32(uint32x2_t __a, uint32x2_t __b) {
-  return __a | __b; }
-__ai uint64x1_t vorr_u64(uint64x1_t __a, uint64x1_t __b) {
-  return __a | __b; }
-__ai int8x16_t vorrq_s8(int8x16_t __a, int8x16_t __b) {
-  return __a | __b; }
-__ai int16x8_t vorrq_s16(int16x8_t __a, int16x8_t __b) {
-  return __a | __b; }
-__ai int32x4_t vorrq_s32(int32x4_t __a, int32x4_t __b) {
-  return __a | __b; }
-__ai int64x2_t vorrq_s64(int64x2_t __a, int64x2_t __b) {
-  return __a | __b; }
-__ai uint8x16_t vorrq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return __a | __b; }
-__ai uint16x8_t vorrq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return __a | __b; }
-__ai uint32x4_t vorrq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return __a | __b; }
-__ai uint64x2_t vorrq_u64(uint64x2_t __a, uint64x2_t __b) {
-  return __a | __b; }
-
-__ai int16x4_t vpadal_s8(int16x4_t __a, int8x8_t __b) {
-  return (int16x4_t)__builtin_neon_vpadal_v((int8x8_t)__a, __b, 1); }
-__ai int32x2_t vpadal_s16(int32x2_t __a, int16x4_t __b) {
-  return (int32x2_t)__builtin_neon_vpadal_v((int8x8_t)__a, (int8x8_t)__b, 2); }
-__ai int64x1_t vpadal_s32(int64x1_t __a, int32x2_t __b) {
-  return (int64x1_t)__builtin_neon_vpadal_v((int8x8_t)__a, (int8x8_t)__b, 3); }
-__ai uint16x4_t vpadal_u8(uint16x4_t __a, uint8x8_t __b) {
-  return (uint16x4_t)__builtin_neon_vpadal_v((int8x8_t)__a, (int8x8_t)__b, 17); }
-__ai uint32x2_t vpadal_u16(uint32x2_t __a, uint16x4_t __b) {
-  return (uint32x2_t)__builtin_neon_vpadal_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai uint64x1_t vpadal_u32(uint64x1_t __a, uint32x2_t __b) {
-  return (uint64x1_t)__builtin_neon_vpadal_v((int8x8_t)__a, (int8x8_t)__b, 19); }
-__ai int16x8_t vpadalq_s8(int16x8_t __a, int8x16_t __b) {
-  return (int16x8_t)__builtin_neon_vpadalq_v((int8x16_t)__a, __b, 33); }
-__ai int32x4_t vpadalq_s16(int32x4_t __a, int16x8_t __b) {
-  return (int32x4_t)__builtin_neon_vpadalq_v((int8x16_t)__a, (int8x16_t)__b, 34); }
-__ai int64x2_t vpadalq_s32(int64x2_t __a, int32x4_t __b) {
-  return (int64x2_t)__builtin_neon_vpadalq_v((int8x16_t)__a, (int8x16_t)__b, 35); }
-__ai uint16x8_t vpadalq_u8(uint16x8_t __a, uint8x16_t __b) {
-  return (uint16x8_t)__builtin_neon_vpadalq_v((int8x16_t)__a, (int8x16_t)__b, 49); }
-__ai uint32x4_t vpadalq_u16(uint32x4_t __a, uint16x8_t __b) {
-  return (uint32x4_t)__builtin_neon_vpadalq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-__ai uint64x2_t vpadalq_u32(uint64x2_t __a, uint32x4_t __b) {
-  return (uint64x2_t)__builtin_neon_vpadalq_v((int8x16_t)__a, (int8x16_t)__b, 51); }
-
-__ai int8x8_t vpadd_s8(int8x8_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vpadd_v(__a, __b, 0); }
-__ai int16x4_t vpadd_s16(int16x4_t __a, int16x4_t __b) {
-  return (int16x4_t)__builtin_neon_vpadd_v((int8x8_t)__a, (int8x8_t)__b, 1); }
-__ai int32x2_t vpadd_s32(int32x2_t __a, int32x2_t __b) {
-  return (int32x2_t)__builtin_neon_vpadd_v((int8x8_t)__a, (int8x8_t)__b, 2); }
-__ai uint8x8_t vpadd_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vpadd_v((int8x8_t)__a, (int8x8_t)__b, 16); }
-__ai uint16x4_t vpadd_u16(uint16x4_t __a, uint16x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vpadd_v((int8x8_t)__a, (int8x8_t)__b, 17); }
-__ai uint32x2_t vpadd_u32(uint32x2_t __a, uint32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vpadd_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai float32x2_t vpadd_f32(float32x2_t __a, float32x2_t __b) {
-  return (float32x2_t)__builtin_neon_vpadd_v((int8x8_t)__a, (int8x8_t)__b, 9); }
-
-__ai int16x4_t vpaddl_s8(int8x8_t __a) {
-  return (int16x4_t)__builtin_neon_vpaddl_v(__a, 1); }
-__ai int32x2_t vpaddl_s16(int16x4_t __a) {
-  return (int32x2_t)__builtin_neon_vpaddl_v((int8x8_t)__a, 2); }
-__ai int64x1_t vpaddl_s32(int32x2_t __a) {
-  return (int64x1_t)__builtin_neon_vpaddl_v((int8x8_t)__a, 3); }
-__ai uint16x4_t vpaddl_u8(uint8x8_t __a) {
-  return (uint16x4_t)__builtin_neon_vpaddl_v((int8x8_t)__a, 17); }
-__ai uint32x2_t vpaddl_u16(uint16x4_t __a) {
-  return (uint32x2_t)__builtin_neon_vpaddl_v((int8x8_t)__a, 18); }
-__ai uint64x1_t vpaddl_u32(uint32x2_t __a) {
-  return (uint64x1_t)__builtin_neon_vpaddl_v((int8x8_t)__a, 19); }
-__ai int16x8_t vpaddlq_s8(int8x16_t __a) {
-  return (int16x8_t)__builtin_neon_vpaddlq_v(__a, 33); }
-__ai int32x4_t vpaddlq_s16(int16x8_t __a) {
-  return (int32x4_t)__builtin_neon_vpaddlq_v((int8x16_t)__a, 34); }
-__ai int64x2_t vpaddlq_s32(int32x4_t __a) {
-  return (int64x2_t)__builtin_neon_vpaddlq_v((int8x16_t)__a, 35); }
-__ai uint16x8_t vpaddlq_u8(uint8x16_t __a) {
-  return (uint16x8_t)__builtin_neon_vpaddlq_v((int8x16_t)__a, 49); }
-__ai uint32x4_t vpaddlq_u16(uint16x8_t __a) {
-  return (uint32x4_t)__builtin_neon_vpaddlq_v((int8x16_t)__a, 50); }
-__ai uint64x2_t vpaddlq_u32(uint32x4_t __a) {
-  return (uint64x2_t)__builtin_neon_vpaddlq_v((int8x16_t)__a, 51); }
-
-__ai int8x8_t vpmax_s8(int8x8_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vpmax_v(__a, __b, 0); }
-__ai int16x4_t vpmax_s16(int16x4_t __a, int16x4_t __b) {
-  return (int16x4_t)__builtin_neon_vpmax_v((int8x8_t)__a, (int8x8_t)__b, 1); }
-__ai int32x2_t vpmax_s32(int32x2_t __a, int32x2_t __b) {
-  return (int32x2_t)__builtin_neon_vpmax_v((int8x8_t)__a, (int8x8_t)__b, 2); }
-__ai uint8x8_t vpmax_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vpmax_v((int8x8_t)__a, (int8x8_t)__b, 16); }
-__ai uint16x4_t vpmax_u16(uint16x4_t __a, uint16x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vpmax_v((int8x8_t)__a, (int8x8_t)__b, 17); }
-__ai uint32x2_t vpmax_u32(uint32x2_t __a, uint32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vpmax_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai float32x2_t vpmax_f32(float32x2_t __a, float32x2_t __b) {
-  return (float32x2_t)__builtin_neon_vpmax_v((int8x8_t)__a, (int8x8_t)__b, 9); }
-
-__ai int8x8_t vpmin_s8(int8x8_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vpmin_v(__a, __b, 0); }
-__ai int16x4_t vpmin_s16(int16x4_t __a, int16x4_t __b) {
-  return (int16x4_t)__builtin_neon_vpmin_v((int8x8_t)__a, (int8x8_t)__b, 1); }
-__ai int32x2_t vpmin_s32(int32x2_t __a, int32x2_t __b) {
-  return (int32x2_t)__builtin_neon_vpmin_v((int8x8_t)__a, (int8x8_t)__b, 2); }
-__ai uint8x8_t vpmin_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vpmin_v((int8x8_t)__a, (int8x8_t)__b, 16); }
-__ai uint16x4_t vpmin_u16(uint16x4_t __a, uint16x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vpmin_v((int8x8_t)__a, (int8x8_t)__b, 17); }
-__ai uint32x2_t vpmin_u32(uint32x2_t __a, uint32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vpmin_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai float32x2_t vpmin_f32(float32x2_t __a, float32x2_t __b) {
-  return (float32x2_t)__builtin_neon_vpmin_v((int8x8_t)__a, (int8x8_t)__b, 9); }
-
-__ai int8x8_t vqabs_s8(int8x8_t __a) {
-  return (int8x8_t)__builtin_neon_vqabs_v(__a, 0); }
-__ai int16x4_t vqabs_s16(int16x4_t __a) {
-  return (int16x4_t)__builtin_neon_vqabs_v((int8x8_t)__a, 1); }
-__ai int32x2_t vqabs_s32(int32x2_t __a) {
-  return (int32x2_t)__builtin_neon_vqabs_v((int8x8_t)__a, 2); }
-__ai int8x16_t vqabsq_s8(int8x16_t __a) {
-  return (int8x16_t)__builtin_neon_vqabsq_v(__a, 32); }
-__ai int16x8_t vqabsq_s16(int16x8_t __a) {
-  return (int16x8_t)__builtin_neon_vqabsq_v((int8x16_t)__a, 33); }
-__ai int32x4_t vqabsq_s32(int32x4_t __a) {
-  return (int32x4_t)__builtin_neon_vqabsq_v((int8x16_t)__a, 34); }
-
-__ai int8x8_t vqadd_s8(int8x8_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vqadd_v(__a, __b, 0); }
-__ai int16x4_t vqadd_s16(int16x4_t __a, int16x4_t __b) {
-  return (int16x4_t)__builtin_neon_vqadd_v((int8x8_t)__a, (int8x8_t)__b, 1); }
-__ai int32x2_t vqadd_s32(int32x2_t __a, int32x2_t __b) {
-  return (int32x2_t)__builtin_neon_vqadd_v((int8x8_t)__a, (int8x8_t)__b, 2); }
-__ai int64x1_t vqadd_s64(int64x1_t __a, int64x1_t __b) {
-  return (int64x1_t)__builtin_neon_vqadd_v((int8x8_t)__a, (int8x8_t)__b, 3); }
-__ai uint8x8_t vqadd_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vqadd_v((int8x8_t)__a, (int8x8_t)__b, 16); }
-__ai uint16x4_t vqadd_u16(uint16x4_t __a, uint16x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vqadd_v((int8x8_t)__a, (int8x8_t)__b, 17); }
-__ai uint32x2_t vqadd_u32(uint32x2_t __a, uint32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vqadd_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai uint64x1_t vqadd_u64(uint64x1_t __a, uint64x1_t __b) {
-  return (uint64x1_t)__builtin_neon_vqadd_v((int8x8_t)__a, (int8x8_t)__b, 19); }
-__ai int8x16_t vqaddq_s8(int8x16_t __a, int8x16_t __b) {
-  return (int8x16_t)__builtin_neon_vqaddq_v(__a, __b, 32); }
-__ai int16x8_t vqaddq_s16(int16x8_t __a, int16x8_t __b) {
-  return (int16x8_t)__builtin_neon_vqaddq_v((int8x16_t)__a, (int8x16_t)__b, 33); }
-__ai int32x4_t vqaddq_s32(int32x4_t __a, int32x4_t __b) {
-  return (int32x4_t)__builtin_neon_vqaddq_v((int8x16_t)__a, (int8x16_t)__b, 34); }
-__ai int64x2_t vqaddq_s64(int64x2_t __a, int64x2_t __b) {
-  return (int64x2_t)__builtin_neon_vqaddq_v((int8x16_t)__a, (int8x16_t)__b, 35); }
-__ai uint8x16_t vqaddq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vqaddq_v((int8x16_t)__a, (int8x16_t)__b, 48); }
-__ai uint16x8_t vqaddq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint16x8_t)__builtin_neon_vqaddq_v((int8x16_t)__a, (int8x16_t)__b, 49); }
-__ai uint32x4_t vqaddq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vqaddq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-__ai uint64x2_t vqaddq_u64(uint64x2_t __a, uint64x2_t __b) {
-  return (uint64x2_t)__builtin_neon_vqaddq_v((int8x16_t)__a, (int8x16_t)__b, 51); }
-
-__ai int32x4_t vqdmlal_s16(int32x4_t __a, int16x4_t __b, int16x4_t __c) {
-  return (int32x4_t)__builtin_neon_vqdmlal_v((int8x16_t)__a, (int8x8_t)__b, (int8x8_t)__c, 34); }
-__ai int64x2_t vqdmlal_s32(int64x2_t __a, int32x2_t __b, int32x2_t __c) {
-  return (int64x2_t)__builtin_neon_vqdmlal_v((int8x16_t)__a, (int8x8_t)__b, (int8x8_t)__c, 35); }
+#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
 
-#if defined(__aarch64__)
-__ai int32x4_t vqdmlal_high_s16(int32x4_t __a, int16x8_t __b, int16x8_t __c) {
-  return vqdmlal_s16(__a, vget_high_s16(__b), vget_high_s16(__c)); }
-__ai int64x2_t vqdmlal_high_s32(int64x2_t __a, int32x4_t __b, int32x4_t __c) {
-  return vqdmlal_s32(__a, vget_high_s32(__b), vget_high_s32(__c)); }
-
-#define vqdmlal_high_lane_s16(a, b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int16x8_t __b = (b); int16x4_t __c = (c); \
-  vqdmlal_s16(__a, vget_high_s16(__b), __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vqdmlal_high_lane_s32(a, b, c, __d) __extension__ ({ \
-  int64x2_t __a = (a); int32x4_t __b = (b); int32x2_t __c = (c); \
-  vqdmlal_s32(__a, vget_high_s32(__b), __builtin_shufflevector(__c, __c, __d, __d)); })
-
-#define vqdmlal_high_laneq_s16(a, b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int16x8_t __b = (b); int16x8_t __c = (c); \
-  vqdmlal_s16(__a, vget_high_s16(__b), __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vqdmlal_high_laneq_s32(a, b, c, __d) __extension__ ({ \
-  int64x2_t __a = (a); int32x4_t __b = (b); int32x4_t __c = (c); \
-  vqdmlal_s32(__a, vget_high_s32(__b), __builtin_shufflevector(__c, __c, __d, __d)); })
-
-__ai int32x4_t vqdmlal_high_n_s16(int32x4_t __a, int16x8_t __b, int16_t __c) {
-  return vqdmlal_n_s16(__a, vget_high_s16(__b), __c); }
-__ai int64x2_t vqdmlal_high_n_s32(int64x2_t __a, int32x4_t __b, int32_t __c) {
-  return vqdmlal_n_s32(__a, vget_high_s32(__b), __c); }
-
-#endif
-
-#define vqdmlal_lane_s16(a, b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int16x4_t __b = (b); int16x4_t __c = (c); \
-  vqdmlal_s16(__a, __b, __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vqdmlal_lane_s32(a, b, c, __d) __extension__ ({ \
-  int64x2_t __a = (a); int32x2_t __b = (b); int32x2_t __c = (c); \
-  vqdmlal_s32(__a, __b, __builtin_shufflevector(__c, __c, __d, __d)); })
+#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
 
-#if defined(__aarch64__)
-#define vqdmlal_laneq_s16(a, b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int16x4_t __b = (b); int16x8_t __c = (c); \
-  vqdmlal_s16(__a, __b, __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vqdmlal_laneq_s32(a, b, c, __d) __extension__ ({ \
-  int64x2_t __a = (a); int32x2_t __b = (b); int32x4_t __c = (c); \
-  vqdmlal_s32(__a, __b, __builtin_shufflevector(__c, __c, __d, __d)); })
+#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
 
-__ai int32x4_t vqdmlsl_s16(int32x4_t __a, int16x4_t __b, int16x4_t __c) {
-  return (int32x4_t)__builtin_neon_vqdmlsl_v((int8x16_t)__a, (int8x8_t)__b, (int8x8_t)__c, 34); }
-__ai int64x2_t vqdmlsl_s32(int64x2_t __a, int32x2_t __b, int32x2_t __c) {
-  return (int64x2_t)__builtin_neon_vqdmlsl_v((int8x16_t)__a, (int8x8_t)__b, (int8x8_t)__c, 35); }
+#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
 
-#if defined(__aarch64__)
-__ai int32x4_t vqdmlsl_high_s16(int32x4_t __a, int16x8_t __b, int16x8_t __c) {
-  return vqdmlsl_s16(__a, vget_high_s16(__b), vget_high_s16(__c)); }
-__ai int64x2_t vqdmlsl_high_s32(int64x2_t __a, int32x4_t __b, int32x4_t __c) {
-  return vqdmlsl_s32(__a, vget_high_s32(__b), vget_high_s32(__c)); }
-
-#define vqdmlsl_high_lane_s16(a, b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int16x8_t __b = (b); int16x4_t __c = (c); \
-  vqdmlsl_s16(__a, vget_high_s16(__b), __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vqdmlsl_high_lane_s32(a, b, c, __d) __extension__ ({ \
-  int64x2_t __a = (a); int32x4_t __b = (b); int32x2_t __c = (c); \
-  vqdmlsl_s32(__a, vget_high_s32(__b), __builtin_shufflevector(__c, __c, __d, __d)); })
-
-#define vqdmlsl_high_laneq_s16(a, b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int16x8_t __b = (b); int16x8_t __c = (c); \
-  vqdmlsl_s16(__a, vget_high_s16(__b), __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vqdmlsl_high_laneq_s32(a, b, c, __d) __extension__ ({ \
-  int64x2_t __a = (a); int32x4_t __b = (b); int32x4_t __c = (c); \
-  vqdmlsl_s32(__a, vget_high_s32(__b), __builtin_shufflevector(__c, __c, __d, __d)); })
-
-__ai int32x4_t vqdmlsl_high_n_s16(int32x4_t __a, int16x8_t __b, int16_t __c) {
-  return vqdmlsl_n_s16(__a, vget_high_s16(__b), __c); }
-__ai int64x2_t vqdmlsl_high_n_s32(int64x2_t __a, int32x4_t __b, int32_t __c) {
-  return vqdmlsl_n_s32(__a, vget_high_s32(__b), __c); }
-
-#endif
-
-#define vqdmlsl_lane_s16(a, b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int16x4_t __b = (b); int16x4_t __c = (c); \
-  vqdmlsl_s16(__a, __b, __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vqdmlsl_lane_s32(a, b, c, __d) __extension__ ({ \
-  int64x2_t __a = (a); int32x2_t __b = (b); int32x2_t __c = (c); \
-  vqdmlsl_s32(__a, __b, __builtin_shufflevector(__c, __c, __d, __d)); })
+#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
 
-#if defined(__aarch64__)
-#define vqdmlsl_laneq_s16(a, b, c, __d) __extension__ ({ \
-  int32x4_t __a = (a); int16x4_t __b = (b); int16x8_t __c = (c); \
-  vqdmlsl_s16(__a, __b, __builtin_shufflevector(__c, __c, __d, __d, __d, __d)); })
-#define vqdmlsl_laneq_s32(a, b, c, __d) __extension__ ({ \
-  int64x2_t __a = (a); int32x2_t __b = (b); int32x4_t __c = (c); \
-  vqdmlsl_s32(__a, __b, __builtin_shufflevector(__c, __c, __d, __d)); })
-
-#endif
-
-__ai int16x4_t vqdmulh_s16(int16x4_t __a, int16x4_t __b) {
-  return (int16x4_t)__builtin_neon_vqdmulh_v((int8x8_t)__a, (int8x8_t)__b, 1); }
-__ai int32x2_t vqdmulh_s32(int32x2_t __a, int32x2_t __b) {
-  return (int32x2_t)__builtin_neon_vqdmulh_v((int8x8_t)__a, (int8x8_t)__b, 2); }
-__ai int16x8_t vqdmulhq_s16(int16x8_t __a, int16x8_t __b) {
-  return (int16x8_t)__builtin_neon_vqdmulhq_v((int8x16_t)__a, (int8x16_t)__b, 33); }
-__ai int32x4_t vqdmulhq_s32(int32x4_t __a, int32x4_t __b) {
-  return (int32x4_t)__builtin_neon_vqdmulhq_v((int8x16_t)__a, (int8x16_t)__b, 34); }
-
-#define vqdmulh_lane_s16(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int16x4_t __b = (b); \
-  vqdmulh_s16(__a, __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
-#define vqdmulh_lane_s32(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int32x2_t __b = (b); \
-  vqdmulh_s32(__a, __builtin_shufflevector(__b, __b, __c, __c)); })
-#define vqdmulhq_lane_s16(a, b, __c) __extension__ ({ \
-  int16x8_t __a = (a); int16x4_t __b = (b); \
-  vqdmulhq_s16(__a, __builtin_shufflevector(__b, __b, __c, __c, __c, __c, __c, __c, __c, __c)); })
-#define vqdmulhq_lane_s32(a, b, __c) __extension__ ({ \
-  int32x4_t __a = (a); int32x2_t __b = (b); \
-  vqdmulhq_s32(__a, __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
+#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
 
-#if defined(__aarch64__)
-#define vqdmulh_laneq_s16(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int16x8_t __b = (b); \
-  vqdmulh_s16(__a, __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
-#define vqdmulh_laneq_s32(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int32x4_t __b = (b); \
-  vqdmulh_s32(__a, __builtin_shufflevector(__b, __b, __c, __c)); })
-#define vqdmulhq_laneq_s16(a, b, __c) __extension__ ({ \
-  int16x8_t __a = (a); int16x8_t __b = (b); \
-  vqdmulhq_s16(__a, __builtin_shufflevector(__b, __b, __c, __c, __c, __c, __c, __c, __c, __c)); })
-#define vqdmulhq_laneq_s32(a, b, __c) __extension__ ({ \
-  int32x4_t __a = (a); int32x4_t __b = (b); \
-  vqdmulhq_s32(__a, __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
-
-#endif
-
-__ai int16x4_t vqdmulh_n_s16(int16x4_t __a, int16_t __b) {
-  return (int16x4_t)__builtin_neon_vqdmulh_v((int8x8_t)__a, (int8x8_t)(int16x4_t){ __b, __b, __b, __b }, 1); }
-__ai int32x2_t vqdmulh_n_s32(int32x2_t __a, int32_t __b) {
-  return (int32x2_t)__builtin_neon_vqdmulh_v((int8x8_t)__a, (int8x8_t)(int32x2_t){ __b, __b }, 2); }
-__ai int16x8_t vqdmulhq_n_s16(int16x8_t __a, int16_t __b) {
-  return (int16x8_t)__builtin_neon_vqdmulhq_v((int8x16_t)__a, (int8x16_t)(int16x8_t){ __b, __b, __b, __b, __b, __b, __b, __b }, 33); }
-__ai int32x4_t vqdmulhq_n_s32(int32x4_t __a, int32_t __b) {
-  return (int32x4_t)__builtin_neon_vqdmulhq_v((int8x16_t)__a, (int8x16_t)(int32x4_t){ __b, __b, __b, __b }, 34); }
-
-__ai int32x4_t vqdmull_s16(int16x4_t __a, int16x4_t __b) {
-  return (int32x4_t)__builtin_neon_vqdmull_v((int8x8_t)__a, (int8x8_t)__b, 34); }
-__ai int64x2_t vqdmull_s32(int32x2_t __a, int32x2_t __b) {
-  return (int64x2_t)__builtin_neon_vqdmull_v((int8x8_t)__a, (int8x8_t)__b, 35); }
+#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
 
-#if defined(__aarch64__)
-__ai int32x4_t vqdmull_high_s16(int16x8_t __a, int16x8_t __b) {
-  return vqdmull_s16(vget_high_s16(__a), vget_high_s16(__b)); }
-__ai int64x2_t vqdmull_high_s32(int32x4_t __a, int32x4_t __b) {
-  return vqdmull_s32(vget_high_s32(__a), vget_high_s32(__b)); }
-
-#define vqdmull_high_lane_s16(a, b, __c) __extension__ ({ \
-  int16x8_t __a = (a); int16x4_t __b = (b); \
-  vqdmull_s16(vget_high_s16(__a), __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
-#define vqdmull_high_lane_s32(a, b, __c) __extension__ ({ \
-  int32x4_t __a = (a); int32x2_t __b = (b); \
-  vqdmull_s32(vget_high_s32(__a), __builtin_shufflevector(__b, __b, __c, __c)); })
-
-#define vqdmull_high_laneq_s16(a, b, __c) __extension__ ({ \
-  int16x8_t __a = (a); int16x8_t __b = (b); \
-  vqdmull_s16(vget_high_s16(__a), __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
-#define vqdmull_high_laneq_s32(a, b, __c) __extension__ ({ \
-  int32x4_t __a = (a); int32x4_t __b = (b); \
-  vqdmull_s32(vget_high_s32(__a), __builtin_shufflevector(__b, __b, __c, __c)); })
-
-__ai int32x4_t vqdmull_high_n_s16(int16x8_t __a, int16_t __b) {
-  return vqdmull_n_s16(vget_high_s16(__a), __b); }
-__ai int64x2_t vqdmull_high_n_s32(int32x4_t __a, int32_t __b) {
-  return vqdmull_n_s32(vget_high_s32(__a), __b); }
-
-#endif
-
-#define vqdmull_lane_s16(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int16x4_t __b = (b); \
-  vqdmull_s16(__a, __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
-#define vqdmull_lane_s32(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int32x2_t __b = (b); \
-  vqdmull_s32(__a, __builtin_shufflevector(__b, __b, __c, __c)); })
+#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
 
-#if defined(__aarch64__)
-#define vqdmull_laneq_s16(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int16x8_t __b = (b); \
-  vqdmull_s16(__a, __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
-#define vqdmull_laneq_s32(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int32x4_t __b = (b); \
-  vqdmull_s32(__a, __builtin_shufflevector(__b, __b, __c, __c)); })
-
-#endif
-
-__ai int8x8_t vqneg_s8(int8x8_t __a) {
-  return (int8x8_t)__builtin_neon_vqneg_v(__a, 0); }
-__ai int16x4_t vqneg_s16(int16x4_t __a) {
-  return (int16x4_t)__builtin_neon_vqneg_v((int8x8_t)__a, 1); }
-__ai int32x2_t vqneg_s32(int32x2_t __a) {
-  return (int32x2_t)__builtin_neon_vqneg_v((int8x8_t)__a, 2); }
-__ai int8x16_t vqnegq_s8(int8x16_t __a) {
-  return (int8x16_t)__builtin_neon_vqnegq_v(__a, 32); }
-__ai int16x8_t vqnegq_s16(int16x8_t __a) {
-  return (int16x8_t)__builtin_neon_vqnegq_v((int8x16_t)__a, 33); }
-__ai int32x4_t vqnegq_s32(int32x4_t __a) {
-  return (int32x4_t)__builtin_neon_vqnegq_v((int8x16_t)__a, 34); }
-
-__ai int16x4_t vqrdmulh_s16(int16x4_t __a, int16x4_t __b) {
-  return (int16x4_t)__builtin_neon_vqrdmulh_v((int8x8_t)__a, (int8x8_t)__b, 1); }
-__ai int32x2_t vqrdmulh_s32(int32x2_t __a, int32x2_t __b) {
-  return (int32x2_t)__builtin_neon_vqrdmulh_v((int8x8_t)__a, (int8x8_t)__b, 2); }
-__ai int16x8_t vqrdmulhq_s16(int16x8_t __a, int16x8_t __b) {
-  return (int16x8_t)__builtin_neon_vqrdmulhq_v((int8x16_t)__a, (int8x16_t)__b, 33); }
-__ai int32x4_t vqrdmulhq_s32(int32x4_t __a, int32x4_t __b) {
-  return (int32x4_t)__builtin_neon_vqrdmulhq_v((int8x16_t)__a, (int8x16_t)__b, 34); }
-
-#define vqrdmulh_lane_s16(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int16x4_t __b = (b); \
-  vqrdmulh_s16(__a, __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
-#define vqrdmulh_lane_s32(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int32x2_t __b = (b); \
-  vqrdmulh_s32(__a, __builtin_shufflevector(__b, __b, __c, __c)); })
-#define vqrdmulhq_lane_s16(a, b, __c) __extension__ ({ \
-  int16x8_t __a = (a); int16x4_t __b = (b); \
-  vqrdmulhq_s16(__a, __builtin_shufflevector(__b, __b, __c, __c, __c, __c, __c, __c, __c, __c)); })
-#define vqrdmulhq_lane_s32(a, b, __c) __extension__ ({ \
-  int32x4_t __a = (a); int32x2_t __b = (b); \
-  vqrdmulhq_s32(__a, __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
+#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
 
