diff options
| author | Carlos Hernandez <chernand@google.com> | 2014-08-05 17:53:32 -0700 |
|---|---|---|
| committer | Carlos Hernandez <chernand@google.com> | 2014-08-05 17:54:05 -0700 |
| commit | 7faaa9f3f0df9d23790277834d426c3d992ac3ba (patch) | |
| tree | b788ae3b96daf9f5a79d8ec434e1e9edd56b3a72 /unsupported | |
| parent | 810535bb0c575a003b32076e5352ab8fd3f23a1c (diff) | |
| download | eigen-lollipop-mr1-dev.tar.gz | |
Update Eigen to the latest stable release, 3.2.2android-wear-5.1.1_r1android-wear-5.1.0_r1android-wear-5.0.0_r1android-l-preview_r2android-cts-5.1_r9android-cts-5.1_r8android-cts-5.1_r7android-cts-5.1_r6android-cts-5.1_r5android-cts-5.1_r4android-cts-5.1_r3android-cts-5.1_r28android-cts-5.1_r27android-cts-5.1_r26android-cts-5.1_r25android-cts-5.1_r24android-cts-5.1_r23android-cts-5.1_r22android-cts-5.1_r21android-cts-5.1_r20android-cts-5.1_r2android-cts-5.1_r19android-cts-5.1_r18android-cts-5.1_r17android-cts-5.1_r16android-cts-5.1_r15android-cts-5.1_r14android-cts-5.1_r13android-cts-5.1_r10android-cts-5.1_r1android-cts-5.0_r9android-cts-5.0_r8android-cts-5.0_r7android-cts-5.0_r6android-cts-5.0_r5android-cts-5.0_r4android-cts-5.0_r3android-5.1.1_r9android-5.1.1_r8android-5.1.1_r7android-5.1.1_r6android-5.1.1_r5android-5.1.1_r4android-5.1.1_r38android-5.1.1_r37android-5.1.1_r36android-5.1.1_r35android-5.1.1_r34android-5.1.1_r33android-5.1.1_r30android-5.1.1_r3android-5.1.1_r29android-5.1.1_r28android-5.1.1_r26android-5.1.1_r25android-5.1.1_r24android-5.1.1_r23android-5.1.1_r22android-5.1.1_r20android-5.1.1_r2android-5.1.1_r19android-5.1.1_r18android-5.1.1_r17android-5.1.1_r16android-5.1.1_r15android-5.1.1_r14android-5.1.1_r13android-5.1.1_r12android-5.1.1_r10android-5.1.1_r1android-5.1.0_r5android-5.1.0_r4android-5.1.0_r3android-5.1.0_r1android-5.0.2_r3android-5.0.2_r1android-5.0.1_r1android-5.0.0_r7android-5.0.0_r6android-5.0.0_r5.1android-5.0.0_r5android-5.0.0_r4android-5.0.0_r3android-5.0.0_r2android-5.0.0_r1lollipop-wear-releaselollipop-releaselollipop-mr1-wfc-releaselollipop-mr1-releaselollipop-mr1-fi-releaselollipop-mr1-devlollipop-mr1-cts-releaselollipop-devlollipop-cts-releasel-preview
./Eigen/src/Core/util/NonMPL2.h is left untouched, so that
usage of non MPL2 code is disabled.
Change-Id: I86fc9257b3c30d0ca15b268d4ef07bf038bba7ca
Diffstat (limited to 'unsupported')
102 files changed, 10816 insertions, 10539 deletions
diff --git a/unsupported/Eigen/AdolcForward b/unsupported/Eigen/AdolcForward index b96f9450e..2627decd0 100644 --- a/unsupported/Eigen/AdolcForward +++ b/unsupported/Eigen/AdolcForward @@ -44,7 +44,7 @@ namespace Eigen { -/** \ingroup Unsupported_modules +/** * \defgroup AdolcForward_Module Adolc forward module * This module provides support for adolc's adouble type in forward mode. * ADOL-C is a C++ automatic differentiation library, diff --git a/unsupported/Eigen/AlignedVector3 b/unsupported/Eigen/AlignedVector3 index 8ad0eb477..7b45e6cce 100644 --- a/unsupported/Eigen/AlignedVector3 +++ b/unsupported/Eigen/AlignedVector3 @@ -14,7 +14,7 @@ namespace Eigen { -/** \ingroup Unsupported_modules +/** * \defgroup AlignedVector3_Module Aligned vector3 module * * \code @@ -167,7 +167,8 @@ template<typename _Scalar> class AlignedVector3 inline Scalar norm() const { - return internal::sqrt(squaredNorm()); + using std::sqrt; + return sqrt(squaredNorm()); } inline AlignedVector3 cross(const AlignedVector3& other) const diff --git a/unsupported/Eigen/ArpackSupport b/unsupported/Eigen/ArpackSupport new file mode 100644 index 000000000..37a2799ef --- /dev/null +++ b/unsupported/Eigen/ArpackSupport @@ -0,0 +1,31 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + +#ifndef EIGEN_ARPACKSUPPORT_MODULE_H +#define EIGEN_ARPACKSUPPORT_MODULE_H + +#include <Eigen/Core> + +#include <Eigen/src/Core/util/DisableStupidWarnings.h> + +/** \defgroup ArpackSupport_Module Arpack support module + * + * This module provides a wrapper to Arpack, a library for sparse eigenvalue decomposition. + * + * \code + * #include <Eigen/ArpackSupport> + * \endcode + */ + +#include <Eigen/SparseCholesky> +#include "src/Eigenvalues/ArpackSelfAdjointEigenSolver.h" + +#include <Eigen/src/Core/util/ReenableStupidWarnings.h> + +#endif // EIGEN_ARPACKSUPPORT_MODULE_H +/* vim: set filetype=cpp et sw=2 ts=2 ai: */ diff --git a/unsupported/Eigen/AutoDiff b/unsupported/Eigen/AutoDiff index 3c73b424e..abf5b7d67 100644 --- a/unsupported/Eigen/AutoDiff +++ b/unsupported/Eigen/AutoDiff @@ -12,7 +12,7 @@ namespace Eigen { -/** \ingroup Unsupported_modules +/** * \defgroup AutoDiff_Module Auto Diff module * * This module features forward automatic differentation via a simple diff --git a/unsupported/Eigen/BVH b/unsupported/Eigen/BVH index 860a7dd89..0161a5402 100644 --- a/unsupported/Eigen/BVH +++ b/unsupported/Eigen/BVH @@ -18,7 +18,7 @@ namespace Eigen { -/** \ingroup Unsupported_modules +/** * \defgroup BVH_Module BVH module * \brief This module provides generic bounding volume hierarchy algorithms * and reference tree implementations. diff --git a/unsupported/Eigen/CMakeLists.txt b/unsupported/Eigen/CMakeLists.txt index e961e72c5..e06f1238b 100644 --- a/unsupported/Eigen/CMakeLists.txt +++ b/unsupported/Eigen/CMakeLists.txt @@ -1,6 +1,6 @@ set(Eigen_HEADERS AdolcForward BVH IterativeSolvers MatrixFunctions MoreVectorization AutoDiff AlignedVector3 Polynomials FFT NonLinearOptimization SparseExtra IterativeSolvers - NumericalDiff Skyline MPRealSupport OpenGLSupport KroneckerProduct Splines + NumericalDiff Skyline MPRealSupport OpenGLSupport KroneckerProduct Splines LevenbergMarquardt ) install(FILES diff --git a/unsupported/Eigen/FFT b/unsupported/Eigen/FFT index d233c6d5f..2c45b3999 100644 --- a/unsupported/Eigen/FFT +++ b/unsupported/Eigen/FFT @@ -16,7 +16,7 @@ #include <Eigen/Core> -/** \ingroup Unsupported_modules +/** * \defgroup FFT_Module Fast Fourier Transform module * * \code diff --git a/unsupported/Eigen/IterativeSolvers b/unsupported/Eigen/IterativeSolvers index 6c6946d91..aa15403db 100644 --- a/unsupported/Eigen/IterativeSolvers +++ b/unsupported/Eigen/IterativeSolvers @@ -12,7 +12,7 @@ #include <Eigen/Sparse> -/** \ingroup Unsupported_modules +/** * \defgroup IterativeSolvers_Module Iterative solvers module * This module aims to provide various iterative linear and non linear solver algorithms. * It currently provides: @@ -27,13 +27,18 @@ #include "../../Eigen/src/misc/Solve.h" #include "../../Eigen/src/misc/SparseSolve.h" +#ifndef EIGEN_MPL2_ONLY #include "src/IterativeSolvers/IterationController.h" #include "src/IterativeSolvers/ConstrainedConjGrad.h" +#endif + #include "src/IterativeSolvers/IncompleteLU.h" #include "../../Eigen/Jacobi" #include "../../Eigen/Householder" #include "src/IterativeSolvers/GMRES.h" +#include "src/IterativeSolvers/IncompleteCholesky.h" //#include "src/IterativeSolvers/SSORPreconditioner.h" +#include "src/IterativeSolvers/MINRES.h" //@} diff --git a/unsupported/Eigen/KroneckerProduct b/unsupported/Eigen/KroneckerProduct index 796e386ad..c932c06a6 100644 --- a/unsupported/Eigen/KroneckerProduct +++ b/unsupported/Eigen/KroneckerProduct @@ -1,3 +1,11 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + #ifndef EIGEN_KRONECKER_PRODUCT_MODULE_H #define EIGEN_KRONECKER_PRODUCT_MODULE_H @@ -7,7 +15,7 @@ namespace Eigen { -/** \ingroup Unsupported_modules +/** * \defgroup KroneckerProduct_Module KroneckerProduct module * * This module contains an experimental Kronecker product implementation. diff --git a/unsupported/Eigen/LevenbergMarquardt b/unsupported/Eigen/LevenbergMarquardt new file mode 100644 index 000000000..0fe2680ba --- /dev/null +++ b/unsupported/Eigen/LevenbergMarquardt @@ -0,0 +1,45 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2009 Thomas Capricelli <orzel@freehackers.org> +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + +#ifndef EIGEN_LEVENBERGMARQUARDT_MODULE +#define EIGEN_LEVENBERGMARQUARDT_MODULE + +// #include <vector> + +#include <Eigen/Core> +#include <Eigen/Jacobi> +#include <Eigen/QR> +#include <unsupported/Eigen/NumericalDiff> + +#include <Eigen/SparseQR> + +/** + * \defgroup LevenbergMarquardt_Module Levenberg-Marquardt module + * + * \code + * #include </Eigen/LevenbergMarquardt> + * \endcode + * + * + */ + +#include "Eigen/SparseCore" +#ifndef EIGEN_PARSED_BY_DOXYGEN + +#include "src/LevenbergMarquardt/LMqrsolv.h" +#include "src/LevenbergMarquardt/LMcovar.h" +#include "src/LevenbergMarquardt/LMpar.h" + +#endif + +#include "src/LevenbergMarquardt/LevenbergMarquardt.h" +#include "src/LevenbergMarquardt/LMonestep.h" + + +#endif // EIGEN_LEVENBERGMARQUARDT_MODULE diff --git a/unsupported/Eigen/MPRealSupport b/unsupported/Eigen/MPRealSupport index 3895623fe..d4b03647d 100644 --- a/unsupported/Eigen/MPRealSupport +++ b/unsupported/Eigen/MPRealSupport @@ -1,41 +1,36 @@ // This file is part of a joint effort between Eigen, a lightweight C++ template library // for linear algebra, and MPFR C++, a C++ interface to MPFR library (http://www.holoborodko.com/pavel/) // -// Copyright (C) 2010 Pavel Holoborodko <pavel@holoborodko.com> +// Copyright (C) 2010-2012 Pavel Holoborodko <pavel@holoborodko.com> // Copyright (C) 2010 Konstantin Holoborodko <konstantin@holoborodko.com> // Copyright (C) 2010 Gael Guennebaud <gael.guennebaud@inria.fr> // // This Source Code Form is subject to the terms of the Mozilla // Public License v. 2.0. If a copy of the MPL was not distributed // with this file, You can obtain one at http://mozilla.org/MPL/2.0/. -// -// Contributors: -// Brian Gladman, Helmut Jarausch, Fokko Beekhof, Ulrich Mutze, Heinz van Saanen, Pere Constans #ifndef EIGEN_MPREALSUPPORT_MODULE_H #define EIGEN_MPREALSUPPORT_MODULE_H -#include <ctime> -#include <mpreal.h> #include <Eigen/Core> +#include <mpreal.h> namespace Eigen { - /** \ingroup Unsupported_modules - * \defgroup MPRealSupport_Module MPFRC++ Support module - * - * \code - * #include <Eigen/MPRealSupport> - * \endcode - * - * This module provides support for multi precision floating point numbers - * via the <a href="http://www.holoborodko.com/pavel/mpfr">MPFR C++</a> - * library which itself is built upon <a href="http://www.mpfr.org/">MPFR</a>/<a href="http://gmplib.org/">GMP</a>. - * - * You can find a copy of MPFR C++ that is known to be compatible in the unsupported/test/mpreal folder. - * - * Here is an example: - * +/** + * \defgroup MPRealSupport_Module MPFRC++ Support module + * \code + * #include <Eigen/MPRealSupport> + * \endcode + * + * This module provides support for multi precision floating point numbers + * via the <a href="http://www.holoborodko.com/pavel/mpfr">MPFR C++</a> + * library which itself is built upon <a href="http://www.mpfr.org/">MPFR</a>/<a href="http://gmplib.org/">GMP</a>. + * + * You can find a copy of MPFR C++ that is known to be compatible in the unsupported/test/mpreal folder. + * + * Here is an example: + * \code #include <iostream> #include <Eigen/MPRealSupport> @@ -59,9 +54,9 @@ int main() return 0; } \endcode - * - */ - + * + */ + template<> struct NumTraits<mpfr::mpreal> : GenericNumTraits<mpfr::mpreal> { @@ -77,72 +72,132 @@ int main() typedef mpfr::mpreal Real; typedef mpfr::mpreal NonInteger; - - inline static mpfr::mpreal highest() { return mpfr::mpreal_max(mpfr::mpreal::get_default_prec()); } - inline static mpfr::mpreal lowest() { return -mpfr::mpreal_max(mpfr::mpreal::get_default_prec()); } - - inline static Real epsilon() - { - return mpfr::machine_epsilon(mpfr::mpreal::get_default_prec()); - } - inline static Real dummy_precision() - { - unsigned int weak_prec = ((mpfr::mpreal::get_default_prec()-1)*90)/100; - return mpfr::machine_epsilon(weak_prec); + + inline static Real highest (long Precision = mpfr::mpreal::get_default_prec()) { return mpfr::maxval(Precision); } + inline static Real lowest (long Precision = mpfr::mpreal::get_default_prec()) { return -mpfr::maxval(Precision); } + + // Constants + inline static Real Pi (long Precision = mpfr::mpreal::get_default_prec()) { return mpfr::const_pi(Precision); } + inline static Real Euler (long Precision = mpfr::mpreal::get_default_prec()) { return mpfr::const_euler(Precision); } + inline static Real Log2 (long Precision = mpfr::mpreal::get_default_prec()) { return mpfr::const_log2(Precision); } + inline static Real Catalan (long Precision = mpfr::mpreal::get_default_prec()) { return mpfr::const_catalan(Precision); } + + inline static Real epsilon (long Precision = mpfr::mpreal::get_default_prec()) { return mpfr::machine_epsilon(Precision); } + inline static Real epsilon (const Real& x) { return mpfr::machine_epsilon(x); } + + inline static Real dummy_precision() + { + unsigned int weak_prec = ((mpfr::mpreal::get_default_prec()-1) * 90) / 100; + return mpfr::machine_epsilon(weak_prec); } }; -namespace internal { + namespace internal { - template<> mpfr::mpreal random<mpfr::mpreal>() + template<> inline mpfr::mpreal random<mpfr::mpreal>() { -#if (MPFR_VERSION >= MPFR_VERSION_NUM(3,0,0)) - static gmp_randstate_t state; - static bool isFirstTime = true; - - if(isFirstTime) - { - gmp_randinit_default(state); - gmp_randseed_ui(state,(unsigned)time(NULL)); - isFirstTime = false; - } - - return mpfr::urandom(state)*2-1; -#else - return mpfr::mpreal(random<double>()); -#endif + return mpfr::random(); } - template<> mpfr::mpreal random<mpfr::mpreal>(const mpfr::mpreal& a, const mpfr::mpreal& b) + template<> inline mpfr::mpreal random<mpfr::mpreal>(const mpfr::mpreal& a, const mpfr::mpreal& b) { return a + (b-a) * random<mpfr::mpreal>(); } - bool isMuchSmallerThan(const mpfr::mpreal& a, const mpfr::mpreal& b, const mpfr::mpreal& prec) + inline bool isMuchSmallerThan(const mpfr::mpreal& a, const mpfr::mpreal& b, const mpfr::mpreal& eps) { - return mpfr::abs(a) <= mpfr::abs(b) * prec; + return mpfr::abs(a) <= mpfr::abs(b) * eps; } - inline bool isApprox(const mpfr::mpreal& a, const mpfr::mpreal& b, const mpfr::mpreal& prec) + inline bool isApprox(const mpfr::mpreal& a, const mpfr::mpreal& b, const mpfr::mpreal& eps) { - return mpfr::abs(a - b) <= (mpfr::min)(mpfr::abs(a), mpfr::abs(b)) * prec; + return mpfr::isEqualFuzzy(a,b,eps); } - inline bool isApproxOrLessThan(const mpfr::mpreal& a, const mpfr::mpreal& b, const mpfr::mpreal& prec) + inline bool isApproxOrLessThan(const mpfr::mpreal& a, const mpfr::mpreal& b, const mpfr::mpreal& eps) { - return a <= b || isApprox(a, b, prec); + return a <= b || mpfr::isEqualFuzzy(a,b,eps); } - + template<> inline long double cast<mpfr::mpreal,long double>(const mpfr::mpreal& x) { return x.toLDouble(); } + template<> inline double cast<mpfr::mpreal,double>(const mpfr::mpreal& x) { return x.toDouble(); } + template<> inline long cast<mpfr::mpreal,long>(const mpfr::mpreal& x) { return x.toLong(); } + template<> inline int cast<mpfr::mpreal,int>(const mpfr::mpreal& x) { return int(x.toLong()); } -} // end namespace internal + // Specialize GEBP kernel and traits for mpreal (no need for peeling, nor complicated stuff) + // This also permits to directly call mpfr's routines and avoid many temporaries produced by mpreal + template<> + class gebp_traits<mpfr::mpreal, mpfr::mpreal, false, false> + { + public: + typedef mpfr::mpreal ResScalar; + enum { + nr = 2, // must be 2 for proper packing... + mr = 1, + WorkSpaceFactor = nr, + LhsProgress = 1, + RhsProgress = 1 + }; + }; + + template<typename Index, int mr, int nr, bool ConjugateLhs, bool ConjugateRhs> + struct gebp_kernel<mpfr::mpreal,mpfr::mpreal,Index,mr,nr,ConjugateLhs,ConjugateRhs> + { + typedef mpfr::mpreal mpreal; + + EIGEN_DONT_INLINE + void operator()(mpreal* res, Index resStride, const mpreal* blockA, const mpreal* blockB, Index rows, Index depth, Index cols, mpreal alpha, + Index strideA=-1, Index strideB=-1, Index offsetA=0, Index offsetB=0, mpreal* /*unpackedB*/ = 0) + { + mpreal acc1, acc2, tmp; + + if(strideA==-1) strideA = depth; + if(strideB==-1) strideB = depth; + + for(Index j=0; j<cols; j+=nr) + { + Index actual_nr = (std::min<Index>)(nr,cols-j); + mpreal *C1 = res + j*resStride; + mpreal *C2 = res + (j+1)*resStride; + for(Index i=0; i<rows; i++) + { + mpreal *B = const_cast<mpreal*>(blockB) + j*strideB + offsetB*actual_nr; + mpreal *A = const_cast<mpreal*>(blockA) + i*strideA + offsetA; + acc1 = 0; + acc2 = 0; + for(Index k=0; k<depth; k++) + { + mpfr_mul(tmp.mpfr_ptr(), A[k].mpfr_ptr(), B[0].mpfr_ptr(), mpreal::get_default_rnd()); + mpfr_add(acc1.mpfr_ptr(), acc1.mpfr_ptr(), tmp.mpfr_ptr(), mpreal::get_default_rnd()); + + if(actual_nr==2) { + mpfr_mul(tmp.mpfr_ptr(), A[k].mpfr_ptr(), B[1].mpfr_ptr(), mpreal::get_default_rnd()); + mpfr_add(acc2.mpfr_ptr(), acc2.mpfr_ptr(), tmp.mpfr_ptr(), mpreal::get_default_rnd()); + } + + B+=actual_nr; + } + + mpfr_mul(acc1.mpfr_ptr(), acc1.mpfr_ptr(), alpha.mpfr_ptr(), mpreal::get_default_rnd()); + mpfr_add(C1[i].mpfr_ptr(), C1[i].mpfr_ptr(), acc1.mpfr_ptr(), mpreal::get_default_rnd()); + + if(actual_nr==2) { + mpfr_mul(acc2.mpfr_ptr(), acc2.mpfr_ptr(), alpha.mpfr_ptr(), mpreal::get_default_rnd()); + mpfr_add(C2[i].mpfr_ptr(), C2[i].mpfr_ptr(), acc2.mpfr_ptr(), mpreal::get_default_rnd()); + } + } + } + } + }; + + } // end namespace internal } #endif // EIGEN_MPREALSUPPORT_MODULE_H diff --git a/unsupported/Eigen/MatrixFunctions b/unsupported/Eigen/MatrixFunctions index 56ab71cd3..0991817d5 100644 --- a/unsupported/Eigen/MatrixFunctions +++ b/unsupported/Eigen/MatrixFunctions @@ -2,6 +2,7 @@ // for linear algebra. // // Copyright (C) 2009 Jitse Niesen <jitse@maths.leeds.ac.uk> +// Copyright (C) 2012 Chen-Pang He <jdh8@ms63.hinet.net> // // This Source Code Form is subject to the terms of the Mozilla // Public License v. 2.0. If a copy of the MPL was not distributed @@ -19,7 +20,7 @@ #include <Eigen/LU> #include <Eigen/Eigenvalues> -/** \ingroup Unsupported_modules +/** * \defgroup MatrixFunctions_Module Matrix functions module * \brief This module aims to provide various methods for the computation of * matrix functions. @@ -35,6 +36,7 @@ * - \ref matrixbase_cosh "MatrixBase::cosh()", for computing the matrix hyperbolic cosine * - \ref matrixbase_exp "MatrixBase::exp()", for computing the matrix exponential * - \ref matrixbase_log "MatrixBase::log()", for computing the matrix logarithm + * - \ref matrixbase_pow "MatrixBase::pow()", for computing the matrix power * - \ref matrixbase_matrixfunction "MatrixBase::matrixFunction()", for computing general matrix functions * - \ref matrixbase_sin "MatrixBase::sin()", for computing the matrix sine * - \ref matrixbase_sinh "MatrixBase::sinh()", for computing the matrix hyperbolic sine @@ -57,19 +59,21 @@ #include "src/MatrixFunctions/MatrixFunction.h" #include "src/MatrixFunctions/MatrixSquareRoot.h" #include "src/MatrixFunctions/MatrixLogarithm.h" - +#include "src/MatrixFunctions/MatrixPower.h" /** -\page matrixbaseextra MatrixBase methods defined in the MatrixFunctions module +\page matrixbaseextra_page \ingroup MatrixFunctions_Module +\section matrixbaseextra MatrixBase methods defined in the MatrixFunctions module + The remainder of the page documents the following MatrixBase methods which are defined in the MatrixFunctions module. -\section matrixbase_cos MatrixBase::cos() +\subsection matrixbase_cos MatrixBase::cos() Compute the matrix cosine. @@ -86,7 +90,7 @@ This function calls \ref matrixbase_matrixfunction "matrixFunction()" with StdSt -\section matrixbase_cosh MatrixBase::cosh() +\subsection matrixbase_cosh MatrixBase::cosh() Compute the matrix hyberbolic cosine. @@ -103,7 +107,7 @@ This function calls \ref matrixbase_matrixfunction "matrixFunction()" with StdSt -\section matrixbase_exp MatrixBase::exp() +\subsection matrixbase_exp MatrixBase::exp() Compute the matrix exponential. @@ -158,7 +162,7 @@ Output: \verbinclude MatrixExponential.out \c complex<float>, \c complex<double>, or \c complex<long double> . -\section matrixbase_log MatrixBase::log() +\subsection matrixbase_log MatrixBase::log() Compute the matrix logarithm. @@ -209,14 +213,77 @@ documentation of \ref matrixbase_exp "exp()". \include MatrixLogarithm.cpp Output: \verbinclude MatrixLogarithm.out -\note \p M has to be a matrix of \c float, \c double, \c long double -\c complex<float>, \c complex<double>, or \c complex<long double> . +\note \p M has to be a matrix of \c float, \c double, <tt>long +double</tt>, \c complex<float>, \c complex<double>, or \c complex<long +double> . \sa MatrixBase::exp(), MatrixBase::matrixFunction(), class MatrixLogarithmAtomic, MatrixBase::sqrt(). -\section matrixbase_matrixfunction MatrixBase::matrixFunction() +\subsection matrixbase_pow MatrixBase::pow() + +Compute the matrix raised to arbitrary real power. + +\code +const MatrixPowerReturnValue<Derived> MatrixBase<Derived>::pow(RealScalar p) const +\endcode + +\param[in] M base of the matrix power, should be a square matrix. +\param[in] p exponent of the matrix power, should be real. + +The matrix power \f$ M^p \f$ is defined as \f$ \exp(p \log(M)) \f$, +where exp denotes the matrix exponential, and log denotes the matrix +logarithm. + +The matrix \f$ M \f$ should meet the conditions to be an argument of +matrix logarithm. If \p p is not of the real scalar type of \p M, it +is casted into the real scalar type of \p M. + +This function computes the matrix power using the Schur-Padé +algorithm as implemented by class MatrixPower. The exponent is split +into integral part and fractional part, where the fractional part is +in the interval \f$ (-1, 1) \f$. The main diagonal and the first +super-diagonal is directly computed. + +Details of the algorithm can be found in: Nicholas J. Higham and +Lijing Lin, "A Schur-Padé algorithm for fractional powers of a +matrix," <em>SIAM J. %Matrix Anal. Applic.