-#if defined(__aarch64__)
-#define vqrdmulh_laneq_s16(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int16x8_t __b = (b); \
-  vqrdmulh_s16(__a, __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
-#define vqrdmulh_laneq_s32(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int32x4_t __b = (b); \
-  vqrdmulh_s32(__a, __builtin_shufflevector(__b, __b, __c, __c)); })
-#define vqrdmulhq_laneq_s16(a, b, __c) __extension__ ({ \
-  int16x8_t __a = (a); int16x8_t __b = (b); \
-  vqrdmulhq_s16(__a, __builtin_shufflevector(__b, __b, __c, __c, __c, __c, __c, __c, __c, __c)); })
-#define vqrdmulhq_laneq_s32(a, b, __c) __extension__ ({ \
-  int32x4_t __a = (a); int32x4_t __b = (b); \
-  vqrdmulhq_s32(__a, __builtin_shufflevector(__b, __b, __c, __c, __c, __c)); })
-
-#endif
-
-__ai int16x4_t vqrdmulh_n_s16(int16x4_t __a, int16_t __b) {
-  return (int16x4_t)__builtin_neon_vqrdmulh_v((int8x8_t)__a, (int8x8_t)(int16x4_t){ __b, __b, __b, __b }, 1); }
-__ai int32x2_t vqrdmulh_n_s32(int32x2_t __a, int32_t __b) {
-  return (int32x2_t)__builtin_neon_vqrdmulh_v((int8x8_t)__a, (int8x8_t)(int32x2_t){ __b, __b }, 2); }
-__ai int16x8_t vqrdmulhq_n_s16(int16x8_t __a, int16_t __b) {
-  return (int16x8_t)__builtin_neon_vqrdmulhq_v((int8x16_t)__a, (int8x16_t)(int16x8_t){ __b, __b, __b, __b, __b, __b, __b, __b }, 33); }
-__ai int32x4_t vqrdmulhq_n_s32(int32x4_t __a, int32_t __b) {
-  return (int32x4_t)__builtin_neon_vqrdmulhq_v((int8x16_t)__a, (int8x16_t)(int32x4_t){ __b, __b, __b, __b }, 34); }
-
-__ai int8x8_t vqrshl_s8(int8x8_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vqrshl_v(__a, __b, 0); }
-__ai int16x4_t vqrshl_s16(int16x4_t __a, int16x4_t __b) {
-  return (int16x4_t)__builtin_neon_vqrshl_v((int8x8_t)__a, (int8x8_t)__b, 1); }
-__ai int32x2_t vqrshl_s32(int32x2_t __a, int32x2_t __b) {
-  return (int32x2_t)__builtin_neon_vqrshl_v((int8x8_t)__a, (int8x8_t)__b, 2); }
-__ai int64x1_t vqrshl_s64(int64x1_t __a, int64x1_t __b) {
-  return (int64x1_t)__builtin_neon_vqrshl_v((int8x8_t)__a, (int8x8_t)__b, 3); }
-__ai uint8x8_t vqrshl_u8(uint8x8_t __a, int8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vqrshl_v((int8x8_t)__a, __b, 16); }
-__ai uint16x4_t vqrshl_u16(uint16x4_t __a, int16x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vqrshl_v((int8x8_t)__a, (int8x8_t)__b, 17); }
-__ai uint32x2_t vqrshl_u32(uint32x2_t __a, int32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vqrshl_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai uint64x1_t vqrshl_u64(uint64x1_t __a, int64x1_t __b) {
-  return (uint64x1_t)__builtin_neon_vqrshl_v((int8x8_t)__a, (int8x8_t)__b, 19); }
-__ai int8x16_t vqrshlq_s8(int8x16_t __a, int8x16_t __b) {
-  return (int8x16_t)__builtin_neon_vqrshlq_v(__a, __b, 32); }
-__ai int16x8_t vqrshlq_s16(int16x8_t __a, int16x8_t __b) {
-  return (int16x8_t)__builtin_neon_vqrshlq_v((int8x16_t)__a, (int8x16_t)__b, 33); }
-__ai int32x4_t vqrshlq_s32(int32x4_t __a, int32x4_t __b) {
-  return (int32x4_t)__builtin_neon_vqrshlq_v((int8x16_t)__a, (int8x16_t)__b, 34); }
-__ai int64x2_t vqrshlq_s64(int64x2_t __a, int64x2_t __b) {
-  return (int64x2_t)__builtin_neon_vqrshlq_v((int8x16_t)__a, (int8x16_t)__b, 35); }
-__ai uint8x16_t vqrshlq_u8(uint8x16_t __a, int8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vqrshlq_v((int8x16_t)__a, __b, 48); }
-__ai uint16x8_t vqrshlq_u16(uint16x8_t __a, int16x8_t __b) {
-  return (uint16x8_t)__builtin_neon_vqrshlq_v((int8x16_t)__a, (int8x16_t)__b, 49); }
-__ai uint32x4_t vqrshlq_u32(uint32x4_t __a, int32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vqrshlq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-__ai uint64x2_t vqrshlq_u64(uint64x2_t __a, int64x2_t __b) {
-  return (uint64x2_t)__builtin_neon_vqrshlq_v((int8x16_t)__a, (int8x16_t)__b, 51); }
-
-#define vqrshrn_n_s16(a, __b) __extension__ ({ \
-  int16x8_t __a = (a); \
-  (int8x8_t)__builtin_neon_vqrshrn_n_v((int8x16_t)__a, __b, 0); })
-#define vqrshrn_n_s32(a, __b) __extension__ ({ \
-  int32x4_t __a = (a); \
-  (int16x4_t)__builtin_neon_vqrshrn_n_v((int8x16_t)__a, __b, 1); })
-#define vqrshrn_n_s64(a, __b) __extension__ ({ \
-  int64x2_t __a = (a); \
-  (int32x2_t)__builtin_neon_vqrshrn_n_v((int8x16_t)__a, __b, 2); })
-#define vqrshrn_n_u16(a, __b) __extension__ ({ \
-  uint16x8_t __a = (a); \
-  (uint8x8_t)__builtin_neon_vqrshrn_n_v((int8x16_t)__a, __b, 16); })
-#define vqrshrn_n_u32(a, __b) __extension__ ({ \
-  uint32x4_t __a = (a); \
-  (uint16x4_t)__builtin_neon_vqrshrn_n_v((int8x16_t)__a, __b, 17); })
-#define vqrshrn_n_u64(a, __b) __extension__ ({ \
-  uint64x2_t __a = (a); \
-  (uint32x2_t)__builtin_neon_vqrshrn_n_v((int8x16_t)__a, __b, 18); })
-
-#define vqrshrun_n_s16(a, __b) __extension__ ({ \
-  int16x8_t __a = (a); \
-  (uint8x8_t)__builtin_neon_vqrshrun_n_v((int8x16_t)__a, __b, 16); })
-#define vqrshrun_n_s32(a, __b) __extension__ ({ \
-  int32x4_t __a = (a); \
-  (uint16x4_t)__builtin_neon_vqrshrun_n_v((int8x16_t)__a, __b, 17); })
-#define vqrshrun_n_s64(a, __b) __extension__ ({ \
-  int64x2_t __a = (a); \
-  (uint32x2_t)__builtin_neon_vqrshrun_n_v((int8x16_t)__a, __b, 18); })
-
-__ai int8x8_t vqshl_s8(int8x8_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vqshl_v(__a, __b, 0); }
-__ai int16x4_t vqshl_s16(int16x4_t __a, int16x4_t __b) {
-  return (int16x4_t)__builtin_neon_vqshl_v((int8x8_t)__a, (int8x8_t)__b, 1); }
-__ai int32x2_t vqshl_s32(int32x2_t __a, int32x2_t __b) {
-  return (int32x2_t)__builtin_neon_vqshl_v((int8x8_t)__a, (int8x8_t)__b, 2); }
-__ai int64x1_t vqshl_s64(int64x1_t __a, int64x1_t __b) {
-  return (int64x1_t)__builtin_neon_vqshl_v((int8x8_t)__a, (int8x8_t)__b, 3); }
-__ai uint8x8_t vqshl_u8(uint8x8_t __a, int8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vqshl_v((int8x8_t)__a, __b, 16); }
-__ai uint16x4_t vqshl_u16(uint16x4_t __a, int16x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vqshl_v((int8x8_t)__a, (int8x8_t)__b, 17); }
-__ai uint32x2_t vqshl_u32(uint32x2_t __a, int32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vqshl_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai uint64x1_t vqshl_u64(uint64x1_t __a, int64x1_t __b) {
-  return (uint64x1_t)__builtin_neon_vqshl_v((int8x8_t)__a, (int8x8_t)__b, 19); }
-__ai int8x16_t vqshlq_s8(int8x16_t __a, int8x16_t __b) {
-  return (int8x16_t)__builtin_neon_vqshlq_v(__a, __b, 32); }
-__ai int16x8_t vqshlq_s16(int16x8_t __a, int16x8_t __b) {
-  return (int16x8_t)__builtin_neon_vqshlq_v((int8x16_t)__a, (int8x16_t)__b, 33); }
-__ai int32x4_t vqshlq_s32(int32x4_t __a, int32x4_t __b) {
-  return (int32x4_t)__builtin_neon_vqshlq_v((int8x16_t)__a, (int8x16_t)__b, 34); }
-__ai int64x2_t vqshlq_s64(int64x2_t __a, int64x2_t __b) {
-  return (int64x2_t)__builtin_neon_vqshlq_v((int8x16_t)__a, (int8x16_t)__b, 35); }
-__ai uint8x16_t vqshlq_u8(uint8x16_t __a, int8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vqshlq_v((int8x16_t)__a, __b, 48); }
-__ai uint16x8_t vqshlq_u16(uint16x8_t __a, int16x8_t __b) {
-  return (uint16x8_t)__builtin_neon_vqshlq_v((int8x16_t)__a, (int8x16_t)__b, 49); }
-__ai uint32x4_t vqshlq_u32(uint32x4_t __a, int32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vqshlq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-__ai uint64x2_t vqshlq_u64(uint64x2_t __a, int64x2_t __b) {
-  return (uint64x2_t)__builtin_neon_vqshlq_v((int8x16_t)__a, (int8x16_t)__b, 51); }
-
-#define vqshlu_n_s8(a, __b) __extension__ ({ \
-  int8x8_t __a = (a); \
-  (uint8x8_t)__builtin_neon_vqshlu_n_v(__a, __b, 16); })
-#define vqshlu_n_s16(a, __b) __extension__ ({ \
-  int16x4_t __a = (a); \
-  (uint16x4_t)__builtin_neon_vqshlu_n_v((int8x8_t)__a, __b, 17); })
-#define vqshlu_n_s32(a, __b) __extension__ ({ \
-  int32x2_t __a = (a); \
-  (uint32x2_t)__builtin_neon_vqshlu_n_v((int8x8_t)__a, __b, 18); })
-#define vqshlu_n_s64(a, __b) __extension__ ({ \
-  int64x1_t __a = (a); \
-  (uint64x1_t)__builtin_neon_vqshlu_n_v((int8x8_t)__a, __b, 19); })
-#define vqshluq_n_s8(a, __b) __extension__ ({ \
-  int8x16_t __a = (a); \
-  (uint8x16_t)__builtin_neon_vqshluq_n_v(__a, __b, 48); })
-#define vqshluq_n_s16(a, __b) __extension__ ({ \
-  int16x8_t __a = (a); \
-  (uint16x8_t)__builtin_neon_vqshluq_n_v((int8x16_t)__a, __b, 49); })
-#define vqshluq_n_s32(a, __b) __extension__ ({ \
-  int32x4_t __a = (a); \
-  (uint32x4_t)__builtin_neon_vqshluq_n_v((int8x16_t)__a, __b, 50); })
-#define vqshluq_n_s64(a, __b) __extension__ ({ \
-  int64x2_t __a = (a); \
-  (uint64x2_t)__builtin_neon_vqshluq_n_v((int8x16_t)__a, __b, 51); })
-
-#define vqshl_n_s8(a, __b) __extension__ ({ \
-  int8x8_t __a = (a); \
-  (int8x8_t)__builtin_neon_vqshl_n_v(__a, __b, 0); })
-#define vqshl_n_s16(a, __b) __extension__ ({ \
-  int16x4_t __a = (a); \
-  (int16x4_t)__builtin_neon_vqshl_n_v((int8x8_t)__a, __b, 1); })
-#define vqshl_n_s32(a, __b) __extension__ ({ \
-  int32x2_t __a = (a); \
-  (int32x2_t)__builtin_neon_vqshl_n_v((int8x8_t)__a, __b, 2); })
-#define vqshl_n_s64(a, __b) __extension__ ({ \
-  int64x1_t __a = (a); \
-  (int64x1_t)__builtin_neon_vqshl_n_v((int8x8_t)__a, __b, 3); })
-#define vqshl_n_u8(a, __b) __extension__ ({ \
-  uint8x8_t __a = (a); \
-  (uint8x8_t)__builtin_neon_vqshl_n_v((int8x8_t)__a, __b, 16); })
-#define vqshl_n_u16(a, __b) __extension__ ({ \
-  uint16x4_t __a = (a); \
-  (uint16x4_t)__builtin_neon_vqshl_n_v((int8x8_t)__a, __b, 17); })
-#define vqshl_n_u32(a, __b) __extension__ ({ \
-  uint32x2_t __a = (a); \
-  (uint32x2_t)__builtin_neon_vqshl_n_v((int8x8_t)__a, __b, 18); })
-#define vqshl_n_u64(a, __b) __extension__ ({ \
-  uint64x1_t __a = (a); \
-  (uint64x1_t)__builtin_neon_vqshl_n_v((int8x8_t)__a, __b, 19); })
-#define vqshlq_n_s8(a, __b) __extension__ ({ \
-  int8x16_t __a = (a); \
-  (int8x16_t)__builtin_neon_vqshlq_n_v(__a, __b, 32); })
-#define vqshlq_n_s16(a, __b) __extension__ ({ \
-  int16x8_t __a = (a); \
-  (int16x8_t)__builtin_neon_vqshlq_n_v((int8x16_t)__a, __b, 33); })
-#define vqshlq_n_s32(a, __b) __extension__ ({ \
-  int32x4_t __a = (a); \
-  (int32x4_t)__builtin_neon_vqshlq_n_v((int8x16_t)__a, __b, 34); })
-#define vqshlq_n_s64(a, __b) __extension__ ({ \
-  int64x2_t __a = (a); \
-  (int64x2_t)__builtin_neon_vqshlq_n_v((int8x16_t)__a, __b, 35); })
-#define vqshlq_n_u8(a, __b) __extension__ ({ \
-  uint8x16_t __a = (a); \
-  (uint8x16_t)__builtin_neon_vqshlq_n_v((int8x16_t)__a, __b, 48); })
-#define vqshlq_n_u16(a, __b) __extension__ ({ \
-  uint16x8_t __a = (a); \
-  (uint16x8_t)__builtin_neon_vqshlq_n_v((int8x16_t)__a, __b, 49); })
-#define vqshlq_n_u32(a, __b) __extension__ ({ \
-  uint32x4_t __a = (a); \
-  (uint32x4_t)__builtin_neon_vqshlq_n_v((int8x16_t)__a, __b, 50); })
-#define vqshlq_n_u64(a, __b) __extension__ ({ \
-  uint64x2_t __a = (a); \
-  (uint64x2_t)__builtin_neon_vqshlq_n_v((int8x16_t)__a, __b, 51); })
-
-#define vqshrn_n_s16(a, __b) __extension__ ({ \
-  int16x8_t __a = (a); \
-  (int8x8_t)__builtin_neon_vqshrn_n_v((int8x16_t)__a, __b, 0); })
-#define vqshrn_n_s32(a, __b) __extension__ ({ \
-  int32x4_t __a = (a); \
-  (int16x4_t)__builtin_neon_vqshrn_n_v((int8x16_t)__a, __b, 1); })
-#define vqshrn_n_s64(a, __b) __extension__ ({ \
-  int64x2_t __a = (a); \
-  (int32x2_t)__builtin_neon_vqshrn_n_v((int8x16_t)__a, __b, 2); })
-#define vqshrn_n_u16(a, __b) __extension__ ({ \
-  uint16x8_t __a = (a); \
-  (uint8x8_t)__builtin_neon_vqshrn_n_v((int8x16_t)__a, __b, 16); })
-#define vqshrn_n_u32(a, __b) __extension__ ({ \
-  uint32x4_t __a = (a); \
-  (uint16x4_t)__builtin_neon_vqshrn_n_v((int8x16_t)__a, __b, 17); })
-#define vqshrn_n_u64(a, __b) __extension__ ({ \
-  uint64x2_t __a = (a); \
-  (uint32x2_t)__builtin_neon_vqshrn_n_v((int8x16_t)__a, __b, 18); })
-
-#define vqshrun_n_s16(a, __b) __extension__ ({ \
-  int16x8_t __a = (a); \
-  (uint8x8_t)__builtin_neon_vqshrun_n_v((int8x16_t)__a, __b, 16); })
-#define vqshrun_n_s32(a, __b) __extension__ ({ \
-  int32x4_t __a = (a); \
-  (uint16x4_t)__builtin_neon_vqshrun_n_v((int8x16_t)__a, __b, 17); })
-#define vqshrun_n_s64(a, __b) __extension__ ({ \
-  int64x2_t __a = (a); \
-  (uint32x2_t)__builtin_neon_vqshrun_n_v((int8x16_t)__a, __b, 18); })
-
-__ai int8x8_t vqsub_s8(int8x8_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vqsub_v(__a, __b, 0); }
-__ai int16x4_t vqsub_s16(int16x4_t __a, int16x4_t __b) {
-  return (int16x4_t)__builtin_neon_vqsub_v((int8x8_t)__a, (int8x8_t)__b, 1); }
-__ai int32x2_t vqsub_s32(int32x2_t __a, int32x2_t __b) {
-  return (int32x2_t)__builtin_neon_vqsub_v((int8x8_t)__a, (int8x8_t)__b, 2); }
-__ai int64x1_t vqsub_s64(int64x1_t __a, int64x1_t __b) {
-  return (int64x1_t)__builtin_neon_vqsub_v((int8x8_t)__a, (int8x8_t)__b, 3); }
-__ai uint8x8_t vqsub_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vqsub_v((int8x8_t)__a, (int8x8_t)__b, 16); }
-__ai uint16x4_t vqsub_u16(uint16x4_t __a, uint16x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vqsub_v((int8x8_t)__a, (int8x8_t)__b, 17); }
-__ai uint32x2_t vqsub_u32(uint32x2_t __a, uint32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vqsub_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai uint64x1_t vqsub_u64(uint64x1_t __a, uint64x1_t __b) {
-  return (uint64x1_t)__builtin_neon_vqsub_v((int8x8_t)__a, (int8x8_t)__b, 19); }
-__ai int8x16_t vqsubq_s8(int8x16_t __a, int8x16_t __b) {
-  return (int8x16_t)__builtin_neon_vqsubq_v(__a, __b, 32); }
-__ai int16x8_t vqsubq_s16(int16x8_t __a, int16x8_t __b) {
-  return (int16x8_t)__builtin_neon_vqsubq_v((int8x16_t)__a, (int8x16_t)__b, 33); }
-__ai int32x4_t vqsubq_s32(int32x4_t __a, int32x4_t __b) {
-  return (int32x4_t)__builtin_neon_vqsubq_v((int8x16_t)__a, (int8x16_t)__b, 34); }
-__ai int64x2_t vqsubq_s64(int64x2_t __a, int64x2_t __b) {
-  return (int64x2_t)__builtin_neon_vqsubq_v((int8x16_t)__a, (int8x16_t)__b, 35); }
-__ai uint8x16_t vqsubq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vqsubq_v((int8x16_t)__a, (int8x16_t)__b, 48); }
-__ai uint16x8_t vqsubq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint16x8_t)__builtin_neon_vqsubq_v((int8x16_t)__a, (int8x16_t)__b, 49); }
-__ai uint32x4_t vqsubq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vqsubq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-__ai uint64x2_t vqsubq_u64(uint64x2_t __a, uint64x2_t __b) {
-  return (uint64x2_t)__builtin_neon_vqsubq_v((int8x16_t)__a, (int8x16_t)__b, 51); }
+#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
 
-#if defined(__aarch64__)
-__ai uint8x8_t vqtbl1_u8(uint8x16_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vqtbl1_v((int8x16_t)__a, (int8x8_t)__b, 16); }
-__ai int8x8_t vqtbl1_s8(int8x16_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vqtbl1_v(__a, __b, 0); }
-__ai poly8x8_t vqtbl1_p8(poly8x16_t __a, uint8x8_t __b) {
-  return (poly8x8_t)__builtin_neon_vqtbl1_v((int8x16_t)__a, (int8x8_t)__b, 4); }
-__ai uint8x16_t vqtbl1q_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vqtbl1q_v((int8x16_t)__a, (int8x16_t)__b, 48); }
-__ai int8x16_t vqtbl1q_s8(int8x16_t __a, int8x16_t __b) {
-  return (int8x16_t)__builtin_neon_vqtbl1q_v(__a, __b, 32); }
-__ai poly8x16_t vqtbl1q_p8(poly8x16_t __a, uint8x16_t __b) {
-  return (poly8x16_t)__builtin_neon_vqtbl1q_v((int8x16_t)__a, (int8x16_t)__b, 36); }
-
-__ai uint8x8_t vqtbl2_u8(uint8x16x2_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vqtbl2_v((int8x16_t)__a.val[0], (int8x16_t)__a.val[1], (int8x8_t)__b, 16); }
-__ai int8x8_t vqtbl2_s8(int8x16x2_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vqtbl2_v(__a.val[0], __a.val[1], __b, 0); }
-__ai poly8x8_t vqtbl2_p8(poly8x16x2_t __a, uint8x8_t __b) {
-  return (poly8x8_t)__builtin_neon_vqtbl2_v((int8x16_t)__a.val[0], (int8x16_t)__a.val[1], (int8x8_t)__b, 4); }
-__ai uint8x16_t vqtbl2q_u8(uint8x16x2_t __a, uint8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vqtbl2q_v((int8x16_t)__a.val[0], (int8x16_t)__a.val[1], (int8x16_t)__b, 48); }
-__ai int8x16_t vqtbl2q_s8(int8x16x2_t __a, int8x16_t __b) {
-  return (int8x16_t)__builtin_neon_vqtbl2q_v(__a.val[0], __a.val[1], __b, 32); }
-__ai poly8x16_t vqtbl2q_p8(poly8x16x2_t __a, uint8x16_t __b) {
-  return (poly8x16_t)__builtin_neon_vqtbl2q_v((int8x16_t)__a.val[0], (int8x16_t)__a.val[1], (int8x16_t)__b, 36); }
-
-__ai uint8x8_t vqtbl3_u8(uint8x16x3_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vqtbl3_v((int8x16_t)__a.val[0], (int8x16_t)__a.val[1], (int8x16_t)__a.val[2], (int8x8_t)__b, 16); }
-__ai int8x8_t vqtbl3_s8(int8x16x3_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vqtbl3_v(__a.val[0], __a.val[1], __a.val[2], __b, 0); }
-__ai poly8x8_t vqtbl3_p8(poly8x16x3_t __a, uint8x8_t __b) {
-  return (poly8x8_t)__builtin_neon_vqtbl3_v((int8x16_t)__a.val[0], (int8x16_t)__a.val[1], (int8x16_t)__a.val[2], (int8x8_t)__b, 4); }
-__ai uint8x16_t vqtbl3q_u8(uint8x16x3_t __a, uint8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vqtbl3q_v((int8x16_t)__a.val[0], (int8x16_t)__a.val[1], (int8x16_t)__a.val[2], (int8x16_t)__b, 48); }
-__ai int8x16_t vqtbl3q_s8(int8x16x3_t __a, int8x16_t __b) {
-  return (int8x16_t)__builtin_neon_vqtbl3q_v(__a.val[0], __a.val[1], __a.val[2], __b, 32); }
-__ai poly8x16_t vqtbl3q_p8(poly8x16x3_t __a, uint8x16_t __b) {
-  return (poly8x16_t)__builtin_neon_vqtbl3q_v((int8x16_t)__a.val[0], (int8x16_t)__a.val[1], (int8x16_t)__a.val[2], (int8x16_t)__b, 36); }
-
-__ai uint8x8_t vqtbl4_u8(uint8x16x4_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vqtbl4_v((int8x16_t)__a.val[0], (int8x16_t)__a.val[1], (int8x16_t)__a.val[2], (int8x16_t)__a.val[3], (int8x8_t)__b, 16); }
-__ai int8x8_t vqtbl4_s8(int8x16x4_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vqtbl4_v(__a.val[0], __a.val[1], __a.val[2], __a.val[3], __b, 0); }
-__ai poly8x8_t vqtbl4_p8(poly8x16x4_t __a, uint8x8_t __b) {
-  return (poly8x8_t)__builtin_neon_vqtbl4_v((int8x16_t)__a.val[0], (int8x16_t)__a.val[1], (int8x16_t)__a.val[2], (int8x16_t)__a.val[3], (int8x8_t)__b, 4); }
-__ai uint8x16_t vqtbl4q_u8(uint8x16x4_t __a, uint8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vqtbl4q_v((int8x16_t)__a.val[0], (int8x16_t)__a.val[1], (int8x16_t)__a.val[2], (int8x16_t)__a.val[3], (int8x16_t)__b, 48); }
-__ai int8x16_t vqtbl4q_s8(int8x16x4_t __a, int8x16_t __b) {
-  return (int8x16_t)__builtin_neon_vqtbl4q_v(__a.val[0], __a.val[1], __a.val[2], __a.val[3], __b, 32); }
-__ai poly8x16_t vqtbl4q_p8(poly8x16x4_t __a, uint8x16_t __b) {
-  return (poly8x16_t)__builtin_neon_vqtbl4q_v((int8x16_t)__a.val[0], (int8x16_t)__a.val[1], (int8x16_t)__a.val[2], (int8x16_t)__a.val[3], (int8x16_t)__b, 36); }
-
-__ai uint8x8_t vqtbx1_u8(uint8x8_t __a, uint8x16_t __b, uint8x8_t __c) {
-  return (uint8x8_t)__builtin_neon_vqtbx1_v((int8x8_t)__a, (int8x16_t)__b, (int8x8_t)__c, 16); }
-__ai int8x8_t vqtbx1_s8(int8x8_t __a, int8x16_t __b, int8x8_t __c) {
-  return (int8x8_t)__builtin_neon_vqtbx1_v(__a, __b, __c, 0); }
-__ai poly8x8_t vqtbx1_p8(poly8x8_t __a, poly8x16_t __b, uint8x8_t __c) {
-  return (poly8x8_t)__builtin_neon_vqtbx1_v((int8x8_t)__a, (int8x16_t)__b, (int8x8_t)__c, 4); }
-__ai uint8x16_t vqtbx1q_u8(uint8x16_t __a, uint8x16_t __b, uint8x16_t __c) {
-  return (uint8x16_t)__builtin_neon_vqtbx1q_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 48); }
-__ai int8x16_t vqtbx1q_s8(int8x16_t __a, int8x16_t __b, int8x16_t __c) {
-  return (int8x16_t)__builtin_neon_vqtbx1q_v(__a, __b, __c, 32); }
-__ai poly8x16_t vqtbx1q_p8(poly8x16_t __a, poly8x16_t __b, uint8x16_t __c) {
-  return (poly8x16_t)__builtin_neon_vqtbx1q_v((int8x16_t)__a, (int8x16_t)__b, (int8x16_t)__c, 36); }
-
-__ai uint8x8_t vqtbx2_u8(uint8x8_t __a, uint8x16x2_t __b, uint8x8_t __c) {
-  return (uint8x8_t)__builtin_neon_vqtbx2_v((int8x8_t)__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x8_t)__c, 16); }
-__ai int8x8_t vqtbx2_s8(int8x8_t __a, int8x16x2_t __b, int8x8_t __c) {
-  return (int8x8_t)__builtin_neon_vqtbx2_v(__a, __b.val[0], __b.val[1], __c, 0); }
-__ai poly8x8_t vqtbx2_p8(poly8x8_t __a, poly8x16x2_t __b, uint8x8_t __c) {
-  return (poly8x8_t)__builtin_neon_vqtbx2_v((int8x8_t)__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x8_t)__c, 4); }
-__ai uint8x16_t vqtbx2q_u8(uint8x16_t __a, uint8x16x2_t __b, uint8x16_t __c) {
-  return (uint8x16_t)__builtin_neon_vqtbx2q_v((int8x16_t)__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__c, 48); }
-__ai int8x16_t vqtbx2q_s8(int8x16_t __a, int8x16x2_t __b, int8x16_t __c) {
-  return (int8x16_t)__builtin_neon_vqtbx2q_v(__a, __b.val[0], __b.val[1], __c, 32); }
-__ai poly8x16_t vqtbx2q_p8(poly8x16_t __a, poly8x16x2_t __b, uint8x16_t __c) {
-  return (poly8x16_t)__builtin_neon_vqtbx2q_v((int8x16_t)__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__c, 36); }
-
-__ai uint8x8_t vqtbx3_u8(uint8x8_t __a, uint8x16x3_t __b, uint8x8_t __c) {
-  return (uint8x8_t)__builtin_neon_vqtbx3_v((int8x8_t)__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x8_t)__c, 16); }
-__ai int8x8_t vqtbx3_s8(int8x8_t __a, int8x16x3_t __b, int8x8_t __c) {
-  return (int8x8_t)__builtin_neon_vqtbx3_v(__a, __b.val[0], __b.val[1], __b.val[2], __c, 0); }
-__ai poly8x8_t vqtbx3_p8(poly8x8_t __a, poly8x16x3_t __b, uint8x8_t __c) {
-  return (poly8x8_t)__builtin_neon_vqtbx3_v((int8x8_t)__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x8_t)__c, 4); }
-__ai uint8x16_t vqtbx3q_u8(uint8x16_t __a, uint8x16x3_t __b, uint8x16_t __c) {
-  return (uint8x16_t)__builtin_neon_vqtbx3q_v((int8x16_t)__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__c, 48); }
-__ai int8x16_t vqtbx3q_s8(int8x16_t __a, int8x16x3_t __b, int8x16_t __c) {
-  return (int8x16_t)__builtin_neon_vqtbx3q_v(__a, __b.val[0], __b.val[1], __b.val[2], __c, 32); }
-__ai poly8x16_t vqtbx3q_p8(poly8x16_t __a, poly8x16x3_t __b, uint8x16_t __c) {
-  return (poly8x16_t)__builtin_neon_vqtbx3q_v((int8x16_t)__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__c, 36); }
-
-__ai uint8x8_t vqtbx4_u8(uint8x8_t __a, uint8x16x4_t __b, uint8x8_t __c) {
-  return (uint8x8_t)__builtin_neon_vqtbx4_v((int8x8_t)__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], (int8x8_t)__c, 16); }
-__ai int8x8_t vqtbx4_s8(int8x8_t __a, int8x16x4_t __b, int8x8_t __c) {
-  return (int8x8_t)__builtin_neon_vqtbx4_v(__a, __b.val[0], __b.val[1], __b.val[2], __b.val[3], __c, 0); }
-__ai poly8x8_t vqtbx4_p8(poly8x8_t __a, poly8x16x4_t __b, uint8x8_t __c) {
-  return (poly8x8_t)__builtin_neon_vqtbx4_v((int8x8_t)__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], (int8x8_t)__c, 4); }
-__ai uint8x16_t vqtbx4q_u8(uint8x16_t __a, uint8x16x4_t __b, uint8x16_t __c) {
-  return (uint8x16_t)__builtin_neon_vqtbx4q_v((int8x16_t)__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], (int8x16_t)__c, 48); }
-__ai int8x16_t vqtbx4q_s8(int8x16_t __a, int8x16x4_t __b, int8x16_t __c) {
-  return (int8x16_t)__builtin_neon_vqtbx4q_v(__a, __b.val[0], __b.val[1], __b.val[2], __b.val[3], __c, 32); }
-__ai poly8x16_t vqtbx4q_p8(poly8x16_t __a, poly8x16x4_t __b, uint8x16_t __c) {
-  return (poly8x16_t)__builtin_neon_vqtbx4q_v((int8x16_t)__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], (int8x16_t)__c, 36); }
-
-#endif
-
-__ai int8x8_t vraddhn_s16(int16x8_t __a, int16x8_t __b) {
-  return (int8x8_t)__builtin_neon_vraddhn_v((int8x16_t)__a, (int8x16_t)__b, 0); }
-__ai int16x4_t vraddhn_s32(int32x4_t __a, int32x4_t __b) {
-  return (int16x4_t)__builtin_neon_vraddhn_v((int8x16_t)__a, (int8x16_t)__b, 1); }
-__ai int32x2_t vraddhn_s64(int64x2_t __a, int64x2_t __b) {
-  return (int32x2_t)__builtin_neon_vraddhn_v((int8x16_t)__a, (int8x16_t)__b, 2); }
-__ai uint8x8_t vraddhn_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vraddhn_v((int8x16_t)__a, (int8x16_t)__b, 16); }
-__ai uint16x4_t vraddhn_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vraddhn_v((int8x16_t)__a, (int8x16_t)__b, 17); }
-__ai uint32x2_t vraddhn_u64(uint64x2_t __a, uint64x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vraddhn_v((int8x16_t)__a, (int8x16_t)__b, 18); }
+#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
 