</em>, +<b>32(3)</b>:1056–1078, 2011. + +Example: The following program checks that +\f[ \left[ \begin{array}{ccc} + \cos1 & -\sin1 & 0 \\ + \sin1 & \cos1 & 0 \\ + 0 & 0 & 1 + \end{array} \right]^{\frac14\pi} = \left[ \begin{array}{ccc} + \frac12\sqrt2 & -\frac12\sqrt2 & 0 \\ + \frac12\sqrt2 & \frac12\sqrt2 & 0 \\ + 0 & 0 & 1 + \end{array} \right]. \f] +This corresponds to \f$ \frac14\pi \f$ rotations of 1 radian around +the z-axis. + +\include MatrixPower.cpp +Output: \verbinclude MatrixPower.out + +MatrixBase::pow() is user-friendly. However, there are some +circumstances under which you should use class MatrixPower directly. +MatrixPower can save the result of Schur decomposition, so it's +better for computing various powers for the same matrix. + +Example: +\include MatrixPower_optimal.cpp +Output: \verbinclude MatrixPower_optimal.out + +\note \p M has to be a matrix of \c float, \c double, <tt>long +double</tt>, \c complex<float>, \c complex<double>, or \c complex<long +double> . + +\sa MatrixBase::exp(), MatrixBase::log(), class MatrixPower. + + +\subsection matrixbase_matrixfunction MatrixBase::matrixFunction() Compute a matrix function. @@ -272,7 +339,7 @@ A.matrixFunction(StdStemFunctions<std::complex<double> >::exp, &B); -\section matrixbase_sin MatrixBase::sin() +\subsection matrixbase_sin MatrixBase::sin() Compute the matrix sine. @@ -290,7 +357,7 @@ Output: \verbinclude MatrixSine.out -\section matrixbase_sinh MatrixBase::sinh() +\subsection matrixbase_sinh MatrixBase::sinh() Compute the matrix hyperbolic sine. @@ -307,7 +374,7 @@ Example: \include MatrixSinh.cpp Output: \verbinclude MatrixSinh.out -\section matrixbase_sqrt MatrixBase::sqrt() +\subsection matrixbase_sqrt MatrixBase::sqrt() Compute the matrix square root. diff --git a/unsupported/Eigen/MoreVectorization b/unsupported/Eigen/MoreVectorization index 9f0a39f75..470e72430 100644 --- a/unsupported/Eigen/MoreVectorization +++ b/unsupported/Eigen/MoreVectorization @@ -1,3 +1,11 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + #ifndef EIGEN_MOREVECTORIZATION_MODULE_H #define EIGEN_MOREVECTORIZATION_MODULE_H @@ -5,7 +13,7 @@ namespace Eigen { -/** \ingroup Unsupported_modules +/** * \defgroup MoreVectorization More vectorization module */ diff --git a/unsupported/Eigen/NonLinearOptimization b/unsupported/Eigen/NonLinearOptimization index cf6ca58f8..600ab4c12 100644 --- a/unsupported/Eigen/NonLinearOptimization +++ b/unsupported/Eigen/NonLinearOptimization @@ -1,4 +1,4 @@ -// This file is part of Eugenio, a lightweight C++ template library +// This file is part of Eigen, a lightweight C++ template library // for linear algebra. // // Copyright (C) 2009 Thomas Capricelli <orzel@freehackers.org> @@ -17,7 +17,7 @@ #include <Eigen/QR> #include <unsupported/Eigen/NumericalDiff> -/** \ingroup Unsupported_modules +/** * \defgroup NonLinearOptimization_Module Non linear optimization module * * \code diff --git a/unsupported/Eigen/NumericalDiff b/unsupported/Eigen/NumericalDiff index b3480312d..433334ca8 100644 --- a/unsupported/Eigen/NumericalDiff +++ b/unsupported/Eigen/NumericalDiff @@ -14,7 +14,7 @@ namespace Eigen { -/** \ingroup Unsupported_modules +/** * \defgroup NumericalDiff_Module Numerical differentiation module * * \code diff --git a/unsupported/Eigen/OpenGLSupport b/unsupported/Eigen/OpenGLSupport index e66a425f8..c4090ab11 100644 --- a/unsupported/Eigen/OpenGLSupport +++ b/unsupported/Eigen/OpenGLSupport @@ -11,11 +11,16 @@ #define EIGEN_OPENGL_MODULE #include <Eigen/Geometry> -#include <GL/gl.h> + +#if defined(__APPLE_CC__) + #include <OpenGL/gl.h> +#else + #include <GL/gl.h> +#endif namespace Eigen { -/** \ingroup Unsupported_modules +/** * \defgroup OpenGLSUpport_Module OpenGL Support module * * This module provides wrapper functions for a couple of OpenGL functions diff --git a/unsupported/Eigen/Polynomials b/unsupported/Eigen/Polynomials index fa58b006d..cece56337 100644 --- a/unsupported/Eigen/Polynomials +++ b/unsupported/Eigen/Polynomials @@ -1,3 +1,11 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + #ifndef EIGEN_POLYNOMIALS_MODULE_H #define EIGEN_POLYNOMIALS_MODULE_H @@ -16,11 +24,8 @@ #undef EIGEN_HIDE_HEAVY_CODE #endif -/** \ingroup Unsupported_modules +/** * \defgroup Polynomials_Module Polynomials module - * - * - * * \brief This module provides a QR based polynomial solver. * * To use this module, add diff --git a/unsupported/Eigen/SVD b/unsupported/Eigen/SVD new file mode 100644 index 000000000..7cc059280 --- /dev/null +++ b/unsupported/Eigen/SVD @@ -0,0 +1,39 @@ +#ifndef EIGEN_SVD_MODULE_H +#define EIGEN_SVD_MODULE_H + +#include <Eigen/QR> +#include <Eigen/Householder> +#include <Eigen/Jacobi> + +#include "../../Eigen/src/Core/util/DisableStupidWarnings.h" + +/** \defgroup SVD_Module SVD module + * + * + * + * This module provides SVD decomposition for matrices (both real and complex). + * This decomposition is accessible via the following MatrixBase method: + * - MatrixBase::jacobiSvd() + * + * \code + * #include <Eigen/SVD> + * \endcode + */ + +#include "../../Eigen/src/misc/Solve.h" +#include "../../Eigen/src/SVD/UpperBidiagonalization.h" +#include "src/SVD/SVDBase.h" +#include "src/SVD/JacobiSVD.h" +#include "src/SVD/BDCSVD.h" +#if defined(EIGEN_USE_LAPACKE) && !defined(EIGEN_USE_LAPACKE_STRICT) +#include "../../Eigen/src/SVD/JacobiSVD_MKL.h" +#endif + +#ifdef EIGEN2_SUPPORT +#include "../../Eigen/src/Eigen2Support/SVD.h" +#endif + +#include "../../Eigen/src/Core/util/ReenableStupidWarnings.h" + +#endif // EIGEN_SVD_MODULE_H +/* vim: set filetype=cpp et sw=2 ts=2 ai: */ diff --git a/unsupported/Eigen/Skyline b/unsupported/Eigen/Skyline index c9823f358..71a68cb42 100644 --- a/unsupported/Eigen/Skyline +++ b/unsupported/Eigen/Skyline @@ -1,3 +1,11 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + #ifndef EIGEN_SKYLINE_MODULE_H #define EIGEN_SKYLINE_MODULE_H @@ -11,7 +19,7 @@ #include <cstring> #include <algorithm> -/** \ingroup Unsupported_modules +/** * \defgroup Skyline_Module Skyline module * * diff --git a/unsupported/Eigen/SparseExtra b/unsupported/Eigen/SparseExtra index 340c34736..b5597902a 100644 --- a/unsupported/Eigen/SparseExtra +++ b/unsupported/Eigen/SparseExtra @@ -1,3 +1,12 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2008-2009 Gael Guennebaud <g.gael@free.fr> +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + #ifndef EIGEN_SPARSE_EXTRA_MODULE_H #define EIGEN_SPARSE_EXTRA_MODULE_H @@ -17,7 +26,7 @@ #include <google/dense_hash_map> #endif -/** \ingroup Unsupported_modules +/** * \defgroup SparseExtra_Module SparseExtra module * * This module contains some experimental features extending the sparse module. diff --git a/unsupported/Eigen/Splines b/unsupported/Eigen/Splines index 801cec1a1..322e6b9f5 100644 --- a/unsupported/Eigen/Splines +++ b/unsupported/Eigen/Splines @@ -12,7 +12,7 @@ namespace Eigen { -/** \ingroup Unsupported_modules +/** * \defgroup Splines_Module Spline and spline fitting module * * This module provides a simple multi-dimensional spline class while diff --git a/unsupported/Eigen/src/AutoDiff/AutoDiffScalar.h b/unsupported/Eigen/src/AutoDiff/AutoDiffScalar.h index b833df3c0..8d42e69b9 100644 --- a/unsupported/Eigen/src/AutoDiff/AutoDiffScalar.h +++ b/unsupported/Eigen/src/AutoDiff/AutoDiffScalar.h @@ -47,8 +47,8 @@ template<typename _DerType, bool Enable> struct auto_diff_special_op; * * It supports the following list of global math function: * - std::abs, std::sqrt, std::pow, std::exp, std::log, std::sin, std::cos, - * - internal::abs, internal::sqrt, internal::pow, internal::exp, internal::log, internal::sin, internal::cos, - * - internal::conj, internal::real, internal::imag, internal::abs2. + * - internal::abs, internal::sqrt, numext::pow, internal::exp, internal::log, internal::sin, internal::cos, + * - internal::conj, internal::real, internal::imag, numext::abs2. * * AutoDiffScalar can be used as the scalar type of an Eigen::Matrix object. However, * in that case, the expression template mechanism only occurs at the top Matrix level, @@ -489,20 +489,32 @@ struct make_coherent_impl<Matrix<A_Scalar, A_Rows, A_Cols, A_Options, A_MaxRows, } }; -template<typename A_Scalar, int A_Rows, int A_Cols, int A_Options, int A_MaxRows, int A_MaxCols> struct scalar_product_traits<Matrix<A_Scalar, A_Rows, A_Cols, A_Options, A_MaxRows, A_MaxCols>,A_Scalar> +template<typename A_Scalar, int A_Rows, int A_Cols, int A_Options, int A_MaxRows, int A_MaxCols> +struct scalar_product_traits<Matrix<A_Scalar, A_Rows, A_Cols, A_Options, A_MaxRows, A_MaxCols>,A_Scalar> { - typedef Matrix<A_Scalar, A_Rows, A_Cols, A_Options, A_MaxRows, A_MaxCols> ReturnType; + enum { Defined = 1 }; + typedef Matrix<A_Scalar, A_Rows, A_Cols, A_Options, A_MaxRows, A_MaxCols> ReturnType; }; -template<typename A_Scalar, int A_Rows, int A_Cols, int A_Options, int A_MaxRows, int A_MaxCols> struct scalar_product_traits<A_Scalar, Matrix<A_Scalar, A_Rows, A_Cols, A_Options, A_MaxRows, A_MaxCols> > +template<typename A_Scalar, int A_Rows, int A_Cols, int A_Options, int A_MaxRows, int A_MaxCols> +struct scalar_product_traits<A_Scalar, Matrix<A_Scalar, A_Rows, A_Cols, A_Options, A_MaxRows, A_MaxCols> > { - typedef Matrix<A_Scalar, A_Rows, A_Cols, A_Options, A_MaxRows, A_MaxCols> ReturnType; + enum { Defined = 1 }; + typedef Matrix<A_Scalar, A_Rows, A_Cols, A_Options, A_MaxRows, A_MaxCols> ReturnType; }; template<typename DerType> struct scalar_product_traits<AutoDiffScalar<DerType>,typename DerType::Scalar> { - typedef AutoDiffScalar<DerType> ReturnType; + enum { Defined = 1 }; + typedef AutoDiffScalar<DerType> ReturnType; +}; + +template<typename DerType> +struct scalar_product_traits<typename DerType::Scalar,AutoDiffScalar<DerType> > +{ + enum { Defined = 1 }; + typedef AutoDiffScalar<DerType> ReturnType; }; } // end namespace internal @@ -532,14 +544,12 @@ inline AutoDiffScalar<DerType> (min)(const T& x, const AutoDiffScalar<DerType>& template<typename DerType, typename T> inline AutoDiffScalar<DerType> (max)(const T& x, const AutoDiffScalar<DerType>& y) { return (x > y ? x : y); } -#define sign(x) x >= 0 ? 1 : -1 // required for abs function below - EIGEN_AUTODIFF_DECLARE_GLOBAL_UNARY(abs, using std::abs; - return ReturnType(abs(x.value()), x.derivatives() * (sign(x.value())));) + return ReturnType(abs(x.value()), x.derivatives() * (x.value()<0 ? -1 : 1) );) EIGEN_AUTODIFF_DECLARE_GLOBAL_UNARY(abs2, - using internal::abs2; + using numext::abs2; return ReturnType(abs2(x.value()), x.derivatives() * (Scalar(2)*x.value()));) EIGEN_AUTODIFF_DECLARE_GLOBAL_UNARY(sqrt, @@ -602,17 +612,17 @@ atan2(const AutoDiffScalar<DerTypeA>& a, const AutoDiffScalar<DerTypeB>& b) EIGEN_AUTODIFF_DECLARE_GLOBAL_UNARY(tan, using std::tan; using std::cos; - return ReturnType(tan(x.value()),x.derivatives() * (Scalar(1)/internal::abs2(cos(x.value()))));) + return ReturnType(tan(x.value()),x.derivatives() * (Scalar(1)/numext::abs2(cos(x.value()))));) EIGEN_AUTODIFF_DECLARE_GLOBAL_UNARY(asin, using std::sqrt; using std::asin; - return ReturnType(asin(x.value()),x.derivatives() * (Scalar(1)/sqrt(1-internal::abs2(x.value()))));) + return ReturnType(asin(x.value()),x.derivatives() * (Scalar(1)/sqrt(1-numext::abs2(x.value()))));) EIGEN_AUTODIFF_DECLARE_GLOBAL_UNARY(acos, using std::sqrt; using std::acos; - return ReturnType(acos(x.value()),x.derivatives() * (Scalar(-1)/sqrt(1-internal::abs2(x.value()))));) + return ReturnType(acos(x.value()),x.derivatives() * (Scalar(-1)/sqrt(1-numext::abs2(x.value()))));) #undef EIGEN_AUTODIFF_DECLARE_GLOBAL_UNARY diff --git a/unsupported/Eigen/src/AutoDiff/AutoDiffVector.h b/unsupported/Eigen/src/AutoDiff/AutoDiffVector.h index 0540add0a..8c2d04830 100644 --- a/unsupported/Eigen/src/AutoDiff/AutoDiffVector.h +++ b/unsupported/Eigen/src/AutoDiff/AutoDiffVector.h @@ -21,8 +21,8 @@ namespace Eigen { * * It supports the following list of global math function: * - std::abs, std::sqrt, std::pow, std::exp, std::log, std::sin, std::cos, - * - internal::abs, internal::sqrt, internal::pow, internal::exp, internal::log, internal::sin, internal::cos, - * - internal::conj, internal::real, internal::imag, internal::abs2. + * - internal::abs, internal::sqrt, numext::pow, internal::exp, internal::log, internal::sin, internal::cos, + * - internal::conj, internal::real, internal::imag, numext::abs2. * * AutoDiffScalar can be used as the scalar type of an Eigen::Matrix object. However, * in that case, the expression template mechanism only occurs at the top Matrix level, diff --git a/unsupported/Eigen/src/BVH/BVAlgorithms.h b/unsupported/Eigen/src/BVH/BVAlgorithms.h index e5b51decb..994c8af54 100644 --- a/unsupported/Eigen/src/BVH/BVAlgorithms.h +++ b/unsupported/Eigen/src/BVH/BVAlgorithms.h @@ -189,7 +189,7 @@ struct minimizer_helper1 Object2 stored; Minimizer &minimizer; private: - minimizer_helper1& operator=(const minimizer_helper1&) {} + minimizer_helper1& operator=(const minimizer_helper1&); }; template<typename Volume2, typename Object2, typename Object1, typename Minimizer> diff --git a/unsupported/Eigen/src/Eigenvalues/ArpackSelfAdjointEigenSolver.h b/unsupported/Eigen/src/Eigenvalues/ArpackSelfAdjointEigenSolver.h new file mode 100644 index 000000000..3b6a69aff --- /dev/null +++ b/unsupported/Eigen/src/Eigenvalues/ArpackSelfAdjointEigenSolver.h @@ -0,0 +1,805 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2012 David Harmon <dharmon@gmail.com> +// +// Eigen is free software; you can redistribute it and/or +// modify it under the terms of the GNU Lesser General Public +// License as published by the Free Software Foundation; either +// version 3 of the License, or (at your option) any later version. +// +// Alternatively, you can redistribute it and/or +// modify it under the terms of the GNU General Public License as +// published by the Free Software Foundation; either version 2 of +// the License, or (at your option) any later version. +// +// Eigen is distributed in the hope that it will be useful, but WITHOUT ANY +// WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS +// FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License or the +// GNU General Public License for more details. +// +// You should have received a copy of the GNU Lesser General Public +// License and a copy of the GNU General Public License along with +// Eigen. If not, see <http://www.gnu.org/licenses/>. + +#ifndef EIGEN_ARPACKGENERALIZEDSELFADJOINTEIGENSOLVER_H +#define EIGEN_ARPACKGENERALIZEDSELFADJOINTEIGENSOLVER_H + +#include <Eigen/Dense> + +namespace Eigen { + +namespace internal { + template<typename Scalar, typename RealScalar> struct arpack_wrapper; + template<typename MatrixSolver, typename MatrixType, typename Scalar, bool BisSPD> struct OP; +} + + + +template<typename MatrixType, typename MatrixSolver=SimplicialLLT<MatrixType>, bool BisSPD=false> +class ArpackGeneralizedSelfAdjointEigenSolver +{ +public: + //typedef typename MatrixSolver::MatrixType MatrixType; + + /** \brief Scalar type for matrices of type \p MatrixType. */ + typedef typename MatrixType::Scalar Scalar; + typedef typename MatrixType::Index Index; + + /** \brief Real scalar type for \p MatrixType. + * + * This is just \c Scalar if #Scalar is real (e.g., \c float or + * \c Scalar), and the type of the real part of \c Scalar if #Scalar is + * complex. + */ + typedef typename NumTraits<Scalar>::Real RealScalar; + + /** \brief Type for vector of eigenvalues as returned by eigenvalues(). + * + * This is a column vector with entries of type #RealScalar. + * The length of the vector is the size of \p nbrEigenvalues. + */ + typedef typename internal::plain_col_type<MatrixType, RealScalar>::type RealVectorType; + + /** \brief Default constructor. + * + * The default constructor is for cases in which the user intends to + * perform decompositions via compute(). + * + */ + ArpackGeneralizedSelfAdjointEigenSolver() + : m_eivec(), + m_eivalues(), + m_isInitialized(false), + m_eigenvectorsOk(false), + m_nbrConverged(0), + m_nbrIterations(0) + { } + + /** \brief Constructor; computes generalized eigenvalues of given matrix with respect to another matrix. + * + * \param[in] A Self-adjoint matrix whose eigenvalues / eigenvectors will + * computed. By default, the upper triangular part is used, but can be changed + * through the template parameter. + * \param[in] B Self-adjoint matrix for the generalized eigenvalue problem. + * \param[in] nbrEigenvalues The number of eigenvalues / eigenvectors to compute. + * Must be less than the size of the input matrix, or an error is returned. + * \param[in] eigs_sigma String containing either "LM", "SM", "LA", or "SA", with + * respective meanings to find the largest magnitude , smallest magnitude, + * largest algebraic, or smallest algebraic eigenvalues. Alternatively, this + * value can contain floating point value in string form, in which case the + * eigenvalues closest to this value will be found. + * \param[in] options Can be #ComputeEigenvectors (default) or #EigenvaluesOnly. + * \param[in] tol What tolerance to find the eigenvalues to. Default is 0, which + * means machine precision. + * + * This constructor calls compute(const MatrixType&, const MatrixType&, Index, string, int, RealScalar) + * to compute the eigenvalues of the matrix \p A with respect to \p B. The eigenvectors are computed if + * \p options equals #ComputeEigenvectors. + * + */ + ArpackGeneralizedSelfAdjointEigenSolver(const MatrixType& A, const MatrixType& B, + Index nbrEigenvalues, std::string eigs_sigma="LM", + int options=ComputeEigenvectors, RealScalar tol=0.0) + : m_eivec(), + m_eivalues(), + m_isInitialized(false), + m_eigenvectorsOk(false), + m_nbrConverged(0), + m_nbrIterations(0) + { + compute(A, B, nbrEigenvalues, eigs_sigma, options, tol); + } + + /** \brief Constructor; computes eigenvalues of given matrix. + * + * \param[in] A Self-adjoint matrix whose eigenvalues / eigenvectors will + * computed. By default, the upper triangular part is used, but can be changed + * through the template parameter. + * \param[in] nbrEigenvalues The number of eigenvalues / eigenvectors to compute. + * Must be less than the size of the input matrix, or an error is returned. + * \param[in] eigs_sigma String containing either "LM", "SM", "LA", or "SA", with + * respective meanings to find the largest magnitude , smallest magnitude, + * largest algebraic, or smallest algebraic eigenvalues. Alternatively, this + * value can contain floating point value in string form, in which case the + * eigenvalues closest to this value will be found. + * \param[in] options Can be #ComputeEigenvectors (default) or #EigenvaluesOnly. + * \param[in] tol What tolerance to find the eigenvalues to. Default is 0, which + * means machine precision. + * + * This constructor calls compute(const MatrixType&, Index, string, int, RealScalar) + * to compute the eigenvalues of the matrix \p A. The eigenvectors are computed if + * \p options equals #ComputeEigenvectors. + * + */ + + ArpackGeneralizedSelfAdjointEigenSolver(const MatrixType& A, + Index nbrEigenvalues, std::string eigs_sigma="LM", + int options=ComputeEigenvectors, RealScalar tol=0.0) + : m_eivec(), + m_eivalues(), + m_isInitialized(false), + m_eigenvectorsOk(false), + m_nbrConverged(0), + m_nbrIterations(0) + { + compute(A, nbrEigenvalues, eigs_sigma, options, tol); + } + + + /** \brief Computes generalized eigenvalues / eigenvectors of given matrix using the external ARPACK library. + * + * \param[in] A Selfadjoint matrix whose eigendecomposition is to be computed. + * \param[in] B Selfadjoint matrix for generalized eigenvalues. + * \param[in] nbrEigenvalues The number of eigenvalues / eigenvectors to compute. + * Must be less than the size of the input matrix, or an error is returned. + * \param[in] eigs_sigma String containing either "LM", "SM", "LA", or "SA", with + * respective meanings to find the largest magnitude , smallest magnitude, + * largest algebraic, or smallest algebraic eigenvalues. Alternatively, this + * value can contain floating point value in string form, in which case the + * eigenvalues closest to this value will be found. + * \param[in] options Can be #ComputeEigenvectors (default) or #EigenvaluesOnly. + * \param[in] tol What tolerance to find the eigenvalues to. Default is 0, which + * means machine precision. + * + * \returns Reference to \c *this + * + * This function computes the generalized eigenvalues of \p A with respect to \p B using ARPACK. The eigenvalues() + * function can be used to retrieve them. If \p options equals #ComputeEigenvectors, + * then the eigenvectors are also computed and can be retrieved by + * calling eigenvectors(). + * + */ + ArpackGeneralizedSelfAdjointEigenSolver& compute(const MatrixType& A, const MatrixType& B, + Index nbrEigenvalues, std::string eigs_sigma="LM", + int options=ComputeEigenvectors, RealScalar tol=0.0); + + /** \brief Computes eigenvalues / eigenvectors of given matrix using the external ARPACK library. + * + * \param[in] A Selfadjoint matrix whose eigendecomposition is to be computed. + * \param[in] nbrEigenvalues The number of eigenvalues / eigenvectors to compute. + * Must be less than the size of the input matrix, or an error is returned. + * \param[in] eigs_sigma String containing either "LM", "SM", "LA", or "SA", with + * respective meanings to find the largest magnitude , smallest magnitude, + * largest algebraic, or smallest algebraic eigenvalues. Alternatively, this + * value can contain floating point value in string form, in which case the + * eigenvalues closest to this value will be