-#if defined(__aarch64__)
-__ai int8x16_t vraddhn_high_s16(int8x8_t __a, int16x8_t __b, int16x8_t __c) {
-  return vcombine_s8(__a, vraddhn_s16(__b, __c)); }
-__ai int16x8_t vraddhn_high_s32(int16x4_t __a, int32x4_t __b, int32x4_t __c) {
-  return vcombine_s16(__a, vraddhn_s32(__b, __c)); }
-__ai int32x4_t vraddhn_high_s64(int32x2_t __a, int64x2_t __b, int64x2_t __c) {
-  return vcombine_s32(__a, vraddhn_s64(__b, __c)); }
-__ai uint8x16_t vraddhn_high_u16(uint8x8_t __a, uint16x8_t __b, uint16x8_t __c) {
-  return vcombine_u8(__a, vraddhn_u16(__b, __c)); }
-__ai uint16x8_t vraddhn_high_u32(uint16x4_t __a, uint32x4_t __b, uint32x4_t __c) {
-  return vcombine_u16(__a, vraddhn_u32(__b, __c)); }
-__ai uint32x4_t vraddhn_high_u64(uint32x2_t __a, uint64x2_t __b, uint64x2_t __c) {
-  return vcombine_u32(__a, vraddhn_u64(__b, __c)); }
-
-#endif
-
-__ai float32x2_t vrecpe_f32(float32x2_t __a) {
-  return (float32x2_t)__builtin_neon_vrecpe_v((int8x8_t)__a, 9); }
-__ai uint32x2_t vrecpe_u32(uint32x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vrecpe_v((int8x8_t)__a, 18); }
-__ai float32x4_t vrecpeq_f32(float32x4_t __a) {
-  return (float32x4_t)__builtin_neon_vrecpeq_v((int8x16_t)__a, 41); }
-__ai uint32x4_t vrecpeq_u32(uint32x4_t __a) {
-  return (uint32x4_t)__builtin_neon_vrecpeq_v((int8x16_t)__a, 50); }
-
-__ai float32x2_t vrecps_f32(float32x2_t __a, float32x2_t __b) {
-  return (float32x2_t)__builtin_neon_vrecps_v((int8x8_t)__a, (int8x8_t)__b, 9); }
-__ai float32x4_t vrecpsq_f32(float32x4_t __a, float32x4_t __b) {
-  return (float32x4_t)__builtin_neon_vrecpsq_v((int8x16_t)__a, (int8x16_t)__b, 41); }
-
-__ai int8x8_t vreinterpret_s8_s16(int16x4_t __a) {
-  return (int8x8_t)__a; }
-__ai int8x8_t vreinterpret_s8_s32(int32x2_t __a) {
-  return (int8x8_t)__a; }
-__ai int8x8_t vreinterpret_s8_s64(int64x1_t __a) {
-  return (int8x8_t)__a; }
-__ai int8x8_t vreinterpret_s8_u8(uint8x8_t __a) {
-  return (int8x8_t)__a; }
-__ai int8x8_t vreinterpret_s8_u16(uint16x4_t __a) {
-  return (int8x8_t)__a; }
-__ai int8x8_t vreinterpret_s8_u32(uint32x2_t __a) {
-  return (int8x8_t)__a; }
-__ai int8x8_t vreinterpret_s8_u64(uint64x1_t __a) {
-  return (int8x8_t)__a; }
-__ai int8x8_t vreinterpret_s8_f16(float16x4_t __a) {
-  return (int8x8_t)__a; }
-__ai int8x8_t vreinterpret_s8_f32(float32x2_t __a) {
-  return (int8x8_t)__a; }
-__ai int8x8_t vreinterpret_s8_p8(poly8x8_t __a) {
-  return (int8x8_t)__a; }
-__ai int8x8_t vreinterpret_s8_p16(poly16x4_t __a) {
-  return (int8x8_t)__a; }
-__ai int16x4_t vreinterpret_s16_s8(int8x8_t __a) {
-  return (int16x4_t)__a; }
-__ai int16x4_t vreinterpret_s16_s32(int32x2_t __a) {
-  return (int16x4_t)__a; }
-__ai int16x4_t vreinterpret_s16_s64(int64x1_t __a) {
-  return (int16x4_t)__a; }
-__ai int16x4_t vreinterpret_s16_u8(uint8x8_t __a) {
-  return (int16x4_t)__a; }
-__ai int16x4_t vreinterpret_s16_u16(uint16x4_t __a) {
-  return (int16x4_t)__a; }
-__ai int16x4_t vreinterpret_s16_u32(uint32x2_t __a) {
-  return (int16x4_t)__a; }
-__ai int16x4_t vreinterpret_s16_u64(uint64x1_t __a) {
-  return (int16x4_t)__a; }
-__ai int16x4_t vreinterpret_s16_f16(float16x4_t __a) {
-  return (int16x4_t)__a; }
-__ai int16x4_t vreinterpret_s16_f32(float32x2_t __a) {
-  return (int16x4_t)__a; }
-__ai int16x4_t vreinterpret_s16_p8(poly8x8_t __a) {
-  return (int16x4_t)__a; }
-__ai int16x4_t vreinterpret_s16_p16(poly16x4_t __a) {
-  return (int16x4_t)__a; }
-__ai int32x2_t vreinterpret_s32_s8(int8x8_t __a) {
-  return (int32x2_t)__a; }
-__ai int32x2_t vreinterpret_s32_s16(int16x4_t __a) {
-  return (int32x2_t)__a; }
-__ai int32x2_t vreinterpret_s32_s64(int64x1_t __a) {
-  return (int32x2_t)__a; }
-__ai int32x2_t vreinterpret_s32_u8(uint8x8_t __a) {
-  return (int32x2_t)__a; }
-__ai int32x2_t vreinterpret_s32_u16(uint16x4_t __a) {
-  return (int32x2_t)__a; }
-__ai int32x2_t vreinterpret_s32_u32(uint32x2_t __a) {
-  return (int32x2_t)__a; }
-__ai int32x2_t vreinterpret_s32_u64(uint64x1_t __a) {
-  return (int32x2_t)__a; }
-__ai int32x2_t vreinterpret_s32_f16(float16x4_t __a) {
-  return (int32x2_t)__a; }
-__ai int32x2_t vreinterpret_s32_f32(float32x2_t __a) {
-  return (int32x2_t)__a; }
-__ai int32x2_t vreinterpret_s32_p8(poly8x8_t __a) {
-  return (int32x2_t)__a; }
-__ai int32x2_t vreinterpret_s32_p16(poly16x4_t __a) {
-  return (int32x2_t)__a; }
-__ai int64x1_t vreinterpret_s64_s8(int8x8_t __a) {
-  return (int64x1_t)__a; }
-__ai int64x1_t vreinterpret_s64_s16(int16x4_t __a) {
-  return (int64x1_t)__a; }
-__ai int64x1_t vreinterpret_s64_s32(int32x2_t __a) {
-  return (int64x1_t)__a; }
-__ai int64x1_t vreinterpret_s64_u8(uint8x8_t __a) {
-  return (int64x1_t)__a; }
-__ai int64x1_t vreinterpret_s64_u16(uint16x4_t __a) {
-  return (int64x1_t)__a; }
-__ai int64x1_t vreinterpret_s64_u32(uint32x2_t __a) {
-  return (int64x1_t)__a; }
-__ai int64x1_t vreinterpret_s64_u64(uint64x1_t __a) {
-  return (int64x1_t)__a; }
-__ai int64x1_t vreinterpret_s64_f16(float16x4_t __a) {
-  return (int64x1_t)__a; }
-__ai int64x1_t vreinterpret_s64_f32(float32x2_t __a) {
-  return (int64x1_t)__a; }
-__ai int64x1_t vreinterpret_s64_p8(poly8x8_t __a) {
-  return (int64x1_t)__a; }
-__ai int64x1_t vreinterpret_s64_p16(poly16x4_t __a) {
-  return (int64x1_t)__a; }
-__ai uint8x8_t vreinterpret_u8_s8(int8x8_t __a) {
-  return (uint8x8_t)__a; }
-__ai uint8x8_t vreinterpret_u8_s16(int16x4_t __a) {
-  return (uint8x8_t)__a; }
-__ai uint8x8_t vreinterpret_u8_s32(int32x2_t __a) {
-  return (uint8x8_t)__a; }
-__ai uint8x8_t vreinterpret_u8_s64(int64x1_t __a) {
-  return (uint8x8_t)__a; }
-__ai uint8x8_t vreinterpret_u8_u16(uint16x4_t __a) {
-  return (uint8x8_t)__a; }
-__ai uint8x8_t vreinterpret_u8_u32(uint32x2_t __a) {
-  return (uint8x8_t)__a; }
-__ai uint8x8_t vreinterpret_u8_u64(uint64x1_t __a) {
-  return (uint8x8_t)__a; }
-__ai uint8x8_t vreinterpret_u8_f16(float16x4_t __a) {
-  return (uint8x8_t)__a; }
-__ai uint8x8_t vreinterpret_u8_f32(float32x2_t __a) {
-  return (uint8x8_t)__a; }
-__ai uint8x8_t vreinterpret_u8_p8(poly8x8_t __a) {
-  return (uint8x8_t)__a; }
-__ai uint8x8_t vreinterpret_u8_p16(poly16x4_t __a) {
-  return (uint8x8_t)__a; }
-__ai uint16x4_t vreinterpret_u16_s8(int8x8_t __a) {
-  return (uint16x4_t)__a; }
-__ai uint16x4_t vreinterpret_u16_s16(int16x4_t __a) {
-  return (uint16x4_t)__a; }
-__ai uint16x4_t vreinterpret_u16_s32(int32x2_t __a) {
-  return (uint16x4_t)__a; }
-__ai uint16x4_t vreinterpret_u16_s64(int64x1_t __a) {
-  return (uint16x4_t)__a; }
-__ai uint16x4_t vreinterpret_u16_u8(uint8x8_t __a) {
-  return (uint16x4_t)__a; }
-__ai uint16x4_t vreinterpret_u16_u32(uint32x2_t __a) {
-  return (uint16x4_t)__a; }
-__ai uint16x4_t vreinterpret_u16_u64(uint64x1_t __a) {
-  return (uint16x4_t)__a; }
-__ai uint16x4_t vreinterpret_u16_f16(float16x4_t __a) {
-  return (uint16x4_t)__a; }
-__ai uint16x4_t vreinterpret_u16_f32(float32x2_t __a) {
-  return (uint16x4_t)__a; }
-__ai uint16x4_t vreinterpret_u16_p8(poly8x8_t __a) {
-  return (uint16x4_t)__a; }
-__ai uint16x4_t vreinterpret_u16_p16(poly16x4_t __a) {
-  return (uint16x4_t)__a; }
-__ai uint32x2_t vreinterpret_u32_s8(int8x8_t __a) {
-  return (uint32x2_t)__a; }
-__ai uint32x2_t vreinterpret_u32_s16(int16x4_t __a) {
-  return (uint32x2_t)__a; }
-__ai uint32x2_t vreinterpret_u32_s32(int32x2_t __a) {
-  return (uint32x2_t)__a; }
-__ai uint32x2_t vreinterpret_u32_s64(int64x1_t __a) {
-  return (uint32x2_t)__a; }
-__ai uint32x2_t vreinterpret_u32_u8(uint8x8_t __a) {
-  return (uint32x2_t)__a; }
-__ai uint32x2_t vreinterpret_u32_u16(uint16x4_t __a) {
-  return (uint32x2_t)__a; }
-__ai uint32x2_t vreinterpret_u32_u64(uint64x1_t __a) {
-  return (uint32x2_t)__a; }
-__ai uint32x2_t vreinterpret_u32_f16(float16x4_t __a) {
-  return (uint32x2_t)__a; }
-__ai uint32x2_t vreinterpret_u32_f32(float32x2_t __a) {
-  return (uint32x2_t)__a; }
-__ai uint32x2_t vreinterpret_u32_p8(poly8x8_t __a) {
-  return (uint32x2_t)__a; }
-__ai uint32x2_t vreinterpret_u32_p16(poly16x4_t __a) {
-  return (uint32x2_t)__a; }
-__ai uint64x1_t vreinterpret_u64_s8(int8x8_t __a) {
-  return (uint64x1_t)__a; }
-__ai uint64x1_t vreinterpret_u64_s16(int16x4_t __a) {
-  return (uint64x1_t)__a; }
-__ai uint64x1_t vreinterpret_u64_s32(int32x2_t __a) {
-  return (uint64x1_t)__a; }
-__ai uint64x1_t vreinterpret_u64_s64(int64x1_t __a) {
-  return (uint64x1_t)__a; }
-__ai uint64x1_t vreinterpret_u64_u8(uint8x8_t __a) {
-  return (uint64x1_t)__a; }
-__ai uint64x1_t vreinterpret_u64_u16(uint16x4_t __a) {
-  return (uint64x1_t)__a; }
-__ai uint64x1_t vreinterpret_u64_u32(uint32x2_t __a) {
-  return (uint64x1_t)__a; }
-__ai uint64x1_t vreinterpret_u64_f16(float16x4_t __a) {
-  return (uint64x1_t)__a; }
-__ai uint64x1_t vreinterpret_u64_f32(float32x2_t __a) {
-  return (uint64x1_t)__a; }
-__ai uint64x1_t vreinterpret_u64_p8(poly8x8_t __a) {
-  return (uint64x1_t)__a; }
-__ai uint64x1_t vreinterpret_u64_p16(poly16x4_t __a) {
-  return (uint64x1_t)__a; }
-__ai float16x4_t vreinterpret_f16_s8(int8x8_t __a) {
-  return (float16x4_t)__a; }
-__ai float16x4_t vreinterpret_f16_s16(int16x4_t __a) {
-  return (float16x4_t)__a; }
-__ai float16x4_t vreinterpret_f16_s32(int32x2_t __a) {
-  return (float16x4_t)__a; }
-__ai float16x4_t vreinterpret_f16_s64(int64x1_t __a) {
-  return (float16x4_t)__a; }
-__ai float16x4_t vreinterpret_f16_u8(uint8x8_t __a) {
-  return (float16x4_t)__a; }
-__ai float16x4_t vreinterpret_f16_u16(uint16x4_t __a) {
-  return (float16x4_t)__a; }
-__ai float16x4_t vreinterpret_f16_u32(uint32x2_t __a) {
-  return (float16x4_t)__a; }
-__ai float16x4_t vreinterpret_f16_u64(uint64x1_t __a) {
-  return (float16x4_t)__a; }
-__ai float16x4_t vreinterpret_f16_f32(float32x2_t __a) {
-  return (float16x4_t)__a; }
-__ai float16x4_t vreinterpret_f16_p8(poly8x8_t __a) {
-  return (float16x4_t)__a; }
-__ai float16x4_t vreinterpret_f16_p16(poly16x4_t __a) {
-  return (float16x4_t)__a; }
-__ai float32x2_t vreinterpret_f32_s8(int8x8_t __a) {
-  return (float32x2_t)__a; }
-__ai float32x2_t vreinterpret_f32_s16(int16x4_t __a) {
-  return (float32x2_t)__a; }
-__ai float32x2_t vreinterpret_f32_s32(int32x2_t __a) {
-  return (float32x2_t)__a; }
-__ai float32x2_t vreinterpret_f32_s64(int64x1_t __a) {
-  return (float32x2_t)__a; }
-__ai float32x2_t vreinterpret_f32_u8(uint8x8_t __a) {
-  return (float32x2_t)__a; }
-__ai float32x2_t vreinterpret_f32_u16(uint16x4_t __a) {
-  return (float32x2_t)__a; }
-__ai float32x2_t vreinterpret_f32_u32(uint32x2_t __a) {
-  return (float32x2_t)__a; }
-__ai float32x2_t vreinterpret_f32_u64(uint64x1_t __a) {
-  return (float32x2_t)__a; }
-__ai float32x2_t vreinterpret_f32_f16(float16x4_t __a) {
-  return (float32x2_t)__a; }
-__ai float32x2_t vreinterpret_f32_p8(poly8x8_t __a) {
-  return (float32x2_t)__a; }
-__ai float32x2_t vreinterpret_f32_p16(poly16x4_t __a) {
-  return (float32x2_t)__a; }
-__ai poly8x8_t vreinterpret_p8_s8(int8x8_t __a) {
-  return (poly8x8_t)__a; }
-__ai poly8x8_t vreinterpret_p8_s16(int16x4_t __a) {
-  return (poly8x8_t)__a; }
-__ai poly8x8_t vreinterpret_p8_s32(int32x2_t __a) {
-  return (poly8x8_t)__a; }
-__ai poly8x8_t vreinterpret_p8_s64(int64x1_t __a) {
-  return (poly8x8_t)__a; }
-__ai poly8x8_t vreinterpret_p8_u8(uint8x8_t __a) {
-  return (poly8x8_t)__a; }
-__ai poly8x8_t vreinterpret_p8_u16(uint16x4_t __a) {
-  return (poly8x8_t)__a; }
-__ai poly8x8_t vreinterpret_p8_u32(uint32x2_t __a) {
-  return (poly8x8_t)__a; }
-__ai poly8x8_t vreinterpret_p8_u64(uint64x1_t __a) {
-  return (poly8x8_t)__a; }
-__ai poly8x8_t vreinterpret_p8_f16(float16x4_t __a) {
-  return (poly8x8_t)__a; }
-__ai poly8x8_t vreinterpret_p8_f32(float32x2_t __a) {
-  return (poly8x8_t)__a; }
-__ai poly8x8_t vreinterpret_p8_p16(poly16x4_t __a) {
-  return (poly8x8_t)__a; }
-__ai poly16x4_t vreinterpret_p16_s8(int8x8_t __a) {
-  return (poly16x4_t)__a; }
-__ai poly16x4_t vreinterpret_p16_s16(int16x4_t __a) {
-  return (poly16x4_t)__a; }
-__ai poly16x4_t vreinterpret_p16_s32(int32x2_t __a) {
-  return (poly16x4_t)__a; }
-__ai poly16x4_t vreinterpret_p16_s64(int64x1_t __a) {
-  return (poly16x4_t)__a; }
-__ai poly16x4_t vreinterpret_p16_u8(uint8x8_t __a) {
-  return (poly16x4_t)__a; }
-__ai poly16x4_t vreinterpret_p16_u16(uint16x4_t __a) {
-  return (poly16x4_t)__a; }
-__ai poly16x4_t vreinterpret_p16_u32(uint32x2_t __a) {
-  return (poly16x4_t)__a; }
-__ai poly16x4_t vreinterpret_p16_u64(uint64x1_t __a) {
-  return (poly16x4_t)__a; }
-__ai poly16x4_t vreinterpret_p16_f16(float16x4_t __a) {
-  return (poly16x4_t)__a; }
-__ai poly16x4_t vreinterpret_p16_f32(float32x2_t __a) {
-  return (poly16x4_t)__a; }
-__ai poly16x4_t vreinterpret_p16_p8(poly8x8_t __a) {
-  return (poly16x4_t)__a; }
-__ai int8x16_t vreinterpretq_s8_s16(int16x8_t __a) {
-  return (int8x16_t)__a; }
-__ai int8x16_t vreinterpretq_s8_s32(int32x4_t __a) {
-  return (int8x16_t)__a; }
-__ai int8x16_t vreinterpretq_s8_s64(int64x2_t __a) {
-  return (int8x16_t)__a; }
-__ai int8x16_t vreinterpretq_s8_u8(uint8x16_t __a) {
-  return (int8x16_t)__a; }
-__ai int8x16_t vreinterpretq_s8_u16(uint16x8_t __a) {
-  return (int8x16_t)__a; }
-__ai int8x16_t vreinterpretq_s8_u32(uint32x4_t __a) {
-  return (int8x16_t)__a; }
-__ai int8x16_t vreinterpretq_s8_u64(uint64x2_t __a) {
-  return (int8x16_t)__a; }
-__ai int8x16_t vreinterpretq_s8_f16(float16x8_t __a) {
-  return (int8x16_t)__a; }
-__ai int8x16_t vreinterpretq_s8_f32(float32x4_t __a) {
-  return (int8x16_t)__a; }
-__ai int8x16_t vreinterpretq_s8_p8(poly8x16_t __a) {
-  return (int8x16_t)__a; }
-__ai int8x16_t vreinterpretq_s8_p16(poly16x8_t __a) {
-  return (int8x16_t)__a; }
-__ai int16x8_t vreinterpretq_s16_s8(int8x16_t __a) {
-  return (int16x8_t)__a; }
-__ai int16x8_t vreinterpretq_s16_s32(int32x4_t __a) {
-  return (int16x8_t)__a; }
-__ai int16x8_t vreinterpretq_s16_s64(int64x2_t __a) {
-  return (int16x8_t)__a; }
-__ai int16x8_t vreinterpretq_s16_u8(uint8x16_t __a) {
-  return (int16x8_t)__a; }
-__ai int16x8_t vreinterpretq_s16_u16(uint16x8_t __a) {
-  return (int16x8_t)__a; }
-__ai int16x8_t vreinterpretq_s16_u32(uint32x4_t __a) {
-  return (int16x8_t)__a; }
-__ai int16x8_t vreinterpretq_s16_u64(uint64x2_t __a) {
-  return (int16x8_t)__a; }
-__ai int16x8_t vreinterpretq_s16_f16(float16x8_t __a) {
-  return (int16x8_t)__a; }
-__ai int16x8_t vreinterpretq_s16_f32(float32x4_t __a) {
-  return (int16x8_t)__a; }
-__ai int16x8_t vreinterpretq_s16_p8(poly8x16_t __a) {
-  return (int16x8_t)__a; }
-__ai int16x8_t vreinterpretq_s16_p16(poly16x8_t __a) {
-  return (int16x8_t)__a; }
-__ai int32x4_t vreinterpretq_s32_s8(int8x16_t __a) {
-  return (int32x4_t)__a; }
-__ai int32x4_t vreinterpretq_s32_s16(int16x8_t __a) {
-  return (int32x4_t)__a; }
-__ai int32x4_t vreinterpretq_s32_s64(int64x2_t __a) {
-  return (int32x4_t)__a; }
-__ai int32x4_t vreinterpretq_s32_u8(uint8x16_t __a) {
-  return (int32x4_t)__a; }
-__ai int32x4_t vreinterpretq_s32_u16(uint16x8_t __a) {
-  return (int32x4_t)__a; }
-__ai int32x4_t vreinterpretq_s32_u32(uint32x4_t __a) {
-  return (int32x4_t)__a; }
-__ai int32x4_t vreinterpretq_s32_u64(uint64x2_t __a) {
-  return (int32x4_t)__a; }
-__ai int32x4_t vreinterpretq_s32_f16(float16x8_t __a) {
-  return (int32x4_t)__a; }
-__ai int32x4_t vreinterpretq_s32_f32(float32x4_t __a) {
-  return (int32x4_t)__a; }
-__ai int32x4_t vreinterpretq_s32_p8(poly8x16_t __a) {
-  return (int32x4_t)__a; }
-__ai int32x4_t vreinterpretq_s32_p16(poly16x8_t __a) {
-  return (int32x4_t)__a; }
-__ai int64x2_t vreinterpretq_s64_s8(int8x16_t __a) {
-  return (int64x2_t)__a; }
-__ai int64x2_t vreinterpretq_s64_s16(int16x8_t __a) {
-  return (int64x2_t)__a; }
-__ai int64x2_t vreinterpretq_s64_s32(int32x4_t __a) {
-  return (int64x2_t)__a; }
-__ai int64x2_t vreinterpretq_s64_u8(uint8x16_t __a) {
-  return (int64x2_t)__a; }
-__ai int64x2_t vreinterpretq_s64_u16(uint16x8_t __a) {
-  return (int64x2_t)__a; }
-__ai int64x2_t vreinterpretq_s64_u32(uint32x4_t __a) {
-  return (int64x2_t)__a; }
-__ai int64x2_t vreinterpretq_s64_u64(uint64x2_t __a) {
-  return (int64x2_t)__a; }
-__ai int64x2_t vreinterpretq_s64_f16(float16x8_t __a) {
-  return (int64x2_t)__a; }
-__ai int64x2_t vreinterpretq_s64_f32(float32x4_t __a) {
-  return (int64x2_t)__a; }
-__ai int64x2_t vreinterpretq_s64_p8(poly8x16_t __a) {
-  return (int64x2_t)__a; }
-__ai int64x2_t vreinterpretq_s64_p16(poly16x8_t __a) {
-  return (int64x2_t)__a; }
-__ai uint8x16_t vreinterpretq_u8_s8(int8x16_t __a) {
-  return (uint8x16_t)__a; }
-__ai uint8x16_t vreinterpretq_u8_s16(int16x8_t __a) {
-  return (uint8x16_t)__a; }
-__ai uint8x16_t vreinterpretq_u8_s32(int32x4_t __a) {
-  return (uint8x16_t)__a; }
-__ai uint8x16_t vreinterpretq_u8_s64(int64x2_t __a) {
-  return (uint8x16_t)__a; }
-__ai uint8x16_t vreinterpretq_u8_u16(uint16x8_t __a) {
-  return (uint8x16_t)__a; }
-__ai uint8x16_t vreinterpretq_u8_u32(uint32x4_t __a) {
-  return (uint8x16_t)__a; }
-__ai uint8x16_t vreinterpretq_u8_u64(uint64x2_t __a) {
-  return (uint8x16_t)__a; }
-__ai uint8x16_t vreinterpretq_u8_f16(float16x8_t __a) {
-  return (uint8x16_t)__a; }
-__ai uint8x16_t vreinterpretq_u8_f32(float32x4_t __a) {
-  return (uint8x16_t)__a; }
-__ai uint8x16_t vreinterpretq_u8_p8(poly8x16_t __a) {
-  return (uint8x16_t)__a; }
-__ai uint8x16_t vreinterpretq_u8_p16(poly16x8_t __a) {
-  return (uint8x16_t)__a; }
-__ai uint16x8_t vreinterpretq_u16_s8(int8x16_t __a) {
-  return (uint16x8_t)__a; }
-__ai uint16x8_t vreinterpretq_u16_s16(int16x8_t __a) {
-  return (uint16x8_t)__a; }
-__ai uint16x8_t vreinterpretq_u16_s32(int32x4_t __a) {
-  return (uint16x8_t)__a; }
-__ai uint16x8_t vreinterpretq_u16_s64(int64x2_t __a) {
-  return (uint16x8_t)__a; }
-__ai uint16x8_t vreinterpretq_u16_u8(uint8x16_t __a) {
-  return (uint16x8_t)__a; }
-__ai uint16x8_t vreinterpretq_u16_u32(uint32x4_t __a) {
-  return (uint16x8_t)__a; }
-__ai uint16x8_t vreinterpretq_u16_u64(uint64x2_t __a) {
-  return (uint16x8_t)__a; }
-__ai uint16x8_t vreinterpretq_u16_f16(float16x8_t __a) {
-  return (uint16x8_t)__a; }
-__ai uint16x8_t vreinterpretq_u16_f32(float32x4_t __a) {
-  return (uint16x8_t)__a; }
-__ai uint16x8_t vreinterpretq_u16_p8(poly8x16_t __a) {
-  return (uint16x8_t)__a; }
-__ai uint16x8_t vreinterpretq_u16_p16(poly16x8_t __a) {
-  return (uint16x8_t)__a; }
-__ai uint32x4_t vreinterpretq_u32_s8(int8x16_t __a) {
-  return (uint32x4_t)__a; }
-__ai uint32x4_t vreinterpretq_u32_s16(int16x8_t __a) {
-  return (uint32x4_t)__a; }
-__ai uint32x4_t vreinterpretq_u32_s32(int32x4_t __a) {
-  return (uint32x4_t)__a; }
-__ai uint32x4_t vreinterpretq_u32_s64(int64x2_t __a) {
-  return (uint32x4_t)__a; }
-__ai uint32x4_t vreinterpretq_u32_u8(uint8x16_t __a) {
-  return (uint32x4_t)__a; }
-__ai uint32x4_t vreinterpretq_u32_u16(uint16x8_t __a) {
-  return (uint32x4_t)__a; }
-__ai uint32x4_t vreinterpretq_u32_u64(uint64x2_t __a) {
-  return (uint32x4_t)__a; }
-__ai uint32x4_t vreinterpretq_u32_f16(float16x8_t __a) {
-  return (uint32x4_t)__a; }
-__ai uint32x4_t vreinterpretq_u32_f32(float32x4_t __a) {
-  return (uint32x4_t)__a; }
-__ai uint32x4_t vreinterpretq_u32_p8(poly8x16_t __a) {
-  return (uint32x4_t)__a; }
-__ai uint32x4_t vreinterpretq_u32_p16(poly16x8_t __a) {
-  return (uint32x4_t)__a; }
-__ai uint64x2_t vreinterpretq_u64_s8(int8x16_t __a) {
-  return (uint64x2_t)__a; }
-__ai uint64x2_t vreinterpretq_u64_s16(int16x8_t __a) {
-  return (uint64x2_t)__a; }
-__ai uint64x2_t vreinterpretq_u64_s32(int32x4_t __a) {
-  return (uint64x2_t)__a; }
-__ai uint64x2_t vreinterpretq_u64_s64(int64x2_t __a) {
-  return (uint64x2_t)__a; }
-__ai uint64x2_t vreinterpretq_u64_u8(uint8x16_t __a) {
-  return (uint64x2_t)__a; }
-__ai uint64x2_t vreinterpretq_u64_u16(uint16x8_t __a) {
-  return (uint64x2_t)__a; }
-__ai uint64x2_t vreinterpretq_u64_u32(uint32x4_t __a) {
-  return (uint64x2_t)__a; }
-__ai uint64x2_t vreinterpretq_u64_f16(float16x8_t __a) {
-  return (uint64x2_t)__a; }
-__ai uint64x2_t vreinterpretq_u64_f32(float32x4_t __a) {
-  return (uint64x2_t)__a; }
-__ai uint64x2_t vreinterpretq_u64_p8(poly8x16_t __a) {
-  return (uint64x2_t)__a; }
-__ai uint64x2_t vreinterpretq_u64_p16(poly16x8_t __a) {
-  return (uint64x2_t)__a; }
-__ai float16x8_t vreinterpretq_f16_s8(int8x16_t __a) {
-  return (float16x8_t)__a; }
-__ai float16x8_t vreinterpretq_f16_s16(int16x8_t __a) {
-  return (float16x8_t)__a; }
-__ai float16x8_t vreinterpretq_f16_s32(int32x4_t __a) {
-  return (float16x8_t)__a; }
-__ai float16x8_t vreinterpretq_f16_s64(int64x2_t __a) {
-  return (float16x8_t)__a; }
-__ai float16x8_t vreinterpretq_f16_u8(uint8x16_t __a) {
-  return (float16x8_t)__a; }
-__ai float16x8_t vreinterpretq_f16_u16(uint16x8_t __a) {
-  return (float16x8_t)__a; }
-__ai float16x8_t vreinterpretq_f16_u32(uint32x4_t __a) {
-  return (float16x8_t)__a; }
-__ai float16x8_t vreinterpretq_f16_u64(uint64x2_t __a) {
-  return (float16x8_t)__a; }
-__ai float16x8_t vreinterpretq_f16_f32(float32x4_t __a) {
-  return (float16x8_t)__a; }
-__ai float16x8_t vreinterpretq_f16_p8(poly8x16_t __a) {
-  return (float16x8_t)__a; }
-__ai float16x8_t vreinterpretq_f16_p16(poly16x8_t __a) {
-  return (float16x8_t)__a; }
-__ai float32x4_t vreinterpretq_f32_s8(int8x16_t __a) {
-  return (float32x4_t)__a; }
-__ai float32x4_t vreinterpretq_f32_s16(int16x8_t __a) {
-  return (float32x4_t)__a; }
-__ai float32x4_t vreinterpretq_f32_s32(int32x4_t __a) {
-  return (float32x4_t)__a; }
-__ai float32x4_t vreinterpretq_f32_s64(int64x2_t __a) {
-  return (float32x4_t)__a; }
-__ai float32x4_t vreinterpretq_f32_u8(uint8x16_t __a) {
-  return (float32x4_t)__a; }
-__ai float32x4_t vreinterpretq_f32_u16(uint16x8_t __a) {
-  return (float32x4_t)__a; }
-__ai float32x4_t vreinterpretq_f32_u32(uint32x4_t __a) {
-  return (float32x4_t)__a; }
-__ai float32x4_t vreinterpretq_f32_u64(uint64x2_t __a) {
-  return (float32x4_t)__a; }
-__ai float32x4_t vreinterpretq_f32_f16(float16x8_t __a) {
-  return (float32x4_t)__a; }
-__ai float32x4_t vreinterpretq_f32_p8(poly8x16_t __a) {
-  return (float32x4_t)__a; }
-__ai float32x4_t vreinterpretq_f32_p16(poly16x8_t __a) {
-  return (float32x4_t)__a; }
-__ai poly8x16_t vreinterpretq_p8_s8(int8x16_t __a) {
-  return (poly8x16_t)__a; }
-__ai poly8x16_t vreinterpretq_p8_s16(int16x8_t __a) {
-  return (poly8x16_t)__a; }
-__ai poly8x16_t vreinterpretq_p8_s32(int32x4_t __a) {
-  return (poly8x16_t)__a; }
-__ai poly8x16_t vreinterpretq_p8_s64(int64x2_t __a) {
-  return (poly8x16_t)__a; }
-__ai poly8x16_t vreinterpretq_p8_u8(uint8x16_t __a) {
-  return (poly8x16_t)__a; }
-__ai poly8x16_t vreinterpretq_p8_u16(uint16x8_t __a) {
-  return (poly8x16_t)__a; }
-__ai poly8x16_t vreinterpretq_p8_u32(uint32x4_t __a) {
-  return (poly8x16_t)__a; }
-__ai poly8x16_t vreinterpretq_p8_u64(uint64x2_t __a) {
-  return (poly8x16_t)__a; }
-__ai poly8x16_t vreinterpretq_p8_f16(float16x8_t __a) {
-  return (poly8x16_t)__a; }
-__ai poly8x16_t vreinterpretq_p8_f32(float32x4_t __a) {
-  return (poly8x16_t)__a; }
-__ai poly8x16_t vreinterpretq_p8_p16(poly16x8_t __a) {
-  return (poly8x16_t)__a; }
-__ai poly16x8_t vreinterpretq_p16_s8(int8x16_t __a) {
-  return (poly16x8_t)__a; }
-__ai poly16x8_t vreinterpretq_p16_s16(int16x8_t __a) {
-  return (poly16x8_t)__a; }
-__ai poly16x8_t vreinterpretq_p16_s32(int32x4_t __a) {
-  return (poly16x8_t)__a; }
-__ai poly16x8_t vreinterpretq_p16_s64(int64x2_t __a) {
-  return (poly16x8_t)__a; }
-__ai poly16x8_t vreinterpretq_p16_u8(uint8x16_t __a) {
-  return (poly16x8_t)__a; }
-__ai poly16x8_t vreinterpretq_p16_u16(uint16x8_t __a) {
-  return (poly16x8_t)__a; }
-__ai poly16x8_t vreinterpretq_p16_u32(uint32x4_t __a) {
-  return (poly16x8_t)__a; }
-__ai poly16x8_t vreinterpretq_p16_u64(uint64x2_t __a) {
-  return (poly16x8_t)__a; }
-__ai poly16x8_t vreinterpretq_p16_f16(float16x8_t __a) {
-  return (poly16x8_t)__a; }
-__ai poly16x8_t vreinterpretq_p16_f32(float32x4_t __a) {
-  return (poly16x8_t)__a; }
-__ai poly16x8_t vreinterpretq_p16_p8(poly8x16_t __a) {
-  return (poly16x8_t)__a; }
-
-__ai int8x8_t vrev16_s8(int8x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 1, 0, 3, 2, 5, 4, 7, 6); }
-__ai uint8x8_t vrev16_u8(uint8x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 1, 0, 3, 2, 5, 4, 7, 6); }
-__ai poly8x8_t vrev16_p8(poly8x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 1, 0, 3, 2, 5, 4, 7, 6); }
-__ai int8x16_t vrev16q_s8(int8x16_t __a) {
-  return __builtin_shufflevector(__a, __a, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14); }
-__ai uint8x16_t vrev16q_u8(uint8x16_t __a) {
-  return __builtin_shufflevector(__a, __a, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14); }
-__ai poly8x16_t vrev16q_p8(poly8x16_t __a) {
-  return __builtin_shufflevector(__a, __a, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14); }
-
-__ai int8x8_t vrev32_s8(int8x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 3, 2, 1, 0, 7, 6, 5, 4); }
-__ai int16x4_t vrev32_s16(int16x4_t __a) {
-  return __builtin_shufflevector(__a, __a, 1, 0, 3, 2); }
-__ai uint8x8_t vrev32_u8(uint8x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 3, 2, 1, 0, 7, 6, 5, 4); }
-__ai uint16x4_t vrev32_u16(uint16x4_t __a) {
-  return __builtin_shufflevector(__a, __a, 1, 0, 3, 2); }
-__ai poly8x8_t vrev32_p8(poly8x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 3, 2, 1, 0, 7, 6, 5, 4); }
-__ai poly16x4_t vrev32_p16(poly16x4_t __a) {
-  return __builtin_shufflevector(__a, __a, 1, 0, 3, 2); }
-__ai int8x16_t vrev32q_s8(int8x16_t __a) {
-  return __builtin_shufflevector(__a, __a, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12); }
-__ai int16x8_t vrev32q_s16(int16x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 1, 0, 3, 2, 5, 4, 7, 6); }
-__ai uint8x16_t vrev32q_u8(uint8x16_t __a) {
-  return __builtin_shufflevector(__a, __a, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12); }
-__ai uint16x8_t vrev32q_u16(uint16x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 1, 0, 3, 2, 5, 4, 7, 6); }
-__ai poly8x16_t vrev32q_p8(poly8x16_t __a) {
-  return __builtin_shufflevector(__a, __a, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12); }
-__ai poly16x8_t vrev32q_p16(poly16x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 1, 0, 3, 2, 5, 4, 7, 6); }
-
-__ai int8x8_t vrev64_s8(int8x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 7, 6, 5, 4, 3, 2, 1, 0); }
-__ai int16x4_t vrev64_s16(int16x4_t __a) {
-  return __builtin_shufflevector(__a, __a, 3, 2, 1, 0); }
-__ai int32x2_t vrev64_s32(int32x2_t __a) {
-  return __builtin_shufflevector(__a, __a, 1, 0); }
-__ai uint8x8_t vrev64_u8(uint8x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 7, 6, 5, 4, 3, 2, 1, 0); }
-__ai uint16x4_t vrev64_u16(uint16x4_t __a) {
-  return __builtin_shufflevector(__a, __a, 3, 2, 1, 0); }
-__ai uint32x2_t vrev64_u32(uint32x2_t __a) {
-  return __builtin_shufflevector(__a, __a, 1, 0); }
-__ai poly8x8_t vrev64_p8(poly8x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 7, 6, 5, 4, 3, 2, 1, 0); }
-__ai poly16x4_t vrev64_p16(poly16x4_t __a) {
-  return __builtin_shufflevector(__a, __a, 3, 2, 1, 0); }
-__ai float32x2_t vrev64_f32(float32x2_t __a) {
-  return __builtin_shufflevector(__a, __a, 1, 0); }
-__ai int8x16_t vrev64q_s8(int8x16_t __a) {
-  return __builtin_shufflevector(__a, __a, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8); }
-__ai int16x8_t vrev64q_s16(int16x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 3, 2, 1, 0, 7, 6, 5, 4); }
-__ai int32x4_t vrev64q_s32(int32x4_t __a) {
-  return __builtin_shufflevector(__a, __a, 1, 0, 3, 2); }
-__ai uint8x16_t vrev64q_u8(uint8x16_t __a) {
-  return __builtin_shufflevector(__a, __a, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8); }
-__ai uint16x8_t vrev64q_u16(uint16x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 3, 2, 1, 0, 7, 6, 5, 4); }
-__ai uint32x4_t vrev64q_u32(uint32x4_t __a) {
-  return __builtin_shufflevector(__a, __a, 1, 0, 3, 2); }
-__ai poly8x16_t vrev64q_p8(poly8x16_t __a) {
-  return __builtin_shufflevector(__a, __a, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8); }
-__ai poly16x8_t vrev64q_p16(poly16x8_t __a) {
-  return __builtin_shufflevector(__a, __a, 3, 2, 1, 0, 7, 6, 5, 4); }
-__ai float32x4_t vrev64q_f32(float32x4_t __a) {
-  return __builtin_shufflevector(__a, __a, 1, 0, 3, 2); }
-
-__ai int8x8_t vrhadd_s8(int8x8_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vrhadd_v(__a, __b, 0); }
-__ai int16x4_t vrhadd_s16(int16x4_t __a, int16x4_t __b) {
-  return (int16x4_t)__builtin_neon_vrhadd_v((int8x8_t)__a, (int8x8_t)__b, 1); }
-__ai int32x2_t vrhadd_s32(int32x2_t __a, int32x2_t __b) {
-  return (int32x2_t)__builtin_neon_vrhadd_v((int8x8_t)__a, (int8x8_t)__b, 2); }
-__ai uint8x8_t vrhadd_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vrhadd_v((int8x8_t)__a, (int8x8_t)__b, 16); }
-__ai uint16x4_t vrhadd_u16(uint16x4_t __a, uint16x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vrhadd_v((int8x8_t)__a, (int8x8_t)__b, 17); }
-__ai uint32x2_t vrhadd_u32(uint32x2_t __a, uint32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vrhadd_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai int8x16_t vrhaddq_s8(int8x16_t __a, int8x16_t __b) {
-  return (int8x16_t)__builtin_neon_vrhaddq_v(__a, __b, 32); }
-__ai int16x8_t vrhaddq_s16(int16x8_t __a, int16x8_t __b) {
-  return (int16x8_t)__builtin_neon_vrhaddq_v((int8x16_t)__a, (int8x16_t)__b, 33); }
-__ai int32x4_t vrhaddq_s32(int32x4_t __a, int32x4_t __b) {
-  return (int32x4_t)__builtin_neon_vrhaddq_v((int8x16_t)__a, (int8x16_t)__b, 34); }
-__ai uint8x16_t vrhaddq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vrhaddq_v((int8x16_t)__a, (int8x16_t)__b, 48); }
-__ai uint16x8_t vrhaddq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint16x8_t)__builtin_neon_vrhaddq_v((int8x16_t)__a, (int8x16_t)__b, 49); }
-__ai uint32x4_t vrhaddq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vrhaddq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-
-__ai int8x8_t vrshl_s8(int8x8_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vrshl_v(__a, __b, 0); }
-__ai int16x4_t vrshl_s16(int16x4_t __a, int16x4_t __b) {
-  return (int16x4_t)__builtin_neon_vrshl_v((int8x8_t)__a, (int8x8_t)__b, 1); }
-__ai int32x2_t vrshl_s32(int32x2_t __a, int32x2_t __b) {
-  return (int32x2_t)__builtin_neon_vrshl_v((int8x8_t)__a, (int8x8_t)__b, 2); }
-__ai int64x1_t vrshl_s64(int64x1_t __a, int64x1_t __b) {
-  return (int64x1_t)__builtin_neon_vrshl_v((int8x8_t)__a, (int8x8_t)__b, 3); }
-__ai uint8x8_t vrshl_u8(uint8x8_t __a, int8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vrshl_v((int8x8_t)__a, __b, 16); }
-__ai uint16x4_t vrshl_u16(uint16x4_t __a, int16x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vrshl_v((int8x8_t)__a, (int8x8_t)__b, 17); }
-__ai uint32x2_t vrshl_u32(uint32x2_t __a, int32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vrshl_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai uint64x1_t vrshl_u64(uint64x1_t __a, int64x1_t __b) {
-  return (uint64x1_t)__builtin_neon_vrshl_v((int8x8_t)__a, (int8x8_t)__b, 19); }
-__ai int8x16_t vrshlq_s8(int8x16_t __a, int8x16_t __b) {
-  return (int8x16_t)__builtin_neon_vrshlq_v(__a, __b, 32); }
-__ai int16x8_t vrshlq_s16(int16x8_t __a, int16x8_t __b) {
-  return (int16x8_t)__builtin_neon_vrshlq_v((int8x16_t)__a, (int8x16_t)__b, 33); }
-__ai int32x4_t vrshlq_s32(int32x4_t __a, int32x4_t __b) {
-  return (int32x4_t)__builtin_neon_vrshlq_v((int8x16_t)__a, (int8x16_t)__b, 34); }
-__ai int64x2_t vrshlq_s64(int64x2_t __a, int64x2_t __b) {
-  return (int64x2_t)__builtin_neon_vrshlq_v((int8x16_t)__a, (int8x16_t)__b, 35); }
-__ai uint8x16_t vrshlq_u8(uint8x16_t __a, int8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vrshlq_v((int8x16_t)__a, __b, 48); }
-__ai uint16x8_t vrshlq_u16(uint16x8_t __a, int16x8_t __b) {
-  return (uint16x8_t)__builtin_neon_vrshlq_v((int8x16_t)__a, (int8x16_t)__b, 49); }
-__ai uint32x4_t vrshlq_u32(uint32x4_t __a, int32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vrshlq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-__ai uint64x2_t vrshlq_u64(uint64x2_t __a, int64x2_t __b) {
-  return (uint64x2_t)__builtin_neon_vrshlq_v((int8x16_t)__a, (int8x16_t)__b, 51); }
-
-#define vrshrn_n_s16(a, __b) __extension__ ({ \
-  int16x8_t __a = (a); \
-  (int8x8_t)__builtin_neon_vrshrn_n_v((int8x16_t)__a, __b, 0); })
-#define vrshrn_n_s32(a, __b) __extension__ ({ \
-  int32x4_t __a = (a); \
-  (int16x4_t)__builtin_neon_vrshrn_n_v((int8x16_t)__a, __b, 1); })
-#define vrshrn_n_s64(a, __b) __extension__ ({ \
-  int64x2_t __a = (a); \
-  (int32x2_t)__builtin_neon_vrshrn_n_v((int8x16_t)__a, __b, 2); })
-#define vrshrn_n_u16(a, __b) __extension__ ({ \
-  uint16x8_t __a = (a); \
-  (uint8x8_t)__builtin_neon_vrshrn_n_v((int8x16_t)__a, __b, 16); })
-#define vrshrn_n_u32(a, __b) __extension__ ({ \
-  uint32x4_t __a = (a); \
-  (uint16x4_t)__builtin_neon_vrshrn_n_v((int8x16_t)__a, __b, 17); })
-#define vrshrn_n_u64(a, __b) __extension__ ({ \
-  uint64x2_t __a = (a); \
-  (uint32x2_t)__builtin_neon_vrshrn_n_v((int8x16_t)__a, __b, 18); })
-
-#define vrshr_n_s8(a, __b) __extension__ ({ \
-  int8x8_t __a = (a); \
-  (int8x8_t)__builtin_neon_vrshr_n_v(__a, __b, 0); })
-#define vrshr_n_s16(a, __b) __extension__ ({ \
-  int16x4_t __a = (a); \
-  (int16x4_t)__builtin_neon_vrshr_n_v((int8x8_t)__a, __b, 1); })
-#define vrshr_n_s32(a, __b) __extension__ ({ \
-  int32x2_t __a = (a); \
-  (int32x2_t)__builtin_neon_vrshr_n_v((int8x8_t)__a, __b, 2); })
-#define vrshr_n_s64(a, __b) __extension__ ({ \
-  int64x1_t __a = (a); \
-  (int64x1_t)__builtin_neon_vrshr_n_v((int8x8_t)__a, __b, 3); })
-#define vrshr_n_u8(a, __b) __extension__ ({ \
-  uint8x8_t __a = (a); \
-  (uint8x8_t)__builtin_neon_vrshr_n_v((int8x8_t)__a, __b, 16); })
-#define vrshr_n_u16(a, __b) __extension__ ({ \
-  uint16x4_t __a = (a); \
-  (uint16x4_t)__builtin_neon_vrshr_n_v((int8x8_t)__a, __b, 17); })
-#define vrshr_n_u32(a, __b) __extension__ ({ \
-  uint32x2_t __a = (a); \
-  (uint32x2_t)__builtin_neon_vrshr_n_v((int8x8_t)__a, __b, 18); })
-#define vrshr_n_u64(a, __b) __extension__ ({ \
-  uint64x1_t __a = (a); \
-  (uint64x1_t)__builtin_neon_vrshr_n_v((int8x8_t)__a, __b, 19); })
-#define vrshrq_n_s8(a, __b) __extension__ ({ \
-  int8x16_t __a = (a); \
-  (int8x16_t)__builtin_neon_vrshrq_n_v(__a, __b, 32); })
-#define vrshrq_n_s16(a, __b) __extension__ ({ \
-  int16x8_t __a = (a); \
-  (int16x8_t)__builtin_neon_vrshrq_n_v((int8x16_t)__a, __b, 33); })
-#define vrshrq_n_s32(a, __b) __extension__ ({ \
-  int32x4_t __a = (a); \
-  (int32x4_t)__builtin_neon_vrshrq_n_v((int8x16_t)__a, __b, 34); })
-#define vrshrq_n_s64(a, __b) __extension__ ({ \
-  int64x2_t __a = (a); \
-  (int64x2_t)__builtin_neon_vrshrq_n_v((int8x16_t)__a, __b, 35); })
-#define vrshrq_n_u8(a, __b) __extension__ ({ \
-  uint8x16_t __a = (a); \
-  (uint8x16_t)__builtin_neon_vrshrq_n_v((int8x16_t)__a, __b, 48); })
-#define vrshrq_n_u16(a, __b) __extension__ ({ \
-  uint16x8_t __a = (a); \
-  (uint16x8_t)__builtin_neon_vrshrq_n_v((int8x16_t)__a, __b, 49); })
-#define vrshrq_n_u32(a, __b) __extension__ ({ \
-  uint32x4_t __a = (a); \
-  (uint32x4_t)__builtin_neon_vrshrq_n_v((int8x16_t)__a, __b, 50); })
-#define vrshrq_n_u64(a, __b) __extension__ ({ \
-  uint64x2_t __a = (a); \
-  (uint64x2_t)__builtin_neon_vrshrq_n_v((int8x16_t)__a, __b, 51); })
-
-__ai float32x2_t vrsqrte_f32(float32x2_t __a) {
-  return (float32x2_t)__builtin_neon_vrsqrte_v((int8x8_t)__a, 9); }
-__ai uint32x2_t vrsqrte_u32(uint32x2_t __a) {
-  return (uint32x2_t)__builtin_neon_vrsqrte_v((int8x8_t)__a, 18); }
-__ai float32x4_t vrsqrteq_f32(float32x4_t __a) {
-  return (float32x4_t)__builtin_neon_vrsqrteq_v((int8x16_t)__a, 41); }
-__ai uint32x4_t vrsqrteq_u32(uint32x4_t __a) {
-  return (uint32x4_t)__builtin_neon_vrsqrteq_v((int8x16_t)__a, 50); }
-
-__ai float32x2_t vrsqrts_f32(float32x2_t __a, float32x2_t __b) {
-  return (float32x2_t)__builtin_neon_vrsqrts_v((int8x8_t)__a, (int8x8_t)__b, 9); }
-__ai float32x4_t vrsqrtsq_f32(float32x4_t __a, float32x4_t __b) {
-  return (float32x4_t)__builtin_neon_vrsqrtsq_v((int8x16_t)__a, (int8x16_t)__b, 41); }
-
-#define vrsra_n_s8(a, b, __c) __extension__ ({ \
-  int8x8_t __a = (a); int8x8_t __b = (b); \
-  (int8x8_t)__builtin_neon_vrsra_n_v(__a, __b, __c, 0); })
-#define vrsra_n_s16(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int16x4_t __b = (b); \
-  (int16x4_t)__builtin_neon_vrsra_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 1); })
-#define vrsra_n_s32(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int32x2_t __b = (b); \
-  (int32x2_t)__builtin_neon_vrsra_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 2); })
-#define vrsra_n_s64(a, b, __c) __extension__ ({ \
-  int64x1_t __a = (a); int64x1_t __b = (b); \
-  (int64x1_t)__builtin_neon_vrsra_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 3); })
-#define vrsra_n_u8(a, b, __c) __extension__ ({ \
-  uint8x8_t __a = (a); uint8x8_t __b = (b); \
-  (uint8x8_t)__builtin_neon_vrsra_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 16); })
-#define vrsra_n_u16(a, b, __c) __extension__ ({ \
-  uint16x4_t __a = (a); uint16x4_t __b = (b); \
-  (uint16x4_t)__builtin_neon_vrsra_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 17); })
-#define vrsra_n_u32(a, b, __c) __extension__ ({ \
-  uint32x2_t __a = (a); uint32x2_t __b = (b); \
-  (uint32x2_t)__builtin_neon_vrsra_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 18); })
-#define vrsra_n_u64(a, b, __c) __extension__ ({ \
-  uint64x1_t __a = (a); uint64x1_t __b = (b); \
-  (uint64x1_t)__builtin_neon_vrsra_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 19); })
-#define vrsraq_n_s8(a, b, __c) __extension__ ({ \
-  int8x16_t __a = (a); int8x16_t __b = (b); \
-  (int8x16_t)__builtin_neon_vrsraq_n_v(__a, __b, __c, 32); })
-#define vrsraq_n_s16(a, b, __c) __extension__ ({ \
-  int16x8_t __a = (a); int16x8_t __b = (b); \
-  (int16x8_t)__builtin_neon_vrsraq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 33); })
-#define vrsraq_n_s32(a, b, __c) __extension__ ({ \
-  int32x4_t __a = (a); int32x4_t __b = (b); \
-  (int32x4_t)__builtin_neon_vrsraq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 34); })
-#define vrsraq_n_s64(a, b, __c) __extension__ ({ \
-  int64x2_t __a = (a); int64x2_t __b = (b); \
-  (int64x2_t)__builtin_neon_vrsraq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 35); })
-#define vrsraq_n_u8(a, b, __c) __extension__ ({ \
-  uint8x16_t __a = (a); uint8x16_t __b = (b); \
-  (uint8x16_t)__builtin_neon_vrsraq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 48); })
-#define vrsraq_n_u16(a, b, __c) __extension__ ({ \
-  uint16x8_t __a = (a); uint16x8_t __b = (b); \
-  (uint16x8_t)__builtin_neon_vrsraq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 49); })
-#define vrsraq_n_u32(a, b, __c) __extension__ ({ \
-  uint32x4_t __a = (a); uint32x4_t __b = (b); \
-  (uint32x4_t)__builtin_neon_vrsraq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 50); })
-#define vrsraq_n_u64(a, b, __c) __extension__ ({ \
-  uint64x2_t __a = (a); uint64x2_t __b = (b); \
-  (uint64x2_t)__builtin_neon_vrsraq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 51); })
-
-__ai int8x8_t vrsubhn_s16(int16x8_t __a, int16x8_t __b) {
-  return (int8x8_t)__builtin_neon_vrsubhn_v((int8x16_t)__a, (int8x16_t)__b, 0); }
-__ai int16x4_t vrsubhn_s32(int32x4_t __a, int32x4_t __b) {
-  return (int16x4_t)__builtin_neon_vrsubhn_v((int8x16_t)__a, (int8x16_t)__b, 1); }
-__ai int32x2_t vrsubhn_s64(int64x2_t __a, int64x2_t __b) {
-  return (int32x2_t)__builtin_neon_vrsubhn_v((int8x16_t)__a, (int8x16_t)__b, 2); }
-__ai uint8x8_t vrsubhn_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vrsubhn_v((int8x16_t)__a, (int8x16_t)__b, 16); }
-__ai uint16x4_t vrsubhn_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vrsubhn_v((int8x16_t)__a, (int8x16_t)__b, 17); }
-__ai uint32x2_t vrsubhn_u64(uint64x2_t __a, uint64x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vrsubhn_v((int8x16_t)__a, (int8x16_t)__b, 18); }
+#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
 