found. + * \param[in] options Can be #ComputeEigenvectors (default) or #EigenvaluesOnly. + * \param[in] tol What tolerance to find the eigenvalues to. Default is 0, which + * means machine precision. + * + * \returns Reference to \c *this + * + * This function computes the eigenvalues of \p A using ARPACK. The eigenvalues() + * function can be used to retrieve them. If \p options equals #ComputeEigenvectors, + * then the eigenvectors are also computed and can be retrieved by + * calling eigenvectors(). + * + */ + ArpackGeneralizedSelfAdjointEigenSolver& compute(const MatrixType& A, + Index nbrEigenvalues, std::string eigs_sigma="LM", + int options=ComputeEigenvectors, RealScalar tol=0.0); + + + /** \brief Returns the eigenvectors of given matrix. + * + * \returns A const reference to the matrix whose columns are the eigenvectors. + * + * \pre The eigenvectors have been computed before. + * + * Column \f$ k \f$ of the returned matrix is an eigenvector corresponding + * to eigenvalue number \f$ k \f$ as returned by eigenvalues(). The + * eigenvectors are normalized to have (Euclidean) norm equal to one. If + * this object was used to solve the eigenproblem for the selfadjoint + * matrix \f$ A \f$, then the matrix returned by this function is the + * matrix \f$ V \f$ in the eigendecomposition \f$ A V = D V \f$. + * For the generalized eigenproblem, the matrix returned is the solution \f$ A V = D B V \f$ + * + * Example: \include SelfAdjointEigenSolver_eigenvectors.cpp + * Output: \verbinclude SelfAdjointEigenSolver_eigenvectors.out + * + * \sa eigenvalues() + */ + const Matrix<Scalar, Dynamic, Dynamic>& eigenvectors() const + { + eigen_assert(m_isInitialized && "ArpackGeneralizedSelfAdjointEigenSolver is not initialized."); + eigen_assert(m_eigenvectorsOk && "The eigenvectors have not been computed together with the eigenvalues."); + return m_eivec; + } + + /** \brief Returns the eigenvalues of given matrix. + * + * \returns A const reference to the column vector containing the eigenvalues. + * + * \pre The eigenvalues have been computed before. + * + * The eigenvalues are repeated according to their algebraic multiplicity, + * so there are as many eigenvalues as rows in the matrix. The eigenvalues + * are sorted in increasing order. + * + * Example: \include SelfAdjointEigenSolver_eigenvalues.cpp + * Output: \verbinclude SelfAdjointEigenSolver_eigenvalues.out + * + * \sa eigenvectors(), MatrixBase::eigenvalues() + */ + const Matrix<Scalar, Dynamic, 1>& eigenvalues() const + { + eigen_assert(m_isInitialized && "ArpackGeneralizedSelfAdjointEigenSolver is not initialized."); + return m_eivalues; + } + + /** \brief Computes the positive-definite square root of the matrix. + * + * \returns the positive-definite square root of the matrix + * + * \pre The eigenvalues and eigenvectors of a positive-definite matrix + * have been computed before. + * + * The square root of a positive-definite matrix \f$ A \f$ is the + * positive-definite matrix whose square equals \f$ A \f$. This function + * uses the eigendecomposition \f$ A = V D V^{-1} \f$ to compute the + * square root as \f$ A^{1/2} = V D^{1/2} V^{-1} \f$. + * + * Example: \include SelfAdjointEigenSolver_operatorSqrt.cpp + * Output: \verbinclude SelfAdjointEigenSolver_operatorSqrt.out + * + * \sa operatorInverseSqrt(), + * \ref MatrixFunctions_Module "MatrixFunctions Module" + */ + Matrix<Scalar, Dynamic, Dynamic> operatorSqrt() const + { + eigen_assert(m_isInitialized && "SelfAdjointEigenSolver is not initialized."); + eigen_assert(m_eigenvectorsOk && "The eigenvectors have not been computed together with the eigenvalues."); + return m_eivec * m_eivalues.cwiseSqrt().asDiagonal() * m_eivec.adjoint(); + } + + /** \brief Computes the inverse square root of the matrix. + * + * \returns the inverse positive-definite square root of the matrix + * + * \pre The eigenvalues and eigenvectors of a positive-definite matrix + * have been computed before. + * + * This function uses the eigendecomposition \f$ A = V D V^{-1} \f$ to + * compute the inverse square root as \f$ V D^{-1/2} V^{-1} \f$. This is + * cheaper than first computing the square root with operatorSqrt() and + * then its inverse with MatrixBase::inverse(). + * + * Example: \include SelfAdjointEigenSolver_operatorInverseSqrt.cpp + * Output: \verbinclude SelfAdjointEigenSolver_operatorInverseSqrt.out + * + * \sa operatorSqrt(), MatrixBase::inverse(), + * \ref MatrixFunctions_Module "MatrixFunctions Module" + */ + Matrix<Scalar, Dynamic, Dynamic> operatorInverseSqrt() const + { + eigen_assert(m_isInitialized && "SelfAdjointEigenSolver is not initialized."); + eigen_assert(m_eigenvectorsOk && "The eigenvectors have not been computed together with the eigenvalues."); + return m_eivec * m_eivalues.cwiseInverse().cwiseSqrt().asDiagonal() * m_eivec.adjoint(); + } + + /** \brief Reports whether previous computation was successful. + * + * \returns \c Success if computation was succesful, \c NoConvergence otherwise. + */ + ComputationInfo info() const + { + eigen_assert(m_isInitialized && "ArpackGeneralizedSelfAdjointEigenSolver is not initialized."); + return m_info; + } + + size_t getNbrConvergedEigenValues() const + { return m_nbrConverged; } + + size_t getNbrIterations() const + { return m_nbrIterations; } + +protected: + Matrix<Scalar, Dynamic, Dynamic> m_eivec; + Matrix<Scalar, Dynamic, 1> m_eivalues; + ComputationInfo m_info; + bool m_isInitialized; + bool m_eigenvectorsOk; + + size_t m_nbrConverged; + size_t m_nbrIterations; +}; + + + + + +template<typename MatrixType, typename MatrixSolver, bool BisSPD> +ArpackGeneralizedSelfAdjointEigenSolver<MatrixType, MatrixSolver, BisSPD>& + ArpackGeneralizedSelfAdjointEigenSolver<MatrixType, MatrixSolver, BisSPD> +::compute(const MatrixType& A, Index nbrEigenvalues, + std::string eigs_sigma, int options, RealScalar tol) +{ + MatrixType B(0,0); + compute(A, B, nbrEigenvalues, eigs_sigma, options, tol); + + return *this; +} + + +template<typename MatrixType, typename MatrixSolver, bool BisSPD> +ArpackGeneralizedSelfAdjointEigenSolver<MatrixType, MatrixSolver, BisSPD>& + ArpackGeneralizedSelfAdjointEigenSolver<MatrixType, MatrixSolver, BisSPD> +::compute(const MatrixType& A, const MatrixType& B, Index nbrEigenvalues, + std::string eigs_sigma, int options, RealScalar tol) +{ + eigen_assert(A.cols() == A.rows()); + eigen_assert(B.cols() == B.rows()); + eigen_assert(B.rows() == 0 || A.cols() == B.rows()); + eigen_assert((options &~ (EigVecMask | GenEigMask)) == 0 + && (options & EigVecMask) != EigVecMask + && "invalid option parameter"); + + bool isBempty = (B.rows() == 0) || (B.cols() == 0); + + // For clarity, all parameters match their ARPACK name + // + // Always 0 on the first call + // + int ido = 0; + + int n = (int)A.cols(); + + // User options: "LA", "SA", "SM", "LM", "BE" + // + char whch[3] = "LM"; + + // Specifies the shift if iparam[6] = { 3, 4, 5 }, not used if iparam[6] = { 1, 2 } + // + RealScalar sigma = 0.0; + + if (eigs_sigma.length() >= 2 && isalpha(eigs_sigma[0]) && isalpha(eigs_sigma[1])) + { + eigs_sigma[0] = toupper(eigs_sigma[0]); + eigs_sigma[1] = toupper(eigs_sigma[1]); + + // In the following special case we're going to invert the problem, since solving + // for larger magnitude is much much faster + // i.e., if 'SM' is specified, we're going to really use 'LM', the default + // + if (eigs_sigma.substr(0,2) != "SM") + { + whch[0] = eigs_sigma[0]; + whch[1] = eigs_sigma[1]; + } + } + else + { + eigen_assert(false && "Specifying clustered eigenvalues is not yet supported!"); + + // If it's not scalar values, then the user may be explicitly + // specifying the sigma value to cluster the evs around + // + sigma = atof(eigs_sigma.c_str()); + + // If atof fails, it returns 0.0, which is a fine default + // + } + + // "I" means normal eigenvalue problem, "G" means generalized + // + char bmat[2] = "I"; + if (eigs_sigma.substr(0,2) == "SM" || !(isalpha(eigs_sigma[0]) && isalpha(eigs_sigma[1])) || (!isBempty && !BisSPD)) + bmat[0] = 'G'; + + // Now we determine the mode to use + // + int mode = (bmat[0] == 'G') + 1; + if (eigs_sigma.substr(0,2) == "SM" || !(isalpha(eigs_sigma[0]) && isalpha(eigs_sigma[1]))) + { + // We're going to use shift-and-invert mode, and basically find + // the largest eigenvalues of the inverse operator + // + mode = 3; + } + + // The user-specified number of eigenvalues/vectors to compute + // + int nev = (int)nbrEigenvalues; + + // Allocate space for ARPACK to store the residual + // + Scalar *resid = new Scalar[n]; + + // Number of Lanczos vectors, must satisfy nev < ncv <= n + // Note that this indicates that nev != n, and we cannot compute + // all eigenvalues of a mtrix + // + int ncv = std::min(std::max(2*nev, 20), n); + + // The working n x ncv matrix, also store the final eigenvectors (if computed) + // + Scalar *v = new Scalar[n*ncv]; + int ldv = n; + + // Working space + // + Scalar *workd = new Scalar[3*n]; + int lworkl = ncv*ncv+8*ncv; // Must be at least this length + Scalar *workl = new Scalar[lworkl]; + + int *iparam= new int[11]; + iparam[0] = 1; // 1 means we let ARPACK perform the shifts, 0 means we'd have to do it + iparam[2] = std::max(300, (int)std::ceil(2*n/std::max(ncv,1))); + iparam[6] = mode; // The mode, 1 is standard ev problem, 2 for generalized ev, 3 for shift-and-invert + + // Used during reverse communicate to notify where arrays start + // + int *ipntr = new int[11]; + + // Error codes are returned in here, initial value of 0 indicates a random initial + // residual vector is used, any other values means resid contains the initial residual + // vector, possibly from a previous run + // + int info = 0; + + Scalar scale = 1.0; + //if (!isBempty) + //{ + //Scalar scale = B.norm() / std::sqrt(n); + //scale = std::pow(2, std::floor(std::log(scale+1))); + ////M /= scale; + //for (size_t i=0; i<(size_t)B.outerSize(); i++) + // for (typename MatrixType::InnerIterator it(B, i); it; ++it) + // it.valueRef() /= scale; + //} + + MatrixSolver OP; + if (mode == 1 || mode == 2) + { + if (!isBempty) + OP.compute(B); + } + else if (mode == 3) + { + if (sigma == 0.0) + { + OP.compute(A); + } + else + { + // Note: We will never enter here because sigma must be 0.0 + // + if (isBempty) + { + MatrixType AminusSigmaB(A); + for (Index i=0; i<A.rows(); ++i) + AminusSigmaB.coeffRef(i,i) -= sigma; + + OP.compute(AminusSigmaB); + } + else + { + MatrixType AminusSigmaB = A - sigma * B; + OP.compute(AminusSigmaB); + } + } + } + + if (!(mode == 1 && isBempty) && !(mode == 2 && isBempty) && OP.info() != Success) + std::cout << "Error factoring matrix" << std::endl; + + do + { + internal::arpack_wrapper<Scalar, RealScalar>::saupd(&ido, bmat, &n, whch, &nev, &tol, resid, + &ncv, v, &ldv, iparam, ipntr, workd, workl, + &lworkl, &info); + + if (ido == -1 || ido == 1) + { + Scalar *in = workd + ipntr[0] - 1; + Scalar *out = workd + ipntr[1] - 1; + + if (ido == 1 && mode != 2) + { + Scalar *out2 = workd + ipntr[2] - 1; + if (isBempty || mode == 1) + Matrix<Scalar, Dynamic, 1>::Map(out2, n) = Matrix<Scalar, Dynamic, 1>::Map(in, n); + else + Matrix<Scalar, Dynamic, 1>::Map(out2, n) = B * Matrix<Scalar, Dynamic, 1>::Map(in, n); + + in = workd + ipntr[2] - 1; + } + + if (mode == 1) + { + if (isBempty) + { + // OP = A + // + Matrix<Scalar, Dynamic, 1>::Map(out, n) = A * Matrix<Scalar, Dynamic, 1>::Map(in, n); + } + else + { + // OP = L^{-1}AL^{-T} + // + internal::OP<MatrixSolver, MatrixType, Scalar, BisSPD>::applyOP(OP, A, n, in, out); + } + } + else if (mode == 2) + { + if (ido == 1) + Matrix<Scalar, Dynamic, 1>::Map(in, n) = A * Matrix<Scalar, Dynamic, 1>::Map(in, n); + + // OP = B^{-1} A + // + Matrix<Scalar, Dynamic, 1>::Map(out, n) = OP.solve(Matrix<Scalar, Dynamic, 1>::Map(in, n)); + } + else if (mode == 3) + { + // OP = (A-\sigmaB)B (\sigma could be 0, and B could be I) + // The B * in is already computed and stored at in if ido == 1 + // + if (ido == 1 || isBempty) + Matrix<Scalar, Dynamic, 1>::Map(out, n) = OP.solve(Matrix<Scalar, Dynamic, 1>::Map(in, n)); + else + Matrix<Scalar, Dynamic, 1>::Map(out, n) = OP.solve(B * Matrix<Scalar, Dynamic, 1>::Map(in, n)); + } + } + else if (ido == 2) + { + Scalar *in = workd + ipntr[0] - 1; + Scalar *out = workd + ipntr[1] - 1; + + if (isBempty || mode == 1) + Matrix<Scalar, Dynamic, 1>::Map(out, n) = Matrix<Scalar, Dynamic, 1>::Map(in, n); + else + Matrix<Scalar, Dynamic, 1>::Map(out, n) = B * Matrix<Scalar, Dynamic, 1>::Map(in, n); + } + } while (ido != 99); + + if (info == 1) + m_info = NoConvergence; + else if (info == 3) + m_info = NumericalIssue; + else if (info < 0) + m_info = InvalidInput; + else if (info != 0) + eigen_assert(false && "Unknown ARPACK return value!"); + else + { + // Do we compute eigenvectors or not? + // + int rvec = (options & ComputeEigenvectors) == ComputeEigenvectors; + + // "A" means "All", use "S" to choose specific eigenvalues (not yet supported in ARPACK)) + // + char howmny[2] = "A"; + + // if howmny == "S", specifies the eigenvalues to compute (not implemented in ARPACK) + // + int *select = new int[ncv]; + + // Final eigenvalues + // + m_eivalues.resize(nev, 1); + + internal::arpack_wrapper<Scalar, RealScalar>::seupd(&rvec, howmny, select, m_eivalues.data(), v, &ldv, + &sigma, bmat, &n, whch, &nev, &tol, resid, &ncv, + v, &ldv, iparam, ipntr, workd, workl, &lworkl, &info); + + if (info == -14) + m_info = NoConvergence; + else if (info != 0) + m_info = InvalidInput; + else + { + if (rvec) + { + m_eivec.resize(A.rows(), nev); + for (int i=0; i<nev; i++) + for (int j=0; j<n; j++) + m_eivec(j,i) = v[i*n+j] / scale; + + if (mode == 1 && !isBempty && BisSPD) + internal::OP<MatrixSolver, MatrixType, Scalar, BisSPD>::project(OP, n, nev, m_eivec.data()); + + m_eigenvectorsOk = true; + } + + m_nbrIterations = iparam[2]; + m_nbrConverged = iparam[4]; + + m_info = Success; + } + + delete select; + } + + delete v; + delete iparam; + delete ipntr; + delete workd; + delete workl; + delete resid; + + m_isInitialized = true; + + return *this; +} + + +// Single precision +// +extern "C" void ssaupd_(int *ido, char *bmat, int *n, char *which, + int *nev, float *tol, float *resid, int *ncv, + float *v, int *ldv, int *iparam, int *ipntr, + float *workd, float *workl, int *lworkl, + int *info); + +extern "C" void sseupd_(int *rvec, char *All, int *select, float *d, + float *z, int *ldz, float *sigma, + char *bmat, int *n, char *which, int *nev, + float *tol, float *resid, int *ncv, float *v, + int *ldv, int *iparam, int *ipntr, float *workd, + float *workl, int *lworkl, int *ierr); + +// Double precision +// +extern "C" void dsaupd_(int *ido, char *bmat, int *n, char *which, + int *nev, double *tol, double *resid, int *ncv, + double *v, int *ldv, int *iparam, int *ipntr, + double *workd, double *workl, int *lworkl, + int *info); + +extern "C" void dseupd_(int *rvec, char *All, int *select, double *d, + double *z, int *ldz, double *sigma, + char *bmat, int *n, char *which, int *nev, + double *tol, double *resid, int *ncv, double *v, + int *ldv, int *iparam, int *ipntr, double *workd, + double *workl, int *lworkl, int *ierr); + + +namespace internal { + +template<typename Scalar, typename RealScalar> struct arpack_wrapper +{ + static inline void saupd(int *ido, char *bmat, int *n, char *which, + int *nev, RealScalar *tol, Scalar *resid, int *ncv, + Scalar *v, int *ldv, int *iparam, int *ipntr, + Scalar *workd, Scalar *workl, int *lworkl, int *info) + { + EIGEN_STATIC_ASSERT(!NumTraits<Scalar>::IsComplex, NUMERIC_TYPE_MUST_BE_REAL) + } + + static inline void seupd(int *rvec, char *All, int *select, Scalar *d, + Scalar *z, int *ldz, RealScalar *sigma, + char *bmat, int *n, char *which, int *nev, + RealScalar *tol, Scalar *resid, int *ncv, Scalar *v, + int *ldv, int *iparam, int *ipntr, Scalar *workd, + Scalar *workl, int *lworkl, int *ierr) + { + EIGEN_STATIC_ASSERT(!NumTraits<Scalar>::IsComplex, NUMERIC_TYPE_MUST_BE_REAL) + } +}; + +template <> struct arpack_wrapper<float, float> +{ + static inline void saupd(int *ido, char *bmat, int *n, char *which, + int *nev, float *tol, float *resid, int *ncv, + float *v, int *ldv, int *iparam, int *ipntr, + float *workd, float *workl, int *lworkl, int *info) + { + ssaupd_(ido, bmat, n, which, nev, tol, resid, ncv, v, ldv, iparam, ipntr, workd, workl, lworkl, info); + } + + static inline void seupd(int *rvec, char *All, int *select, float *d, + float *z, int *ldz, float *sigma, + char *bmat, int *n, char *which, int *nev, + float *tol, float *resid, int *ncv, float *v, + int *ldv, int *iparam, int *ipntr, float *workd, + float *workl, int *lworkl, int *ierr) + { + sseupd_(rvec, All, select, d, z, ldz, sigma, bmat, n, which, nev, tol, resid, ncv, v, ldv, iparam, ipntr, + workd, workl, lworkl, ierr); + } +}; + +template <> struct arpack_wrapper<double, double> +{ + static inline void saupd(int *ido, char *bmat, int *n, char *which, + int *nev, double *tol, double *resid, int *ncv, + double *v, int *ldv, int *iparam, int *ipntr, + double *workd, double *workl, int *lworkl, int *info) + { + dsaupd_(ido, bmat, n, which, nev, tol, resid, ncv, v, ldv, iparam, ipntr, workd, workl, lworkl, info); + } + + static inline void seupd(int *rvec, char *All, int *select, double *d, + double *z, int *ldz, double *sigma, + char *bmat, int *n, char *which, int *nev, + double *tol, double *resid, int *ncv, double *v, + int *ldv, int *iparam, int *ipntr, double *workd, + double *workl, int *lworkl, int *ierr) + { + dseupd_(rvec, All, select, d, v, ldv, sigma, bmat, n, which, nev, tol, resid, ncv, v, ldv, iparam, ipntr, + workd, workl, lworkl, ierr); + } +}; + + +template<typename MatrixSolver, typename MatrixType, typename Scalar, bool BisSPD> +struct OP +{ + static inline void applyOP(MatrixSolver &OP, const MatrixType &A, int n, Scalar *in, Scalar *out); + static inline void project(MatrixSolver &OP, int n, int k, Scalar *vecs); +}; + +template<typename MatrixSolver, typename MatrixType, typename Scalar> +struct OP<MatrixSolver, MatrixType, Scalar, true> +{ + static inline void applyOP(MatrixSolver &OP, const MatrixType &A, int n, Scalar *in, Scalar *out) +{ + // OP = L^{-1} A L^{-T} (B = LL^T) + // + // First solve L^T out = in + // + Matrix<Scalar, Dynamic, 1>::Map(out, n) = OP.matrixU().solve(Matrix<Scalar, Dynamic, 1>::Map(in, n)); + Matrix<Scalar, Dynamic, 1>::Map(out, n) = OP.permutationPinv() * Matrix<Scalar, Dynamic, 1>::Map(out, n); + + // Then compute out = A out + // + Matrix<Scalar, Dynamic, 1>::Map(out, n) = A * Matrix<Scalar, Dynamic, 1>::Map(out, n); + + // Then solve L out = out + // + Matrix<Scalar, Dynamic, 1>::Map(out, n) = OP.permutationP() * Matrix<Scalar, Dynamic, 1>::Map(out, n); + Matrix<Scalar, Dynamic, 1>::Map(out, n) = OP.matrixL().solve(Matrix<Scalar, Dynamic, 1>::Map(out, n)); +} + + static inline void project(MatrixSolver &OP, int n, int k, Scalar *vecs) +{ + // Solve L^T out = in + // + Matrix<Scalar, Dynamic, Dynamic>::Map(vecs, n, k) = OP.matrixU().solve(Matrix<Scalar, Dynamic, Dynamic>::Map(vecs, n, k)); + Matrix<Scalar, Dynamic, Dynamic>::Map(vecs, n, k) = OP.permutationPinv() * Matrix<Scalar, Dynamic, Dynamic>::Map(vecs, n, k); +} + +}; + +template<typename MatrixSolver, typename MatrixType, typename Scalar> +struct OP<MatrixSolver, MatrixType, Scalar, false> +{ + static inline void applyOP(MatrixSolver &OP, const MatrixType &A, int n, Scalar *in, Scalar *out) +{ + eigen_assert(false && "Should never be in here..."); +} + + static inline void project(MatrixSolver &OP, int n, int k, Scalar *vecs) +{ + eigen_assert(false && "Should never be in here..."); +} + +}; + +} // end namespace internal + +} // end namespace Eigen + +#endif // EIGEN_ARPACKSELFADJOINTEIGENSOLVER_H + diff --git a/unsupported/Eigen/src/FFT/ei_kissfft_impl.h b/unsupported/Eigen/src/FFT/ei_kissfft_impl.h index 37f5af4c1..be51b4e6f 100644 --- a/unsupported/Eigen/src/FFT/ei_kissfft_impl.h +++ b/unsupported/Eigen/src/FFT/ei_kissfft_impl.h @@ -28,6 +28,7 @@ struct kiss_cpx_fft inline void make_twiddles(int nfft,bool inverse) { + using std::acos; m_inverse = inverse; m_twiddles.resize(nfft); Scalar phinc = (inverse?2:-2)* acos( (Scalar) -1) / nfft; @@ -399,6 +400,7 @@ struct kissfft_impl inline Complex * real_twiddles(int ncfft2) { + using std::acos; std::vector<Complex> & twidref = m_realTwiddles[ncfft2];// creates new if not there if ( (int)twidref.size() != ncfft2 ) { twidref.resize(ncfft2); diff --git a/unsupported/Eigen/src/IterativeSolvers/ConstrainedConjGrad.h b/unsupported/Eigen/src/IterativeSolvers/ConstrainedConjGrad.h index b83bf7aef..dc0093eb9 100644 --- a/unsupported/Eigen/src/IterativeSolvers/ConstrainedConjGrad.h +++ b/unsupported/Eigen/src/IterativeSolvers/ConstrainedConjGrad.h @@ -2,10 +2,6 @@ // for linear algebra. // // Copyright (C) 2008 Gael Guennebaud <gael.guennebaud@inria.fr> -// -// This Source Code Form is subject to the terms of the Mozilla -// Public License v. 2.0. If a copy of the MPL was not distributed -// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. /* NOTE The functions of this file have been adapted from the GMM++ library */ @@ -62,7 +58,9 @@ void pseudo_inverse(const CMatrix &C, CINVMatrix &CINV) Scalar rho, rho_1, alpha; d.setZero(); - CINV.startFill(); // FIXME estimate the number of non-zeros + typedef Triplet<double> T; + std::vector<T> tripletList; + for (Index i = 0; i < rows; ++i) { d[i] = 1.0; @@ -88,11 +86,12 @@ void pseudo_inverse(const CMatrix &C, CINVMatrix &CINV) // FIXME add a generic "prune/filter" expression for both dense and sparse object to sparse for (Index j=0; j<l.size(); ++j) if (l[j]<1e-15) - CINV.fill(i,j) = l[j]; + tripletList.push_back(T(i,j,l(j))); + d[i] = 0.0; } - CINV.endFill(); + CINV.setFromTriplets(tripletList.begin(), tripletList.end()); } @@ -107,6 +106,7 @@ template<typename TMatrix, typename CMatrix, void constrained_cg(const TMatrix& A, const CMatrix& C, VectorX& x, const VectorB& b, const VectorF& f, IterationController &iter) { + using std::sqrt; typedef typename TMatrix::Scalar Scalar; typedef typename TMatrix::Index Index; typedef Matrix<Scalar,Dynamic,1> TmpVec; diff --git a/unsupported/Eigen/src/IterativeSolvers/DGMRES.h b/unsupported/Eigen/src/IterativeSolvers/DGMRES.h new file mode 100644 index 000000000..9fcc8a8d9 --- /dev/null +++ b/unsupported/Eigen/src/IterativeSolvers/DGMRES.h @@ -0,0 +1,542 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2012 Désiré Nuentsa-Wakam <desire.nuentsa_wakam@inria.fr> +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + +#ifndef EIGEN_DGMRES_H +#define EIGEN_DGMRES_H + +#include <Eigen/Eigenvalues> + +namespace Eigen { + +template< typename _MatrixType, + typename _Preconditioner = DiagonalPreconditioner<typename _MatrixType::Scalar> > +class DGMRES; + +namespace internal { + +template< typename _MatrixType, typename _Preconditioner> +struct traits<DGMRES<_MatrixType,_Preconditioner> > +{ + typedef _MatrixType MatrixType; + typedef _Preconditioner Preconditioner; +}; + +/** \brief Computes a permutation vector to have a sorted sequence + * \param vec The vector to reorder. + * \param perm gives the sorted sequence on output. Must be initialized with 0..n-1 + * \param ncut Put the ncut smallest elements at the end of the vector + * WARNING This is an expensive sort, so should be used only + * for small size vectors + * TODO Use modified QuickSplit or std::nth_element to get the smallest values + */ +template <typename VectorType, typename IndexType> +void sortWithPermutation (VectorType& vec, IndexType& perm, typename IndexType::Scalar& ncut) +{ + eigen_assert(vec.size() == perm.size()); + typedef typename IndexType::Scalar Index; + typedef typename VectorType::Scalar Scalar; + bool flag; + for (Index k = 0; k < ncut; k++) + { + flag = false; + for (Index j = 0; j < vec.size()-1; j++) + { + if ( vec(perm(j)) < vec(perm(j+1)) ) + { + std::swap(perm(j),perm(j+1)); + flag = true; + } + if (!flag) break; // The vector is in sorted order + } + } +} + +} +/** + * \ingroup IterativeLInearSolvers_Module + * \brief A Restarted GMRES with deflation. + * This class implements a modification of the GMRES solver for + * sparse linear systems. The basis is built with modified + * Gram-Schmidt. At each restart, a few approximated eigenvectors + * corresponding to the smallest eigenvalues are used to build a + * preconditioner for the next cycle. This preconditioner + * for deflation can be combined with any other preconditioner, + * the IncompleteLUT for instance. The preconditioner is applied + * at right of the matrix and the combination is multiplicative. + * + * \tparam _MatrixType the type of the sparse matrix A, can be a dense or a sparse matrix. + * \tparam _Preconditioner the type of the preconditioner. Default is DiagonalPreconditioner + * Typical usage : + * \code + * SparseMatrix<double> A; + * VectorXd x, b; + * //Fill A and b ... + * DGMRES<SparseMatrix<double> > solver; + * solver.set_restart(30); // Set restarting value + * solver.setEigenv(1); // Set the number of eigenvalues to deflate + * solver.compute(A); + * x = solver.solve(b); + * \endcode + * + * References : + * [1] D. NUENTSA WAKAM and F. PACULL, Memory Efficient Hybrid + * Algebraic Solvers for Linear Systems Arising from Compressible + * Flows, Computers and Fluids, In Press, + * http://dx.doi.org/10.1016/j.compfluid.2012.03.023 + * [2] K. Burrage and J. Erhel, On the performance of various + * adaptive preconditioned GMRES strategies, 5(1998), 101-121. + * [3] J. Erhel, K. Burrage and B. Pohl, Restarted GMRES + * preconditioned by deflation,J. Computational and Applied + * Mathematics, 69(1996), 303-318. + + * + */ +template< typename _MatrixType, typename _Preconditioner> +class DGMRES : public IterativeSolverBase<DGMRES<_MatrixType,_Preconditioner> > +{ + typedef IterativeSolverBase<DGMRES> Base; + using Base::mp_matrix; + using Base::m_error; + using Base::m_iterations; + using Base::m_info; + using Base::m_isInitialized; + using Base::m_tolerance; + public: + typedef _MatrixType MatrixType; + typedef typename MatrixType::Scalar Scalar; + typedef typename MatrixType::Index Index; + typedef typename MatrixType::RealScalar RealScalar; + typedef _Preconditioner Preconditioner; + typedef Matrix<Scalar,Dynamic,Dynamic> DenseMatrix; + typedef Matrix<RealScalar,Dynamic,Dynamic> DenseRealMatrix; + typedef Matrix<Scalar,Dynamic,1> DenseVector; + typedef Matrix<RealScalar,Dynamic,1> DenseRealVector; + typedef Matrix<std::complex<RealScalar>, Dynamic, 1> ComplexVector; + + + /** Default constructor. */ + DGMRES() : Base(),m_restart(30),m_neig(0),m_r(0),m_maxNeig(5),m_isDeflAllocated(false),m_isDeflInitialized(false) {} + + /** Initialize the solver with matrix \a A for further \c Ax=b solving. + * + * This constructor is a shortcut for the default constructor followed + * by a call to compute(). + * + * \warning this class stores a reference to the matrix A as well as some + * precomputed values that depend on it. Therefore, if \a A is changed + * this class becomes invalid. Call compute() to update it with the new + * matrix A, or modify a copy of A. + */ + DGMRES(const MatrixType& A) : Base(A),m_restart(30),m_neig(0),m_r(0),m_maxNeig(5),m_isDeflAllocated(false),m_isDeflInitialized(false) + {} + + ~DGMRES() {} + + /** \returns the solution x of \f$ A x = b \f$ using the current decomposition of A + * \a x0 as an initial solution. + * + * \sa compute() + */ + template<typename Rhs,typename Guess> + inline const internal::solve_retval_with_guess<DGMRES, Rhs, Guess> + solveWithGuess(const MatrixBase<Rhs>& b, const Guess& x0) const + { + eigen_assert(m_isInitialized && "DGMRES is not initialized."); + eigen_assert(Base::rows()==b.rows() + && "DGMRES::solve(): invalid number of rows of the right hand side matrix b"); + return internal::solve_retval_with_guess + <DGMRES, Rhs, Guess>(*this, b.derived(), x0); + } + + /** \internal */ + template<typename Rhs,typename Dest> + void _solveWithGuess(const Rhs& b, Dest& x) const + { + bool failed = false; + for(int j=0; j<b.cols(); ++j) + { + m_iterations = Base::maxIterations(); + m_error = Base::m_tolerance; + + typename Dest::ColXpr xj(x,j); + dgmres(*mp_matrix, b.col(j), xj, Base::m_preconditioner); + } + m_info = failed ? NumericalIssue + : m_error <= Base::m_tolerance ? Success + : NoConvergence; + m_isInitialized = true; + } + + /** \internal */ + template<typename Rhs,typename Dest> + void _solve(const Rhs& b, Dest& x) const + { + x = b; + _solveWithGuess(b,x); + } + /** + * Get the restart value + */ + int restart() { return m_restart; } + + /** + * Set the restart value (default is 30) + */ + void set_restart(const int restart) { m_restart=restart; } + + /** + * Set the number of eigenvalues to deflate at each restart + */ + void setEigenv(const int neig) + { + m_neig = neig; + if (neig+1 > m_maxNeig) m_maxNeig = neig+1; // To allow for complex conjugates + } + + /** + * Get the size of the deflation subspace size + */ + int deflSize() {return m_r; } + + /** + * Set the maximum size of the deflation subspace + */ + void setMaxEigenv(const int maxNeig) { m_maxNeig = maxNeig; } + + protected: + // DGMRES algorithm + template<typename Rhs, typename Dest> + void dgmres(const MatrixType& mat,const Rhs& rhs, Dest& x, const Preconditioner& precond) const; + // Perform one cycle of GMRES + template<typename Dest> + int dgmresCycle(const MatrixType& mat, const Preconditioner& precond, Dest& x, DenseVector& r0, RealScalar& beta, const RealScalar& normRhs, int& nbIts) const; + // Compute data to use for deflation + int dgmresComputeDeflationData(const MatrixType& mat, const Preconditioner& precond, const Index& it, Index& neig) const; + // Apply deflation to a vector + template<typename RhsType, typename DestType> + int dgmresApplyDeflation(const RhsType& In, DestType& Out) const; + ComplexVector schurValues(const ComplexSchur<DenseMatrix>& schurofH) const; + ComplexVector schurValues(const RealSchur<DenseMatrix>& schurofH) const; + // Init data for deflation + void dgmresInitDeflation(Index& rows) const; + mutable DenseMatrix m_V; // Krylov basis vectors + mutable DenseMatrix m_H; // Hessenberg matrix + mutable DenseMatrix m_Hes; // Initial hessenberg matrix wihout Givens rotations applied + mutable Index m_restart; // Maximum size of the Krylov subspace + mutable DenseMatrix m_U; // Vectors that form the basis of the invariant subspace + mutable DenseMatrix m_MU; // matrix operator applied to m_U (for next cycles) + mutable DenseMatrix m_T; /* T=U^T*M^{-1}*A*U */ + mutable PartialPivLU<DenseMatrix> m_luT; // LU factorization of m_T + mutable int m_neig; //Number of eigenvalues to extract at each restart + mutable int m_r; // Current number of deflated eigenvalues, size of m_U + mutable int m_maxNeig; // Maximum number of eigenvalues to deflate + mutable RealScalar m_lambdaN; //Modulus of the largest eigenvalue of A + mutable bool m_isDeflAllocated; + mutable bool m_isDeflInitialized; + + //Adaptive strategy + mutable RealScalar m_smv; // Smaller multiple of the remaining number of steps allowed + mutable bool m_force; // Force the use of deflation at each restart + +}; +/** + * \brief Perform several cycles of restarted GMRES with modified Gram Schmidt, + * + * A right preconditioner is used combined with deflation. + * + */ +template< typename _MatrixType, typename _Preconditioner> +template<typename Rhs, typename Dest> +void DGMRES<_MatrixType, _Preconditioner>::dgmres(const MatrixType& mat,const Rhs& rhs, Dest& x, + const Preconditioner& precond) const +{ + //Initialization + int n = mat.rows(); + DenseVector r0(n); + int nbIts = 0; + m_H.resize(m_restart+1, m_restart); + m_Hes.resize(m_restart, m_restart); + m_V.resize(n,m_restart+1); + //Initial residual vector and intial norm + x = precond.solve(x); + r0 = rhs - mat * x; + RealScalar beta = r0.norm(); + RealScalar normRhs = rhs.norm(); + m_error = beta/normRhs; + if(m_error < m_tolerance) + m_info = Success; + else + m_info = NoConvergence; + + // Iterative process + while (nbIts < m_iterations && m_info == NoConvergence) + { + dgmresCycle(mat, precond, x, r0, beta, normRhs, nbIts); + + // Compute the new residual vector for the restart + if (nbIts < m_iterations && m_info == NoConvergence) + r0 = rhs - mat * x; + } +} + +/** + * \brief Perform one restart cycle of DGMRES + * \param mat The coefficient matrix + * \param precond The preconditioner + * \param x the new approximated solution + * \param r0 The initial residual vector + * \param beta The norm of the residual computed so far + * \param normRhs The norm of the right hand side vector + * \param nbIts The number of iterations + */ +template< typename _MatrixType, typename _Preconditioner> +template<typename Dest> +int DGMRES<_MatrixType, _Preconditioner>::dgmresCycle(const MatrixType& mat, const Preconditioner& precond, Dest& x, DenseVector& r0, RealScalar& beta, const RealScalar& normRhs, int& nbIts) const +{ + //Initialization + DenseVector g(m_restart+1); // Right hand side of the least square problem + g.setZero(); + g(0) = Scalar(beta); + m_V.col(0) = r0/beta; + m_info = NoConvergence; + std::vector<JacobiRotation<Scalar> >gr(m_restart); // Givens rotations + int it = 0; // Number of inner iterations + int n = mat.rows(); + DenseVector tv1(n), tv2(n); //Temporary vectors + while (m_info == NoConvergence && it < m_restart && nbIts < m_iterations) + { + // Apply preconditioner(s) at right + if (m_isDeflInitialized ) + { + dgmresApplyDeflation(m_V.col(it), tv1); // Deflation + tv2 = precond.solve(tv1); + } + else + { + tv2 = precond.solve(m_V.col(it)); // User's selected preconditioner + } + tv1 = mat * tv2; + + // Orthogonalize it with the previous basis in the basis using modified Gram-Schmidt + Scalar coef; + for (int i = 0; i <= it; ++i) + { + coef = tv1.dot(m_V.col(i)); + tv1 = tv1 - coef * m_V.col(i); + m_H(i,it) = coef; + m_Hes(i,it) = coef; + } + // Normalize the vector + coef = tv1.norm(); + m_V.col(it+1) = tv1/coef; + m_H(it+1, it) = coef; +// m_Hes(it+1,it) = coef; + + // FIXME Check for happy breakdown + + // Update Hessenberg matrix with Givens rotations + for (int i = 1; i <= it; ++i) + { + m_H.col(it).applyOnTheLeft(i-1,i,gr[i-1].adjoint()); + } + // Compute the new plane rotation + gr[it].makeGivens(m_H(it, it), m_H(it+1,it)); + // Apply the new rotation + m_H.col(it).applyOnTheLeft(it,it+1,gr[it].adjoint()); + g.applyOnTheLeft(it,it+1, gr[it].adjoint()); + + beta = std::abs(g(it+1)); + m_error = beta/normRhs; + std::cerr << nbIts << " Relative Residual Norm " << m_error << std::endl; + it++; nbIts++; + + if (m_error < m_tolerance) + { + // The method has converged + m_info = Success; + break; + } + } + + // Compute the new coefficients by solving the least square problem +// it++; + //FIXME Check first if the matrix is singular ... zero diagonal + DenseVector nrs(m_restart); + nrs = m_H.topLeftCorner(it,it).template triangularView<Upper>().solve(g.head(it)); + + // Form the new solution + if (m_isDeflInitialized) + { + tv1 = m_V.leftCols(it) * nrs; + dgmresApplyDeflation(tv1, tv2); + x = x + precond.solve(tv2); + } + else + x = x + precond.solve(m_V.leftCols(it) * nrs); + + // Go for a new cycle and compute data for deflation + if(nbIts < m_iterations && m_info == NoConvergence && m_neig > 0 && (m_r+m_neig) < m_maxNeig) + dgmresComputeDeflationData(mat, precond, it, m_neig); + return 0; + +} + + +template< typename _MatrixType, typename _Preconditioner> +void DGMRES<_MatrixType, _Preconditioner>::dgmresInitDeflation(Index& rows) const +{ + m_U.resize(rows, m_maxNeig); + m_MU.resize(rows, m_maxNeig); + m_T.resize(m_maxNeig, m_maxNeig); + m_lambdaN = 0.0; + m_isDeflAllocated = true; +} + +template< typename _MatrixType, typename _Preconditioner> +inline typename DGMRES<_MatrixType, _Preconditioner>::ComplexVector DGMRES<_MatrixType, _Preconditioner>::schurValues(const ComplexSchur<DenseMatrix>& schurofH) const +{ + return schurofH.matrixT().diagonal(); +} + +template< typename _MatrixType, typename _Preconditioner> +inline typename DGMRES<_MatrixType, _Preconditioner>::ComplexVector DGMRES<_MatrixType, _Preconditioner>::schurValues(const RealSchur<DenseMatrix>& schurofH) const +{ + typedef typename MatrixType::Index Index; + const DenseMatrix& T = schurofH.matrixT(); + Index it = T.rows(); + ComplexVector eig(it); + Index j = 0; + while (j < it-1) + { + if (T(j+1,j) ==Scalar(0)) + { + eig(j) = std::complex<RealScalar>(T(j,j),RealScalar(0)); + j++; + } + else + { + eig(j) = std::complex<RealScalar>(T(j,j),T(j+1,j)); + eig(j+1) = std::complex<RealScalar>(T(j,j+1),T(j+1,j+1)); + j++; + } + } + if (j < it-1) eig(j) = std::complex<RealScalar>(T(j,j),RealScalar(0)); + return eig; +} + +template< typename _MatrixType, typename _Preconditioner> +int DGMRES<_MatrixType, _Preconditioner>::dgmresComputeDeflationData(const MatrixType& mat, const Preconditioner& precond, const Index& it, Index& neig) const +{ + // First, find the Schur form of the Hessenberg matrix H + typename internal::conditional<NumTraits<Scalar>::IsComplex, ComplexSchur<DenseMatrix>, RealSchur<DenseMatrix> >::type schurofH; + bool computeU = true; + DenseMatrix matrixQ(it,it); + matrixQ.setIdentity(); + schurofH.computeFromHessenberg(m_Hes.topLeftCorner(it,it), matrixQ, computeU); + + ComplexVector eig(it); + Matrix<Index,Dynamic,1>perm(it); + eig = this->schurValues(schurofH); + + // Reorder the absolute values of Schur values + DenseRealVector modulEig(it); + for (int j=0; j<it; ++j) modulEig(j) = std::abs(eig(j)); + perm.setLinSpaced(it,0,it-1); + internal::sortWithPermutation(modulEig, perm, neig); + + if (!m_lambdaN) + { + m_lambdaN = (std::max)(modulEig.maxCoeff(), m_lambdaN); + } + //Count the real number of extracted eigenvalues (with complex conjugates) + int nbrEig = 0; + while (nbrEig < neig) + { + if(eig(perm(it-nbrEig-1)).imag() == RealScalar(0)) nbrEig++; + else nbrEig += 2; + } + // Extract the Schur vectors corresponding to the smallest Ritz values + DenseMatrix Sr(it, nbrEig); + Sr.setZero(); + for (int j = 0; j < nbrEig; j++) + { + Sr.col(j) = schurofH.matrixU().col(perm(it-j-1)); + } + + // Form the Schur vectors of the initial matrix using the Krylov basis + DenseMatrix X; + X = m_V.leftCols(it) * Sr; + if (m_r) + { + // Orthogonalize X against m_U using modified Gram-Schmidt + for (int j = 0; j < nbrEig; j++) + for (int k =0; k < m_r; k++) + X.col(j) = X.col(j) - (m_U.col(k).dot(X.col(j)))*m_U.col(k); + } + + // Compute m_MX = A * M^-1 * X + Index m = m_V.rows(); + if (!m_isDeflAllocated) + dgmresInitDeflation(m); + DenseMatrix MX(m, nbrEig); + DenseVector tv1(m); + for (int j = 0; j < nbrEig; j++) + { + tv1 = mat * X.col(j); + MX.col(j) = precond.solve(tv1); + } + + //Update m_T = [U'MU U'MX; X'MU X'MX] + m_T.block(m_r, m_r, nbrEig, nbrEig) = X.transpose() * MX; + if(m_r) + { + m_T.block(0, m_r, m_r, nbrEig) = m_U.leftCols(m_r).transpose() * MX; + m_T.block(m_r, 0, nbrEig, m_r) = X.transpose() * m_MU.leftCols(m_r); + } + + // Save X into m_U and m_MX in m_MU + for (int j = 0; j < nbrEig; j++) m_U.col(m_r+j) = X.col(j); + for (int j = 0; j < nbrEig; j++) m_MU.col(m_r+j) = MX.col(j); + // Increase the size of the invariant subspace + m_r += nbrEig; + + // Factorize m_T into m_luT + m_luT.compute(m_T.topLeftCorner(m_r, m_r)); + + //FIXME CHeck if the factorization was correctly done (nonsingular matrix) + m_isDeflInitialized = true; + return 0; +} +template<typename _MatrixType, typename _Preconditioner> +template<typename RhsType, typename DestType> +int DGMRES<_MatrixType, _Preconditioner>::dgmresApplyDeflation(const RhsType &x, DestType &y) const +{ + DenseVector x1 = m_U.leftCols(m_r).transpose() * x; + y = x + m_U.leftCols(m_r) * ( m_lambdaN * m_luT.solve(x1) - x1); + return 0; +} + +namespace internal { + + template<typename _MatrixType, typename _Preconditioner, typename Rhs> +struct solve_retval<DGMRES<_MatrixType, _Preconditioner>, Rhs> + : solve_retval_base<DGMRES<_MatrixType, _Preconditioner>, Rhs> +{ + typedef DGMRES<_MatrixType, _Preconditioner> Dec; + EIGEN_MAKE_SOLVE_HELPERS(Dec,Rhs) + + template<typename Dest> void evalTo(Dest& dst) const + { + dec()._solve(rhs(),dst); + } +}; +} // end namespace internal + +} // end namespace Eigen +#endif diff --git a/unsupported/Eigen/src/IterativeSolvers/GMRES.h b/unsupported/Eigen/src/IterativeSolvers/GMRES.h index 34e67db82..c8c84069e 100644 --- a/unsupported/Eigen/src/IterativeSolvers/GMRES.h +++ b/unsupported/Eigen/src/IterativeSolvers/GMRES.h @@ -2,7 +2,7 @@ // for linear algebra. // // Copyright (C) 2011 Gael Guennebaud <gael.guennebaud@inria.fr> -// Copyright (C) 2012 Kolja Brix <brix@igpm.rwth-aaachen.de> +// Copyright (C) 2012, 2014 Kolja Brix <brix@igpm.rwth-aaachen.de> // // This Source Code Form is subject to the terms of the Mozilla // Public License v. 2.0. If a copy of the MPL was not distributed @@ -61,7 +61,6 @@ bool gmres(const MatrixType & mat, const Rhs & rhs, Dest & x, const Precondition typedef typename Dest::RealScalar RealScalar; typedef typename Dest::Scalar Scalar; - typedef Matrix < RealScalar, Dynamic, 1 > RealVectorType; typedef Matrix < Scalar, Dynamic, 1 > VectorType; typedef Matrix < Scalar, Dynamic, Dynamic > FMatrixType; @@ -73,16 +72,20 @@ bool gmres(const MatrixType & mat, const Rhs & rhs, Dest & x, const Precondition VectorType p0 = rhs - mat*x; VectorType r0 = precond.solve(p0); -// RealScalar r0_sqnorm = r0.squaredNorm(); + + // is initial guess already good enough? + if(abs(r0.norm()) < tol) { + return true; + } VectorType w = VectorType::Zero(restart + 1); - FMatrixType H = FMatrixType::Zero(m, restart + 1); + FMatrixType H = FMatrixType::Zero(m, restart + 1); // Hessenberg matrix VectorType tau = VectorType::Zero(restart + 1); std::vector < JacobiRotation < Scalar > > G(restart); // generate first Householder vector - VectorType e; + VectorType e(m-1); RealScalar beta; r0.makeHouseholder(e, tau.coeffRef(0), beta); w(0)=(Scalar) beta; @@ -348,7 +351,8 @@ public: template<typename Rhs,typename Dest> void _solve(const Rhs& b, Dest& x) const { - x.setZero(); + x = b; + if(x.squaredNorm() == 0) return; // Check Zero right hand side _solveWithGuess(b,x); } diff --git a/unsupported/Eigen/src/IterativeSolvers/IncompleteCholesky.h b/unsupported/Eigen/src/IterativeSolvers/IncompleteCholesky.h new file mode 100644 index 000000000..661c1f2e0 --- /dev/null +++ b/unsupported/Eigen/src/IterativeSolvers/IncompleteCholesky.h @@ -0,0 +1,278 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2012 Désiré Nuentsa-Wakam <desire.nuentsa_wakam@inria.fr> +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + +#ifndef EIGEN_INCOMPLETE_CHOlESKY_H +#define EIGEN_INCOMPLETE_CHOlESKY_H +#include "Eigen/src/IterativeLinearSolvers/IncompleteLUT.h" +#include <Eigen/OrderingMethods> +#include <list> + +namespace Eigen { +/** + * \brief Modified Incomplete Cholesky with dual threshold + * + * References : C-J. Lin and J. J. Moré, Incomplete Cholesky Factorizations with + * Limited memory, SIAM J. Sci. Comput. 