-#if defined(__aarch64__)
-__ai int8x16_t vrsubhn_high_s16(int8x8_t __a, int16x8_t __b, int16x8_t __c) {
-  return vcombine_s8(__a, vrsubhn_s16(__b, __c)); }
-__ai int16x8_t vrsubhn_high_s32(int16x4_t __a, int32x4_t __b, int32x4_t __c) {
-  return vcombine_s16(__a, vrsubhn_s32(__b, __c)); }
-__ai int32x4_t vrsubhn_high_s64(int32x2_t __a, int64x2_t __b, int64x2_t __c) {
-  return vcombine_s32(__a, vrsubhn_s64(__b, __c)); }
-__ai uint8x16_t vrsubhn_high_u16(uint8x8_t __a, uint16x8_t __b, uint16x8_t __c) {
-  return vcombine_u8(__a, vrsubhn_u16(__b, __c)); }
-__ai uint16x8_t vrsubhn_high_u32(uint16x4_t __a, uint32x4_t __b, uint32x4_t __c) {
-  return vcombine_u16(__a, vrsubhn_u32(__b, __c)); }
-__ai uint32x4_t vrsubhn_high_u64(uint32x2_t __a, uint64x2_t __b, uint64x2_t __c) {
-  return vcombine_u32(__a, vrsubhn_u64(__b, __c)); }
-
-#endif
-
-#define vset_lane_u8(a, b, __c) __extension__ ({ \
-  uint8_t __a = (a); uint8x8_t __b = (b); \
-  (uint8x8_t)__builtin_neon_vset_lane_i8(__a, (int8x8_t)__b, __c); })
-#define vset_lane_u16(a, b, __c) __extension__ ({ \
-  uint16_t __a = (a); uint16x4_t __b = (b); \
-  (uint16x4_t)__builtin_neon_vset_lane_i16(__a, (int16x4_t)__b, __c); })
-#define vset_lane_u32(a, b, __c) __extension__ ({ \
-  uint32_t __a = (a); uint32x2_t __b = (b); \
-  (uint32x2_t)__builtin_neon_vset_lane_i32(__a, (int32x2_t)__b, __c); })
-#define vset_lane_s8(a, b, __c) __extension__ ({ \
-  int8_t __a = (a); int8x8_t __b = (b); \
-  (int8x8_t)__builtin_neon_vset_lane_i8(__a, __b, __c); })
-#define vset_lane_s16(a, b, __c) __extension__ ({ \
-  int16_t __a = (a); int16x4_t __b = (b); \
-  (int16x4_t)__builtin_neon_vset_lane_i16(__a, __b, __c); })
-#define vset_lane_s32(a, b, __c) __extension__ ({ \
-  int32_t __a = (a); int32x2_t __b = (b); \
-  (int32x2_t)__builtin_neon_vset_lane_i32(__a, __b, __c); })
-#define vset_lane_p8(a, b, __c) __extension__ ({ \
-  poly8_t __a = (a); poly8x8_t __b = (b); \
-  (poly8x8_t)__builtin_neon_vset_lane_i8(__a, (int8x8_t)__b, __c); })
-#define vset_lane_p16(a, b, __c) __extension__ ({ \
-  poly16_t __a = (a); poly16x4_t __b = (b); \
-  (poly16x4_t)__builtin_neon_vset_lane_i16(__a, (int16x4_t)__b, __c); })
-#define vset_lane_f32(a, b, __c) __extension__ ({ \
-  float32_t __a = (a); float32x2_t __b = (b); \
-  (float32x2_t)__builtin_neon_vset_lane_f32(__a, __b, __c); })
-#define vsetq_lane_u8(a, b, __c) __extension__ ({ \
-  uint8_t __a = (a); uint8x16_t __b = (b); \
-  (uint8x16_t)__builtin_neon_vsetq_lane_i8(__a, (int8x16_t)__b, __c); })
-#define vsetq_lane_u16(a, b, __c) __extension__ ({ \
-  uint16_t __a = (a); uint16x8_t __b = (b); \
-  (uint16x8_t)__builtin_neon_vsetq_lane_i16(__a, (int16x8_t)__b, __c); })
-#define vsetq_lane_u32(a, b, __c) __extension__ ({ \
-  uint32_t __a = (a); uint32x4_t __b = (b); \
-  (uint32x4_t)__builtin_neon_vsetq_lane_i32(__a, (int32x4_t)__b, __c); })
-#define vsetq_lane_s8(a, b, __c) __extension__ ({ \
-  int8_t __a = (a); int8x16_t __b = (b); \
-  (int8x16_t)__builtin_neon_vsetq_lane_i8(__a, __b, __c); })
-#define vsetq_lane_s16(a, b, __c) __extension__ ({ \
-  int16_t __a = (a); int16x8_t __b = (b); \
-  (int16x8_t)__builtin_neon_vsetq_lane_i16(__a, __b, __c); })
-#define vsetq_lane_s32(a, b, __c) __extension__ ({ \
-  int32_t __a = (a); int32x4_t __b = (b); \
-  (int32x4_t)__builtin_neon_vsetq_lane_i32(__a, __b, __c); })
-#define vsetq_lane_p8(a, b, __c) __extension__ ({ \
-  poly8_t __a = (a); poly8x16_t __b = (b); \
-  (poly8x16_t)__builtin_neon_vsetq_lane_i8(__a, (int8x16_t)__b, __c); })
-#define vsetq_lane_p16(a, b, __c) __extension__ ({ \
-  poly16_t __a = (a); poly16x8_t __b = (b); \
-  (poly16x8_t)__builtin_neon_vsetq_lane_i16(__a, (int16x8_t)__b, __c); })
-#define vsetq_lane_f32(a, b, __c) __extension__ ({ \
-  float32_t __a = (a); float32x4_t __b = (b); \
-  (float32x4_t)__builtin_neon_vsetq_lane_f32(__a, __b, __c); })
-#define vset_lane_s64(a, b, __c) __extension__ ({ \
-  int64_t __a = (a); int64x1_t __b = (b); \
-  (int64x1_t)__builtin_neon_vset_lane_i64(__a, __b, __c); })
-#define vset_lane_u64(a, b, __c) __extension__ ({ \
-  uint64_t __a = (a); uint64x1_t __b = (b); \
-  (uint64x1_t)__builtin_neon_vset_lane_i64(__a, (int64x1_t)__b, __c); })
-#define vsetq_lane_s64(a, b, __c) __extension__ ({ \
-  int64_t __a = (a); int64x2_t __b = (b); \
-  (int64x2_t)__builtin_neon_vsetq_lane_i64(__a, __b, __c); })
-#define vsetq_lane_u64(a, b, __c) __extension__ ({ \
-  uint64_t __a = (a); uint64x2_t __b = (b); \
-  (uint64x2_t)__builtin_neon_vsetq_lane_i64(__a, (int64x2_t)__b, __c); })
-
-__ai int8x8_t vshl_s8(int8x8_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vshl_v(__a, __b, 0); }
-__ai int16x4_t vshl_s16(int16x4_t __a, int16x4_t __b) {
-  return (int16x4_t)__builtin_neon_vshl_v((int8x8_t)__a, (int8x8_t)__b, 1); }
-__ai int32x2_t vshl_s32(int32x2_t __a, int32x2_t __b) {
-  return (int32x2_t)__builtin_neon_vshl_v((int8x8_t)__a, (int8x8_t)__b, 2); }
-__ai int64x1_t vshl_s64(int64x1_t __a, int64x1_t __b) {
-  return (int64x1_t)__builtin_neon_vshl_v((int8x8_t)__a, (int8x8_t)__b, 3); }
-__ai uint8x8_t vshl_u8(uint8x8_t __a, int8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vshl_v((int8x8_t)__a, __b, 16); }
-__ai uint16x4_t vshl_u16(uint16x4_t __a, int16x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vshl_v((int8x8_t)__a, (int8x8_t)__b, 17); }
-__ai uint32x2_t vshl_u32(uint32x2_t __a, int32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vshl_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai uint64x1_t vshl_u64(uint64x1_t __a, int64x1_t __b) {
-  return (uint64x1_t)__builtin_neon_vshl_v((int8x8_t)__a, (int8x8_t)__b, 19); }
-__ai int8x16_t vshlq_s8(int8x16_t __a, int8x16_t __b) {
-  return (int8x16_t)__builtin_neon_vshlq_v(__a, __b, 32); }
-__ai int16x8_t vshlq_s16(int16x8_t __a, int16x8_t __b) {
-  return (int16x8_t)__builtin_neon_vshlq_v((int8x16_t)__a, (int8x16_t)__b, 33); }
-__ai int32x4_t vshlq_s32(int32x4_t __a, int32x4_t __b) {
-  return (int32x4_t)__builtin_neon_vshlq_v((int8x16_t)__a, (int8x16_t)__b, 34); }
-__ai int64x2_t vshlq_s64(int64x2_t __a, int64x2_t __b) {
-  return (int64x2_t)__builtin_neon_vshlq_v((int8x16_t)__a, (int8x16_t)__b, 35); }
-__ai uint8x16_t vshlq_u8(uint8x16_t __a, int8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vshlq_v((int8x16_t)__a, __b, 48); }
-__ai uint16x8_t vshlq_u16(uint16x8_t __a, int16x8_t __b) {
-  return (uint16x8_t)__builtin_neon_vshlq_v((int8x16_t)__a, (int8x16_t)__b, 49); }
-__ai uint32x4_t vshlq_u32(uint32x4_t __a, int32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vshlq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-__ai uint64x2_t vshlq_u64(uint64x2_t __a, int64x2_t __b) {
-  return (uint64x2_t)__builtin_neon_vshlq_v((int8x16_t)__a, (int8x16_t)__b, 51); }
-
-#define vshl_n_s8(a, __b) __extension__ ({ \
-  int8x8_t __a = (a); \
-  (int8x8_t)__builtin_neon_vshl_n_v(__a, __b, 0); })
-#define vshl_n_s16(a, __b) __extension__ ({ \
-  int16x4_t __a = (a); \
-  (int16x4_t)__builtin_neon_vshl_n_v((int8x8_t)__a, __b, 1); })
-#define vshl_n_s32(a, __b) __extension__ ({ \
-  int32x2_t __a = (a); \
-  (int32x2_t)__builtin_neon_vshl_n_v((int8x8_t)__a, __b, 2); })
-#define vshl_n_s64(a, __b) __extension__ ({ \
-  int64x1_t __a = (a); \
-  (int64x1_t)__builtin_neon_vshl_n_v((int8x8_t)__a, __b, 3); })
-#define vshl_n_u8(a, __b) __extension__ ({ \
-  uint8x8_t __a = (a); \
-  (uint8x8_t)__builtin_neon_vshl_n_v((int8x8_t)__a, __b, 16); })
-#define vshl_n_u16(a, __b) __extension__ ({ \
-  uint16x4_t __a = (a); \
-  (uint16x4_t)__builtin_neon_vshl_n_v((int8x8_t)__a, __b, 17); })
-#define vshl_n_u32(a, __b) __extension__ ({ \
-  uint32x2_t __a = (a); \
-  (uint32x2_t)__builtin_neon_vshl_n_v((int8x8_t)__a, __b, 18); })
-#define vshl_n_u64(a, __b) __extension__ ({ \
-  uint64x1_t __a = (a); \
-  (uint64x1_t)__builtin_neon_vshl_n_v((int8x8_t)__a, __b, 19); })
-#define vshlq_n_s8(a, __b) __extension__ ({ \
-  int8x16_t __a = (a); \
-  (int8x16_t)__builtin_neon_vshlq_n_v(__a, __b, 32); })
-#define vshlq_n_s16(a, __b) __extension__ ({ \
-  int16x8_t __a = (a); \
-  (int16x8_t)__builtin_neon_vshlq_n_v((int8x16_t)__a, __b, 33); })
-#define vshlq_n_s32(a, __b) __extension__ ({ \
-  int32x4_t __a = (a); \
-  (int32x4_t)__builtin_neon_vshlq_n_v((int8x16_t)__a, __b, 34); })
-#define vshlq_n_s64(a, __b) __extension__ ({ \
-  int64x2_t __a = (a); \
-  (int64x2_t)__builtin_neon_vshlq_n_v((int8x16_t)__a, __b, 35); })
-#define vshlq_n_u8(a, __b) __extension__ ({ \
-  uint8x16_t __a = (a); \
-  (uint8x16_t)__builtin_neon_vshlq_n_v((int8x16_t)__a, __b, 48); })
-#define vshlq_n_u16(a, __b) __extension__ ({ \
-  uint16x8_t __a = (a); \
-  (uint16x8_t)__builtin_neon_vshlq_n_v((int8x16_t)__a, __b, 49); })
-#define vshlq_n_u32(a, __b) __extension__ ({ \
-  uint32x4_t __a = (a); \
-  (uint32x4_t)__builtin_neon_vshlq_n_v((int8x16_t)__a, __b, 50); })
-#define vshlq_n_u64(a, __b) __extension__ ({ \
-  uint64x2_t __a = (a); \
-  (uint64x2_t)__builtin_neon_vshlq_n_v((int8x16_t)__a, __b, 51); })
-
-#define vshrn_n_s16(a, __b) __extension__ ({ \
-  int16x8_t __a = (a); \
-  (int8x8_t)__builtin_neon_vshrn_n_v((int8x16_t)__a, __b, 0); })
-#define vshrn_n_s32(a, __b) __extension__ ({ \
-  int32x4_t __a = (a); \
-  (int16x4_t)__builtin_neon_vshrn_n_v((int8x16_t)__a, __b, 1); })
-#define vshrn_n_s64(a, __b) __extension__ ({ \
-  int64x2_t __a = (a); \
-  (int32x2_t)__builtin_neon_vshrn_n_v((int8x16_t)__a, __b, 2); })
-#define vshrn_n_u16(a, __b) __extension__ ({ \
-  uint16x8_t __a = (a); \
-  (uint8x8_t)__builtin_neon_vshrn_n_v((int8x16_t)__a, __b, 16); })
-#define vshrn_n_u32(a, __b) __extension__ ({ \
-  uint32x4_t __a = (a); \
-  (uint16x4_t)__builtin_neon_vshrn_n_v((int8x16_t)__a, __b, 17); })
-#define vshrn_n_u64(a, __b) __extension__ ({ \
-  uint64x2_t __a = (a); \
-  (uint32x2_t)__builtin_neon_vshrn_n_v((int8x16_t)__a, __b, 18); })
-
-#define vshr_n_s8(a, __b) __extension__ ({ \
-  int8x8_t __a = (a); \
-  (int8x8_t)__builtin_neon_vshr_n_v(__a, __b, 0); })
-#define vshr_n_s16(a, __b) __extension__ ({ \
-  int16x4_t __a = (a); \
-  (int16x4_t)__builtin_neon_vshr_n_v((int8x8_t)__a, __b, 1); })
-#define vshr_n_s32(a, __b) __extension__ ({ \
-  int32x2_t __a = (a); \
-  (int32x2_t)__builtin_neon_vshr_n_v((int8x8_t)__a, __b, 2); })
-#define vshr_n_s64(a, __b) __extension__ ({ \
-  int64x1_t __a = (a); \
-  (int64x1_t)__builtin_neon_vshr_n_v((int8x8_t)__a, __b, 3); })
-#define vshr_n_u8(a, __b) __extension__ ({ \
-  uint8x8_t __a = (a); \
-  (uint8x8_t)__builtin_neon_vshr_n_v((int8x8_t)__a, __b, 16); })
-#define vshr_n_u16(a, __b) __extension__ ({ \
-  uint16x4_t __a = (a); \
-  (uint16x4_t)__builtin_neon_vshr_n_v((int8x8_t)__a, __b, 17); })
-#define vshr_n_u32(a, __b) __extension__ ({ \
-  uint32x2_t __a = (a); \
-  (uint32x2_t)__builtin_neon_vshr_n_v((int8x8_t)__a, __b, 18); })
-#define vshr_n_u64(a, __b) __extension__ ({ \
-  uint64x1_t __a = (a); \
-  (uint64x1_t)__builtin_neon_vshr_n_v((int8x8_t)__a, __b, 19); })
-#define vshrq_n_s8(a, __b) __extension__ ({ \
-  int8x16_t __a = (a); \
-  (int8x16_t)__builtin_neon_vshrq_n_v(__a, __b, 32); })
-#define vshrq_n_s16(a, __b) __extension__ ({ \
-  int16x8_t __a = (a); \
-  (int16x8_t)__builtin_neon_vshrq_n_v((int8x16_t)__a, __b, 33); })
-#define vshrq_n_s32(a, __b) __extension__ ({ \
-  int32x4_t __a = (a); \
-  (int32x4_t)__builtin_neon_vshrq_n_v((int8x16_t)__a, __b, 34); })
-#define vshrq_n_s64(a, __b) __extension__ ({ \
-  int64x2_t __a = (a); \
-  (int64x2_t)__builtin_neon_vshrq_n_v((int8x16_t)__a, __b, 35); })
-#define vshrq_n_u8(a, __b) __extension__ ({ \
-  uint8x16_t __a = (a); \
-  (uint8x16_t)__builtin_neon_vshrq_n_v((int8x16_t)__a, __b, 48); })
-#define vshrq_n_u16(a, __b) __extension__ ({ \
-  uint16x8_t __a = (a); \
-  (uint16x8_t)__builtin_neon_vshrq_n_v((int8x16_t)__a, __b, 49); })
-#define vshrq_n_u32(a, __b) __extension__ ({ \
-  uint32x4_t __a = (a); \
-  (uint32x4_t)__builtin_neon_vshrq_n_v((int8x16_t)__a, __b, 50); })
-#define vshrq_n_u64(a, __b) __extension__ ({ \
-  uint64x2_t __a = (a); \
-  (uint64x2_t)__builtin_neon_vshrq_n_v((int8x16_t)__a, __b, 51); })
-
-#define vsli_n_s8(a, b, __c) __extension__ ({ \
-  int8x8_t __a = (a); int8x8_t __b = (b); \
-  (int8x8_t)__builtin_neon_vsli_n_v(__a, __b, __c, 0); })
-#define vsli_n_s16(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int16x4_t __b = (b); \
-  (int16x4_t)__builtin_neon_vsli_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 1); })
-#define vsli_n_s32(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int32x2_t __b = (b); \
-  (int32x2_t)__builtin_neon_vsli_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 2); })
-#define vsli_n_s64(a, b, __c) __extension__ ({ \
-  int64x1_t __a = (a); int64x1_t __b = (b); \
-  (int64x1_t)__builtin_neon_vsli_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 3); })
-#define vsli_n_u8(a, b, __c) __extension__ ({ \
-  uint8x8_t __a = (a); uint8x8_t __b = (b); \
-  (uint8x8_t)__builtin_neon_vsli_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 16); })
-#define vsli_n_u16(a, b, __c) __extension__ ({ \
-  uint16x4_t __a = (a); uint16x4_t __b = (b); \
-  (uint16x4_t)__builtin_neon_vsli_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 17); })
-#define vsli_n_u32(a, b, __c) __extension__ ({ \
-  uint32x2_t __a = (a); uint32x2_t __b = (b); \
-  (uint32x2_t)__builtin_neon_vsli_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 18); })
-#define vsli_n_u64(a, b, __c) __extension__ ({ \
-  uint64x1_t __a = (a); uint64x1_t __b = (b); \
-  (uint64x1_t)__builtin_neon_vsli_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 19); })
-#define vsli_n_p8(a, b, __c) __extension__ ({ \
-  poly8x8_t __a = (a); poly8x8_t __b = (b); \
-  (poly8x8_t)__builtin_neon_vsli_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 4); })
-#define vsli_n_p16(a, b, __c) __extension__ ({ \
-  poly16x4_t __a = (a); poly16x4_t __b = (b); \
-  (poly16x4_t)__builtin_neon_vsli_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 5); })
-#define vsliq_n_s8(a, b, __c) __extension__ ({ \
-  int8x16_t __a = (a); int8x16_t __b = (b); \
-  (int8x16_t)__builtin_neon_vsliq_n_v(__a, __b, __c, 32); })
-#define vsliq_n_s16(a, b, __c) __extension__ ({ \
-  int16x8_t __a = (a); int16x8_t __b = (b); \
-  (int16x8_t)__builtin_neon_vsliq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 33); })
-#define vsliq_n_s32(a, b, __c) __extension__ ({ \
-  int32x4_t __a = (a); int32x4_t __b = (b); \
-  (int32x4_t)__builtin_neon_vsliq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 34); })
-#define vsliq_n_s64(a, b, __c) __extension__ ({ \
-  int64x2_t __a = (a); int64x2_t __b = (b); \
-  (int64x2_t)__builtin_neon_vsliq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 35); })
-#define vsliq_n_u8(a, b, __c) __extension__ ({ \
-  uint8x16_t __a = (a); uint8x16_t __b = (b); \
-  (uint8x16_t)__builtin_neon_vsliq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 48); })
-#define vsliq_n_u16(a, b, __c) __extension__ ({ \
-  uint16x8_t __a = (a); uint16x8_t __b = (b); \
-  (uint16x8_t)__builtin_neon_vsliq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 49); })
-#define vsliq_n_u32(a, b, __c) __extension__ ({ \
-  uint32x4_t __a = (a); uint32x4_t __b = (b); \
-  (uint32x4_t)__builtin_neon_vsliq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 50); })
-#define vsliq_n_u64(a, b, __c) __extension__ ({ \
-  uint64x2_t __a = (a); uint64x2_t __b = (b); \
-  (uint64x2_t)__builtin_neon_vsliq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 51); })
-#define vsliq_n_p8(a, b, __c) __extension__ ({ \
-  poly8x16_t __a = (a); poly8x16_t __b = (b); \
-  (poly8x16_t)__builtin_neon_vsliq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 36); })
-#define vsliq_n_p16(a, b, __c) __extension__ ({ \
-  poly16x8_t __a = (a); poly16x8_t __b = (b); \
-  (poly16x8_t)__builtin_neon_vsliq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 37); })
-
-#define vsra_n_s8(a, b, __c) __extension__ ({ \
-  int8x8_t __a = (a); int8x8_t __b = (b); \
-  (int8x8_t)__builtin_neon_vsra_n_v(__a, __b, __c, 0); })
-#define vsra_n_s16(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int16x4_t __b = (b); \
-  (int16x4_t)__builtin_neon_vsra_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 1); })
-#define vsra_n_s32(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int32x2_t __b = (b); \
-  (int32x2_t)__builtin_neon_vsra_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 2); })
-#define vsra_n_s64(a, b, __c) __extension__ ({ \
-  int64x1_t __a = (a); int64x1_t __b = (b); \
-  (int64x1_t)__builtin_neon_vsra_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 3); })
-#define vsra_n_u8(a, b, __c) __extension__ ({ \
-  uint8x8_t __a = (a); uint8x8_t __b = (b); \
-  (uint8x8_t)__builtin_neon_vsra_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 16); })
-#define vsra_n_u16(a, b, __c) __extension__ ({ \
-  uint16x4_t __a = (a); uint16x4_t __b = (b); \
-  (uint16x4_t)__builtin_neon_vsra_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 17); })
-#define vsra_n_u32(a, b, __c) __extension__ ({ \
-  uint32x2_t __a = (a); uint32x2_t __b = (b); \
-  (uint32x2_t)__builtin_neon_vsra_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 18); })
-#define vsra_n_u64(a, b, __c) __extension__ ({ \
-  uint64x1_t __a = (a); uint64x1_t __b = (b); \
-  (uint64x1_t)__builtin_neon_vsra_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 19); })
-#define vsraq_n_s8(a, b, __c) __extension__ ({ \
-  int8x16_t __a = (a); int8x16_t __b = (b); \
-  (int8x16_t)__builtin_neon_vsraq_n_v(__a, __b, __c, 32); })
-#define vsraq_n_s16(a, b, __c) __extension__ ({ \
-  int16x8_t __a = (a); int16x8_t __b = (b); \
-  (int16x8_t)__builtin_neon_vsraq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 33); })
-#define vsraq_n_s32(a, b, __c) __extension__ ({ \
-  int32x4_t __a = (a); int32x4_t __b = (b); \
-  (int32x4_t)__builtin_neon_vsraq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 34); })
-#define vsraq_n_s64(a, b, __c) __extension__ ({ \
-  int64x2_t __a = (a); int64x2_t __b = (b); \
-  (int64x2_t)__builtin_neon_vsraq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 35); })
-#define vsraq_n_u8(a, b, __c) __extension__ ({ \
-  uint8x16_t __a = (a); uint8x16_t __b = (b); \
-  (uint8x16_t)__builtin_neon_vsraq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 48); })
-#define vsraq_n_u16(a, b, __c) __extension__ ({ \
-  uint16x8_t __a = (a); uint16x8_t __b = (b); \
-  (uint16x8_t)__builtin_neon_vsraq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 49); })
-#define vsraq_n_u32(a, b, __c) __extension__ ({ \
-  uint32x4_t __a = (a); uint32x4_t __b = (b); \
-  (uint32x4_t)__builtin_neon_vsraq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 50); })
-#define vsraq_n_u64(a, b, __c) __extension__ ({ \
-  uint64x2_t __a = (a); uint64x2_t __b = (b); \
-  (uint64x2_t)__builtin_neon_vsraq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 51); })
-
-#define vsri_n_s8(a, b, __c) __extension__ ({ \
-  int8x8_t __a = (a); int8x8_t __b = (b); \
-  (int8x8_t)__builtin_neon_vsri_n_v(__a, __b, __c, 0); })
-#define vsri_n_s16(a, b, __c) __extension__ ({ \
-  int16x4_t __a = (a); int16x4_t __b = (b); \
-  (int16x4_t)__builtin_neon_vsri_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 1); })
-#define vsri_n_s32(a, b, __c) __extension__ ({ \
-  int32x2_t __a = (a); int32x2_t __b = (b); \
-  (int32x2_t)__builtin_neon_vsri_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 2); })
-#define vsri_n_s64(a, b, __c) __extension__ ({ \
-  int64x1_t __a = (a); int64x1_t __b = (b); \
-  (int64x1_t)__builtin_neon_vsri_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 3); })
-#define vsri_n_u8(a, b, __c) __extension__ ({ \
-  uint8x8_t __a = (a); uint8x8_t __b = (b); \
-  (uint8x8_t)__builtin_neon_vsri_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 16); })
-#define vsri_n_u16(a, b, __c) __extension__ ({ \
-  uint16x4_t __a = (a); uint16x4_t __b = (b); \
-  (uint16x4_t)__builtin_neon_vsri_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 17); })
-#define vsri_n_u32(a, b, __c) __extension__ ({ \
-  uint32x2_t __a = (a); uint32x2_t __b = (b); \
-  (uint32x2_t)__builtin_neon_vsri_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 18); })
-#define vsri_n_u64(a, b, __c) __extension__ ({ \
-  uint64x1_t __a = (a); uint64x1_t __b = (b); \
-  (uint64x1_t)__builtin_neon_vsri_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 19); })
-#define vsri_n_p8(a, b, __c) __extension__ ({ \
-  poly8x8_t __a = (a); poly8x8_t __b = (b); \
-  (poly8x8_t)__builtin_neon_vsri_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 4); })
-#define vsri_n_p16(a, b, __c) __extension__ ({ \
-  poly16x4_t __a = (a); poly16x4_t __b = (b); \
-  (poly16x4_t)__builtin_neon_vsri_n_v((int8x8_t)__a, (int8x8_t)__b, __c, 5); })
-#define vsriq_n_s8(a, b, __c) __extension__ ({ \
-  int8x16_t __a = (a); int8x16_t __b = (b); \
-  (int8x16_t)__builtin_neon_vsriq_n_v(__a, __b, __c, 32); })
-#define vsriq_n_s16(a, b, __c) __extension__ ({ \
-  int16x8_t __a = (a); int16x8_t __b = (b); \
-  (int16x8_t)__builtin_neon_vsriq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 33); })
-#define vsriq_n_s32(a, b, __c) __extension__ ({ \
-  int32x4_t __a = (a); int32x4_t __b = (b); \
-  (int32x4_t)__builtin_neon_vsriq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 34); })
-#define vsriq_n_s64(a, b, __c) __extension__ ({ \
-  int64x2_t __a = (a); int64x2_t __b = (b); \
-  (int64x2_t)__builtin_neon_vsriq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 35); })
-#define vsriq_n_u8(a, b, __c) __extension__ ({ \
-  uint8x16_t __a = (a); uint8x16_t __b = (b); \
-  (uint8x16_t)__builtin_neon_vsriq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 48); })
-#define vsriq_n_u16(a, b, __c) __extension__ ({ \
-  uint16x8_t __a = (a); uint16x8_t __b = (b); \
-  (uint16x8_t)__builtin_neon_vsriq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 49); })
-#define vsriq_n_u32(a, b, __c) __extension__ ({ \
-  uint32x4_t __a = (a); uint32x4_t __b = (b); \
-  (uint32x4_t)__builtin_neon_vsriq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 50); })
-#define vsriq_n_u64(a, b, __c) __extension__ ({ \
-  uint64x2_t __a = (a); uint64x2_t __b = (b); \
-  (uint64x2_t)__builtin_neon_vsriq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 51); })
-#define vsriq_n_p8(a, b, __c) __extension__ ({ \
-  poly8x16_t __a = (a); poly8x16_t __b = (b); \
-  (poly8x16_t)__builtin_neon_vsriq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 36); })
-#define vsriq_n_p16(a, b, __c) __extension__ ({ \
-  poly16x8_t __a = (a); poly16x8_t __b = (b); \
-  (poly16x8_t)__builtin_neon_vsriq_n_v((int8x16_t)__a, (int8x16_t)__b, __c, 37); })
-
-#define vst1q_u8(__a, b) __extension__ ({ \
-  uint8x16_t __b = (b); \
-  __builtin_neon_vst1q_v(__a, (int8x16_t)__b, 48); })
-#define vst1q_u16(__a, b) __extension__ ({ \
-  uint16x8_t __b = (b); \
-  __builtin_neon_vst1q_v(__a, (int8x16_t)__b, 49); })
-#define vst1q_u32(__a, b) __extension__ ({ \
-  uint32x4_t __b = (b); \
-  __builtin_neon_vst1q_v(__a, (int8x16_t)__b, 50); })
-#define vst1q_u64(__a, b) __extension__ ({ \
-  uint64x2_t __b = (b); \
-  __builtin_neon_vst1q_v(__a, (int8x16_t)__b, 51); })
-#define vst1q_s8(__a, b) __extension__ ({ \
-  int8x16_t __b = (b); \
-  __builtin_neon_vst1q_v(__a, __b, 32); })
-#define vst1q_s16(__a, b) __extension__ ({ \
-  int16x8_t __b = (b); \
-  __builtin_neon_vst1q_v(__a, (int8x16_t)__b, 33); })
-#define vst1q_s32(__a, b) __extension__ ({ \
-  int32x4_t __b = (b); \
-  __builtin_neon_vst1q_v(__a, (int8x16_t)__b, 34); })
-#define vst1q_s64(__a, b) __extension__ ({ \
-  int64x2_t __b = (b); \
-  __builtin_neon_vst1q_v(__a, (int8x16_t)__b, 35); })
-#define vst1q_f16(__a, b) __extension__ ({ \
-  float16x8_t __b = (b); \
-  __builtin_neon_vst1q_v(__a, (int8x16_t)__b, 40); })
-#define vst1q_f32(__a, b) __extension__ ({ \
-  float32x4_t __b = (b); \
-  __builtin_neon_vst1q_v(__a, (int8x16_t)__b, 41); })
-#define vst1q_p8(__a, b) __extension__ ({ \
-  poly8x16_t __b = (b); \
-  __builtin_neon_vst1q_v(__a, (int8x16_t)__b, 36); })
-#define vst1q_p16(__a, b) __extension__ ({ \
-  poly16x8_t __b = (b); \
-  __builtin_neon_vst1q_v(__a, (int8x16_t)__b, 37); })
-#define vst1_u8(__a, b) __extension__ ({ \
-  uint8x8_t __b = (b); \
-  __builtin_neon_vst1_v(__a, (int8x8_t)__b, 16); })
-#define vst1_u16(__a, b) __extension__ ({ \
-  uint16x4_t __b = (b); \
-  __builtin_neon_vst1_v(__a, (int8x8_t)__b, 17); })
-#define vst1_u32(__a, b) __extension__ ({ \
-  uint32x2_t __b = (b); \
-  __builtin_neon_vst1_v(__a, (int8x8_t)__b, 18); })
-#define vst1_u64(__a, b) __extension__ ({ \
-  uint64x1_t __b = (b); \
-  __builtin_neon_vst1_v(__a, (int8x8_t)__b, 19); })
-#define vst1_s8(__a, b) __extension__ ({ \
-  int8x8_t __b = (b); \
-  __builtin_neon_vst1_v(__a, __b, 0); })
-#define vst1_s16(__a, b) __extension__ ({ \
-  int16x4_t __b = (b); \
-  __builtin_neon_vst1_v(__a, (int8x8_t)__b, 1); })
-#define vst1_s32(__a, b) __extension__ ({ \
-  int32x2_t __b = (b); \
-  __builtin_neon_vst1_v(__a, (int8x8_t)__b, 2); })
-#define vst1_s64(__a, b) __extension__ ({ \
-  int64x1_t __b = (b); \
-  __builtin_neon_vst1_v(__a, (int8x8_t)__b, 3); })
-#define vst1_f16(__a, b) __extension__ ({ \
-  float16x4_t __b = (b); \
-  __builtin_neon_vst1_v(__a, (int8x8_t)__b, 8); })
-#define vst1_f32(__a, b) __extension__ ({ \
-  float32x2_t __b = (b); \
-  __builtin_neon_vst1_v(__a, (int8x8_t)__b, 9); })
-#define vst1_p8(__a, b) __extension__ ({ \
-  poly8x8_t __b = (b); \
-  __builtin_neon_vst1_v(__a, (int8x8_t)__b, 4); })
-#define vst1_p16(__a, b) __extension__ ({ \
-  poly16x4_t __b = (b); \
-  __builtin_neon_vst1_v(__a, (int8x8_t)__b, 5); })
-
-#define vst1q_lane_u8(__a, b, __c) __extension__ ({ \
-  uint8x16_t __b = (b); \
-  __builtin_neon_vst1q_lane_v(__a, (int8x16_t)__b, __c, 48); })
-#define vst1q_lane_u16(__a, b, __c) __extension__ ({ \
-  uint16x8_t __b = (b); \
-  __builtin_neon_vst1q_lane_v(__a, (int8x16_t)__b, __c, 49); })
-#define vst1q_lane_u32(__a, b, __c) __extension__ ({ \
-  uint32x4_t __b = (b); \
-  __builtin_neon_vst1q_lane_v(__a, (int8x16_t)__b, __c, 50); })
-#define vst1q_lane_u64(__a, b, __c) __extension__ ({ \
-  uint64x2_t __b = (b); \
-  __builtin_neon_vst1q_lane_v(__a, (int8x16_t)__b, __c, 51); })
-#define vst1q_lane_s8(__a, b, __c) __extension__ ({ \
-  int8x16_t __b = (b); \
-  __builtin_neon_vst1q_lane_v(__a, __b, __c, 32); })
-#define vst1q_lane_s16(__a, b, __c) __extension__ ({ \
-  int16x8_t __b = (b); \
-  __builtin_neon_vst1q_lane_v(__a, (int8x16_t)__b, __c, 33); })
-#define vst1q_lane_s32(__a, b, __c) __extension__ ({ \
-  int32x4_t __b = (b); \
-  __builtin_neon_vst1q_lane_v(__a, (int8x16_t)__b, __c, 34); })
-#define vst1q_lane_s64(__a, b, __c) __extension__ ({ \
-  int64x2_t __b = (b); \
-  __builtin_neon_vst1q_lane_v(__a, (int8x16_t)__b, __c, 35); })
-#define vst1q_lane_f16(__a, b, __c) __extension__ ({ \
-  float16x8_t __b = (b); \
-  __builtin_neon_vst1q_lane_v(__a, (int8x16_t)__b, __c, 40); })
-#define vst1q_lane_f32(__a, b, __c) __extension__ ({ \
-  float32x4_t __b = (b); \
-  __builtin_neon_vst1q_lane_v(__a, (int8x16_t)__b, __c, 41); })
-#define vst1q_lane_p8(__a, b, __c) __extension__ ({ \
-  poly8x16_t __b = (b); \
-  __builtin_neon_vst1q_lane_v(__a, (int8x16_t)__b, __c, 36); })
-#define vst1q_lane_p16(__a, b, __c) __extension__ ({ \
-  poly16x8_t __b = (b); \
-  __builtin_neon_vst1q_lane_v(__a, (int8x16_t)__b, __c, 37); })
-#define vst1_lane_u8(__a, b, __c) __extension__ ({ \
-  uint8x8_t __b = (b); \
-  __builtin_neon_vst1_lane_v(__a, (int8x8_t)__b, __c, 16); })
-#define vst1_lane_u16(__a, b, __c) __extension__ ({ \
-  uint16x4_t __b = (b); \
-  __builtin_neon_vst1_lane_v(__a, (int8x8_t)__b, __c, 17); })
-#define vst1_lane_u32(__a, b, __c) __extension__ ({ \
-  uint32x2_t __b = (b); \
-  __builtin_neon_vst1_lane_v(__a, (int8x8_t)__b, __c, 18); })
-#define vst1_lane_u64(__a, b, __c) __extension__ ({ \
-  uint64x1_t __b = (b); \
-  __builtin_neon_vst1_lane_v(__a, (int8x8_t)__b, __c, 19); })
-#define vst1_lane_s8(__a, b, __c) __extension__ ({ \
-  int8x8_t __b = (b); \
-  __builtin_neon_vst1_lane_v(__a, __b, __c, 0); })
-#define vst1_lane_s16(__a, b, __c) __extension__ ({ \
-  int16x4_t __b = (b); \
-  __builtin_neon_vst1_lane_v(__a, (int8x8_t)__b, __c, 1); })
-#define vst1_lane_s32(__a, b, __c) __extension__ ({ \
-  int32x2_t __b = (b); \
-  __builtin_neon_vst1_lane_v(__a, (int8x8_t)__b, __c, 2); })
-#define vst1_lane_s64(__a, b, __c) __extension__ ({ \
-  int64x1_t __b = (b); \
-  __builtin_neon_vst1_lane_v(__a, (int8x8_t)__b, __c, 3); })
-#define vst1_lane_f16(__a, b, __c) __extension__ ({ \
-  float16x4_t __b = (b); \
-  __builtin_neon_vst1_lane_v(__a, (int8x8_t)__b, __c, 8); })
-#define vst1_lane_f32(__a, b, __c) __extension__ ({ \
-  float32x2_t __b = (b); \
-  __builtin_neon_vst1_lane_v(__a, (int8x8_t)__b, __c, 9); })
-#define vst1_lane_p8(__a, b, __c) __extension__ ({ \
-  poly8x8_t __b = (b); \
-  __builtin_neon_vst1_lane_v(__a, (int8x8_t)__b, __c, 4); })
-#define vst1_lane_p16(__a, b, __c) __extension__ ({ \
-  poly16x4_t __b = (b); \
-  __builtin_neon_vst1_lane_v(__a, (int8x8_t)__b, __c, 5); })
-
-#define vst2q_u8(__a, b) __extension__ ({ \
-  uint8x16x2_t __b = (b); \
-  __builtin_neon_vst2q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 48); })
-#define vst2q_u16(__a, b) __extension__ ({ \
-  uint16x8x2_t __b = (b); \
-  __builtin_neon_vst2q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 49); })
-#define vst2q_u32(__a, b) __extension__ ({ \
-  uint32x4x2_t __b = (b); \
-  __builtin_neon_vst2q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 50); })
-#define vst2q_s8(__a, b) __extension__ ({ \
-  int8x16x2_t __b = (b); \
-  __builtin_neon_vst2q_v(__a, __b.val[0], __b.val[1], 32); })
-#define vst2q_s16(__a, b) __extension__ ({ \
-  int16x8x2_t __b = (b); \
-  __builtin_neon_vst2q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 33); })
-#define vst2q_s32(__a, b) __extension__ ({ \
-  int32x4x2_t __b = (b); \
-  __builtin_neon_vst2q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 34); })
-#define vst2q_f16(__a, b) __extension__ ({ \
-  float16x8x2_t __b = (b); \
-  __builtin_neon_vst2q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 40); })
-#define vst2q_f32(__a, b) __extension__ ({ \
-  float32x4x2_t __b = (b); \
-  __builtin_neon_vst2q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 41); })
-#define vst2q_p8(__a, b) __extension__ ({ \
-  poly8x16x2_t __b = (b); \
-  __builtin_neon_vst2q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 36); })
-#define vst2q_p16(__a, b) __extension__ ({ \
-  poly16x8x2_t __b = (b); \
-  __builtin_neon_vst2q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], 37); })
-#define vst2_u8(__a, b) __extension__ ({ \
-  uint8x8x2_t __b = (b); \
-  __builtin_neon_vst2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 16); })
-#define vst2_u16(__a, b) __extension__ ({ \
-  uint16x4x2_t __b = (b); \
-  __builtin_neon_vst2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 17); })
-#define vst2_u32(__a, b) __extension__ ({ \
-  uint32x2x2_t __b = (b); \
-  __builtin_neon_vst2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 18); })
-#define vst2_u64(__a, b) __extension__ ({ \
-  uint64x1x2_t __b = (b); \
-  __builtin_neon_vst2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 19); })
-#define vst2_s8(__a, b) __extension__ ({ \
-  int8x8x2_t __b = (b); \
-  __builtin_neon_vst2_v(__a, __b.val[0], __b.val[1], 0); })
-#define vst2_s16(__a, b) __extension__ ({ \
-  int16x4x2_t __b = (b); \
-  __builtin_neon_vst2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 1); })
-#define vst2_s32(__a, b) __extension__ ({ \
-  int32x2x2_t __b = (b); \
-  __builtin_neon_vst2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 2); })
-#define vst2_s64(__a, b) __extension__ ({ \
-  int64x1x2_t __b = (b); \
-  __builtin_neon_vst2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 3); })
-#define vst2_f16(__a, b) __extension__ ({ \
-  float16x4x2_t __b = (b); \
-  __builtin_neon_vst2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 8); })
-#define vst2_f32(__a, b) __extension__ ({ \
-  float32x2x2_t __b = (b); \
-  __builtin_neon_vst2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 9); })
-#define vst2_p8(__a, b) __extension__ ({ \
-  poly8x8x2_t __b = (b); \
-  __builtin_neon_vst2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 4); })
-#define vst2_p16(__a, b) __extension__ ({ \
-  poly16x4x2_t __b = (b); \
-  __builtin_neon_vst2_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], 5); })
-
-#define vst2q_lane_u16(__a, b, __c) __extension__ ({ \
-  uint16x8x2_t __b = (b); \
-  __builtin_neon_vst2q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 49); })
-#define vst2q_lane_u32(__a, b, __c) __extension__ ({ \
-  uint32x4x2_t __b = (b); \
-  __builtin_neon_vst2q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 50); })
-#define vst2q_lane_s16(__a, b, __c) __extension__ ({ \
-  int16x8x2_t __b = (b); \
-  __builtin_neon_vst2q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 33); })
-#define vst2q_lane_s32(__a, b, __c) __extension__ ({ \
-  int32x4x2_t __b = (b); \
-  __builtin_neon_vst2q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 34); })
-#define vst2q_lane_f16(__a, b, __c) __extension__ ({ \
-  float16x8x2_t __b = (b); \
-  __builtin_neon_vst2q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 40); })
-#define vst2q_lane_f32(__a, b, __c) __extension__ ({ \
-  float32x4x2_t __b = (b); \
-  __builtin_neon_vst2q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 41); })
-#define vst2q_lane_p16(__a, b, __c) __extension__ ({ \
-  poly16x8x2_t __b = (b); \
-  __builtin_neon_vst2q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], __c, 37); })
-#define vst2_lane_u8(__a, b, __c) __extension__ ({ \
-  uint8x8x2_t __b = (b); \
-  __builtin_neon_vst2_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 16); })
-#define vst2_lane_u16(__a, b, __c) __extension__ ({ \
-  uint16x4x2_t __b = (b); \
-  __builtin_neon_vst2_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 17); })
-#define vst2_lane_u32(__a, b, __c) __extension__ ({ \
-  uint32x2x2_t __b = (b); \
-  __builtin_neon_vst2_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 18); })
-#define vst2_lane_s8(__a, b, __c) __extension__ ({ \
-  int8x8x2_t __b = (b); \
-  __builtin_neon_vst2_lane_v(__a, __b.val[0], __b.val[1], __c, 0); })
-#define vst2_lane_s16(__a, b, __c) __extension__ ({ \
-  int16x4x2_t __b = (b); \
-  __builtin_neon_vst2_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 1); })
-#define vst2_lane_s32(__a, b, __c) __extension__ ({ \
-  int32x2x2_t __b = (b); \
-  __builtin_neon_vst2_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 2); })
-#define vst2_lane_f16(__a, b, __c) __extension__ ({ \
-  float16x4x2_t __b = (b); \
-  __builtin_neon_vst2_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 8); })
-#define vst2_lane_f32(__a, b, __c) __extension__ ({ \
-  float32x2x2_t __b = (b); \
-  __builtin_neon_vst2_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 9); })
-#define vst2_lane_p8(__a, b, __c) __extension__ ({ \
-  poly8x8x2_t __b = (b); \
-  __builtin_neon_vst2_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 4); })
-#define vst2_lane_p16(__a, b, __c) __extension__ ({ \
-  poly16x4x2_t __b = (b); \
-  __builtin_neon_vst2_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], __c, 5); })
-
-#define vst3q_u8(__a, b) __extension__ ({ \
-  uint8x16x3_t __b = (b); \
-  __builtin_neon_vst3q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 48); })
-#define vst3q_u16(__a, b) __extension__ ({ \
-  uint16x8x3_t __b = (b); \
-  __builtin_neon_vst3q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 49); })
-#define vst3q_u32(__a, b) __extension__ ({ \
-  uint32x4x3_t __b = (b); \
-  __builtin_neon_vst3q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 50); })
-#define vst3q_s8(__a, b) __extension__ ({ \
-  int8x16x3_t __b = (b); \
-  __builtin_neon_vst3q_v(__a, __b.val[0], __b.val[1], __b.val[2], 32); })
-#define vst3q_s16(__a, b) __extension__ ({ \
-  int16x8x3_t __b = (b); \
-  __builtin_neon_vst3q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 33); })
-#define vst3q_s32(__a, b) __extension__ ({ \
-  int32x4x3_t __b = (b); \
-  __builtin_neon_vst3q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 34); })
-#define vst3q_f16(__a, b) __extension__ ({ \
-  float16x8x3_t __b = (b); \
-  __builtin_neon_vst3q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 40); })
-#define vst3q_f32(__a, b) __extension__ ({ \
-  float32x4x3_t __b = (b); \
-  __builtin_neon_vst3q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 41); })
-#define vst3q_p8(__a, b) __extension__ ({ \
-  poly8x16x3_t __b = (b); \
-  __builtin_neon_vst3q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 36); })
-#define vst3q_p16(__a, b) __extension__ ({ \
-  poly16x8x3_t __b = (b); \
-  __builtin_neon_vst3q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], 37); })
-#define vst3_u8(__a, b) __extension__ ({ \
-  uint8x8x3_t __b = (b); \
-  __builtin_neon_vst3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 16); })
-#define vst3_u16(__a, b) __extension__ ({ \
-  uint16x4x3_t __b = (b); \
-  __builtin_neon_vst3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 17); })
-#define vst3_u32(__a, b) __extension__ ({ \
-  uint32x2x3_t __b = (b); \
-  __builtin_neon_vst3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 18); })
-#define vst3_u64(__a, b) __extension__ ({ \
-  uint64x1x3_t __b = (b); \
-  __builtin_neon_vst3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 19); })
-#define vst3_s8(__a, b) __extension__ ({ \
-  int8x8x3_t __b = (b); \
-  __builtin_neon_vst3_v(__a, __b.val[0], __b.val[1], __b.val[2], 0); })
-#define vst3_s16(__a, b) __extension__ ({ \
-  int16x4x3_t __b = (b); \
-  __builtin_neon_vst3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 1); })
-#define vst3_s32(__a, b) __extension__ ({ \
-  int32x2x3_t __b = (b); \
-  __builtin_neon_vst3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 2); })
-#define vst3_s64(__a, b) __extension__ ({ \
-  int64x1x3_t __b = (b); \
-  __builtin_neon_vst3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 3); })
-#define vst3_f16(__a, b) __extension__ ({ \
-  float16x4x3_t __b = (b); \
-  __builtin_neon_vst3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 8); })
-#define vst3_f32(__a, b) __extension__ ({ \
-  float32x2x3_t __b = (b); \
-  __builtin_neon_vst3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 9); })
-#define vst3_p8(__a, b) __extension__ ({ \
-  poly8x8x3_t __b = (b); \
-  __builtin_neon_vst3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 4); })
-#define vst3_p16(__a, b) __extension__ ({ \
-  poly16x4x3_t __b = (b); \
-  __builtin_neon_vst3_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], 5); })
-
-#define vst3q_lane_u16(__a, b, __c) __extension__ ({ \
-  uint16x8x3_t __b = (b); \
-  __builtin_neon_vst3q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 49); })
-#define vst3q_lane_u32(__a, b, __c) __extension__ ({ \
-  uint32x4x3_t __b = (b); \
-  __builtin_neon_vst3q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 50); })
-#define vst3q_lane_s16(__a, b, __c) __extension__ ({ \
-  int16x8x3_t __b = (b); \
-  __builtin_neon_vst3q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 33); })
-#define vst3q_lane_s32(__a, b, __c) __extension__ ({ \
-  int32x4x3_t __b = (b); \
-  __builtin_neon_vst3q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 34); })
-#define vst3q_lane_f16(__a, b, __c) __extension__ ({ \
-  float16x8x3_t __b = (b); \
-  __builtin_neon_vst3q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 40); })
-#define vst3q_lane_f32(__a, b, __c) __extension__ ({ \
-  float32x4x3_t __b = (b); \
-  __builtin_neon_vst3q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 41); })
-#define vst3q_lane_p16(__a, b, __c) __extension__ ({ \
-  poly16x8x3_t __b = (b); \
-  __builtin_neon_vst3q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], __c, 37); })
-#define vst3_lane_u8(__a, b, __c) __extension__ ({ \
-  uint8x8x3_t __b = (b); \
-  __builtin_neon_vst3_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 16); })
-#define vst3_lane_u16(__a, b, __c) __extension__ ({ \
-  uint16x4x3_t __b = (b); \
-  __builtin_neon_vst3_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 17); })
-#define vst3_lane_u32(__a, b, __c) __extension__ ({ \
-  uint32x2x3_t __b = (b); \
-  __builtin_neon_vst3_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 18); })
-#define vst3_lane_s8(__a, b, __c) __extension__ ({ \
-  int8x8x3_t __b = (b); \
-  __builtin_neon_vst3_lane_v(__a, __b.val[0], __b.val[1], __b.val[2], __c, 0); })
-#define vst3_lane_s16(__a, b, __c) __extension__ ({ \
-  int16x4x3_t __b = (b); \
-  __builtin_neon_vst3_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 1); })
-#define vst3_lane_s32(__a, b, __c) __extension__ ({ \
-  int32x2x3_t __b = (b); \
-  __builtin_neon_vst3_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 2); })
-#define vst3_lane_f16(__a, b, __c) __extension__ ({ \
-  float16x4x3_t __b = (b); \
-  __builtin_neon_vst3_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 8); })
-#define vst3_lane_f32(__a, b, __c) __extension__ ({ \
-  float32x2x3_t __b = (b); \
-  __builtin_neon_vst3_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 9); })
-#define vst3_lane_p8(__a, b, __c) __extension__ ({ \
-  poly8x8x3_t __b = (b); \
-  __builtin_neon_vst3_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 4); })
-#define vst3_lane_p16(__a, b, __c) __extension__ ({ \
-  poly16x4x3_t __b = (b); \
-  __builtin_neon_vst3_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], __c, 5); })
-
-#define vst4q_u8(__a, b) __extension__ ({ \
-  uint8x16x4_t __b = (b); \
-  __builtin_neon_vst4q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 48); })
-#define vst4q_u16(__a, b) __extension__ ({ \
-  uint16x8x4_t __b = (b); \
-  __builtin_neon_vst4q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 49); })
-#define vst4q_u32(__a, b) __extension__ ({ \
-  uint32x4x4_t __b = (b); \
-  __builtin_neon_vst4q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 50); })
-#define vst4q_s8(__a, b) __extension__ ({ \
-  int8x16x4_t __b = (b); \
-  __builtin_neon_vst4q_v(__a, __b.val[0], __b.val[1], __b.val[2], __b.val[3], 32); })
-#define vst4q_s16(__a, b) __extension__ ({ \
-  int16x8x4_t __b = (b); \
-  __builtin_neon_vst4q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 33); })
-#define vst4q_s32(__a, b) __extension__ ({ \
-  int32x4x4_t __b = (b); \
-  __builtin_neon_vst4q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 34); })
-#define vst4q_f16(__a, b) __extension__ ({ \
-  float16x8x4_t __b = (b); \
-  __builtin_neon_vst4q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 40); })
-#define vst4q_f32(__a, b) __extension__ ({ \
-  float32x4x4_t __b = (b); \
-  __builtin_neon_vst4q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 41); })
-#define vst4q_p8(__a, b) __extension__ ({ \
-  poly8x16x4_t __b = (b); \
-  __builtin_neon_vst4q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 36); })
-#define vst4q_p16(__a, b) __extension__ ({ \
-  poly16x8x4_t __b = (b); \
-  __builtin_neon_vst4q_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], 37); })
-#define vst4_u8(__a, b) __extension__ ({ \
-  uint8x8x4_t __b = (b); \
-  __builtin_neon_vst4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 16); })
-#define vst4_u16(__a, b) __extension__ ({ \
-  uint16x4x4_t __b = (b); \
-  __builtin_neon_vst4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 17); })
-#define vst4_u32(__a, b) __extension__ ({ \
-  uint32x2x4_t __b = (b); \
-  __builtin_neon_vst4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 18); })
-#define vst4_u64(__a, b) __extension__ ({ \
-  uint64x1x4_t __b = (b); \
-  __builtin_neon_vst4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 19); })
-#define vst4_s8(__a, b) __extension__ ({ \
-  int8x8x4_t __b = (b); \
-  __builtin_neon_vst4_v(__a, __b.val[0], __b.val[1], __b.val[2], __b.val[3], 0); })
-#define vst4_s16(__a, b) __extension__ ({ \
-  int16x4x4_t __b = (b); \
-  __builtin_neon_vst4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 1); })
-#define vst4_s32(__a, b) __extension__ ({ \
-  int32x2x4_t __b = (b); \
-  __builtin_neon_vst4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 2); })
-#define vst4_s64(__a, b) __extension__ ({ \
-  int64x1x4_t __b = (b); \
-  __builtin_neon_vst4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 3); })
-#define vst4_f16(__a, b) __extension__ ({ \
-  float16x4x4_t __b = (b); \
-  __builtin_neon_vst4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 8); })
-#define vst4_f32(__a, b) __extension__ ({ \
-  float32x2x4_t __b = (b); \
-  __builtin_neon_vst4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 9); })
-#define vst4_p8(__a, b) __extension__ ({ \
-  poly8x8x4_t __b = (b); \
-  __builtin_neon_vst4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 4); })
-#define vst4_p16(__a, b) __extension__ ({ \
-  poly16x4x4_t __b = (b); \
-  __builtin_neon_vst4_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], 5); })
-
-#define vst4q_lane_u16(__a, b, __c) __extension__ ({ \
-  uint16x8x4_t __b = (b); \
-  __builtin_neon_vst4q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 49); })
-#define vst4q_lane_u32(__a, b, __c) __extension__ ({ \
-  uint32x4x4_t __b = (b); \
-  __builtin_neon_vst4q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 50); })
-#define vst4q_lane_s16(__a, b, __c) __extension__ ({ \
-  int16x8x4_t __b = (b); \
-  __builtin_neon_vst4q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 33); })
-#define vst4q_lane_s32(__a, b, __c) __extension__ ({ \
-  int32x4x4_t __b = (b); \
-  __builtin_neon_vst4q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 34); })
-#define vst4q_lane_f16(__a, b, __c) __extension__ ({ \
-  float16x8x4_t __b = (b); \
-  __builtin_neon_vst4q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 40); })
-#define vst4q_lane_f32(__a, b, __c) __extension__ ({ \
-  float32x4x4_t __b = (b); \
-  __builtin_neon_vst4q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 41); })
-#define vst4q_lane_p16(__a, b, __c) __extension__ ({ \
-  poly16x8x4_t __b = (b); \
-  __builtin_neon_vst4q_lane_v(__a, (int8x16_t)__b.val[0], (int8x16_t)__b.val[1], (int8x16_t)__b.val[2], (int8x16_t)__b.val[3], __c, 37); })
-#define vst4_lane_u8(__a, b, __c) __extension__ ({ \
-  uint8x8x4_t __b = (b); \
-  __builtin_neon_vst4_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 16); })
-#define vst4_lane_u16(__a, b, __c) __extension__ ({ \
-  uint16x4x4_t __b = (b); \
-  __builtin_neon_vst4_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 17); })
-#define vst4_lane_u32(__a, b, __c) __extension__ ({ \
-  uint32x2x4_t __b = (b); \
-  __builtin_neon_vst4_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 18); })
-#define vst4_lane_s8(__a, b, __c) __extension__ ({ \
-  int8x8x4_t __b = (b); \
-  __builtin_neon_vst4_lane_v(__a, __b.val[0], __b.val[1], __b.val[2], __b.val[3], __c, 0); })
-#define vst4_lane_s16(__a, b, __c) __extension__ ({ \
-  int16x4x4_t __b = (b); \
-  __builtin_neon_vst4_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 1); })
-#define vst4_lane_s32(__a, b, __c) __extension__ ({ \
-  int32x2x4_t __b = (b); \
-  __builtin_neon_vst4_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 2); })
-#define vst4_lane_f16(__a, b, __c) __extension__ ({ \
-  float16x4x4_t __b = (b); \
-  __builtin_neon_vst4_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 8); })
-#define vst4_lane_f32(__a, b, __c) __extension__ ({ \
-  float32x2x4_t __b = (b); \
-  __builtin_neon_vst4_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 9); })
-#define vst4_lane_p8(__a, b, __c) __extension__ ({ \
-  poly8x8x4_t __b = (b); \
-  __builtin_neon_vst4_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 4); })
-#define vst4_lane_p16(__a, b, __c) __extension__ ({ \
-  poly16x4x4_t __b = (b); \
-  __builtin_neon_vst4_lane_v(__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], __c, 5); })
+#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
 
-#if defined(__aarch64__)
-#define vstrq_p128(__a, b) __extension__ ({ \
-  poly128_t __b = (b); \
-  __builtin_neon_vstrq_p128(__a, __b); })
-
-#endif
-
-__ai int8x8_t vsub_s8(int8x8_t __a, int8x8_t __b) {
-  return __a - __b; }
-__ai int16x4_t vsub_s16(int16x4_t __a, int16x4_t __b) {
-  return __a - __b; }
-__ai int32x2_t vsub_s32(int32x2_t __a, int32x2_t __b) {
-  return __a - __b; }
-__ai int64x1_t vsub_s64(int64x1_t __a, int64x1_t __b) {
-  return __a - __b; }
-__ai float32x2_t vsub_f32(float32x2_t __a, float32x2_t __b) {
-  return __a - __b; }
-__ai uint8x8_t vsub_u8(uint8x8_t __a, uint8x8_t __b) {
-  return __a - __b; }
-__ai uint16x4_t vsub_u16(uint16x4_t __a, uint16x4_t __b) {
-  return __a - __b; }
-__ai uint32x2_t vsub_u32(uint32x2_t __a, uint32x2_t __b) {
-  return __a - __b; }
-__ai uint64x1_t vsub_u64(uint64x1_t __a, uint64x1_t __b) {
-  return __a - __b; }
-__ai int8x16_t vsubq_s8(int8x16_t __a, int8x16_t __b) {
-  return __a - __b; }
-__ai int16x8_t vsubq_s16(int16x8_t __a, int16x8_t __b) {
-  return __a - __b; }
-__ai int32x4_t vsubq_s32(int32x4_t __a, int32x4_t __b) {
-  return __a - __b; }
-__ai int64x2_t vsubq_s64(int64x2_t __a, int64x2_t __b) {
-  return __a - __b; }
-__ai float32x4_t vsubq_f32(float32x4_t __a, float32x4_t __b) {
-  return __a - __b; }
-__ai uint8x16_t vsubq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return __a - __b; }
-__ai uint16x8_t vsubq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return __a - __b; }
-__ai uint32x4_t vsubq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return __a - __b; }
-__ai uint64x2_t vsubq_u64(uint64x2_t __a, uint64x2_t __b) {
-  return __a - __b; }
-
-__ai int8x8_t vsubhn_s16(int16x8_t __a, int16x8_t __b) {
-  return (int8x8_t)__builtin_neon_vsubhn_v((int8x16_t)__a, (int8x16_t)__b, 0); }
-__ai int16x4_t vsubhn_s32(int32x4_t __a, int32x4_t __b) {
-  return (int16x4_t)__builtin_neon_vsubhn_v((int8x16_t)__a, (int8x16_t)__b, 1); }
-__ai int32x2_t vsubhn_s64(int64x2_t __a, int64x2_t __b) {
-  return (int32x2_t)__builtin_neon_vsubhn_v((int8x16_t)__a, (int8x16_t)__b, 2); }
-__ai uint8x8_t vsubhn_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vsubhn_v((int8x16_t)__a, (int8x16_t)__b, 16); }
-__ai uint16x4_t vsubhn_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vsubhn_v((int8x16_t)__a, (int8x16_t)__b, 17); }
-__ai uint32x2_t vsubhn_u64(uint64x2_t __a, uint64x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vsubhn_v((int8x16_t)__a, (int8x16_t)__b, 18); }
+#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
 