21(1), pp. 24-45, 1999 + * + * \tparam _MatrixType The type of the sparse matrix. It should be a symmetric + * matrix. It is advised to give a row-oriented sparse matrix + * \tparam _UpLo The triangular part of the matrix to reference. + * \tparam _OrderingType + */ + +template <typename Scalar, int _UpLo = Lower, typename _OrderingType = NaturalOrdering<int> > +class IncompleteCholesky : internal::noncopyable +{ + public: + typedef SparseMatrix<Scalar,ColMajor> MatrixType; + typedef _OrderingType OrderingType; + typedef typename MatrixType::RealScalar RealScalar; + typedef typename MatrixType::Index Index; + typedef PermutationMatrix<Dynamic, Dynamic, Index> PermutationType; + typedef Matrix<Scalar,Dynamic,1> ScalarType; + typedef Matrix<Index,Dynamic, 1> IndexType; + typedef std::vector<std::list<Index> > VectorList; + enum { UpLo = _UpLo }; + public: + IncompleteCholesky() : m_shift(1),m_factorizationIsOk(false) {} + IncompleteCholesky(const MatrixType& matrix) : m_shift(1),m_factorizationIsOk(false) + { + compute(matrix); + } + + Index rows() const { return m_L.rows(); } + + Index cols() const { return m_L.cols(); } + + + /** \brief Reports whether previous computation was successful. + * + * \returns \c Success if computation was succesful, + * \c NumericalIssue if the matrix appears to be negative. + */ + ComputationInfo info() const + { + eigen_assert(m_isInitialized && "IncompleteLLT is not initialized."); + return m_info; + } + + /** + * \brief Set the initial shift parameter + */ + void setShift( Scalar shift) { m_shift = shift; } + + /** + * \brief Computes the fill reducing permutation vector. + */ + template<typename MatrixType> + void analyzePattern(const MatrixType& mat) + { + OrderingType ord; + ord(mat.template selfadjointView<UpLo>(), m_perm); + m_analysisIsOk = true; + } + + template<typename MatrixType> + void factorize(const MatrixType& amat); + + template<typename MatrixType> + void compute (const MatrixType& matrix) + { + analyzePattern(matrix); + factorize(matrix); + } + + template<typename Rhs, typename Dest> + void _solve(const Rhs& b, Dest& x) const + { + eigen_assert(m_factorizationIsOk && "factorize() should be called first"); + if (m_perm.rows() == b.rows()) + x = m_perm.inverse() * b; + else + x = b; + x = m_scal.asDiagonal() * x; + x = m_L.template triangularView<UnitLower>().solve(x); + x = m_L.adjoint().template triangularView<Upper>().solve(x); + if (m_perm.rows() == b.rows()) + x = m_perm * x; + x = m_scal.asDiagonal() * x; + } + template<typename Rhs> inline const internal::solve_retval<IncompleteCholesky, Rhs> + solve(const MatrixBase<Rhs>& b) const + { + eigen_assert(m_factorizationIsOk && "IncompleteLLT did not succeed"); + eigen_assert(m_isInitialized && "IncompleteLLT is not initialized."); + eigen_assert(cols()==b.rows() + && "IncompleteLLT::solve(): invalid number of rows of the right hand side matrix b"); + return internal::solve_retval<IncompleteCholesky, Rhs>(*this, b.derived()); + } + protected: + SparseMatrix<Scalar,ColMajor> m_L; // The lower part stored in CSC + ScalarType m_scal; // The vector for scaling the matrix + Scalar m_shift; //The initial shift parameter + bool m_analysisIsOk; + bool m_factorizationIsOk; + bool m_isInitialized; + ComputationInfo m_info; + PermutationType m_perm; + + private: + template <typename IdxType, typename SclType> + inline void updateList(const IdxType& colPtr, IdxType& rowIdx, SclType& vals, const Index& col, const Index& jk, IndexType& firstElt, VectorList& listCol); +}; + +template<typename Scalar, int _UpLo, typename OrderingType> +template<typename _MatrixType> +void IncompleteCholesky<Scalar,_UpLo, OrderingType>::factorize(const _MatrixType& mat) +{ + using std::sqrt; + using std::min; + eigen_assert(m_analysisIsOk && "analyzePattern() should be called first"); + + // Dropping strategies : Keep only the p largest elements per column, where p is the number of elements in the column of the original matrix. Other strategies will be added + + // Apply the fill-reducing permutation computed in analyzePattern() + if (m_perm.rows() == mat.rows() ) // To detect the null permutation + m_L.template selfadjointView<Lower>() = mat.template selfadjointView<_UpLo>().twistedBy(m_perm); + else + m_L.template selfadjointView<Lower>() = mat.template selfadjointView<_UpLo>(); + + Index n = m_L.cols(); + Index nnz = m_L.nonZeros(); + Map<ScalarType> vals(m_L.valuePtr(), nnz); //values + Map<IndexType> rowIdx(m_L.innerIndexPtr(), nnz); //Row indices + Map<IndexType> colPtr( m_L.outerIndexPtr(), n+1); // Pointer to the beginning of each row + IndexType firstElt(n-1); // for each j, points to the next entry in vals that will be used in the factorization + VectorList listCol(n); // listCol(j) is a linked list of columns to update column j + ScalarType curCol(n); // Store a nonzero values in each column + IndexType irow(n); // Row indices of nonzero elements in each column + + + // Computes the scaling factors + m_scal.resize(n); + for (int j = 0; j < n; j++) + { + m_scal(j) = m_L.col(j).norm(); + m_scal(j) = sqrt(m_scal(j)); + } + // Scale and compute the shift for the matrix + Scalar mindiag = vals[0]; + for (int j = 0; j < n; j++){ + for (int k = colPtr[j]; k < colPtr[j+1]; k++) + vals[k] /= (m_scal(j) * m_scal(rowIdx[k])); + mindiag = (min)(vals[colPtr[j]], mindiag); + } + + if(mindiag < Scalar(0.)) m_shift = m_shift - mindiag; + // Apply the shift to the diagonal elements of the matrix + for (int j = 0; j < n; j++) + vals[colPtr[j]] += m_shift; + // jki version of the Cholesky factorization + for (int j=0; j < n; ++j) + { + //Left-looking factorize the column j + // First, load the jth column into curCol + Scalar diag = vals[colPtr[j]]; // It is assumed that only the lower part is stored + curCol.setZero(); + irow.setLinSpaced(n,0,n-1); + for (int i = colPtr[j] + 1; i < colPtr[j+1]; i++) + { + curCol(rowIdx[i]) = vals[i]; + irow(rowIdx[i]) = rowIdx[i]; + } + std::list<int>::iterator k; + // Browse all previous columns that will update column j + for(k = listCol[j].begin(); k != listCol[j].end(); k++) + { + int jk = firstElt(*k); // First element to use in the column + jk += 1; + for (int i = jk; i < colPtr[*k+1]; i++) + { + curCol(rowIdx[i]) -= vals[i] * vals[jk] ; + } + updateList(colPtr,rowIdx,vals, *k, jk, firstElt, listCol); + } + + // Scale the current column + if(RealScalar(diag) <= 0) + { + std::cerr << "\nNegative diagonal during Incomplete factorization... "<< j << "\n"; + m_info = NumericalIssue; + return; + } + RealScalar rdiag = sqrt(RealScalar(diag)); + vals[colPtr[j]] = rdiag; + for (int i = j+1; i < n; i++) + { + //Scale + curCol(i) /= rdiag; + //Update the remaining diagonals with curCol + vals[colPtr[i]] -= curCol(i) * curCol(i); + } + // Select the largest p elements + // p is the original number of elements in the column (without the diagonal) + int p = colPtr[j+1] - colPtr[j] - 1 ; + internal::QuickSplit(curCol, irow, p); + // Insert the largest p elements in the matrix + int cpt = 0; + for (int i = colPtr[j]+1; i < colPtr[j+1]; i++) + { + vals[i] = curCol(cpt); + rowIdx[i] = irow(cpt); + cpt ++; + } + // Get the first smallest row index and put it after the diagonal element + Index jk = colPtr(j)+1; + updateList(colPtr,rowIdx,vals,j,jk,firstElt,listCol); + } + m_factorizationIsOk = true; + m_isInitialized = true; + m_info = Success; +} + +template<typename Scalar, int _UpLo, typename OrderingType> +template <typename IdxType, typename SclType> +inline void IncompleteCholesky<Scalar,_UpLo, OrderingType>::updateList(const IdxType& colPtr, IdxType& rowIdx, SclType& vals, const Index& col, const Index& jk, IndexType& firstElt, VectorList& listCol) +{ + if (jk < colPtr(col+1) ) + { + Index p = colPtr(col+1) - jk; + Index minpos; + rowIdx.segment(jk,p).minCoeff(&minpos); + minpos += jk; + if (rowIdx(minpos) != rowIdx(jk)) + { + //Swap + std::swap(rowIdx(jk),rowIdx(minpos)); + std::swap(vals(jk),vals(minpos)); + } + firstElt(col) = jk; + listCol[rowIdx(jk)].push_back(col); + } +} +namespace internal { + +template<typename _Scalar, int _UpLo, typename OrderingType, typename Rhs> +struct solve_retval<IncompleteCholesky<_Scalar, _UpLo, OrderingType>, Rhs> + : solve_retval_base<IncompleteCholesky<_Scalar, _UpLo, OrderingType>, Rhs> +{ + typedef IncompleteCholesky<_Scalar, _UpLo, OrderingType> Dec; + EIGEN_MAKE_SOLVE_HELPERS(Dec,Rhs) + + template<typename Dest> void evalTo(Dest& dst) const + { + dec()._solve(rhs(),dst); + } +}; + +} // end namespace internal + +} // end namespace Eigen + +#endif diff --git a/unsupported/Eigen/src/IterativeSolvers/IterationController.h b/unsupported/Eigen/src/IterativeSolvers/IterationController.h index aaf46d544..c9c1a4be2 100644 --- a/unsupported/Eigen/src/IterativeSolvers/IterationController.h +++ b/unsupported/Eigen/src/IterativeSolvers/IterationController.h @@ -2,10 +2,6 @@ // for linear algebra. // // Copyright (C) 2008-2009 Gael Guennebaud <gael.guennebaud@inria.fr> -// -// This Source Code Form is subject to the terms of the Mozilla -// Public License v. 2.0. If a copy of the MPL was not distributed -// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. /* NOTE The class IterationController has been adapted from the iteration * class of the GMM++ and ITL libraries. @@ -129,7 +125,8 @@ class IterationController bool converged() const { return m_res <= m_rhsn * m_resmax; } bool converged(double nr) { - m_res = internal::abs(nr); + using std::abs; + m_res = abs(nr); m_resminreach = (std::min)(m_resminreach, m_res); return converged(); } diff --git a/unsupported/Eigen/src/IterativeSolvers/MINRES.h b/unsupported/Eigen/src/IterativeSolvers/MINRES.h new file mode 100644 index 000000000..0e56342a8 --- /dev/null +++ b/unsupported/Eigen/src/IterativeSolvers/MINRES.h @@ -0,0 +1,302 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2012 Giacomo Po <gpo@ucla.edu> +// Copyright (C) 2011 Gael Guennebaud <gael.guennebaud@inria.fr> +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + + +#ifndef EIGEN_MINRES_H_ +#define EIGEN_MINRES_H_ + + +namespace Eigen { + + namespace internal { + + /** \internal Low-level MINRES algorithm + * \param mat The matrix A + * \param rhs The right hand side vector b + * \param x On input and initial solution, on output the computed solution. + * \param precond A right preconditioner being able to efficiently solve for an + * approximation of Ax=b (regardless of b) + * \param iters On input the max number of iteration, on output the number of performed iterations. + * \param tol_error On input the tolerance error, on output an estimation of the relative error. + */ + template<typename MatrixType, typename Rhs, typename Dest, typename Preconditioner> + EIGEN_DONT_INLINE + void minres(const MatrixType& mat, const Rhs& rhs, Dest& x, + const Preconditioner& precond, int& iters, + typename Dest::RealScalar& tol_error) + { + using std::sqrt; + typedef typename Dest::RealScalar RealScalar; + typedef typename Dest::Scalar Scalar; + typedef Matrix<Scalar,Dynamic,1> VectorType; + + // initialize + const int maxIters(iters); // initialize maxIters to iters + const int N(mat.cols()); // the size of the matrix + const RealScalar rhsNorm2(rhs.squaredNorm()); + const RealScalar threshold2(tol_error*tol_error*rhsNorm2); // convergence threshold (compared to residualNorm2) + + // Initialize preconditioned Lanczos +// VectorType v_old(N); // will be initialized inside loop + VectorType v( VectorType::Zero(N) ); //initialize v + VectorType v_new(rhs-mat*x); //initialize v_new + RealScalar residualNorm2(v_new.squaredNorm()); +// VectorType w(N); // will be initialized inside loop + VectorType w_new(precond.solve(v_new)); // initialize w_new +// RealScalar beta; // will be initialized inside loop + RealScalar beta_new2(v_new.dot(w_new)); + eigen_assert(beta_new2 >= 0 && "PRECONDITIONER IS NOT POSITIVE DEFINITE"); + RealScalar beta_new(sqrt(beta_new2)); + const RealScalar beta_one(beta_new); + v_new /= beta_new; + w_new /= beta_new; + // Initialize other variables + RealScalar c(1.0); // the cosine of the Givens rotation + RealScalar c_old(1.0); + RealScalar s(0.0); // the sine of the Givens rotation + RealScalar s_old(0.0); // the sine of the Givens rotation +// VectorType p_oold(N); // will be initialized in loop + VectorType p_old(VectorType::Zero(N)); // initialize p_old=0 + VectorType p(p_old); // initialize p=0 + RealScalar eta(1.0); + + iters = 0; // reset iters + while ( iters < maxIters ){ + + // Preconditioned Lanczos + /* Note that there are 4 variants on the Lanczos algorithm. These are + * described in Paige, C. C. (1972). Computational variants of + * the Lanczos method for the eigenproblem. IMA Journal of Applied + * Mathematics, 10(3), 373–381. The current implementation corresponds + * to the case A(2,7) in the paper. It also corresponds to + * algorithm 6.14 in Y. Saad, Iterative Methods for Sparse Linear + * Systems, 2003 p.173. For the preconditioned version see + * A. Greenbaum, Iterative Methods for Solving Linear Systems, SIAM (1987). + */ + const RealScalar beta(beta_new); +// v_old = v; // update: at first time step, this makes v_old = 0 so value of beta doesn't matter + const VectorType v_old(v); // NOT SURE IF CREATING v_old EVERY ITERATION IS EFFICIENT + v = v_new; // update +// w = w_new; // update + const VectorType w(w_new); // NOT SURE IF CREATING w EVERY ITERATION IS EFFICIENT + v_new.noalias() = mat*w - beta*v_old; // compute v_new + const RealScalar alpha = v_new.dot(w); + v_new -= alpha*v; // overwrite v_new + w_new = precond.solve(v_new); // overwrite w_new + beta_new2 = v_new.dot(w_new); // compute beta_new + eigen_assert(beta_new2 >= 0 && "PRECONDITIONER IS NOT POSITIVE DEFINITE"); + beta_new = sqrt(beta_new2); // compute beta_new + v_new /= beta_new; // overwrite v_new for next iteration + w_new /= beta_new; // overwrite w_new for next iteration + + // Givens rotation + const RealScalar r2 =s*alpha+c*c_old*beta; // s, s_old, c and c_old are still from previous iteration + const RealScalar r3 =s_old*beta; // s, s_old, c and c_old are still from previous iteration + const RealScalar r1_hat=c*alpha-c_old*s*beta; + const RealScalar r1 =sqrt( std::pow(r1_hat,2) + std::pow(beta_new,2) ); + c_old = c; // store for next iteration + s_old = s; // store for next iteration + c=r1_hat/r1; // new cosine + s=beta_new/r1; // new sine + + // Update solution +// p_oold = p_old; + const VectorType p_oold(p_old); // NOT SURE IF CREATING p_oold EVERY ITERATION IS EFFICIENT + p_old = p; + p.noalias()=(w-r2*p_old-r3*p_oold) /r1; // IS NOALIAS REQUIRED? + x += beta_one*c*eta*p; + residualNorm2 *= s*s; + + if ( residualNorm2 < threshold2){ + break; + } + + eta=-s*eta; // update eta + iters++; // increment iteration number (for output purposes) + } + tol_error = std::sqrt(residualNorm2 / rhsNorm2); // return error. Note that this is the estimated error. The real error |Ax-b|/|b| may be slightly larger + } + + } + + template< typename _MatrixType, int _UpLo=Lower, + typename _Preconditioner = IdentityPreconditioner> +// typename _Preconditioner = IdentityPreconditioner<typename _MatrixType::Scalar> > // preconditioner must be positive definite + class MINRES; + + namespace internal { + + template< typename _MatrixType, int _UpLo, typename _Preconditioner> + struct traits<MINRES<_MatrixType,_UpLo,_Preconditioner> > + { + typedef _MatrixType MatrixType; + typedef _Preconditioner Preconditioner; + }; + + } + + /** \ingroup IterativeLinearSolvers_Module + * \brief A minimal residual solver for sparse symmetric problems + * + * This class allows to solve for A.x = b sparse linear problems using the MINRES algorithm + * of Paige and Saunders (1975). The sparse matrix A must be symmetric (possibly indefinite). + * The vectors x and b can be either dense or sparse. + * + * \tparam _MatrixType the type of the sparse matrix A, can be a dense or a sparse matrix. + * \tparam _UpLo the triangular part that will be used for the computations. It can be Lower + * or Upper. Default is Lower. + * \tparam _Preconditioner the type of the preconditioner. Default is DiagonalPreconditioner + * + * The maximal number of iterations and tolerance value can be controlled via the setMaxIterations() + * and setTolerance() methods. The defaults are the size of the problem for the maximal number of iterations + * and NumTraits<Scalar>::epsilon() for the tolerance. + * + * This class can be used as the direct solver classes. Here is a typical usage example: + * \code + * int n = 10000; + * VectorXd x(n), b(n); + * SparseMatrix<double> A(n,n); + * // fill A and b + * MINRES<SparseMatrix<double> > mr; + * mr.compute(A); + * x = mr.solve(b); + * std::cout << "#iterations: " << mr.iterations() << std::endl; + * std::cout << "estimated error: " << mr.error() << std::endl; + * // update b, and solve again + * x = mr.solve(b); + * \endcode + * + * By default the iterations start with x=0 as an initial guess of the solution. + * One can control the start using the solveWithGuess() method. Here is a step by + * step execution example starting with a random guess and printing the evolution + * of the estimated error: + * * \code + * x = VectorXd::Random(n); + * mr.setMaxIterations(1); + * int i = 0; + * do { + * x = mr.solveWithGuess(b,x); + * std::cout << i << " : " << mr.error() << std::endl; + * ++i; + * } while (mr.info()!=Success && i<100); + * \endcode + * Note that such a step by step excution is slightly slower. + * + * \sa class ConjugateGradient, BiCGSTAB, SimplicialCholesky, DiagonalPreconditioner, IdentityPreconditioner + */ + template< typename _MatrixType, int _UpLo, typename _Preconditioner> + class MINRES : public IterativeSolverBase<MINRES<_MatrixType,_UpLo,_Preconditioner> > + { + + typedef IterativeSolverBase<MINRES> Base; + using Base::mp_matrix; + using Base::m_error; + using Base::m_iterations; + using Base::m_info; + using Base::m_isInitialized; + public: + typedef _MatrixType MatrixType; + typedef typename MatrixType::Scalar Scalar; + typedef typename MatrixType::Index Index; + typedef typename MatrixType::RealScalar RealScalar; + typedef _Preconditioner Preconditioner; + + enum {UpLo = _UpLo}; + + public: + + /** Default constructor. */ + MINRES() : Base() {} + + /** Initialize the solver with matrix \a A for further \c Ax=b solving. + * + * This constructor is a shortcut for the default constructor followed + * by a call to compute(). + * + * \warning this class stores a reference to the matrix A as well as some + * precomputed values that depend on it. Therefore, if \a A is changed + * this class becomes invalid. Call compute() to update it with the new + * matrix A, or modify a copy of A. + */ + MINRES(const MatrixType& A) : Base(A) {} + + /** Destructor. */ + ~MINRES(){} + + /** \returns the solution x of \f$ A x = b \f$ using the current decomposition of A + * \a x0 as an initial solution. + * + * \sa compute() + */ + template<typename Rhs,typename Guess> + inline const internal::solve_retval_with_guess<MINRES, Rhs, Guess> + solveWithGuess(const MatrixBase<Rhs>& b, const Guess& x0) const + { + eigen_assert(m_isInitialized && "MINRES is not initialized."); + eigen_assert(Base::rows()==b.rows() + && "MINRES::solve(): invalid number of rows of the right hand side matrix b"); + return internal::solve_retval_with_guess + <MINRES, Rhs, Guess>(*this, b.derived(), x0); + } + + /** \internal */ + template<typename Rhs,typename Dest> + void _solveWithGuess(const Rhs& b, Dest& x) const + { + m_iterations = Base::maxIterations(); + m_error = Base::m_tolerance; + + for(int j=0; j<b.cols(); ++j) + { + m_iterations = Base::maxIterations(); + m_error = Base::m_tolerance; + + typename Dest::ColXpr xj(x,j); + internal::minres(mp_matrix->template selfadjointView<UpLo>(), b.col(j), xj, + Base::m_preconditioner, m_iterations, m_error); + } + + m_isInitialized = true; + m_info = m_error <= Base::m_tolerance ? Success : NoConvergence; + } + + /** \internal */ + template<typename Rhs,typename Dest> + void _solve(const Rhs& b, Dest& x) const + { + x.setZero(); + _solveWithGuess(b,x); + } + + protected: + + }; + + namespace internal { + + template<typename _MatrixType, int _UpLo, typename _Preconditioner, typename Rhs> + struct solve_retval<MINRES<_MatrixType,_UpLo,_Preconditioner>, Rhs> + : solve_retval_base<MINRES<_MatrixType,_UpLo,_Preconditioner>, Rhs> + { + typedef MINRES<_MatrixType,_UpLo,_Preconditioner> Dec; + EIGEN_MAKE_SOLVE_HELPERS(Dec,Rhs) + + template<typename Dest> void evalTo(Dest& dst) const + { + dec()._solve(rhs(),dst); + } + }; + + } // end namespace internal + +} // end namespace Eigen + +#endif // EIGEN_MINRES_H + diff --git a/unsupported/Eigen/src/IterativeSolvers/Scaling.h b/unsupported/Eigen/src/IterativeSolvers/Scaling.h index fdef0aca3..4fd439202 100644 --- a/unsupported/Eigen/src/IterativeSolvers/Scaling.h +++ b/unsupported/Eigen/src/IterativeSolvers/Scaling.h @@ -7,8 +7,8 @@ // Public License v. 2.0. If a copy of the MPL was not distributed // with this file, You can obtain one at http://mozilla.org/MPL/2.0/. -#ifndef EIGEN_SCALING_H -#define EIGEN_SCALING_H +#ifndef EIGEN_ITERSCALING_H +#define EIGEN_ITERSCALING_H /** * \ingroup IterativeSolvers_Module * \brief iterative scaling algorithm to equilibrate rows and column norms in matrices @@ -24,7 +24,7 @@ * VectorXd x(n), b(n); * SparseMatrix<double> A; * // fill A and b; - * Scaling<SparseMatrix<double> > scal; + * IterScaling<SparseMatrix<double> > scal; * // Compute the left and right scaling vectors. The matrix is equilibrated at output * scal.computeRef(A); * // Scale the right hand side @@ -41,10 +41,10 @@ * * \sa \ref IncompleteLUT */ +namespace Eigen { using std::abs; -using namespace Eigen; template<typename _MatrixType> -class Scaling +class IterScaling { public: typedef _MatrixType MatrixType; @@ -52,15 +52,15 @@ class Scaling typedef typename MatrixType::Index Index; public: - Scaling() { init(); } + IterScaling() { init(); } - Scaling(const MatrixType& matrix) + IterScaling(const MatrixType& matrix) { init(); compute(matrix); } - ~Scaling() { } + ~IterScaling() { } /** * Compute the left and right diagonal matrices to scale the input matrix @p mat @@ -73,7 +73,7 @@ class Scaling { int m = mat.rows(); int n = mat.cols(); - assert((m>0 && m == n) && "Please give a non - empty matrix"); + eigen_assert((m>0 && m == n) && "Please give a non - empty matrix"); m_left.resize(m); m_right.resize(n); m_left.setOnes(); @@ -181,5 +181,5 @@ class Scaling double m_tol; int m_maxits; // Maximum number of iterations allowed }; - -#endif