-#if defined(__aarch64__)
-__ai int8x16_t vsubhn_high_s16(int8x8_t __a, int16x8_t __b, int16x8_t __c) {
-  return vcombine_s8(__a, vsubhn_s16(__b, __c)); }
-__ai int16x8_t vsubhn_high_s32(int16x4_t __a, int32x4_t __b, int32x4_t __c) {
-  return vcombine_s16(__a, vsubhn_s32(__b, __c)); }
-__ai int32x4_t vsubhn_high_s64(int32x2_t __a, int64x2_t __b, int64x2_t __c) {
-  return vcombine_s32(__a, vsubhn_s64(__b, __c)); }
-__ai uint8x16_t vsubhn_high_u16(uint8x8_t __a, uint16x8_t __b, uint16x8_t __c) {
-  return vcombine_u8(__a, vsubhn_u16(__b, __c)); }
-__ai uint16x8_t vsubhn_high_u32(uint16x4_t __a, uint32x4_t __b, uint32x4_t __c) {
-  return vcombine_u16(__a, vsubhn_u32(__b, __c)); }
-__ai uint32x4_t vsubhn_high_u64(uint32x2_t __a, uint64x2_t __b, uint64x2_t __c) {
-  return vcombine_u32(__a, vsubhn_u64(__b, __c)); }
-
-#endif
-
-__ai int16x8_t vsubl_s8(int8x8_t __a, int8x8_t __b) {
-  return vmovl_s8(__a) - vmovl_s8(__b); }
-__ai int32x4_t vsubl_s16(int16x4_t __a, int16x4_t __b) {
-  return vmovl_s16(__a) - vmovl_s16(__b); }
-__ai int64x2_t vsubl_s32(int32x2_t __a, int32x2_t __b) {
-  return vmovl_s32(__a) - vmovl_s32(__b); }
-__ai uint16x8_t vsubl_u8(uint8x8_t __a, uint8x8_t __b) {
-  return vmovl_u8(__a) - vmovl_u8(__b); }
-__ai uint32x4_t vsubl_u16(uint16x4_t __a, uint16x4_t __b) {
-  return vmovl_u16(__a) - vmovl_u16(__b); }
-__ai uint64x2_t vsubl_u32(uint32x2_t __a, uint32x2_t __b) {
-  return vmovl_u32(__a) - vmovl_u32(__b); }
+#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
 
-#if defined(__aarch64__)
-__ai int16x8_t vsubl_high_s8(int8x16_t __a, int8x16_t __b) {
-  return vmovl_high_s8(__a) - vmovl_high_s8(__b); }
-__ai int32x4_t vsubl_high_s16(int16x8_t __a, int16x8_t __b) {
-  return vmovl_high_s16(__a) - vmovl_high_s16(__b); }
-__ai int64x2_t vsubl_high_s32(int32x4_t __a, int32x4_t __b) {
-  return vmovl_high_s32(__a) - vmovl_high_s32(__b); }
-__ai uint16x8_t vsubl_high_u8(uint8x16_t __a, uint8x16_t __b) {
-  return vmovl_high_u8(__a) - vmovl_high_u8(__b); }
-__ai uint32x4_t vsubl_high_u16(uint16x8_t __a, uint16x8_t __b) {
-  return vmovl_high_u16(__a) - vmovl_high_u16(__b); }
-__ai uint64x2_t vsubl_high_u32(uint32x4_t __a, uint32x4_t __b) {
-  return vmovl_high_u32(__a) - vmovl_high_u32(__b); }
-
-#endif
-
-__ai int16x8_t vsubw_s8(int16x8_t __a, int8x8_t __b) {
-  return __a - vmovl_s8(__b); }
-__ai int32x4_t vsubw_s16(int32x4_t __a, int16x4_t __b) {
-  return __a - vmovl_s16(__b); }
-__ai int64x2_t vsubw_s32(int64x2_t __a, int32x2_t __b) {
-  return __a - vmovl_s32(__b); }
-__ai uint16x8_t vsubw_u8(uint16x8_t __a, uint8x8_t __b) {
-  return __a - vmovl_u8(__b); }
-__ai uint32x4_t vsubw_u16(uint32x4_t __a, uint16x4_t __b) {
-  return __a - vmovl_u16(__b); }
-__ai uint64x2_t vsubw_u32(uint64x2_t __a, uint32x2_t __b) {
-  return __a - vmovl_u32(__b); }
+#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
 
-#if defined(__aarch64__)
-__ai int16x8_t vsubw_high_s8(int16x8_t __a, int8x16_t __b) {
-  return __a - vmovl_high_s8(__b); }
-__ai int32x4_t vsubw_high_s16(int32x4_t __a, int16x8_t __b) {
-  return __a - vmovl_high_s16(__b); }
-__ai int64x2_t vsubw_high_s32(int64x2_t __a, int32x4_t __b) {
-  return __a - vmovl_high_s32(__b); }
-__ai uint16x8_t vsubw_high_u8(uint16x8_t __a, uint8x16_t __b) {
-  return __a - vmovl_high_u8(__b); }
-__ai uint32x4_t vsubw_high_u16(uint32x4_t __a, uint16x8_t __b) {
-  return __a - vmovl_high_u16(__b); }
-__ai uint64x2_t vsubw_high_u32(uint64x2_t __a, uint32x4_t __b) {
-  return __a - vmovl_high_u32(__b); }
-
-#endif
-
-__ai uint8x8_t vtbl1_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vtbl1_v((int8x8_t)__a, (int8x8_t)__b, 16); }
-__ai int8x8_t vtbl1_s8(int8x8_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vtbl1_v(__a, __b, 0); }
-__ai poly8x8_t vtbl1_p8(poly8x8_t __a, uint8x8_t __b) {
-  return (poly8x8_t)__builtin_neon_vtbl1_v((int8x8_t)__a, (int8x8_t)__b, 4); }
-
-__ai uint8x8_t vtbl2_u8(uint8x8x2_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vtbl2_v((int8x8_t)__a.val[0], (int8x8_t)__a.val[1], (int8x8_t)__b, 16); }
-__ai int8x8_t vtbl2_s8(int8x8x2_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vtbl2_v(__a.val[0], __a.val[1], __b, 0); }
-__ai poly8x8_t vtbl2_p8(poly8x8x2_t __a, uint8x8_t __b) {
-  return (poly8x8_t)__builtin_neon_vtbl2_v((int8x8_t)__a.val[0], (int8x8_t)__a.val[1], (int8x8_t)__b, 4); }
-
-__ai uint8x8_t vtbl3_u8(uint8x8x3_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vtbl3_v((int8x8_t)__a.val[0], (int8x8_t)__a.val[1], (int8x8_t)__a.val[2], (int8x8_t)__b, 16); }
-__ai int8x8_t vtbl3_s8(int8x8x3_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vtbl3_v(__a.val[0], __a.val[1], __a.val[2], __b, 0); }
-__ai poly8x8_t vtbl3_p8(poly8x8x3_t __a, uint8x8_t __b) {
-  return (poly8x8_t)__builtin_neon_vtbl3_v((int8x8_t)__a.val[0], (int8x8_t)__a.val[1], (int8x8_t)__a.val[2], (int8x8_t)__b, 4); }
-
-__ai uint8x8_t vtbl4_u8(uint8x8x4_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vtbl4_v((int8x8_t)__a.val[0], (int8x8_t)__a.val[1], (int8x8_t)__a.val[2], (int8x8_t)__a.val[3], (int8x8_t)__b, 16); }
-__ai int8x8_t vtbl4_s8(int8x8x4_t __a, int8x8_t __b) {
-  return (int8x8_t)__builtin_neon_vtbl4_v(__a.val[0], __a.val[1], __a.val[2], __a.val[3], __b, 0); }
-__ai poly8x8_t vtbl4_p8(poly8x8x4_t __a, uint8x8_t __b) {
-  return (poly8x8_t)__builtin_neon_vtbl4_v((int8x8_t)__a.val[0], (int8x8_t)__a.val[1], (int8x8_t)__a.val[2], (int8x8_t)__a.val[3], (int8x8_t)__b, 4); }
-
-__ai uint8x8_t vtbx1_u8(uint8x8_t __a, uint8x8_t __b, uint8x8_t __c) {
-  return (uint8x8_t)__builtin_neon_vtbx1_v((int8x8_t)__a, (int8x8_t)__b, (int8x8_t)__c, 16); }
-__ai int8x8_t vtbx1_s8(int8x8_t __a, int8x8_t __b, int8x8_t __c) {
-  return (int8x8_t)__builtin_neon_vtbx1_v(__a, __b, __c, 0); }
-__ai poly8x8_t vtbx1_p8(poly8x8_t __a, poly8x8_t __b, uint8x8_t __c) {
-  return (poly8x8_t)__builtin_neon_vtbx1_v((int8x8_t)__a, (int8x8_t)__b, (int8x8_t)__c, 4); }
-
-__ai uint8x8_t vtbx2_u8(uint8x8_t __a, uint8x8x2_t __b, uint8x8_t __c) {
-  return (uint8x8_t)__builtin_neon_vtbx2_v((int8x8_t)__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__c, 16); }
-__ai int8x8_t vtbx2_s8(int8x8_t __a, int8x8x2_t __b, int8x8_t __c) {
-  return (int8x8_t)__builtin_neon_vtbx2_v(__a, __b.val[0], __b.val[1], __c, 0); }
-__ai poly8x8_t vtbx2_p8(poly8x8_t __a, poly8x8x2_t __b, uint8x8_t __c) {
-  return (poly8x8_t)__builtin_neon_vtbx2_v((int8x8_t)__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__c, 4); }
-
-__ai uint8x8_t vtbx3_u8(uint8x8_t __a, uint8x8x3_t __b, uint8x8_t __c) {
-  return (uint8x8_t)__builtin_neon_vtbx3_v((int8x8_t)__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__c, 16); }
-__ai int8x8_t vtbx3_s8(int8x8_t __a, int8x8x3_t __b, int8x8_t __c) {
-  return (int8x8_t)__builtin_neon_vtbx3_v(__a, __b.val[0], __b.val[1], __b.val[2], __c, 0); }
-__ai poly8x8_t vtbx3_p8(poly8x8_t __a, poly8x8x3_t __b, uint8x8_t __c) {
-  return (poly8x8_t)__builtin_neon_vtbx3_v((int8x8_t)__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__c, 4); }
-
-__ai uint8x8_t vtbx4_u8(uint8x8_t __a, uint8x8x4_t __b, uint8x8_t __c) {
-  return (uint8x8_t)__builtin_neon_vtbx4_v((int8x8_t)__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], (int8x8_t)__c, 16); }
-__ai int8x8_t vtbx4_s8(int8x8_t __a, int8x8x4_t __b, int8x8_t __c) {
-  return (int8x8_t)__builtin_neon_vtbx4_v(__a, __b.val[0], __b.val[1], __b.val[2], __b.val[3], __c, 0); }
-__ai poly8x8_t vtbx4_p8(poly8x8_t __a, poly8x8x4_t __b, uint8x8_t __c) {
-  return (poly8x8_t)__builtin_neon_vtbx4_v((int8x8_t)__a, (int8x8_t)__b.val[0], (int8x8_t)__b.val[1], (int8x8_t)__b.val[2], (int8x8_t)__b.val[3], (int8x8_t)__c, 4); }
-
-__ai int8x8x2_t vtrn_s8(int8x8_t __a, int8x8_t __b) {
-  int8x8x2_t r; __builtin_neon_vtrn_v(&r, __a, __b, 0); return r; }
-__ai int16x4x2_t vtrn_s16(int16x4_t __a, int16x4_t __b) {
-  int16x4x2_t r; __builtin_neon_vtrn_v(&r, (int8x8_t)__a, (int8x8_t)__b, 1); return r; }
-__ai int32x2x2_t vtrn_s32(int32x2_t __a, int32x2_t __b) {
-  int32x2x2_t r; __builtin_neon_vtrn_v(&r, (int8x8_t)__a, (int8x8_t)__b, 2); return r; }
-__ai uint8x8x2_t vtrn_u8(uint8x8_t __a, uint8x8_t __b) {
-  uint8x8x2_t r; __builtin_neon_vtrn_v(&r, (int8x8_t)__a, (int8x8_t)__b, 16); return r; }
-__ai uint16x4x2_t vtrn_u16(uint16x4_t __a, uint16x4_t __b) {
-  uint16x4x2_t r; __builtin_neon_vtrn_v(&r, (int8x8_t)__a, (int8x8_t)__b, 17); return r; }
-__ai uint32x2x2_t vtrn_u32(uint32x2_t __a, uint32x2_t __b) {
-  uint32x2x2_t r; __builtin_neon_vtrn_v(&r, (int8x8_t)__a, (int8x8_t)__b, 18); return r; }
-__ai float32x2x2_t vtrn_f32(float32x2_t __a, float32x2_t __b) {
-  float32x2x2_t r; __builtin_neon_vtrn_v(&r, (int8x8_t)__a, (int8x8_t)__b, 9); return r; }
-__ai poly8x8x2_t vtrn_p8(poly8x8_t __a, poly8x8_t __b) {
-  poly8x8x2_t r; __builtin_neon_vtrn_v(&r, (int8x8_t)__a, (int8x8_t)__b, 4); return r; }
-__ai poly16x4x2_t vtrn_p16(poly16x4_t __a, poly16x4_t __b) {
-  poly16x4x2_t r; __builtin_neon_vtrn_v(&r, (int8x8_t)__a, (int8x8_t)__b, 5); return r; }
-__ai int8x16x2_t vtrnq_s8(int8x16_t __a, int8x16_t __b) {
-  int8x16x2_t r; __builtin_neon_vtrnq_v(&r, __a, __b, 32); return r; }
-__ai int16x8x2_t vtrnq_s16(int16x8_t __a, int16x8_t __b) {
-  int16x8x2_t r; __builtin_neon_vtrnq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 33); return r; }
-__ai int32x4x2_t vtrnq_s32(int32x4_t __a, int32x4_t __b) {
-  int32x4x2_t r; __builtin_neon_vtrnq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 34); return r; }
-__ai uint8x16x2_t vtrnq_u8(uint8x16_t __a, uint8x16_t __b) {
-  uint8x16x2_t r; __builtin_neon_vtrnq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 48); return r; }
-__ai uint16x8x2_t vtrnq_u16(uint16x8_t __a, uint16x8_t __b) {
-  uint16x8x2_t r; __builtin_neon_vtrnq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 49); return r; }
-__ai uint32x4x2_t vtrnq_u32(uint32x4_t __a, uint32x4_t __b) {
-  uint32x4x2_t r; __builtin_neon_vtrnq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 50); return r; }
-__ai float32x4x2_t vtrnq_f32(float32x4_t __a, float32x4_t __b) {
-  float32x4x2_t r; __builtin_neon_vtrnq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 41); return r; }
-__ai poly8x16x2_t vtrnq_p8(poly8x16_t __a, poly8x16_t __b) {
-  poly8x16x2_t r; __builtin_neon_vtrnq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 36); return r; }
-__ai poly16x8x2_t vtrnq_p16(poly16x8_t __a, poly16x8_t __b) {
-  poly16x8x2_t r; __builtin_neon_vtrnq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 37); return r; }
+#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
 