\ No newline at end of file +} +#endif diff --git a/unsupported/Eigen/src/KroneckerProduct/KroneckerTensorProduct.h b/unsupported/Eigen/src/KroneckerProduct/KroneckerTensorProduct.h index 84fd72fc6..532896c3b 100644 --- a/unsupported/Eigen/src/KroneckerProduct/KroneckerTensorProduct.h +++ b/unsupported/Eigen/src/KroneckerProduct/KroneckerTensorProduct.h @@ -3,153 +3,240 @@ // // Copyright (C) 2011 Kolja Brix <brix@igpm.rwth-aachen.de> // Copyright (C) 2011 Andreas Platen <andiplaten@gmx.de> +// Copyright (C) 2012 Chen-Pang He <jdh8@ms63.hinet.net> // // This Source Code Form is subject to the terms of the Mozilla // Public License v. 2.0. If a copy of the MPL was not distributed // with this file, You can obtain one at http://mozilla.org/MPL/2.0/. - #ifndef KRONECKER_TENSOR_PRODUCT_H #define KRONECKER_TENSOR_PRODUCT_H - namespace Eigen { -namespace internal { +template<typename Scalar, int Options, typename Index> class SparseMatrix; /*! - * Kronecker tensor product helper function for dense matrices + * \brief Kronecker tensor product helper class for dense matrices * - * \param A Dense matrix A - * \param B Dense matrix B - * \param AB_ Kronecker tensor product of A and B + * This class is the return value of kroneckerProduct(MatrixBase, + * MatrixBase). Use the function rather than construct this class + * directly to avoid specifying template prarameters. + * + * \tparam Lhs Type of the left-hand side, a matrix expression. + * \tparam Rhs Type of the rignt-hand side, a matrix expression. */ -template<typename Derived_A, typename Derived_B, typename Derived_AB> -void kroneckerProduct_full(const Derived_A& A, const Derived_B& B, Derived_AB & AB) +template<typename Lhs, typename Rhs> +class KroneckerProduct : public ReturnByValue<KroneckerProduct<Lhs,Rhs> > { - const unsigned int Ar = A.rows(), - Ac = A.cols(), - Br = B.rows(), - Bc = B.cols(); - for (unsigned int i=0; i<Ar; ++i) - for (unsigned int j=0; j<Ac; ++j) - AB.block(i*Br,j*Bc,Br,Bc) = A(i,j)*B; -} + private: + typedef ReturnByValue<KroneckerProduct> Base; + typedef typename Base::Scalar Scalar; + typedef typename Base::Index Index; + + public: + /*! \brief Constructor. */ + KroneckerProduct(const Lhs& A, const Rhs& B) + : m_A(A), m_B(B) + {} + + /*! \brief Evaluate the Kronecker tensor product. */ + template<typename Dest> void evalTo(Dest& dst) const; + + inline Index rows() const { return m_A.rows() * m_B.rows(); } + inline Index cols() const { return m_A.cols() * m_B.cols(); } + + Scalar coeff(Index row, Index col) const + { + return m_A.coeff(row / m_B.rows(), col / m_B.cols()) * + m_B.coeff(row % m_B.rows(), col % m_B.cols()); + } + + Scalar coeff(Index i) const + { + EIGEN_STATIC_ASSERT_VECTOR_ONLY(KroneckerProduct); + return m_A.coeff(i / m_A.size()) * m_B.coeff(i % m_A.size()); + } + private: + typename Lhs::Nested m_A; + typename Rhs::Nested m_B; +}; /*! - * Kronecker tensor product helper function for matrices, where at least one is sparse + * \brief Kronecker tensor product helper class for sparse matrices + * + * If at least one of the operands is a sparse matrix expression, + * then this class is returned and evaluates into a sparse matrix. + * + * This class is the return value of kroneckerProduct(EigenBase, + * EigenBase). Use the function rather than construct this class + * directly to avoid specifying template prarameters. * - * \param A Matrix A - * \param B Matrix B - * \param AB_ Kronecker tensor product of A and B + * \tparam Lhs Type of the left-hand side, a matrix expression. + * \tparam Rhs Type of the rignt-hand side, a matrix expression. */ -template<typename Derived_A, typename Derived_B, typename Derived_AB> -void kroneckerProduct_sparse(const Derived_A &A, const Derived_B &B, Derived_AB &AB) +template<typename Lhs, typename Rhs> +class KroneckerProductSparse : public EigenBase<KroneckerProductSparse<Lhs,Rhs> > { - const unsigned int Ar = A.rows(), - Ac = A.cols(), - Br = B.rows(), - Bc = B.cols(); - AB.resize(Ar*Br,Ac*Bc); - AB.resizeNonZeros(0); - AB.reserve(A.nonZeros()*B.nonZeros()); - - for (int kA=0; kA<A.outerSize(); ++kA) + private: + typedef typename internal::traits<KroneckerProductSparse>::Index Index; + + public: + /*! \brief Constructor. */ + KroneckerProductSparse(const Lhs& A, const Rhs& B) + : m_A(A), m_B(B) + {} + + /*! \brief Evaluate the Kronecker tensor product. */ + template<typename Dest> void evalTo(Dest& dst) const; + + inline Index rows() const { return m_A.rows() * m_B.rows(); } + inline Index cols() const { return m_A.cols() * m_B.cols(); } + + template<typename Scalar, int Options, typename Index> + operator SparseMatrix<Scalar, Options, Index>() + { + SparseMatrix<Scalar, Options, Index> result; + evalTo(result.derived()); + return result; + } + + private: + typename Lhs::Nested m_A; + typename Rhs::Nested m_B; +}; + +template<typename Lhs, typename Rhs> +template<typename Dest> +void KroneckerProduct<Lhs,Rhs>::evalTo(Dest& dst) const +{ + const int BlockRows = Rhs::RowsAtCompileTime, + BlockCols = Rhs::ColsAtCompileTime; + const Index Br = m_B.rows(), + Bc = m_B.cols(); + for (Index i=0; i < m_A.rows(); ++i) + for (Index j=0; j < m_A.cols(); ++j) + Block<Dest,BlockRows,BlockCols>(dst,i*Br,j*Bc,Br,Bc) = m_A.coeff(i,j) * m_B; +} + +template<typename Lhs, typename Rhs> +template<typename Dest> +void KroneckerProductSparse<Lhs,Rhs>::evalTo(Dest& dst) const +{ + const Index Br = m_B.rows(), + Bc = m_B.cols(); + dst.resize(rows(),cols()); + dst.resizeNonZeros(0); + dst.reserve(m_A.nonZeros() * m_B.nonZeros()); + + for (Index kA=0; kA < m_A.outerSize(); ++kA) { - for (int kB=0; kB<B.outerSize(); ++kB) + for (Index kB=0; kB < m_B.outerSize(); ++kB) { - for (typename Derived_A::InnerIterator itA(A,kA); itA; ++itA) + for (typename Lhs::InnerIterator itA(m_A,kA); itA; ++itA) { - for (typename Derived_B::InnerIterator itB(B,kB); itB; ++itB) + for (typename Rhs::InnerIterator itB(m_B,kB); itB; ++itB) { - const unsigned int iA = itA.row(), - jA = itA.col(), - iB = itB.row(), - jB = itB.col(), - i = iA*Br + iB, - j = jA*Bc + jB; - AB.insert(i,j) = itA.value() * itB.value(); + const Index i = itA.row() * Br + itB.row(), + j = itA.col() * Bc + itB.col(); + dst.insert(i,j) = itA.value() * itB.value(); } } } } } -} // end namespace internal +namespace internal { +template<typename _Lhs, typename _Rhs> +struct traits<KroneckerProduct<_Lhs,_Rhs> > +{ + typedef typename remove_all<_Lhs>::type Lhs; + typedef typename remove_all<_Rhs>::type Rhs; + typedef typename scalar_product_traits<typename Lhs::Scalar, typename Rhs::Scalar>::ReturnType Scalar; + + enum { + Rows = size_at_compile_time<traits<Lhs>::RowsAtCompileTime, traits<Rhs>::RowsAtCompileTime>::ret, + Cols = size_at_compile_time<traits<Lhs>::ColsAtCompileTime, traits<Rhs>::ColsAtCompileTime>::ret, + MaxRows = size_at_compile_time<traits<Lhs>::MaxRowsAtCompileTime, traits<Rhs>::MaxRowsAtCompileTime>::ret, + MaxCols = size_at_compile_time<traits<Lhs>::MaxColsAtCompileTime, traits<Rhs>::MaxColsAtCompileTime>::ret, + CoeffReadCost = Lhs::CoeffReadCost + Rhs::CoeffReadCost + NumTraits<Scalar>::MulCost + }; + + typedef Matrix<Scalar,Rows,Cols> ReturnType; +}; + +template<typename _Lhs, typename _Rhs> +struct traits<KroneckerProductSparse<_Lhs,_Rhs> > +{ + typedef MatrixXpr XprKind; + typedef typename remove_all<_Lhs>::type Lhs; + typedef typename remove_all<_Rhs>::type Rhs; + typedef typename scalar_product_traits<typename Lhs::Scalar, typename Rhs::Scalar>::ReturnType Scalar; + typedef typename promote_storage_type<typename traits<Lhs>::StorageKind, typename traits<Rhs>::StorageKind>::ret StorageKind; + typedef typename promote_index_type<typename Lhs::Index, typename Rhs::Index>::type Index; + + enum { + LhsFlags = Lhs::Flags, + RhsFlags = Rhs::Flags, + + RowsAtCompileTime = size_at_compile_time<traits<Lhs>::RowsAtCompileTime, traits<Rhs>::RowsAtCompileTime>::ret, + ColsAtCompileTime = size_at_compile_time<traits<Lhs>::ColsAtCompileTime, traits<Rhs>::ColsAtCompileTime>::ret, + MaxRowsAtCompileTime = size_at_compile_time<traits<Lhs>::MaxRowsAtCompileTime, traits<Rhs>::MaxRowsAtCompileTime>::ret, + MaxColsAtCompileTime = size_at_compile_time<traits<Lhs>::MaxColsAtCompileTime, traits<Rhs>::MaxColsAtCompileTime>::ret, + + EvalToRowMajor = (LhsFlags & RhsFlags & RowMajorBit), + RemovedBits = ~(EvalToRowMajor ? 0 : RowMajorBit), + + Flags = ((LhsFlags | RhsFlags) & HereditaryBits & RemovedBits) + | EvalBeforeNestingBit | EvalBeforeAssigningBit, + CoeffReadCost = Dynamic + }; +}; +} // end namespace internal /*! - * Computes Kronecker tensor product of two dense matrices + * \ingroup KroneckerProduct_Module * - * \param a Dense matrix a - * \param b Dense matrix b - * \param c Kronecker tensor product of a and b - */ -template<typename A,typename B,typename CScalar,int CRows,int CCols, int COptions, int CMaxRows, int CMaxCols> -void kroneckerProduct(const MatrixBase<A>& a, const MatrixBase<B>& b, Matrix<CScalar,CRows,CCols,COptions,CMaxRows,CMaxCols>& c) -{ - c.resize(a.rows()*b.rows(),a.cols()*b.cols()); - internal::kroneckerProduct_full(a.derived(), b.derived(), c); -} - -/*! * Computes Kronecker tensor product of two dense matrices * - * Remark: this function uses the const cast hack and has been - * implemented to make the function call possible, where the - * output matrix is a submatrix, e.g. - * kroneckerProduct(A,B,AB.block(2,5,6,6)); + * \warning If you want to replace a matrix by its Kronecker product + * with some matrix, do \b NOT do this: + * \code + * A = kroneckerProduct(A,B); // bug!!! caused by aliasing effect + * \endcode + * instead, use eval() to work around this: + * \code + * A = kroneckerProduct(A,B).eval(); + * \endcode * * \param a Dense matrix a * \param b Dense matrix b - * \param c Kronecker tensor product of a and b + * \return Kronecker tensor product of a and b */ -template<typename A,typename B,typename C> -void kroneckerProduct(const MatrixBase<A>& a, const MatrixBase<B>& b, MatrixBase<C> const & c_) +template<typename A, typename B> +KroneckerProduct<A,B> kroneckerProduct(const MatrixBase<A>& a, const MatrixBase<B>& b) { - MatrixBase<C>& c = const_cast<MatrixBase<C>& >(c_); - internal::kroneckerProduct_full(a.derived(), b.derived(), c.derived()); + return KroneckerProduct<A, B>(a.derived(), b.derived()); } /*! - * Computes Kronecker tensor product of a dense and a sparse matrix + * \ingroup KroneckerProduct_Module * - * \param a Dense matrix a - * \param b Sparse matrix b - * \param c Kronecker tensor product of a and b - */ -template<typename A,typename B,typename C> -void kroneckerProduct(const MatrixBase<A>& a, const SparseMatrixBase<B>& b, SparseMatrixBase<C>& c) -{ - internal::kroneckerProduct_sparse(a.derived(), b.derived(), c.derived()); -} - -/*! - * Computes Kronecker tensor product of a sparse and a dense matrix - * - * \param a Sparse matrix a - * \param b Dense matrix b - * \param c Kronecker tensor product of a and b - */ -template<typename A,typename B,typename C> -void kroneckerProduct(const SparseMatrixBase<A>& a, const MatrixBase<B>& b, SparseMatrixBase<C>& c) -{ - internal::kroneckerProduct_sparse(a.derived(), b.derived(), c.derived()); -} - -/*! - * Computes Kronecker tensor product of two sparse matrices + * Computes Kronecker tensor product of two matrices, at least one of + * which is sparse * - * \param a Sparse matrix a - * \param b Sparse matrix b - * \param c Kronecker tensor product of a and b + * \param a Dense/sparse matrix a + * \param b Dense/sparse matrix b + * \return Kronecker tensor product of a and b, stored in a sparse + * matrix */ -template<typename A,typename B,typename C> -void kroneckerProduct(const SparseMatrixBase<A>& a, const SparseMatrixBase<B>& b, SparseMatrixBase<C>& c) +template<typename A, typename B> +KroneckerProductSparse<A,B> kroneckerProduct(const EigenBase<A>& a, const EigenBase<B>& b) { - internal::kroneckerProduct_sparse(a.derived(), b.derived(), c.derived()); + return KroneckerProductSparse<A,B>(a.derived(), b.derived()); } } // end namespace Eigen diff --git a/unsupported/Eigen/src/LevenbergMarquardt/CMakeLists.txt b/unsupported/Eigen/src/LevenbergMarquardt/CMakeLists.txt new file mode 100644 index 000000000..8513803ce --- /dev/null +++ b/unsupported/Eigen/src/LevenbergMarquardt/CMakeLists.txt @@ -0,0 +1,6 @@ +FILE(GLOB Eigen_LevenbergMarquardt_SRCS "*.h") + +INSTALL(FILES + ${Eigen_LevenbergMarquardt_SRCS} + DESTINATION ${INCLUDE_INSTALL_DIR}/Eigen/src/LevenbergMarquardt COMPONENT Devel + ) diff --git a/unsupported/Eigen/src/LevenbergMarquardt/CopyrightMINPACK.txt b/unsupported/Eigen/src/LevenbergMarquardt/CopyrightMINPACK.txt new file mode 100644 index 000000000..ae7984dae --- /dev/null +++ b/unsupported/Eigen/src/LevenbergMarquardt/CopyrightMINPACK.txt @@ -0,0 +1,52 @@ +Minpack Copyright Notice (1999) University of Chicago. All rights reserved + +Redistribution and use in source and binary forms, with or +without modification, are permitted provided that the +following conditions are met: + +1. Redistributions of source code must retain the above +copyright notice, this list of conditions and the following +disclaimer. + +2. Redistributions in binary form must reproduce the above +copyright notice, this list of conditions and the following +disclaimer in the documentation and/or other materials +provided with the distribution. + +3. The end-user documentation included with the +redistribution, if any, must include the following +acknowledgment: + + "This product includes software developed by the + University of Chicago, as Operator of Argonne National + Laboratory. + +Alternately, this acknowledgment may appear in the software +itself, if and wherever such third-party acknowledgments +normally appear. + +4. WARRANTY DISCLAIMER. THE SOFTWARE IS SUPPLIED "AS IS" +WITHOUT WARRANTY OF ANY KIND. THE COPYRIGHT HOLDER, THE +UNITED STATES, THE UNITED STATES DEPARTMENT OF ENERGY, AND +THEIR EMPLOYEES: (1) DISCLAIM ANY WARRANTIES, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO ANY IMPLIED WARRANTIES +OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, TITLE +OR NON-INFRINGEMENT, (2) DO NOT ASSUME ANY LEGAL LIABILITY +OR RESPONSIBILITY FOR THE ACCURACY, COMPLETENESS, OR +USEFULNESS OF THE SOFTWARE, (3) DO NOT REPRESENT THAT USE OF +THE SOFTWARE WOULD NOT INFRINGE PRIVATELY OWNED RIGHTS, (4) +DO NOT WARRANT THAT THE SOFTWARE WILL FUNCTION +UNINTERRUPTED, THAT IT IS ERROR-FREE OR THAT ANY ERRORS WILL +BE CORRECTED. + +5. LIMITATION OF LIABILITY. IN NO EVENT WILL THE COPYRIGHT +HOLDER, THE UNITED STATES, THE UNITED STATES DEPARTMENT OF +ENERGY, OR THEIR EMPLOYEES: BE LIABLE FOR ANY INDIRECT, +INCIDENTAL, CONSEQUENTIAL, SPECIAL OR PUNITIVE DAMAGES OF +ANY KIND OR NATURE, INCLUDING BUT NOT LIMITED TO LOSS OF +PROFITS OR LOSS OF DATA, FOR ANY REASON WHATSOEVER, WHETHER +SUCH LIABILITY IS ASSERTED ON THE BASIS OF CONTRACT, TORT +(INCLUDING NEGLIGENCE OR STRICT LIABILITY), OR OTHERWISE, +EVEN IF ANY OF SAID PARTIES HAS BEEN WARNED OF THE +POSSIBILITY OF SUCH LOSS OR DAMAGES. + diff --git a/unsupported/Eigen/src/LevenbergMarquardt/LMcovar.h b/unsupported/Eigen/src/LevenbergMarquardt/LMcovar.h new file mode 100644 index 000000000..32d3ad518 --- /dev/null +++ b/unsupported/Eigen/src/LevenbergMarquardt/LMcovar.h @@ -0,0 +1,85 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// This code initially comes from MINPACK whose original authors are: +// Copyright Jorge More - Argonne National Laboratory +// Copyright Burt Garbow - Argonne National Laboratory +// Copyright Ken Hillstrom - Argonne National Laboratory +// +// This Source Code Form is subject to the terms of the Minpack license +// (a BSD-like license) described in the campaigned CopyrightMINPACK.txt file. + +#ifndef EIGEN_LMCOVAR_H +#define EIGEN_LMCOVAR_H + +namespace Eigen { + +namespace internal { + +template <typename Scalar> +void covar( + Matrix< Scalar, Dynamic, Dynamic > &r, + const VectorXi& ipvt, + Scalar tol = std::sqrt(NumTraits<Scalar>::epsilon()) ) +{ + using std::abs; + typedef DenseIndex Index; + /* Local variables */ + Index i, j, k, l, ii, jj; + bool sing; + Scalar temp; + + /* Function Body */ + const Index n = r.cols(); + const Scalar tolr = tol * abs(r(0,0)); + Matrix< Scalar, Dynamic, 1 > wa(n); + eigen_assert(ipvt.size()==n); + + /* form the inverse of r in the full upper triangle of r. */ + l = -1; + for (k = 0; k < n; ++k) + if (abs(r(k,k)) > tolr) { + r(k,k) = 1. / r(k,k); + for (j = 0; j <= k-1; ++j) { + temp = r(k,k) * r(j,k); + r(j,k) = 0.; + r.col(k).head(j+1) -= r.col(j).head(j+1) * temp; + } + l = k; + } + + /* form the full upper triangle of the inverse of (r transpose)*r */ + /* in the full upper triangle of r. */ + for (k = 0; k <= l; ++k) { + for (j = 0; j <= k-1; ++j) + r.col(j).head(j+1) += r.col(k).head(j+1) * r(j,k); + r.col(k).head(k+1) *= r(k,k); + } + + /* form the full lower triangle of the covariance matrix */ + /* in the strict lower triangle of r and in wa. */ + for (j = 0; j < n; ++j) { + jj = ipvt[j]; + sing = j > l; + for (i = 0; i <= j; ++i) { + if (sing) + r(i,j) = 0.; + ii = ipvt[i]; + if (ii > jj) + r(ii,jj) = r(i,j); + if (ii < jj) + r(jj,ii) = r(i,j); + } + wa[jj] = r(j,j); + } + + /* symmetrize the covariance matrix in r. */ + r.topLeftCorner(n,n).template triangularView<StrictlyUpper>() = r.topLeftCorner(n,n).transpose(); + r.diagonal() = wa; +} + +} // end namespace internal + +} // end namespace Eigen + +#endif // EIGEN_LMCOVAR_H diff --git a/unsupported/Eigen/src/LevenbergMarquardt/LMonestep.h b/unsupported/Eigen/src/LevenbergMarquardt/LMonestep.h new file mode 100644 index 000000000..25b32ec5b --- /dev/null +++ b/unsupported/Eigen/src/LevenbergMarquardt/LMonestep.h @@ -0,0 +1,202 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2009 Thomas Capricelli <orzel@freehackers.org> +// +// This code initially comes from MINPACK whose original authors are: +// Copyright Jorge More - Argonne National Laboratory +// Copyright Burt Garbow - Argonne National Laboratory +// Copyright Ken Hillstrom - Argonne National Laboratory +// +// This Source Code Form is subject to the terms of the Minpack license +// (a BSD-like license) described in the campaigned CopyrightMINPACK.txt file. + +#ifndef EIGEN_LMONESTEP_H +#define EIGEN_LMONESTEP_H + +namespace Eigen { + +template<typename FunctorType> +LevenbergMarquardtSpace::Status +LevenbergMarquardt<FunctorType>::minimizeOneStep(FVectorType &x) +{ + using std::abs; + using std::sqrt; + RealScalar temp, temp1,temp2; + RealScalar ratio; + RealScalar pnorm, xnorm, fnorm1, actred, dirder, prered; + eigen_assert(x.size()==n); // check the caller is not cheating us + + temp = 0.0; xnorm = 0.0; + /* calculate the jacobian matrix. */ + Index df_ret = m_functor.df(x, m_fjac); + if (df_ret<0) + return LevenbergMarquardtSpace::UserAsked; + if (df_ret>0) + // numerical diff, we evaluated the function df_ret times + m_nfev += df_ret; + else m_njev++; + + /* compute the qr factorization of the jacobian. */ + for (int j = 0; j < x.size(); ++j) + m_wa2(j) = m_fjac.col(j).blueNorm(); + QRSolver qrfac(m_fjac); + if(qrfac.info() != Success) { + m_info = NumericalIssue; + return LevenbergMarquardtSpace::ImproperInputParameters; + } + // Make a copy of the first factor with the associated permutation + m_rfactor = qrfac.matrixR(); + m_permutation = (qrfac.colsPermutation()); + + /* on the first iteration and if external scaling is not used, scale according */ + /* to the norms of the columns of the initial jacobian. */ + if (m_iter == 1) { + if (!m_useExternalScaling) + for (Index j = 0; j < n; ++j) + m_diag[j] = (m_wa2[j]==0.)? 1. : m_wa2[j]; + + /* on the first iteration, calculate the norm of the scaled x */ + /* and initialize the step bound m_delta. */ + xnorm = m_diag.cwiseProduct(x).stableNorm(); + m_delta = m_factor * xnorm; + if (m_delta == 0.) + m_delta = m_factor; + } + + /* form (q transpose)*m_fvec and store the first n components in */ + /* m_qtf. */ + m_wa4 = m_fvec; + m_wa4 = qrfac.matrixQ().adjoint() * m_fvec; + m_qtf = m_wa4.head(n); + + /* compute the norm of the scaled gradient. */ + m_gnorm = 0.; + if (m_fnorm != 0.) + for (Index j = 0; j < n; ++j) + if (m_wa2[m_permutation.indices()[j]] != 0.) + m_gnorm = (std::max)(m_gnorm, abs( m_rfactor.col(j).head(j+1).dot(m_qtf.head(j+1)/m_fnorm) / m_wa2[m_permutation.indices()[j]])); + + /* test for convergence of the gradient norm. */ + if (m_gnorm <= m_gtol) { + m_info = Success; + return LevenbergMarquardtSpace::CosinusTooSmall; + } + + /* rescale if necessary. */ + if (!m_useExternalScaling) + m_diag = m_diag.cwiseMax(m_wa2); + + do { + /* determine the levenberg-marquardt parameter. */ + internal::lmpar2(qrfac, m_diag, m_qtf, m_delta, m_par, m_wa1); + + /* store the direction p and x + p. calculate the norm of p. */ + m_wa1 = -m_wa1; + m_wa2 = x + m_wa1; + pnorm = m_diag.cwiseProduct(m_wa1).stableNorm(); + + /* on the first iteration, adjust the initial step bound. */ + if (m_iter == 1) + m_delta = (std::min)(m_delta,pnorm); + + /* evaluate the function at x + p and calculate its norm. */ + if ( m_functor(m_wa2, m_wa4) < 0) + return LevenbergMarquardtSpace::UserAsked; + ++m_nfev; + fnorm1 = m_wa4.stableNorm(); + + /* compute the scaled actual reduction. */ + actred = -1.; + if (Scalar(.1) * fnorm1 < m_fnorm) + actred = 1. - numext::abs2(fnorm1 / m_fnorm); + + /* compute the scaled predicted reduction and */ + /* the scaled directional derivative. */ + m_wa3 = m_rfactor.template triangularView<Upper>() * (m_permutation.inverse() *m_wa1); + temp1 = numext::abs2(m_wa3.stableNorm() / m_fnorm); + temp2 = numext::abs2(sqrt(m_par) * pnorm / m_fnorm); + prered = temp1 + temp2 / Scalar(.5); + dirder = -(temp1 + temp2); + + /* compute the ratio of the actual to the predicted */ + /* reduction. */ + ratio = 0.; + if (prered != 0.) + ratio = actred / prered; + + /* update the step bound. */ + if (ratio <= Scalar(.25)) { + if (actred >= 0.) + temp = RealScalar(.5); + if (actred < 0.) + temp = RealScalar(.5) * dirder / (dirder + RealScalar(.5) * actred); + if (RealScalar(.1) * fnorm1 >= m_fnorm || temp < RealScalar(.1)) + temp = Scalar(.1); + /* Computing MIN */ + m_delta = temp * (std::min)(m_delta, pnorm / RealScalar(.1)); + m_par /= temp; + } else if (!