-#if defined(__aarch64__)
-__ai int8x8_t vtrn1_s8(int8x8_t __a, int8x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 8, 2, 10, 4, 12, 6, 14); }
-__ai int16x4_t vtrn1_s16(int16x4_t __a, int16x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 4, 2, 6); }
-__ai int32x2_t vtrn1_s32(int32x2_t __a, int32x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-__ai uint8x8_t vtrn1_u8(uint8x8_t __a, uint8x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 8, 2, 10, 4, 12, 6, 14); }
-__ai uint16x4_t vtrn1_u16(uint16x4_t __a, uint16x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 4, 2, 6); }
-__ai uint32x2_t vtrn1_u32(uint32x2_t __a, uint32x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-__ai float32x2_t vtrn1_f32(float32x2_t __a, float32x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-__ai poly8x8_t vtrn1_p8(poly8x8_t __a, poly8x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 8, 2, 10, 4, 12, 6, 14); }
-__ai poly16x4_t vtrn1_p16(poly16x4_t __a, poly16x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 4, 2, 6); }
-__ai int8x16_t vtrn1q_s8(int8x16_t __a, int8x16_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30); }
-__ai int16x8_t vtrn1q_s16(int16x8_t __a, int16x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 8, 2, 10, 4, 12, 6, 14); }
-__ai int32x4_t vtrn1q_s32(int32x4_t __a, int32x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 4, 2, 6); }
-__ai int64x2_t vtrn1q_s64(int64x2_t __a, int64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-__ai uint8x16_t vtrn1q_u8(uint8x16_t __a, uint8x16_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30); }
-__ai uint16x8_t vtrn1q_u16(uint16x8_t __a, uint16x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 8, 2, 10, 4, 12, 6, 14); }
-__ai uint32x4_t vtrn1q_u32(uint32x4_t __a, uint32x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 4, 2, 6); }
-__ai uint64x2_t vtrn1q_u64(uint64x2_t __a, uint64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-__ai float32x4_t vtrn1q_f32(float32x4_t __a, float32x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 4, 2, 6); }
-__ai float64x2_t vtrn1q_f64(float64x2_t __a, float64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-__ai poly8x16_t vtrn1q_p8(poly8x16_t __a, poly8x16_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30); }
-__ai poly16x8_t vtrn1q_p16(poly16x8_t __a, poly16x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 8, 2, 10, 4, 12, 6, 14); }
-__ai poly64x2_t vtrn1q_p64(poly64x2_t __a, poly64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-
-__ai int8x8_t vtrn2_s8(int8x8_t __a, int8x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 9, 3, 11, 5, 13, 7, 15); }
-__ai int16x4_t vtrn2_s16(int16x4_t __a, int16x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 5, 3, 7); }
-__ai int32x2_t vtrn2_s32(int32x2_t __a, int32x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-__ai uint8x8_t vtrn2_u8(uint8x8_t __a, uint8x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 9, 3, 11, 5, 13, 7, 15); }
-__ai uint16x4_t vtrn2_u16(uint16x4_t __a, uint16x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 5, 3, 7); }
-__ai uint32x2_t vtrn2_u32(uint32x2_t __a, uint32x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-__ai float32x2_t vtrn2_f32(float32x2_t __a, float32x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-__ai poly8x8_t vtrn2_p8(poly8x8_t __a, poly8x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 9, 3, 11, 5, 13, 7, 15); }
-__ai poly16x4_t vtrn2_p16(poly16x4_t __a, poly16x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 5, 3, 7); }
-__ai int8x16_t vtrn2q_s8(int8x16_t __a, int8x16_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31); }
-__ai int16x8_t vtrn2q_s16(int16x8_t __a, int16x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 9, 3, 11, 5, 13, 7, 15); }
-__ai int32x4_t vtrn2q_s32(int32x4_t __a, int32x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 5, 3, 7); }
-__ai int64x2_t vtrn2q_s64(int64x2_t __a, int64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-__ai uint8x16_t vtrn2q_u8(uint8x16_t __a, uint8x16_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31); }
-__ai uint16x8_t vtrn2q_u16(uint16x8_t __a, uint16x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 9, 3, 11, 5, 13, 7, 15); }
-__ai uint32x4_t vtrn2q_u32(uint32x4_t __a, uint32x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 5, 3, 7); }
-__ai uint64x2_t vtrn2q_u64(uint64x2_t __a, uint64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-__ai float32x4_t vtrn2q_f32(float32x4_t __a, float32x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 5, 3, 7); }
-__ai float64x2_t vtrn2q_f64(float64x2_t __a, float64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-__ai poly8x16_t vtrn2q_p8(poly8x16_t __a, poly8x16_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31); }
-__ai poly16x8_t vtrn2q_p16(poly16x8_t __a, poly16x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 9, 3, 11, 5, 13, 7, 15); }
-__ai poly64x2_t vtrn2q_p64(poly64x2_t __a, poly64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-
-#endif
-
-__ai uint8x8_t vtst_s8(int8x8_t __a, int8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vtst_v(__a, __b, 16); }
-__ai uint16x4_t vtst_s16(int16x4_t __a, int16x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vtst_v((int8x8_t)__a, (int8x8_t)__b, 17); }
-__ai uint32x2_t vtst_s32(int32x2_t __a, int32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vtst_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai uint8x8_t vtst_u8(uint8x8_t __a, uint8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vtst_v((int8x8_t)__a, (int8x8_t)__b, 16); }
-__ai uint16x4_t vtst_u16(uint16x4_t __a, uint16x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vtst_v((int8x8_t)__a, (int8x8_t)__b, 17); }
-__ai uint32x2_t vtst_u32(uint32x2_t __a, uint32x2_t __b) {
-  return (uint32x2_t)__builtin_neon_vtst_v((int8x8_t)__a, (int8x8_t)__b, 18); }
-__ai uint8x8_t vtst_p8(poly8x8_t __a, poly8x8_t __b) {
-  return (uint8x8_t)__builtin_neon_vtst_v((int8x8_t)__a, (int8x8_t)__b, 16); }
-__ai uint16x4_t vtst_p16(poly16x4_t __a, poly16x4_t __b) {
-  return (uint16x4_t)__builtin_neon_vtst_v((int8x8_t)__a, (int8x8_t)__b, 17); }
-__ai uint8x16_t vtstq_s8(int8x16_t __a, int8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vtstq_v(__a, __b, 48); }
-__ai uint16x8_t vtstq_s16(int16x8_t __a, int16x8_t __b) {
-  return (uint16x8_t)__builtin_neon_vtstq_v((int8x16_t)__a, (int8x16_t)__b, 49); }
-__ai uint32x4_t vtstq_s32(int32x4_t __a, int32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vtstq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-__ai uint8x16_t vtstq_u8(uint8x16_t __a, uint8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vtstq_v((int8x16_t)__a, (int8x16_t)__b, 48); }
-__ai uint16x8_t vtstq_u16(uint16x8_t __a, uint16x8_t __b) {
-  return (uint16x8_t)__builtin_neon_vtstq_v((int8x16_t)__a, (int8x16_t)__b, 49); }
-__ai uint32x4_t vtstq_u32(uint32x4_t __a, uint32x4_t __b) {
-  return (uint32x4_t)__builtin_neon_vtstq_v((int8x16_t)__a, (int8x16_t)__b, 50); }
-__ai uint8x16_t vtstq_p8(poly8x16_t __a, poly8x16_t __b) {
-  return (uint8x16_t)__builtin_neon_vtstq_v((int8x16_t)__a, (int8x16_t)__b, 48); }
-__ai uint16x8_t vtstq_p16(poly16x8_t __a, poly16x8_t __b) {
-  return (uint16x8_t)__builtin_neon_vtstq_v((int8x16_t)__a, (int8x16_t)__b, 49); }
-
-__ai int8x8x2_t vuzp_s8(int8x8_t __a, int8x8_t __b) {
-  int8x8x2_t r; __builtin_neon_vuzp_v(&r, __a, __b, 0); return r; }
-__ai int16x4x2_t vuzp_s16(int16x4_t __a, int16x4_t __b) {
-  int16x4x2_t r; __builtin_neon_vuzp_v(&r, (int8x8_t)__a, (int8x8_t)__b, 1); return r; }
-__ai int32x2x2_t vuzp_s32(int32x2_t __a, int32x2_t __b) {
-  int32x2x2_t r; __builtin_neon_vuzp_v(&r, (int8x8_t)__a, (int8x8_t)__b, 2); return r; }
-__ai uint8x8x2_t vuzp_u8(uint8x8_t __a, uint8x8_t __b) {
-  uint8x8x2_t r; __builtin_neon_vuzp_v(&r, (int8x8_t)__a, (int8x8_t)__b, 16); return r; }
-__ai uint16x4x2_t vuzp_u16(uint16x4_t __a, uint16x4_t __b) {
-  uint16x4x2_t r; __builtin_neon_vuzp_v(&r, (int8x8_t)__a, (int8x8_t)__b, 17); return r; }
-__ai uint32x2x2_t vuzp_u32(uint32x2_t __a, uint32x2_t __b) {
-  uint32x2x2_t r; __builtin_neon_vuzp_v(&r, (int8x8_t)__a, (int8x8_t)__b, 18); return r; }
-__ai float32x2x2_t vuzp_f32(float32x2_t __a, float32x2_t __b) {
-  float32x2x2_t r; __builtin_neon_vuzp_v(&r, (int8x8_t)__a, (int8x8_t)__b, 9); return r; }
-__ai poly8x8x2_t vuzp_p8(poly8x8_t __a, poly8x8_t __b) {
-  poly8x8x2_t r; __builtin_neon_vuzp_v(&r, (int8x8_t)__a, (int8x8_t)__b, 4); return r; }
-__ai poly16x4x2_t vuzp_p16(poly16x4_t __a, poly16x4_t __b) {
-  poly16x4x2_t r; __builtin_neon_vuzp_v(&r, (int8x8_t)__a, (int8x8_t)__b, 5); return r; }
-__ai int8x16x2_t vuzpq_s8(int8x16_t __a, int8x16_t __b) {
-  int8x16x2_t r; __builtin_neon_vuzpq_v(&r, __a, __b, 32); return r; }
-__ai int16x8x2_t vuzpq_s16(int16x8_t __a, int16x8_t __b) {
-  int16x8x2_t r; __builtin_neon_vuzpq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 33); return r; }
-__ai int32x4x2_t vuzpq_s32(int32x4_t __a, int32x4_t __b) {
-  int32x4x2_t r; __builtin_neon_vuzpq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 34); return r; }
-__ai uint8x16x2_t vuzpq_u8(uint8x16_t __a, uint8x16_t __b) {
-  uint8x16x2_t r; __builtin_neon_vuzpq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 48); return r; }
-__ai uint16x8x2_t vuzpq_u16(uint16x8_t __a, uint16x8_t __b) {
-  uint16x8x2_t r; __builtin_neon_vuzpq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 49); return r; }
-__ai uint32x4x2_t vuzpq_u32(uint32x4_t __a, uint32x4_t __b) {
-  uint32x4x2_t r; __builtin_neon_vuzpq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 50); return r; }
-__ai float32x4x2_t vuzpq_f32(float32x4_t __a, float32x4_t __b) {
-  float32x4x2_t r; __builtin_neon_vuzpq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 41); return r; }
-__ai poly8x16x2_t vuzpq_p8(poly8x16_t __a, poly8x16_t __b) {
-  poly8x16x2_t r; __builtin_neon_vuzpq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 36); return r; }
-__ai poly16x8x2_t vuzpq_p16(poly16x8_t __a, poly16x8_t __b) {
-  poly16x8x2_t r; __builtin_neon_vuzpq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 37); return r; }
+#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 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 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 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 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 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 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 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 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
+
+#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
+
+#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 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 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 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 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 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 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 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 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 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 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 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 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 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 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 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 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 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 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 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 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 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 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 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 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
+
+#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
+
+#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__)
-__ai int8x8_t vuzp1_s8(int8x8_t __a, int8x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2, 4, 6, 8, 10, 12, 14); }
-__ai int16x4_t vuzp1_s16(int16x4_t __a, int16x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2, 4, 6); }
-__ai int32x2_t vuzp1_s32(int32x2_t __a, int32x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-__ai uint8x8_t vuzp1_u8(uint8x8_t __a, uint8x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2, 4, 6, 8, 10, 12, 14); }
-__ai uint16x4_t vuzp1_u16(uint16x4_t __a, uint16x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2, 4, 6); }
-__ai uint32x2_t vuzp1_u32(uint32x2_t __a, uint32x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-__ai float32x2_t vuzp1_f32(float32x2_t __a, float32x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-__ai poly8x8_t vuzp1_p8(poly8x8_t __a, poly8x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2, 4, 6, 8, 10, 12, 14); }
-__ai poly16x4_t vuzp1_p16(poly16x4_t __a, poly16x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2, 4, 6); }
-__ai int8x16_t vuzp1q_s8(int8x16_t __a, int8x16_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30); }
-__ai int16x8_t vuzp1q_s16(int16x8_t __a, int16x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2, 4, 6, 8, 10, 12, 14); }
-__ai int32x4_t vuzp1q_s32(int32x4_t __a, int32x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2, 4, 6); }
-__ai int64x2_t vuzp1q_s64(int64x2_t __a, int64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-__ai uint8x16_t vuzp1q_u8(uint8x16_t __a, uint8x16_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30); }
-__ai uint16x8_t vuzp1q_u16(uint16x8_t __a, uint16x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2, 4, 6, 8, 10, 12, 14); }
-__ai uint32x4_t vuzp1q_u32(uint32x4_t __a, uint32x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2, 4, 6); }
-__ai uint64x2_t vuzp1q_u64(uint64x2_t __a, uint64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-__ai float32x4_t vuzp1q_f32(float32x4_t __a, float32x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2, 4, 6); }
-__ai float64x2_t vuzp1q_f64(float64x2_t __a, float64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-__ai poly8x16_t vuzp1q_p8(poly8x16_t __a, poly8x16_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30); }
-__ai poly16x8_t vuzp1q_p16(poly16x8_t __a, poly16x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2, 4, 6, 8, 10, 12, 14); }
-__ai poly64x2_t vuzp1q_p64(poly64x2_t __a, poly64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-
-__ai int8x8_t vuzp2_s8(int8x8_t __a, int8x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3, 5, 7, 9, 11, 13, 15); }
-__ai int16x4_t vuzp2_s16(int16x4_t __a, int16x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3, 5, 7); }
-__ai int32x2_t vuzp2_s32(int32x2_t __a, int32x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-__ai uint8x8_t vuzp2_u8(uint8x8_t __a, uint8x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3, 5, 7, 9, 11, 13, 15); }
-__ai uint16x4_t vuzp2_u16(uint16x4_t __a, uint16x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3, 5, 7); }
-__ai uint32x2_t vuzp2_u32(uint32x2_t __a, uint32x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-__ai float32x2_t vuzp2_f32(float32x2_t __a, float32x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-__ai poly8x8_t vuzp2_p8(poly8x8_t __a, poly8x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3, 5, 7, 9, 11, 13, 15); }
-__ai poly16x4_t vuzp2_p16(poly16x4_t __a, poly16x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3, 5, 7); }
-__ai int8x16_t vuzp2q_s8(int8x16_t __a, int8x16_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31); }
-__ai int16x8_t vuzp2q_s16(int16x8_t __a, int16x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3, 5, 7, 9, 11, 13, 15); }
-__ai int32x4_t vuzp2q_s32(int32x4_t __a, int32x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3, 5, 7); }
-__ai int64x2_t vuzp2q_s64(int64x2_t __a, int64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-__ai uint8x16_t vuzp2q_u8(uint8x16_t __a, uint8x16_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31); }
-__ai uint16x8_t vuzp2q_u16(uint16x8_t __a, uint16x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3, 5, 7, 9, 11, 13, 15); }
-__ai uint32x4_t vuzp2q_u32(uint32x4_t __a, uint32x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3, 5, 7); }
-__ai uint64x2_t vuzp2q_u64(uint64x2_t __a, uint64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-__ai float32x4_t vuzp2q_f32(float32x4_t __a, float32x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3, 5, 7); }
-__ai float64x2_t vuzp2q_f64(float64x2_t __a, float64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-__ai poly8x16_t vuzp2q_p8(poly8x16_t __a, poly8x16_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31); }
-__ai poly16x8_t vuzp2q_p16(poly16x8_t __a, poly16x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3, 5, 7, 9, 11, 13, 15); }
-__ai poly64x2_t vuzp2q_p64(poly64x2_t __a, poly64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-
-__ai int8x8_t vreinterpret_s8_f64(float64x1_t __a) {
-  return (int8x8_t)__a; }
-__ai int8x8_t vreinterpret_s8_p64(poly64x1_t __a) {
-  return (int8x8_t)__a; }
-__ai int16x4_t vreinterpret_s16_f64(float64x1_t __a) {
-  return (int16x4_t)__a; }
-__ai int16x4_t vreinterpret_s16_p64(poly64x1_t __a) {
-  return (int16x4_t)__a; }
-__ai int32x2_t vreinterpret_s32_f64(float64x1_t __a) {
-  return (int32x2_t)__a; }
-__ai int32x2_t vreinterpret_s32_p64(poly64x1_t __a) {
-  return (int32x2_t)__a; }
-__ai int64x1_t vreinterpret_s64_f64(float64x1_t __a) {
-  return (int64x1_t)__a; }
-__ai int64x1_t vreinterpret_s64_p64(poly64x1_t __a) {
-  return (int64x1_t)__a; }
-__ai uint8x8_t vreinterpret_u8_f64(float64x1_t __a) {
-  return (uint8x8_t)__a; }
-__ai uint8x8_t vreinterpret_u8_p64(poly64x1_t __a) {
-  return (uint8x8_t)__a; }
-__ai uint16x4_t vreinterpret_u16_f64(float64x1_t __a) {
-  return (uint16x4_t)__a; }
-__ai uint16x4_t vreinterpret_u16_p64(poly64x1_t __a) {
-  return (uint16x4_t)__a; }
-__ai uint32x2_t vreinterpret_u32_f64(float64x1_t __a) {
-  return (uint32x2_t)__a; }
-__ai uint32x2_t vreinterpret_u32_p64(poly64x1_t __a) {
-  return (uint32x2_t)__a; }
-__ai uint64x1_t vreinterpret_u64_f64(float64x1_t __a) {
-  return (uint64x1_t)__a; }
-__ai uint64x1_t vreinterpret_u64_p64(poly64x1_t __a) {
-  return (uint64x1_t)__a; }
-__ai float16x4_t vreinterpret_f16_f64(float64x1_t __a) {
-  return (float16x4_t)__a; }
-__ai float16x4_t vreinterpret_f16_p64(poly64x1_t __a) {
-  return (float16x4_t)__a; }
-__ai float32x2_t vreinterpret_f32_f64(float64x1_t __a) {
-  return (float32x2_t)__a; }
-__ai float32x2_t vreinterpret_f32_p64(poly64x1_t __a) {
-  return (float32x2_t)__a; }
-__ai float64x1_t vreinterpret_f64_s8(int8x8_t __a) {
-  return (float64x1_t)__a; }
-__ai float64x1_t vreinterpret_f64_s16(int16x4_t __a) {
-  return (float64x1_t)__a; }
-__ai float64x1_t vreinterpret_f64_s32(int32x2_t __a) {
-  return (float64x1_t)__a; }
-__ai float64x1_t vreinterpret_f64_s64(int64x1_t __a) {
-  return (float64x1_t)__a; }
-__ai float64x1_t vreinterpret_f64_u8(uint8x8_t __a) {
-  return (float64x1_t)__a; }
-__ai float64x1_t vreinterpret_f64_u16(uint16x4_t __a) {
-  return (float64x1_t)__a; }
-__ai float64x1_t vreinterpret_f64_u32(uint32x2_t __a) {
-  return (float64x1_t)__a; }
-__ai float64x1_t vreinterpret_f64_u64(uint64x1_t __a) {
-  return (float64x1_t)__a; }
-__ai float64x1_t vreinterpret_f64_f16(float16x4_t __a) {
-  return (float64x1_t)__a; }
-__ai float64x1_t vreinterpret_f64_f32(float32x2_t __a) {
-  return (float64x1_t)__a; }
-__ai float64x1_t vreinterpret_f64_p8(poly8x8_t __a) {
-  return (float64x1_t)__a; }
-__ai float64x1_t vreinterpret_f64_p16(poly16x4_t __a) {
-  return (float64x1_t)__a; }
-__ai float64x1_t vreinterpret_f64_p64(poly64x1_t __a) {
-  return (float64x1_t)__a; }
-__ai poly8x8_t vreinterpret_p8_f64(float64x1_t __a) {
-  return (poly8x8_t)__a; }
-__ai poly8x8_t vreinterpret_p8_p64(poly64x1_t __a) {
-  return (poly8x8_t)__a; }
-__ai poly16x4_t vreinterpret_p16_f64(float64x1_t __a) {
-  return (poly16x4_t)__a; }
-__ai poly16x4_t vreinterpret_p16_p64(poly64x1_t __a) {
-  return (poly16x4_t)__a; }
-__ai poly64x1_t vreinterpret_p64_s8(int8x8_t __a) {
-  return (poly64x1_t)__a; }
-__ai poly64x1_t vreinterpret_p64_s16(int16x4_t __a) {
-  return (poly64x1_t)__a; }
-__ai poly64x1_t vreinterpret_p64_s32(int32x2_t __a) {
-  return (poly64x1_t)__a; }
-__ai poly64x1_t vreinterpret_p64_s64(int64x1_t __a) {
-  return (poly64x1_t)__a; }
-__ai poly64x1_t vreinterpret_p64_u8(uint8x8_t __a) {
-  return (poly64x1_t)__a; }
-__ai poly64x1_t vreinterpret_p64_u16(uint16x4_t __a) {
-  return (poly64x1_t)__a; }
-__ai poly64x1_t vreinterpret_p64_u32(uint32x2_t __a) {
-  return (poly64x1_t)__a; }
-__ai poly64x1_t vreinterpret_p64_u64(uint64x1_t __a) {
-  return (poly64x1_t)__a; }
-__ai poly64x1_t vreinterpret_p64_f16(float16x4_t __a) {
-  return (poly64x1_t)__a; }
-__ai poly64x1_t vreinterpret_p64_f32(float32x2_t __a) {
-  return (poly64x1_t)__a; }
-__ai poly64x1_t vreinterpret_p64_f64(float64x1_t __a) {
-  return (poly64x1_t)__a; }
-__ai poly64x1_t vreinterpret_p64_p8(poly8x8_t __a) {
-  return (poly64x1_t)__a; }
-__ai poly64x1_t vreinterpret_p64_p16(poly16x4_t __a) {
-  return (poly64x1_t)__a; }
-__ai int8x16_t vreinterpretq_s8_f64(float64x2_t __a) {
-  return (int8x16_t)__a; }
-__ai int8x16_t vreinterpretq_s8_p64(poly64x2_t __a) {
-  return (int8x16_t)__a; }
-__ai int8x16_t vreinterpretq_s8_p128(poly128_t __a) {
-  return (int8x16_t)__a; }
-__ai int16x8_t vreinterpretq_s16_f64(float64x2_t __a) {
-  return (int16x8_t)__a; }
-__ai int16x8_t vreinterpretq_s16_p64(poly64x2_t __a) {
-  return (int16x8_t)__a; }
-__ai int16x8_t vreinterpretq_s16_p128(poly128_t __a) {
-  return (int16x8_t)__a; }
-__ai int32x4_t vreinterpretq_s32_f64(float64x2_t __a) {
-  return (int32x4_t)__a; }
-__ai int32x4_t vreinterpretq_s32_p64(poly64x2_t __a) {
-  return (int32x4_t)__a; }
-__ai int32x4_t vreinterpretq_s32_p128(poly128_t __a) {
-  return (int32x4_t)__a; }
-__ai int64x2_t vreinterpretq_s64_f64(float64x2_t __a) {
-  return (int64x2_t)__a; }
-__ai int64x2_t vreinterpretq_s64_p64(poly64x2_t __a) {
-  return (int64x2_t)__a; }
-__ai int64x2_t vreinterpretq_s64_p128(poly128_t __a) {
-  return (int64x2_t)__a; }
-__ai uint8x16_t vreinterpretq_u8_f64(float64x2_t __a) {
-  return (uint8x16_t)__a; }
-__ai uint8x16_t vreinterpretq_u8_p64(poly64x2_t __a) {
-  return (uint8x16_t)__a; }
-__ai uint8x16_t vreinterpretq_u8_p128(poly128_t __a) {
-  return (uint8x16_t)__a; }
-__ai uint16x8_t vreinterpretq_u16_f64(float64x2_t __a) {
-  return (uint16x8_t)__a; }
-__ai uint16x8_t vreinterpretq_u16_p64(poly64x2_t __a) {
-  return (uint16x8_t)__a; }
-__ai uint16x8_t vreinterpretq_u16_p128(poly128_t __a) {
-  return (uint16x8_t)__a; }
-__ai uint32x4_t vreinterpretq_u32_f64(float64x2_t __a) {
-  return (uint32x4_t)__a; }
-__ai uint32x4_t vreinterpretq_u32_p64(poly64x2_t __a) {
-  return (uint32x4_t)__a; }
-__ai uint32x4_t vreinterpretq_u32_p128(poly128_t __a) {
-  return (uint32x4_t)__a; }
-__ai uint64x2_t vreinterpretq_u64_f64(float64x2_t __a) {
-  return (uint64x2_t)__a; }
-__ai uint64x2_t vreinterpretq_u64_p64(poly64x2_t __a) {
-  return (uint64x2_t)__a; }
-__ai uint64x2_t vreinterpretq_u64_p128(poly128_t __a) {
-  return (uint64x2_t)__a; }
-__ai float16x8_t vreinterpretq_f16_f64(float64x2_t __a) {
-  return (float16x8_t)__a; }
-__ai float16x8_t vreinterpretq_f16_p64(poly64x2_t __a) {
-  return (float16x8_t)__a; }
-__ai float16x8_t vreinterpretq_f16_p128(poly128_t __a) {
-  return (float16x8_t)__a; }
-__ai float32x4_t vreinterpretq_f32_f64(float64x2_t __a) {
-  return (float32x4_t)__a; }
-__ai float32x4_t vreinterpretq_f32_p64(poly64x2_t __a) {
-  return (float32x4_t)__a; }
-__ai float32x4_t vreinterpretq_f32_p128(poly128_t __a) {
-  return (float32x4_t)__a; }
-__ai float64x2_t vreinterpretq_f64_s8(int8x16_t __a) {
-  return (float64x2_t)__a; }
-__ai float64x2_t vreinterpretq_f64_s16(int16x8_t __a) {
-  return (float64x2_t)__a; }
-__ai float64x2_t vreinterpretq_f64_s32(int32x4_t __a) {
-  return (float64x2_t)__a; }
-__ai float64x2_t vreinterpretq_f64_s64(int64x2_t __a) {
-  return (float64x2_t)__a; }
-__ai float64x2_t vreinterpretq_f64_u8(uint8x16_t __a) {
-  return (float64x2_t)__a; }
-__ai float64x2_t vreinterpretq_f64_u16(uint16x8_t __a) {
-  return (float64x2_t)__a; }
-__ai float64x2_t vreinterpretq_f64_u32(uint32x4_t __a) {
-  return (float64x2_t)__a; }
-__ai float64x2_t vreinterpretq_f64_u64(uint64x2_t __a) {
-  return (float64x2_t)__a; }
-__ai float64x2_t vreinterpretq_f64_f16(float16x8_t __a) {
-  return (float64x2_t)__a; }
-__ai float64x2_t vreinterpretq_f64_f32(float32x4_t __a) {
-  return (float64x2_t)__a; }
-__ai float64x2_t vreinterpretq_f64_p8(poly8x16_t __a) {
-  return (float64x2_t)__a; }
-__ai float64x2_t vreinterpretq_f64_p16(poly16x8_t __a) {
-  return (float64x2_t)__a; }
-__ai float64x2_t vreinterpretq_f64_p64(poly64x2_t __a) {
-  return (float64x2_t)__a; }
-__ai float64x2_t vreinterpretq_f64_p128(poly128_t __a) {
-  return (float64x2_t)__a; }
-__ai poly8x16_t vreinterpretq_p8_f64(float64x2_t __a) {
-  return (poly8x16_t)__a; }
-__ai poly8x16_t vreinterpretq_p8_p64(poly64x2_t __a) {
-  return (poly8x16_t)__a; }
-__ai poly8x16_t vreinterpretq_p8_p128(poly128_t __a) {
-  return (poly8x16_t)__a; }
-__ai poly16x8_t vreinterpretq_p16_f64(float64x2_t __a) {
-  return (poly16x8_t)__a; }
-__ai poly16x8_t vreinterpretq_p16_p64(poly64x2_t __a) {
-  return (poly16x8_t)__a; }
-__ai poly16x8_t vreinterpretq_p16_p128(poly128_t __a) {
-  return (poly16x8_t)__a; }
-__ai poly64x2_t vreinterpretq_p64_s8(int8x16_t __a) {
-  return (poly64x2_t)__a; }
-__ai poly64x2_t vreinterpretq_p64_s16(int16x8_t __a) {
-  return (poly64x2_t)__a; }
-__ai poly64x2_t vreinterpretq_p64_s32(int32x4_t __a) {
-  return (poly64x2_t)__a; }
-__ai poly64x2_t vreinterpretq_p64_s64(int64x2_t __a) {
-  return (poly64x2_t)__a; }
-__ai poly64x2_t vreinterpretq_p64_u8(uint8x16_t __a) {
-  return (poly64x2_t)__a; }
-__ai poly64x2_t vreinterpretq_p64_u16(uint16x8_t __a) {
-  return (poly64x2_t)__a; }
-__ai poly64x2_t vreinterpretq_p64_u32(uint32x4_t __a) {
-  return (poly64x2_t)__a; }
-__ai poly64x2_t vreinterpretq_p64_u64(uint64x2_t __a) {
-  return (poly64x2_t)__a; }
-__ai poly64x2_t vreinterpretq_p64_f16(float16x8_t __a) {
-  return (poly64x2_t)__a; }
-__ai poly64x2_t vreinterpretq_p64_f32(float32x4_t __a) {
-  return (poly64x2_t)__a; }
-__ai poly64x2_t vreinterpretq_p64_f64(float64x2_t __a) {
-  return (poly64x2_t)__a; }
-__ai poly64x2_t vreinterpretq_p64_p8(poly8x16_t __a) {
-  return (poly64x2_t)__a; }
-__ai poly64x2_t vreinterpretq_p64_p16(poly16x8_t __a) {
-  return (poly64x2_t)__a; }
-__ai poly64x2_t vreinterpretq_p64_p128(poly128_t __a) {
-  return (poly64x2_t)__a; }
-__ai poly128_t vreinterpretq_p128_s8(int8x16_t __a) {
-  return (poly128_t)__a; }
-__ai poly128_t vreinterpretq_p128_s16(int16x8_t __a) {
-  return (poly128_t)__a; }
-__ai poly128_t vreinterpretq_p128_s32(int32x4_t __a) {
-  return (poly128_t)__a; }
-__ai poly128_t vreinterpretq_p128_s64(int64x2_t __a) {
-  return (poly128_t)__a; }
-__ai poly128_t vreinterpretq_p128_u8(uint8x16_t __a) {
-  return (poly128_t)__a; }
-__ai poly128_t vreinterpretq_p128_u16(uint16x8_t __a) {
-  return (poly128_t)__a; }
-__ai poly128_t vreinterpretq_p128_u32(uint32x4_t __a) {
-  return (poly128_t)__a; }
-__ai poly128_t vreinterpretq_p128_u64(uint64x2_t __a) {
-  return (poly128_t)__a; }
-__ai poly128_t vreinterpretq_p128_f16(float16x8_t __a) {
-  return (poly128_t)__a; }
-__ai poly128_t vreinterpretq_p128_f32(float32x4_t __a) {
-  return (poly128_t)__a; }
-__ai poly128_t vreinterpretq_p128_f64(float64x2_t __a) {
-  return (poly128_t)__a; }
-__ai poly128_t vreinterpretq_p128_p8(poly8x16_t __a) {
-  return (poly128_t)__a; }
-__ai poly128_t vreinterpretq_p128_p16(poly16x8_t __a) {
-  return (poly128_t)__a; }
-__ai poly128_t vreinterpretq_p128_p64(poly64x2_t __a) {
-  return (poly128_t)__a; }
-
-#endif
-
-__ai int8x8x2_t vzip_s8(int8x8_t __a, int8x8_t __b) {
-  int8x8x2_t r; __builtin_neon_vzip_v(&r, __a, __b, 0); return r; }
-__ai int16x4x2_t vzip_s16(int16x4_t __a, int16x4_t __b) {
-  int16x4x2_t r; __builtin_neon_vzip_v(&r, (int8x8_t)__a, (int8x8_t)__b, 1); return r; }
-__ai int32x2x2_t vzip_s32(int32x2_t __a, int32x2_t __b) {
-  int32x2x2_t r; __builtin_neon_vzip_v(&r, (int8x8_t)__a, (int8x8_t)__b, 2); return r; }
-__ai uint8x8x2_t vzip_u8(uint8x8_t __a, uint8x8_t __b) {
-  uint8x8x2_t r; __builtin_neon_vzip_v(&r, (int8x8_t)__a, (int8x8_t)__b, 16); return r; }
-__ai uint16x4x2_t vzip_u16(uint16x4_t __a, uint16x4_t __b) {
-  uint16x4x2_t r; __builtin_neon_vzip_v(&r, (int8x8_t)__a, (int8x8_t)__b, 17); return r; }
-__ai uint32x2x2_t vzip_u32(uint32x2_t __a, uint32x2_t __b) {
-  uint32x2x2_t r; __builtin_neon_vzip_v(&r, (int8x8_t)__a, (int8x8_t)__b, 18); return r; }
-__ai float32x2x2_t vzip_f32(float32x2_t __a, float32x2_t __b) {
-  float32x2x2_t r; __builtin_neon_vzip_v(&r, (int8x8_t)__a, (int8x8_t)__b, 9); return r; }
-__ai poly8x8x2_t vzip_p8(poly8x8_t __a, poly8x8_t __b) {
-  poly8x8x2_t r; __builtin_neon_vzip_v(&r, (int8x8_t)__a, (int8x8_t)__b, 4); return r; }
-__ai poly16x4x2_t vzip_p16(poly16x4_t __a, poly16x4_t __b) {
-  poly16x4x2_t r; __builtin_neon_vzip_v(&r, (int8x8_t)__a, (int8x8_t)__b, 5); return r; }
-__ai int8x16x2_t vzipq_s8(int8x16_t __a, int8x16_t __b) {
-  int8x16x2_t r; __builtin_neon_vzipq_v(&r, __a, __b, 32); return r; }
-__ai int16x8x2_t vzipq_s16(int16x8_t __a, int16x8_t __b) {
-  int16x8x2_t r; __builtin_neon_vzipq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 33); return r; }
-__ai int32x4x2_t vzipq_s32(int32x4_t __a, int32x4_t __b) {
-  int32x4x2_t r; __builtin_neon_vzipq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 34); return r; }
-__ai uint8x16x2_t vzipq_u8(uint8x16_t __a, uint8x16_t __b) {
-  uint8x16x2_t r; __builtin_neon_vzipq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 48); return r; }
-__ai uint16x8x2_t vzipq_u16(uint16x8_t __a, uint16x8_t __b) {
-  uint16x8x2_t r; __builtin_neon_vzipq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 49); return r; }
-__ai uint32x4x2_t vzipq_u32(uint32x4_t __a, uint32x4_t __b) {
-  uint32x4x2_t r; __builtin_neon_vzipq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 50); return r; }
-__ai float32x4x2_t vzipq_f32(float32x4_t __a, float32x4_t __b) {
-  float32x4x2_t r; __builtin_neon_vzipq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 41); return r; }
-__ai poly8x16x2_t vzipq_p8(poly8x16_t __a, poly8x16_t __b) {
-  poly8x16x2_t r; __builtin_neon_vzipq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 36); return r; }
-__ai poly16x8x2_t vzipq_p16(poly16x8_t __a, poly16x8_t __b) {
-  poly16x8x2_t r; __builtin_neon_vzipq_v(&r, (int8x16_t)__a, (int8x16_t)__b, 37); return r; }
+#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__)
-__ai int8x8_t vzip1_s8(int8x8_t __a, int8x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 8, 1, 9, 2, 10, 3, 11); }
-__ai int16x4_t vzip1_s16(int16x4_t __a, int16x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 4, 1, 5); }
-__ai int32x2_t vzip1_s32(int32x2_t __a, int32x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-__ai uint8x8_t vzip1_u8(uint8x8_t __a, uint8x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 8, 1, 9, 2, 10, 3, 11); }
-__ai uint16x4_t vzip1_u16(uint16x4_t __a, uint16x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 4, 1, 5); }
-__ai uint32x2_t vzip1_u32(uint32x2_t __a, uint32x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-__ai float32x2_t vzip1_f32(float32x2_t __a, float32x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-__ai poly8x8_t vzip1_p8(poly8x8_t __a, poly8x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 8, 1, 9, 2, 10, 3, 11); }
-__ai poly16x4_t vzip1_p16(poly16x4_t __a, poly16x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 4, 1, 5); }
-__ai int8x16_t vzip1q_s8(int8x16_t __a, int8x16_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23); }
-__ai int16x8_t vzip1q_s16(int16x8_t __a, int16x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 8, 1, 9, 2, 10, 3, 11); }
-__ai int32x4_t vzip1q_s32(int32x4_t __a, int32x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 4, 1, 5); }
-__ai int64x2_t vzip1q_s64(int64x2_t __a, int64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-__ai uint8x16_t vzip1q_u8(uint8x16_t __a, uint8x16_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23); }
-__ai uint16x8_t vzip1q_u16(uint16x8_t __a, uint16x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 8, 1, 9, 2, 10, 3, 11); }
-__ai uint32x4_t vzip1q_u32(uint32x4_t __a, uint32x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 4, 1, 5); }
-__ai uint64x2_t vzip1q_u64(uint64x2_t __a, uint64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-__ai float32x4_t vzip1q_f32(float32x4_t __a, float32x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 4, 1, 5); }
-__ai float64x2_t vzip1q_f64(float64x2_t __a, float64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-__ai poly8x16_t vzip1q_p8(poly8x16_t __a, poly8x16_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23); }
-__ai poly16x8_t vzip1q_p16(poly16x8_t __a, poly16x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 8, 1, 9, 2, 10, 3, 11); }
-__ai poly64x2_t vzip1q_p64(poly64x2_t __a, poly64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 0, 2); }
-
-__ai int8x8_t vzip2_s8(int8x8_t __a, int8x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 4, 12, 5, 13, 6, 14, 7, 15); }
-__ai int16x4_t vzip2_s16(int16x4_t __a, int16x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 2, 6, 3, 7); }
-__ai int32x2_t vzip2_s32(int32x2_t __a, int32x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-__ai uint8x8_t vzip2_u8(uint8x8_t __a, uint8x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 4, 12, 5, 13, 6, 14, 7, 15); }
-__ai uint16x4_t vzip2_u16(uint16x4_t __a, uint16x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 2, 6, 3, 7); }
-__ai uint32x2_t vzip2_u32(uint32x2_t __a, uint32x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-__ai float32x2_t vzip2_f32(float32x2_t __a, float32x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-__ai poly8x8_t vzip2_p8(poly8x8_t __a, poly8x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 4, 12, 5, 13, 6, 14, 7, 15); }
-__ai poly16x4_t vzip2_p16(poly16x4_t __a, poly16x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 2, 6, 3, 7); }
-__ai int8x16_t vzip2q_s8(int8x16_t __a, int8x16_t __b) {
-  return __builtin_shufflevector(__a, __b, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31); }
-__ai int16x8_t vzip2q_s16(int16x8_t __a, int16x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 4, 12, 5, 13, 6, 14, 7, 15); }
-__ai int32x4_t vzip2q_s32(int32x4_t __a, int32x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 2, 6, 3, 7); }
-__ai int64x2_t vzip2q_s64(int64x2_t __a, int64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-__ai uint8x16_t vzip2q_u8(uint8x16_t __a, uint8x16_t __b) {
-  return __builtin_shufflevector(__a, __b, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31); }
-__ai uint16x8_t vzip2q_u16(uint16x8_t __a, uint16x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 4, 12, 5, 13, 6, 14, 7, 15); }
-__ai uint32x4_t vzip2q_u32(uint32x4_t __a, uint32x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 2, 6, 3, 7); }
-__ai uint64x2_t vzip2q_u64(uint64x2_t __a, uint64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-__ai float32x4_t vzip2q_f32(float32x4_t __a, float32x4_t __b) {
-  return __builtin_shufflevector(__a, __b, 2, 6, 3, 7); }
-__ai float64x2_t vzip2q_f64(float64x2_t __a, float64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-__ai poly8x16_t vzip2q_p8(poly8x16_t __a, poly8x16_t __b) {
-  return __builtin_shufflevector(__a, __b, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31); }
-__ai poly16x8_t vzip2q_p16(poly16x8_t __a, poly16x8_t __b) {
-  return __builtin_shufflevector(__a, __b, 4, 12, 5, 13, 6, 14, 7, 15); }
-__ai poly64x2_t vzip2q_p64(poly64x2_t __a, poly64x2_t __b) {
-  return __builtin_shufflevector(__a, __b, 1, 3); }
-
-__ai int8x16_t vmovn_high_s16(int8x8_t __a, int16x8_t __b) {
-  int8x8_t __a1 = vmovn_s16(__b);
-  return __builtin_shufflevector(__a, __a1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); }
-__ai int16x8_t vmovn_high_s32(int16x4_t __a, int32x4_t __b) {
-  int16x4_t __a1 = vmovn_s32(__b);
-  return __builtin_shufflevector(__a, __a1, 0, 1, 2, 3, 4, 5, 6, 7); }
-__ai int32x4_t vmovn_high_s64(int32x2_t __a, int64x2_t __b) {
-  int32x2_t __a1 = vmovn_s64(__b);
-  return __builtin_shufflevector(__a, __a1, 0, 1, 2, 3); }
-__ai uint8x16_t vmovn_high_u16(uint8x8_t __a, uint16x8_t __b) {
-  uint8x8_t __a1 = vmovn_u16(__b);
-  return __builtin_shufflevector(__a, __a1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15); }
-__ai uint16x8_t vmovn_high_u32(uint16x4_t __a, uint32x4_t __b) {
-  uint16x4_t __a1 = vmovn_u32(__b);
-  return __builtin_shufflevector(__a, __a1, 0, 1, 2, 3, 4, 5, 6, 7); }
-__ai uint32x4_t vmovn_high_u64(uint32x2_t __a, uint64x2_t __b) {
-  uint32x2_t __a1 = vmovn_u64(__b);
-  return __builtin_shufflevector(__a, __a1, 0, 1, 2, 3); }
+#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
 
diff --git a/lib/clang/3.5/include/avxintrin.h b/lib/clang/3.5/include/avxintrin.h
index 3d50439..4e1044a 100644
--- a/lib/clang/3.5/include/avxintrin.h
+++ b/lib/clang/3.5/include/avxintrin.h
@@ -737,19 +737,22 @@ _mm256_zeroupper(void)
 static __inline __m128 __attribute__((__always_inline__, __nodebug__))
 _mm_broadcast_ss(float const *__a)
 {
-  return (__m128)__builtin_ia32_vbroadcastss(__a);
+  float __f = *__a;
+  return (__m128)(__v4sf){ __f, __f, __f, __f };
 }
 
 static __inline __m256d __attribute__((__always_inline__, __nodebug__))
 _mm256_broadcast_sd(double const *__a)
 {
-  return (__m256d)__builtin_ia32_vbroadcastsd256(__a);
+  double __d = *__a;
+  return (__m256d)(__v4df){ __d, __d, __d, __d };
 }
 
 static __inline __m256 __attribute__((__always_inline__, __nodebug__))
 _mm256_broadcast_ss(float const *__a)
 {
-  return (__m256)__builtin_ia32_vbroadcastss256(__a);
+  float __f = *__a;
+  return (__m256)(__v8sf){ __f, __f, __f, __f, __f, __f, __f, __f };
 }
 
 static __inline __m256d __attribute__((__always_inline__, __nodebug__))
diff --git a/lib/clang/3.5/include/bmiintrin.h b/lib/clang/3.5/include/bmiintrin.h
index 8cb00f5..43c4a5e 100644
--- a/lib/clang/3.5/include/bmiintrin.h
+++ b/lib/clang/3.5/include/bmiintrin.h
@@ -32,6 +32,14 @@
 #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)
 {
@@ -44,12 +52,20 @@ __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)
 {
@@ -75,18 +91,34 @@ __tzcnt_u32(unsigned int __X)
 }
 
 #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)
 {
@@ -110,6 +142,7 @@ __tzcnt_u64(unsigned long long __X)
 {
   return __builtin_ctzll(__X);
 }
-#endif
+
+#endif /* __x86_64__ */
 
 #endif /* __BMIINTRIN_H */
diff --git a/lib/clang/3.5/include/ia32intrin.h b/lib/clang/3.5/include/ia32intrin.h
index 55c2247..5adf3f1 100644
--- a/lib/clang/3.5/include/ia32intrin.h
+++ b/lib/clang/3.5/include/ia32intrin.h
@@ -79,6 +79,11 @@ __writeeflags(unsigned int __f)
 }
 #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) {
diff --git a/lib/clang/3.5/include/stdarg.h b/lib/clang/3.5/include/stdarg.h
index 6110a06..a57e183 100644
--- a/lib/clang/3.5/include/stdarg.h
+++ b/lib/clang/3.5/include/stdarg.h
@@ -44,7 +44,9 @@ typedef __builtin_va_list va_list;
 #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/3.5/include/x86intrin.h b/lib/clang/3.5/include/x86intrin.h
index be9e71d..21a43da 100644
--- a/lib/clang/3.5/include/x86intrin.h
+++ b/lib/clang/3.5/include/x86intrin.h
@@ -76,6 +76,6 @@
 #include <f16cintrin.h>
 #endif
 
-// FIXME: LWP
+/* FIXME: LWP */
 
 #endif /* __X86INTRIN_H */
diff --git a/lib/clang/3.5/include/xmmintrin.h b/lib/clang/3.5/include/xmmintrin.h
index e777ec0..c9befcb 100644
--- a/lib/clang/3.5/include/xmmintrin.h
+++ b/lib/clang/3.5/include/xmmintrin.h
@@ -34,8 +34,8 @@ 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.
+/* 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
@@ -589,7 +589,7 @@ _mm_set1_ps(float __w)
   return (__m128){ __w, __w, __w, __w };
 }
 
-// Microsoft specific.
+/* Microsoft specific. */
 static __inline__ __m128 __attribute__((__always_inline__, __nodebug__))
 _mm_set_ps1(float __w)
 {
diff --git a/lib/libc++.so b/lib/libc++.so
index 96e82ae..9f69843 100755
--- a/lib/libc++.so
+++ b/lib/libc++.so