(m_par != 0. && ratio < RealScalar(.75))) { + m_delta = pnorm / RealScalar(.5); + m_par = RealScalar(.5) * m_par; + } + + /* test for successful iteration. */ + if (ratio >= RealScalar(1e-4)) { + /* successful iteration. update x, m_fvec, and their norms. */ + x = m_wa2; + m_wa2 = m_diag.cwiseProduct(x); + m_fvec = m_wa4; + xnorm = m_wa2.stableNorm(); + m_fnorm = fnorm1; + ++m_iter; + } + + /* tests for convergence. */ + if (abs(actred) <= m_ftol && prered <= m_ftol && Scalar(.5) * ratio <= 1. && m_delta <= m_xtol * xnorm) + { + m_info = Success; + return LevenbergMarquardtSpace::RelativeErrorAndReductionTooSmall; + } + if (abs(actred) <= m_ftol && prered <= m_ftol && Scalar(.5) * ratio <= 1.) + { + m_info = Success; + return LevenbergMarquardtSpace::RelativeReductionTooSmall; + } + if (m_delta <= m_xtol * xnorm) + { + m_info = Success; + return LevenbergMarquardtSpace::RelativeErrorTooSmall; + } + + /* tests for termination and stringent tolerances. */ + if (m_nfev >= m_maxfev) + { + m_info = NoConvergence; + return LevenbergMarquardtSpace::TooManyFunctionEvaluation; + } + if (abs(actred) <= NumTraits<Scalar>::epsilon() && prered <= NumTraits<Scalar>::epsilon() && Scalar(.5) * ratio <= 1.) + { + m_info = Success; + return LevenbergMarquardtSpace::FtolTooSmall; + } + if (m_delta <= NumTraits<Scalar>::epsilon() * xnorm) + { + m_info = Success; + return LevenbergMarquardtSpace::XtolTooSmall; + } + if (m_gnorm <= NumTraits<Scalar>::epsilon()) + { + m_info = Success; + return LevenbergMarquardtSpace::GtolTooSmall; + } + + } while (ratio < Scalar(1e-4)); + + return LevenbergMarquardtSpace::Running; +} + + +} // end namespace Eigen + +#endif // EIGEN_LMONESTEP_H diff --git a/unsupported/Eigen/src/LevenbergMarquardt/LMpar.h b/unsupported/Eigen/src/LevenbergMarquardt/LMpar.h new file mode 100644 index 000000000..9532042d9 --- /dev/null +++ b/unsupported/Eigen/src/LevenbergMarquardt/LMpar.h @@ -0,0 +1,160 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// This code initially comes from MINPACK whose original authors are: +// Copyright Jorge More - Argonne National Laboratory +// Copyright Burt Garbow - Argonne National Laboratory +// Copyright Ken Hillstrom - Argonne National Laboratory +// +// This Source Code Form is subject to the terms of the Minpack license +// (a BSD-like license) described in the campaigned CopyrightMINPACK.txt file. + +#ifndef EIGEN_LMPAR_H +#define EIGEN_LMPAR_H + +namespace Eigen { + +namespace internal { + + template <typename QRSolver, typename VectorType> + void lmpar2( + const QRSolver &qr, + const VectorType &diag, + const VectorType &qtb, + typename VectorType::Scalar m_delta, + typename VectorType::Scalar &par, + VectorType &x) + + { + using std::sqrt; + using std::abs; + typedef typename QRSolver::MatrixType MatrixType; + typedef typename QRSolver::Scalar Scalar; + typedef typename QRSolver::Index Index; + + /* Local variables */ + Index j; + Scalar fp; + Scalar parc, parl; + Index iter; + Scalar temp, paru; + Scalar gnorm; + Scalar dxnorm; + + // Make a copy of the triangular factor. + // This copy is modified during call the qrsolv + MatrixType s; + s = qr.matrixR(); + + /* Function Body */ + const Scalar dwarf = (std::numeric_limits<Scalar>::min)(); + const Index n = qr.matrixR().cols(); + eigen_assert(n==diag.size()); + eigen_assert(n==qtb.size()); + + VectorType wa1, wa2; + + /* compute and store in x the gauss-newton direction. if the */ + /* jacobian is rank-deficient, obtain a least squares solution. */ + + // const Index rank = qr.nonzeroPivots(); // exactly double(0.) + const Index rank = qr.rank(); // use a threshold + wa1 = qtb; + wa1.tail(n-rank).setZero(); + //FIXME There is no solve in place for sparse triangularView + wa1.head(rank) = s.topLeftCorner(rank,rank).template triangularView<Upper>().solve(qtb.head(rank)); + + x = qr.colsPermutation()*wa1; + + /* initialize the iteration counter. */ + /* evaluate the function at the origin, and test */ + /* for acceptance of the gauss-newton direction. */ + iter = 0; + wa2 = diag.cwiseProduct(x); + dxnorm = wa2.blueNorm(); + fp = dxnorm - m_delta; + if (fp <= Scalar(0.1) * m_delta) { + par = 0; + return; + } + + /* if the jacobian is not rank deficient, the newton */ + /* step provides a lower bound, parl, for the zero of */ + /* the function. otherwise set this bound to zero. */ + parl = 0.; + if (rank==n) { + wa1 = qr.colsPermutation().inverse() * diag.cwiseProduct(wa2)/dxnorm; + s.topLeftCorner(n,n).transpose().template triangularView<Lower>().solveInPlace(wa1); + temp = wa1.blueNorm(); + parl = fp / m_delta / temp / temp; + } + + /* calculate an upper bound, paru, for the zero of the function. */ + for (j = 0; j < n; ++j) + wa1[j] = s.col(j).head(j+1).dot(qtb.head(j+1)) / diag[qr.colsPermutation().indices()(j)]; + + gnorm = wa1.stableNorm(); + paru = gnorm / m_delta; + if (paru == 0.) + paru = dwarf / (std::min)(m_delta,Scalar(0.1)); + + /* if the input par lies outside of the interval (parl,paru), */ + /* set par to the closer endpoint. */ + par = (std::max)(par,parl); + par = (std::min)(par,paru); + if (par == 0.) + par = gnorm / dxnorm; + + /* beginning of an iteration. */ + while (true) { + ++iter; + + /* evaluate the function at the current value of par. */ + if (par == 0.) + par = (std::max)(dwarf,Scalar(.001) * paru); /* Computing MAX */ + wa1 = sqrt(par)* diag; + + VectorType sdiag(n); + lmqrsolv(s, qr.colsPermutation(), wa1, qtb, x, sdiag); + + wa2 = diag.cwiseProduct(x); + dxnorm = wa2.blueNorm(); + temp = fp; + fp = dxnorm - m_delta; + + /* if the function is small enough, accept the current value */ + /* of par. also test for the exceptional cases where parl */ + /* is zero or the number of iterations has reached 10. */ + if (abs(fp) <= Scalar(0.1) * m_delta || (parl == 0. && fp <= temp && temp < 0.) || iter == 10) + break; + + /* compute the newton correction. */ + wa1 = qr.colsPermutation().inverse() * diag.cwiseProduct(wa2/dxnorm); + // we could almost use this here, but the diagonal is outside qr, in sdiag[] + for (j = 0; j < n; ++j) { + wa1[j] /= sdiag[j]; + temp = wa1[j]; + for (Index i = j+1; i < n; ++i) + wa1[i] -= s.coeff(i,j) * temp; + } + temp = wa1.blueNorm(); + parc = fp / m_delta / temp / temp; + + /* depending on the sign of the function, update parl or paru. */ + if (fp > 0.) + parl = (std::max)(parl,par); + if (fp < 0.) + paru = (std::min)(paru,par); + + /* compute an improved estimate for par. */ + par = (std::max)(parl,par+parc); + } + if (iter == 0) + par = 0.; + return; + } +} // end namespace internal + +} // end namespace Eigen + +#endif // EIGEN_LMPAR_H diff --git a/unsupported/Eigen/src/LevenbergMarquardt/LMqrsolv.h b/unsupported/Eigen/src/LevenbergMarquardt/LMqrsolv.h new file mode 100644 index 000000000..f5290dee4 --- /dev/null +++ b/unsupported/Eigen/src/LevenbergMarquardt/LMqrsolv.h @@ -0,0 +1,189 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2009 Thomas Capricelli <orzel@freehackers.org> +// Copyright (C) 2012 Desire Nuentsa <desire.nuentsa_wakam@inria.fr> +// +// This code initially comes from MINPACK whose original authors are: +// Copyright Jorge More - Argonne National Laboratory +// Copyright Burt Garbow - Argonne National Laboratory +// Copyright Ken Hillstrom - Argonne National Laboratory +// +// This Source Code Form is subject to the terms of the Minpack license +// (a BSD-like license) described in the campaigned CopyrightMINPACK.txt file. + +#ifndef EIGEN_LMQRSOLV_H +#define EIGEN_LMQRSOLV_H + +namespace Eigen { + +namespace internal { + +template <typename Scalar,int Rows, int Cols, typename Index> +void lmqrsolv( + Matrix<Scalar,Rows,Cols> &s, + const PermutationMatrix<Dynamic,Dynamic,Index> &iPerm, + const Matrix<Scalar,Dynamic,1> &diag, + const Matrix<Scalar,Dynamic,1> &qtb, + Matrix<Scalar,Dynamic,1> &x, + Matrix<Scalar,Dynamic,1> &sdiag) +{ + + /* Local variables */ + Index i, j, k, l; + Scalar temp; + Index n = s.cols(); + Matrix<Scalar,Dynamic,1> wa(n); + JacobiRotation<Scalar> givens; + + /* Function Body */ + // the following will only change the lower triangular part of s, including + // the diagonal, though the diagonal is restored afterward + + /* copy r and (q transpose)*b to preserve input and initialize s. */ + /* in particular, save the diagonal elements of r in x. */ + x = s.diagonal(); + wa = qtb; + + + s.topLeftCorner(n,n).template triangularView<StrictlyLower>() = s.topLeftCorner(n,n).transpose(); + /* eliminate the diagonal matrix d using a givens rotation. */ + for (j = 0; j < n; ++j) { + + /* prepare the row of d to be eliminated, locating the */ + /* diagonal element using p from the qr factorization. */ + l = iPerm.indices()(j); + if (diag[l] == 0.) + break; + sdiag.tail(n-j).setZero(); + sdiag[j] = diag[l]; + + /* the transformations to eliminate the row of d */ + /* modify only a single element of (q transpose)*b */ + /* beyond the first n, which is initially zero. */ + Scalar qtbpj = 0.; + for (k = j; k < n; ++k) { + /* determine a givens rotation which eliminates the */ + /* appropriate element in the current row of d. */ + givens.makeGivens(-s(k,k), sdiag[k]); + + /* compute the modified diagonal element of r and */ + /* the modified element of ((q transpose)*b,0). */ + s(k,k) = givens.c() * s(k,k) + givens.s() * sdiag[k]; + temp = givens.c() * wa[k] + givens.s() * qtbpj; + qtbpj = -givens.s() * wa[k] + givens.c() * qtbpj; + wa[k] = temp; + + /* accumulate the tranformation in the row of s. */ + for (i = k+1; i<n; ++i) { + temp = givens.c() * s(i,k) + givens.s() * sdiag[i]; + sdiag[i] = -givens.s() * s(i,k) + givens.c() * sdiag[i]; + s(i,k) = temp; + } + } + } + + /* solve the triangular system for z. if the system is */ + /* singular, then obtain a least squares solution. */ + Index nsing; + for(nsing=0; nsing<n && sdiag[nsing]!=0; nsing++) {} + + wa.tail(n-nsing).setZero(); + s.topLeftCorner(nsing, nsing).transpose().template triangularView<Upper>().solveInPlace(wa.head(nsing)); + + // restore + sdiag = s.diagonal(); + s.diagonal() = x; + + /* permute the components of z back to components of x. */ + x = iPerm * wa; +} + +template <typename Scalar, int _Options, typename Index> +void lmqrsolv( + SparseMatrix<Scalar,_Options,Index> &s, + const PermutationMatrix<Dynamic,Dynamic> &iPerm, + const Matrix<Scalar,Dynamic,1> &diag, + const Matrix<Scalar,Dynamic,1> &qtb, + Matrix<Scalar,Dynamic,1> &x, + Matrix<Scalar,Dynamic,1> &sdiag) +{ + /* Local variables */ + typedef SparseMatrix<Scalar,RowMajor,Index> FactorType; + Index i, j, k, l; + Scalar temp; + Index n = s.cols(); + Matrix<Scalar,Dynamic,1> wa(n); + JacobiRotation<Scalar> givens; + + /* Function Body */ + // the following will only change the lower triangular part of s, including + // the diagonal, though the diagonal is restored afterward + + /* copy r and (q transpose)*b to preserve input and initialize R. */ + wa = qtb; + FactorType R(s); + // Eliminate the diagonal matrix d using a givens rotation + for (j = 0; j < n; ++j) + { + // Prepare the row of d to be eliminated, locating the + // diagonal element using p from the qr factorization + l = iPerm.indices()(j); + if (diag(l) == Scalar(0)) + break; + sdiag.tail(n-j).setZero(); + sdiag[j] = diag[l]; + // the transformations to eliminate the row of d + // modify only a single element of (q transpose)*b + // beyond the first n, which is initially zero. + + Scalar qtbpj = 0; + // Browse the nonzero elements of row j of the upper triangular s + for (k = j; k < n; ++k) + { + typename FactorType::InnerIterator itk(R,k); + for (; itk; ++itk){ + if (itk.index() < k) continue; + else break; + } + //At this point, we have the diagonal element R(k,k) + // Determine a givens rotation which eliminates + // the appropriate element in the current row of d + givens.makeGivens(-itk.value(), sdiag(k)); + + // Compute the modified diagonal element of r and + // the modified element of ((q transpose)*b,0). + itk.valueRef() = givens.c() * itk.value() + givens.s() * sdiag(k); + temp = givens.c() * wa(k) + givens.s() * qtbpj; + qtbpj = -givens.s() * wa(k) + givens.c() * qtbpj; + wa(k) = temp; + + // Accumulate the transformation in the remaining k row/column of R + for (++itk; itk; ++itk) + { + i = itk.index(); + temp = givens.c() * itk.value() + givens.s() * sdiag(i); + sdiag(i) = -givens.s() * itk.value() + givens.c() * sdiag(i); + itk.valueRef() = temp; + } + } + } + + // Solve the triangular system for z. If the system is + // singular, then obtain a least squares solution + Index nsing; + for(nsing = 0; nsing<n && sdiag(nsing) !=0; nsing++) {} + + wa.tail(n-nsing).setZero(); +// x = wa; + wa.head(nsing) = R.topLeftCorner(nsing,nsing).template triangularView<Upper>().solve/*InPlace*/(wa.head(nsing)); + + sdiag = R.diagonal(); + // Permute the components of z back to components of x + x = iPerm * wa; +} +} // end namespace internal + +} // end namespace Eigen + +#endif // EIGEN_LMQRSOLV_H diff --git a/unsupported/Eigen/src/LevenbergMarquardt/LevenbergMarquardt.h b/unsupported/Eigen/src/LevenbergMarquardt/LevenbergMarquardt.h new file mode 100644 index 000000000..51dd1d3c4 --- /dev/null +++ b/unsupported/Eigen/src/LevenbergMarquardt/LevenbergMarquardt.h @@ -0,0 +1,377 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2009 Thomas Capricelli <orzel@freehackers.org> +// Copyright (C) 2012 Desire Nuentsa <desire.nuentsa_wakam@inria.fr> +// +// The algorithm of this class initially comes from MINPACK whose original authors are: +// Copyright Jorge More - Argonne National Laboratory +// Copyright Burt Garbow - Argonne National Laboratory +// Copyright Ken Hillstrom - Argonne National Laboratory +// +// This Source Code Form is subject to the terms of the Minpack license +// (a BSD-like license) described in the campaigned CopyrightMINPACK.txt file. +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + +#ifndef EIGEN_LEVENBERGMARQUARDT_H +#define EIGEN_LEVENBERGMARQUARDT_H + + +namespace Eigen { +namespace LevenbergMarquardtSpace { + enum Status { + NotStarted = -2, + Running = -1, + ImproperInputParameters = 0, + RelativeReductionTooSmall = 1, + RelativeErrorTooSmall = 2, + RelativeErrorAndReductionTooSmall = 3, + CosinusTooSmall = 4, + TooManyFunctionEvaluation = 5, + FtolTooSmall = 6, + XtolTooSmall = 7, + GtolTooSmall = 8, + UserAsked = 9 + }; +} + +template <typename _Scalar, int NX=Dynamic, int NY=Dynamic> +struct DenseFunctor +{ + typedef _Scalar Scalar; + enum { + InputsAtCompileTime = NX, + ValuesAtCompileTime = NY + }; + typedef Matrix<Scalar,InputsAtCompileTime,1> InputType; + typedef Matrix<Scalar,ValuesAtCompileTime,1> ValueType; + typedef Matrix<Scalar,ValuesAtCompileTime,InputsAtCompileTime> JacobianType; + typedef ColPivHouseholderQR<JacobianType> QRSolver; + const int m_inputs, m_values; + + DenseFunctor() : m_inputs(InputsAtCompileTime), m_values(ValuesAtCompileTime) {} + DenseFunctor(int inputs, int values) : m_inputs(inputs), m_values(values) {} + + int inputs() const { return m_inputs; } + int values() const { return m_values; } + + //int operator()(const InputType &x, ValueType& fvec) { } + // should be defined in derived classes + + //int df(const InputType &x, JacobianType& fjac) { } + // should be defined in derived classes +}; + +template <typename _Scalar, typename _Index> +struct SparseFunctor +{ + typedef _Scalar Scalar; + typedef _Index Index; + typedef Matrix<Scalar,Dynamic,1> InputType; + typedef Matrix<Scalar,Dynamic,1> ValueType; + typedef SparseMatrix<Scalar, ColMajor, Index> JacobianType; + typedef SparseQR<JacobianType, COLAMDOrdering<int> > QRSolver; + enum { + InputsAtCompileTime = Dynamic, + ValuesAtCompileTime = Dynamic + }; + + SparseFunctor(int inputs, int values) : m_inputs(inputs), m_values(values) {} + + int inputs() const { return m_inputs; } + int values() const { return m_values; } + + const int m_inputs, m_values; + //int operator()(const InputType &x, ValueType& fvec) { } + // to be defined in the functor + + //int df(const InputType &x, JacobianType& fjac) { } + // to be defined in the functor if no automatic differentiation + +}; +namespace internal { +template <typename QRSolver, typename VectorType> +void lmpar2(const QRSolver &qr, const VectorType &diag, const VectorType &qtb, + typename VectorType::Scalar m_delta, typename VectorType::Scalar &par, + VectorType &x); + } +/** + * \ingroup NonLinearOptimization_Module + * \brief Performs non linear optimization over a non-linear function, + * using a variant of the Levenberg Marquardt algorithm. + * + * Check wikipedia for more information. + * http://en.wikipedia.org/wiki/Levenberg%E2%80%93Marquardt_algorithm + */ +template<typename _FunctorType> +class LevenbergMarquardt : internal::no_assignment_operator +{ + public: + typedef _FunctorType FunctorType; + typedef typename FunctorType::QRSolver QRSolver; + typedef typename FunctorType::JacobianType JacobianType; + typedef typename JacobianType::Scalar Scalar; + typedef typename JacobianType::RealScalar RealScalar; + typedef typename JacobianType::Index Index; + typedef typename QRSolver::Index PermIndex; + typedef Matrix<Scalar,Dynamic,1> FVectorType; + typedef PermutationMatrix<Dynamic,Dynamic> PermutationType; + public: + LevenbergMarquardt(FunctorType& functor) + : m_functor(functor),m_nfev(0),m_njev(0),m_fnorm(0.0),m_gnorm(0), + m_isInitialized(false),m_info(InvalidInput) + { + resetParameters(); + m_useExternalScaling=false; + } + + LevenbergMarquardtSpace::Status minimize(FVectorType &x); + LevenbergMarquardtSpace::Status minimizeInit(FVectorType &x); + LevenbergMarquardtSpace::Status minimizeOneStep(FVectorType &x); + LevenbergMarquardtSpace::Status lmder1( + FVectorType &x, + const Scalar tol = std::sqrt(NumTraits<Scalar>::epsilon()) + ); + static LevenbergMarquardtSpace::Status lmdif1( + FunctorType &functor, + FVectorType &x, + Index *nfev, + const Scalar tol = std::sqrt(NumTraits<Scalar>::epsilon()) + ); + + /** Sets the default parameters */ + void resetParameters() + { + m_factor = 100.; + m_maxfev = 400; + m_ftol = std::sqrt(NumTraits<RealScalar>::epsilon()); + m_xtol = std::sqrt(NumTraits<RealScalar>::epsilon()); + m_gtol = 0. ; + m_epsfcn = 0. ; + } + + /** Sets the tolerance for the norm of the solution vector*/ + void setXtol(RealScalar xtol) { m_xtol = xtol; } + + /** Sets the tolerance for the norm of the vector function*/ + void setFtol(RealScalar ftol) { m_ftol = ftol; } + + /** Sets the tolerance for the norm of the gradient of the error vector*/ + void setGtol(RealScalar gtol) { m_gtol = gtol; } + + /** Sets the step bound for the diagonal shift */ + void setFactor(RealScalar factor) { m_factor = factor; } + + /** Sets the error precision */ + void setEpsilon (RealScalar epsfcn) { m_epsfcn = epsfcn; } + + /** Sets the maximum number of function evaluation */ + void setMaxfev(Index maxfev) {m_maxfev = maxfev; } + + /** Use an external Scaling. If set to true, pass a nonzero diagonal to diag() */ + void setExternalScaling(bool value) {m_useExternalScaling = value; } + + /** \returns a reference to the diagonal of the jacobian */ + FVectorType& diag() {return m_diag; } + + /** \returns the number of iterations performed */ + Index iterations() { return m_iter; } + + /** \returns the number of functions evaluation */ + Index nfev() { return m_nfev; } + + /** \returns the number of jacobian evaluation */ + Index njev() { return m_njev; } + + /** \returns the norm of current vector function */ + RealScalar fnorm() {return m_fnorm; } + + /** \returns the norm of the gradient of the error */ + RealScalar gnorm() {return m_gnorm; } + + /** \returns the LevenbergMarquardt parameter */ + RealScalar lm_param(void) { return m_par; } + + /** \returns a reference to the current vector function + */ + FVectorType& fvec() {return m_fvec; } + + /** \returns a reference to the matrix where the current Jacobian matrix is stored + */ + JacobianType& jacobian() {return m_fjac; } + + /** \returns a reference to the triangular matrix R from the QR of the jacobian matrix. + * \sa jacobian() + */ + JacobianType& matrixR() {return m_rfactor; } + + /** the permutation used in the QR factorization + */ + PermutationType permutation() {return m_permutation; } + + /** + * \brief Reports whether the minimization was successful + * \returns \c Success if the minimization was succesful, + * \c NumericalIssue if a numerical problem arises during the + * minimization process, for exemple during the QR factorization + * \c NoConvergence if the minimization did not converge after + * the maximum number of function evaluation allowed + * \c InvalidInput if the input matrix is invalid + */ + ComputationInfo info() const + { + + return m_info; + } + private: + JacobianType m_fjac; + JacobianType m_rfactor; // The triangular matrix R from the QR of the jacobian matrix m_fjac + FunctorType &m_functor; + FVectorType m_fvec, m_qtf, m_diag; + Index n; + Index m; + Index m_nfev; + Index m_njev; + RealScalar m_fnorm; // Norm of the current vector function + RealScalar m_gnorm; //Norm of the gradient of the error + RealScalar m_factor; // + Index m_maxfev; // Maximum number of function evaluation + RealScalar m_ftol; //Tolerance in the norm of the vector function + RealScalar m_xtol; // + RealScalar m_gtol; //tolerance of the norm of the error gradient + RealScalar m_epsfcn; // + Index m_iter; // Number of iterations performed + RealScalar m_delta; + bool m_useExternalScaling; + PermutationType m_permutation; + FVectorType m_wa1, m_wa2, m_wa3, m_wa4; //Temporary vectors + RealScalar m_par; + bool m_isInitialized; // Check whether the minimization step has been called + ComputationInfo m_info; +}; + +template<typename FunctorType> +LevenbergMarquardtSpace::Status +LevenbergMarquardt<FunctorType>::minimize(FVectorType &x) +{ + LevenbergMarquardtSpace::Status status = minimizeInit(x); + if (status==LevenbergMarquardtSpace::ImproperInputParameters) { + m_isInitialized = true; + return status; + } + do { +// std::cout << " uv " << x.transpose() << "\n"; + status = minimizeOneStep(x); + } while (status==LevenbergMarquardtSpace::Running); + m_isInitialized = true; + return status; +} + +template<typename FunctorType> +LevenbergMarquardtSpace::Status +LevenbergMarquardt<FunctorType>::minimizeInit(FVectorType &x) +{ + n = x.size(); + m = m_functor.values(); + + m_wa1.resize(n); m_wa2.resize(n); m_wa3.resize(n); + m_wa4.resize(m); + m_fvec.resize(m); + //FIXME Sparse Case : Allocate space for the jacobian + m_fjac.resize(m, n); +// m_fjac.reserve(VectorXi::Constant(n,5)); // FIXME Find a better alternative + if (!m_useExternalScaling) + m_diag.resize(n); + eigen_assert( (!m_useExternalScaling || m_diag.size()==n) || "When m_useExternalScaling is set, the caller must provide a valid 'm_diag'"); + m_qtf.resize(n); + + /* Function Body */ + m_nfev = 0; + m_njev = 0; + + /* check the input parameters for errors. */ + if (n <= 0 || m < n || m_ftol < 0. || m_xtol < 0. || m_gtol < 0. || m_maxfev <= 0 || m_factor <= 0.){ + m_info = InvalidInput; + return LevenbergMarquardtSpace::ImproperInputParameters; + } + + if (m_useExternalScaling) + for (Index j = 0; j < n; ++j) + if (m_diag[j] <= 0.) + { + m_info = InvalidInput; + return LevenbergMarquardtSpace::ImproperInputParameters; + } + + /* evaluate the function at the starting point */ + /* and calculate its norm. */ + m_nfev = 1; + if ( m_functor(x, m_fvec) < 0) + return LevenbergMarquardtSpace::UserAsked; + m_fnorm = m_fvec.stableNorm(); + + /* initialize levenberg-marquardt parameter and iteration counter. */ + m_par = 0.; + m_iter = 1; + + return LevenbergMarquardtSpace::NotStarted; +} + +template<typename FunctorType> +LevenbergMarquardtSpace::Status +LevenbergMarquardt<FunctorType>::lmder1( + FVectorType &x, + const Scalar tol + ) +{ + n = x.size(); + m = m_functor.values(); + + /* check the input parameters for errors. */ + if (n <= 0 || m < n || tol < 0.) + return LevenbergMarquardtSpace::ImproperInputParameters; + + resetParameters(); + m_ftol = tol; + m_xtol = tol; + m_maxfev = 100*(n+1); + + return minimize(x); +} + + +template<typename FunctorType> +LevenbergMarquardtSpace::Status +LevenbergMarquardt<FunctorType>::lmdif1( + FunctorType &functor, + FVectorType &x, + Index *nfev, + const Scalar tol + ) +{ + Index n = x.size(); + Index m = functor.values(); + + /* check the input parameters for errors. */ + if (n <= 0 || m < n || tol < 0.) + return LevenbergMarquardtSpace::ImproperInputParameters; + + NumericalDiff<FunctorType> numDiff(functor); + // embedded LevenbergMarquardt + LevenbergMarquardt<NumericalDiff<FunctorType> > lm(numDiff); + lm.setFtol(tol); + lm.setXtol(tol); + lm.setMaxfev(200*(n+1)); + + LevenbergMarquardtSpace::Status info = LevenbergMarquardtSpace::Status(lm.minimize(x)); + if (nfev) + * nfev = lm.nfev(); + return info; +} + +} // end namespace Eigen + +#endif // EIGEN_LEVENBERGMARQUARDT_H diff --git a/unsupported/Eigen/src/MatrixFunctions/MatrixExponential.h b/unsupported/Eigen/src/MatrixFunctions/MatrixExponential.h index 642916764..6825a7882 100644 --- a/unsupported/Eigen/src/MatrixFunctions/MatrixExponential.h +++ b/unsupported/Eigen/src/MatrixFunctions/MatrixExponential.h @@ -13,12 +13,7 @@ #include "StemFunction.h" -namespace Eigen { - -#if defined(_MSC_VER) || defined(__FreeBSD__) - template <typename Scalar> Scalar log2(Scalar v) { using std::log; return log(v)/log(Scalar(2)); } -#endif - +namespace Eigen { /** \ingroup MatrixFunctions_Module * \brief Class for computing the matrix exponential. @@ -233,7 +228,7 @@ template <typename MatrixType> EIGEN_STRONG_INLINE void MatrixExponential<MatrixType>::pade9(const MatrixType &A) { const RealScalar b[] = {17643225600., 8821612800., 2075673600., 302702400., 30270240., - 2162160., 110880., 3960., 90., 1.}; + 2162160., 110880., 3960., 90., 1.}; MatrixType A2 = A * A; MatrixType A4 = A2 * A2; MatrixType A6 = A4 * A2; @@ -247,8 +242,8 @@ template <typename MatrixType> EIGEN_STRONG_INLINE void MatrixExponential<MatrixType>::pade13(const MatrixType &A) { const RealScalar b[] = {64764752532480000., 32382376266240000., 7771770303897600., - 1187353796428800., 129060195264000., 10559470521600., 670442572800., - 33522128640., 1323241920., 40840800., 960960., 16380., 182., 1.}; + 1187353796428800., 129060195264000., 10559470521600., 670442572800., + 33522128640., 1323241920., 40840800., 960960., 16380., 182., 1.}; MatrixType A2 = A * A; MatrixType A4 = A2 * A2; m_tmp1.noalias() = A4 * A2; @@ -266,11 +261,11 @@ template <typename MatrixType> EIGEN_STRONG_INLINE void MatrixExponential<MatrixType>::pade17(const MatrixType &A) { const RealScalar b[] = {830034394580628357120000.L, 415017197290314178560000.L, - 100610229646136770560000.L, 15720348382208870400000.L, - 1774878043152614400000.L, 153822763739893248000.L, 10608466464820224000.L, - 595373117923584000.L, 27563570274240000.L, 1060137318240000.L, - 33924394183680.L, 899510451840.L, 19554575040.L, 341863200.L, 4651200.L, - 46512.L, 306.L, 1.L}; + 100610229646136770560000.L, 15720348382208870400000.L, + 1774878043152614400000.L, 153822763739893248000.L, 10608466464820224000.L, + 595373117923584000.L, 27563570274240000.L, 1060137318240000.L, + 33924394183680.L, 899510451840.L, 19554575040.L, 341863200.L, 4651200.L, + 46512.L, 306.L, 1.L}; MatrixType A2 = A * A; MatrixType A4 = A2 * A2; MatrixType A6 = A4 * A2; @@ -288,16 +283,16 @@ EIGEN_STRONG_INLINE void MatrixExponential<MatrixType>::pade17(const MatrixType template <typename MatrixType> void MatrixExponential<MatrixType>::computeUV(float) { - using std::max; + using std::frexp; using std::pow; - using std::ceil; if (m_l1norm < 4.258730016922831e-001) { pade3(m_M); } else if (m_l1norm < 1.880152677804762e+000) { pade5(m_M); } else { const float maxnorm = 3.925724783138660f; - m_squarings = (max)(0, (int)ceil(log2(m_l1norm / maxnorm))); + frexp(m_l1norm / maxnorm, &m_squarings); + if (m_squarings < 0) m_squarings = 0; MatrixType A = m_M / pow(Scalar(2), m_squarings); pade7(A); } @@ -306,9 +301,8 @@ void MatrixExponential<MatrixType>::computeUV(float) template <typename MatrixType> void MatrixExponential<MatrixType>::computeUV(double) { - using std::max; + using std::frexp; using std::pow; - using std::ceil; if (m_l1norm < 1.495585217958292e-002) { pade3(m_M); } else if (m_l1norm < 2.539398330063230e-001) { @@ -319,7 +313,8 @@ void MatrixExponential<MatrixType>::computeUV(double) pade9(m_M); } else { const double maxnorm = 5.371920351148152; - m_squarings = (max)(0, (int)ceil(log2(m_l1norm / maxnorm))); + frexp(m_l1norm / maxnorm, &m_squarings); + if (m_squarings < 0) m_squarings = 0; MatrixType A = m_M / pow(Scalar(2), m_squarings); pade13(A); } @@ -328,9 +323,8 @@ void MatrixExponential<MatrixType>::computeUV(double) template <typename MatrixType> void MatrixExponential<MatrixType>::computeUV(long double) { - using std::max; + using std::frexp; using std::pow; - using std::ceil; #if LDBL_MANT_DIG == 53 // double precision computeUV(double()); #elif LDBL_MANT_DIG <= 64 // extended precision @@ -344,7 +338,8 @@ void MatrixExponential<MatrixType>::computeUV(long double) pade9(m_M); } else { const long double maxnorm = 4.0246098906697353063L; - m_squarings = (max)(0, (int)ceil(log2(m_l1norm / maxnorm))); + frexp(m_l1norm / maxnorm, &m_squarings); + if (m_squarings < 0) m_squarings = 0; MatrixType A = m_M / pow(Scalar(2), m_squarings); pade13(A); } @@ -361,7 +356,8 @@ void MatrixExponential<MatrixType>::computeUV(long double) pade13(m_M); } else { const long double maxnorm = 3.2579440895405400856599663723517L; - m_squarings = (max)(0, (int)ceil(log2(m_l1norm / maxnorm))); + frexp(m_l1norm / maxnorm, &m_squarings); + if (m_squarings < 0) m_squarings = 0; MatrixType A = m_M / pow(Scalar(2), m_squarings); pade17(A); } @@ -378,7 +374,8 @@ void MatrixExponential<MatrixType>::computeUV(long double) pade13(m_M); } else { const long double maxnorm = 2.884233277829519311757165057717815L; - m_squarings = (max)(0, (int)ceil(log2(m_l1norm / maxnorm))); + frexp(m_l1norm / maxnorm, &m_squarings); + if (m_squarings < 0) m_squarings = 0; MatrixType A = m_M / pow(Scalar(2), m_squarings); pade17(A); } diff --git a/unsupported/Eigen/src/MatrixFunctions/MatrixFunction.h b/unsupported/Eigen/src/MatrixFunctions/MatrixFunction.h index c57ca87ed..7d426640c 100644 --- a/unsupported/Eigen/src/MatrixFunctions/MatrixFunction.h +++ b/unsupported/Eigen/src/MatrixFunctions/MatrixFunction.h @@ -209,7 +209,7 @@ void MatrixFunction<MatrixType,AtomicType,1>::compute(ResultType& result) permuteSchur(); computeBlockAtomic(); computeOffDiagonal(); - result = m_U * m_fT * m_U.adjoint(); + result = m_U * (m_fT.template triangularView<Upper>() * m_U.adjoint()); } /** \brief Store the Schur decomposition of #m_A in #m_T and #m_U */ @@ -235,6 +235,7 @@ void MatrixFunction<MatrixType,AtomicType,1>::computeSchurDecomposition() template <typename MatrixType, typename AtomicType> void MatrixFunction<MatrixType,AtomicType,1>::partitionEigenvalues() { + using std::abs; const Index rows = m_T.rows(); VectorType diag = m_T.diagonal(); // contains eigenvalues of A @@ -251,14 +252,14 @@ void MatrixFunction<MatrixType,AtomicType,1>::partitionEigenvalues() // Look for other element to add to the set for (Index j=i+1; j<rows; ++j) { - if (internal::abs(diag(j) - diag(i)) <= separation() && std::find(qi->begin(), qi->end(), diag(j)) == qi->end()) { - typename ListOfClusters::iterator qj = findCluster(diag(j)); - if (qj == m_clusters.end()) { - qi->push_back(diag(j)); - } else { - qi->insert(qi->end(), qj->begin(), qj->end()); - m_clusters.erase(qj); - } + if (abs(diag(j) - diag(i)) <= separation() && std::find(qi->begin(), qi->end(), diag(j)) == qi->end()) { + typename ListOfClusters::iterator qj = findCluster(diag(j)); + if (qj == m_clusters.end()) { + qi->push_back(diag(j)); + } else { + qi->insert(qi->end(), qj->begin(), qj->end()); + m_clusters.erase(qj); + } } } } diff --git a/unsupported/Eigen/src/MatrixFunctions/MatrixLogarithm.h b/unsupported/Eigen/src/MatrixFunctions/MatrixLogarithm.h index 3a50514b9..c744fc05f 100644 --- a/unsupported/Eigen/src/MatrixFunctions/MatrixLogarithm.h +++ b/unsupported/Eigen/src/MatrixFunctions/MatrixLogarithm.h @@ -51,7 +51,6 @@ private: void compute2x2(const MatrixType& A, MatrixType& result); void computeBig(const MatrixType& A, MatrixType& result); - static Scalar atanh(Scalar x); int getPadeDegree(float normTminusI); int getPadeDegree(double normTminusI); int getPadeDegree(long double normTminusI); @@ -67,10 +66,11 @@ private: void computePade11(MatrixType& result, const MatrixType& T); static const int minPadeDegree = 3; - static const int maxPadeDegree = std::numeric_limits<RealScalar>::digits<= 24? 5: // single precision - std::numeric_limits<RealScalar>::digits<= 53? 7: // double precision - std::numeric_limits<RealScalar>::digits<= 64? 8: // extended precision - std::numeric_limits<RealScalar>::digits<=106? 10: 11; // double-double or quadruple precision + static const int maxPadeDegree = std::numeric_limits<RealScalar>::digits<= 24? 5: // single precision + std::numeric_limits<RealScalar>::digits<= 53? 7: // double precision + std::numeric_limits<RealScalar>::digits<= 64? 8: // extended precision + std::numeric_limits<RealScalar>::digits<=106? 10: // double-double + 11; // quadruple precision // Prevent copying MatrixLogarithmAtomic(const MatrixLogarithmAtomic&); @@ -92,18 +92,6 @@ MatrixType MatrixLogarithmAtomic<MatrixType>::compute(const MatrixType& A) return result; } -/** \brief Compute atanh (inverse hyperbolic tangent). */ -template <typename MatrixType> -typename MatrixType::Scalar MatrixLogarithmAtomic<MatrixType>::atanh(typename MatrixType::Scalar x) -{ - using std::abs; - using std::sqrt; - if (abs(x) > sqrt(NumTraits<Scalar>::epsilon())) - return Scalar(0.5) * log((Scalar(1) + x) / (Scalar(1) - x)); - else - return x + x*x*x / Scalar(3); -} - /** \brief Compute logarithm of 2x2 triangular matrix. */ template <typename MatrixType> void MatrixLogarithmAtomic<MatrixType>::compute2x2(const MatrixType& A, MatrixType& result) @@ -127,8 +115,8 @@ void MatrixLogarithmAtomic<MatrixType>::compute2x2(const MatrixType& A, MatrixTy } else { // computation in previous branch is inaccurate if A(1,1) \approx A(0,0) int unwindingNumber = static_cast<int>(ceil((imag(logA11 - logA00) - M_PI) / (2*M_PI))); - Scalar z = (A(1,1) - A(0,0)) / (A(1,1) + A(0,0)); - result(0,1) = A(0,1) * (Scalar(2) * atanh(z) + Scalar(0,2*M_PI*unwindingNumber)) / (A(1,1) - A(0,0)); + Scalar y = A(1,1) - A(0,0), x = A(1,1) + A(0,0); + result(0,1) = A(0,1) * (Scalar(2) * numext::atanh2(y,x) + Scalar(0,2*M_PI*unwindingNumber)) / y; } } @@ -137,10 +125,11 @@ void MatrixLogarithmAtomic<MatrixType>::compute2x2(const MatrixType& A, MatrixTy template <typename MatrixType> void MatrixLogarithmAtomic<MatrixType>::computeBig(const MatrixType& A, MatrixType& result) { + using std::pow; int numberOfSquareRoots = 0; int numberOfExtraSquareRoots = 0; int degree; - MatrixType T = A; + MatrixType T = A, sqrtT; const RealScalar maxNormForPade = maxPadeDegree<= 5? 5.3149729967117310e-1: // single precision maxPadeDegree<= 7? 2.6429608311114350e-1: // double precision maxPadeDegree<= 8? 2.32777776523703892094e-1L: // extended precision @@ -153,12 +142,11 @@ void MatrixLogarithmAtomic<MatrixType>::computeBig(const MatrixType& A, MatrixTy degree = getPadeDegree(normTminusI); int degree2 = getPadeDegree(normTminusI / RealScalar(2)); if ((degree - degree2 <= 1) || (numberOfExtraSquareRoots == 1)) - break; + break; ++numberOfExtraSquareRoots; } - MatrixType sqrtT; MatrixSquareRootTriangular<MatrixType>(T).compute(sqrtT); - T = sqrtT; + T = sqrtT.template triangularView<Upper>(); ++numberOfSquareRoots; } @@ -172,10 +160,11 @@ int MatrixLogarithmAtomic<MatrixType>::getPadeDegree(float normTminusI) { const float maxNormForPade[] = { 2.5111573934555054e-1 /* degree = 3 */ , 4.0535837411880493e-1, 5.3149729967117310e-1 }; - for (int degree = 3; degree <= maxPadeDegree; ++degree) + int degree = 3; + for (; degree <= maxPadeDegree; ++degree) if (normTminusI <= maxNormForPade[degree - minPadeDegree]) - return degree; - assert(false); // this line should never be reached + break; + return degree; } /* \brief Get suitable degree for Pade approximation. (specialized for RealScalar = double) */ @@ -184,10 +173,11 @@ int MatrixLogarithmAtomic<MatrixType>::getPadeDegree(double normTminusI) { const double maxNormForPade[] = { 1.6206284795015624e-2 /* degree = 3 */ , 5.3873532631381171e-2, 1.1352802267628681e-1, 1.8662860613541288e-1, 2.642960831111435e-1 }; - for (int degree = 3; degree <= maxPadeDegree; ++degree) + int degree = 3; + for (; degree <= maxPadeDegree; ++degree) if (normTminusI <= maxNormForPade[degree - minPadeDegree]) - return degree; - assert(false); // this line should never be reached + break; + return degree; } /* \brief Get suitable degree for Pade approximation. (specialized for RealScalar = long double) */ @@ -195,29 +185,30 @@ template <typename MatrixType> int MatrixLogarithmAtomic<MatrixType>::getPadeDegree(long double normTminusI) { #if LDBL_MANT_DIG == 53 // double precision - const double maxNormForPade[] = { 1.6206284795015624e-2 /* degree = 3 */ , 5.3873532631381171e-2, - 1.1352802267628681e-1, 1.8662860613541288e-1, 2.642960831111435e-1 }; + const long double maxNormForPade[] = { 1.6206284795015624e-2L /* degree = 3 */ , 5.3873532631381171e-2L, + 1.1352802267628681e-1L, 1.8662860613541288e-1L, 2.642960831111435e-1L }; #elif LDBL_MANT_DIG <= 64 // extended precision - const double maxNormForPade[] = { 5.48256690357782863103e-3 /* degree = 3 */, 2.34559162387971167321e-2, - 5.84603923897347449857e-2, 1.08486423756725170223e-1, 1.68385767881294446649e-1, - 2.32777776523703892094e-1 }; + const long double maxNormForPade[] = { 5.48256690357782863103e-3L /* degree = 3 */, 2.34559162387971167321e-2L, + 5.84603923897347449857e-2L, 1.08486423756725170223e-1L, 1.68385767881294446649e-1L, + 2.32777776523703892094e-1L }; #elif LDBL_MANT_DIG <= 106 // double-double - const double maxNormForPade[] = { 8.58970550342939562202529664318890e-5 /* degree = 3 */, - 9.34074328446359654039446552677759e-4, 4.26117194647672175773064114582860e-3, - 1.21546224740281848743149666560464e-2, 2.61100544998339436713088248557444e-2, - 4.66170074627052749243018566390567e-2, 7.32585144444135027565872014932387e-2, - 1.05026503471351080481093652651105e-1 }; + const long double maxNormForPade[] = { 8.58970550342939562202529664318890e-5L /* degree = 3 */, + 9.34074328446359654039446552677759e-4L, 4.26117194647672175773064114582860e-3L, + 1.21546224740281848743149666560464e-2L, 2.61100544998339436713088248557444e-2L, + 4.66170074627052749243018566390567e-2L, 7.32585144444135027565872014932387e-2L, + 1.05026503471351080481093652651105e-1L }; #else // quadruple precision - const double maxNormForPade[] = { 4.7419931187193005048501568167858103e-5 /* degree = 3 */, - 5.8853168473544560470387769480192666e-4, 2.9216120366601315391789493628113520e-3, - 8.8415758124319434347116734705174308e-3, 1.9850836029449446668518049562565291e-2, - 3.6688019729653446926585242192447447e-2, 5.9290962294020186998954055264528393e-2, - 8.6998436081634343903250580992127677e-2, 1.1880960220216759245467951592883642e-1 }; + const long double maxNormForPade[] = { 4.7419931187193005048501568167858103e-5L /* degree = 3 */, + 5.8853168473544560470387769480192666e-4L, 2.9216120366601315391789493628113520e-3L, + 8.8415758124319434347116734705174308e-3L, 1.9850836029449446668518049562565291e-2L, + 3.6688019729653446926585242192447447e-2L, 5.9290962294020186998954055264528393e-2L, + 8.6998436081634343903250580992127677e-2L, 1.1880960220216759245467951592883642e-1L }; #endif - for (int degree = 3; degree <= maxPadeDegree; ++degree) + int degree = 3; + for (; degree <= maxPadeDegree; ++degree) if (normTminusI <= maxNormForPade[degree - minPadeDegree]) - return degree; - assert(false); // this line should never be reached + break; + return degree; } /* \brief Compute Pade approximation to matrix logarithm */ @@ -246,7 +237,7 @@ void MatrixLogarithmAtomic<MatrixType>::computePade3(MatrixType& result, const M 0.8872983346207416885179265399782400L }; const RealScalar weights[] = { 0.2777777777777777777777777777777778L, 0.4444444444444444444444444444444444L, 0.2777777777777777777777777777777778L }; - assert(degree <= maxPadeDegree); + eigen_assert(degree <= maxPadeDegree); MatrixType TminusI = T - MatrixType::Identity(T.rows(), T.rows()); result.setZero(T.rows(), T.rows()); for (int k = 0; k < degree; ++k) @@ -262,7 +253,7 @@ void MatrixLogarithmAtomic<MatrixType>::computePade4(MatrixType& result, const M 0.6699905217924281324013328795516223L, 0.9305681557970262876119732444464048L }; const RealScalar weights[] = { 0.1739274225687269286865319746109997L, 0.3260725774312730713134680253890003L, 0.3260725774312730713134680253890003L, 0.1739274225687269286865319746109997L }; - assert(degree <= maxPadeDegree); + eigen_assert(degree <= maxPadeDegree); MatrixType TminusI = T - MatrixType::Identity(T.rows(), T.rows()); result.setZero(T.rows(), T.rows()); for (int k = 0; k < degree; ++k) @@ -280,7 +271,7 @@ void MatrixLogarithmAtomic<MatrixType>::computePade5(MatrixType& result, const M const RealScalar weights[] = { 0.1184634425280945437571320203599587L, 0.2393143352496832340206457574178191L, 0.2844444444444444444444444444444444L, 0.2393143352496832340206457574178191L, 0.1184634425280945437571320203599587L }; - assert(degree <= maxPadeDegree); + eigen_assert(degree <= maxPadeDegree); MatrixType TminusI = T - MatrixType::Identity(T.rows(), T.rows()); result.setZero(T.rows(), T.rows()); for (int k = 0; k < degree; ++k) @@ -294,11 +285,11 @@ void MatrixLogarithmAtomic<MatrixType>::computePade6(MatrixType& result, const M const int degree = 6; const RealScalar nodes[] = { 0.0337652428984239860938492227530027L, 0.1693953067668677431693002024900473L, 0.3806904069584015456847491391596440L, 0.6193095930415984543152508608403560L, - 0.8306046932331322568306997975099527L, 0.9662347571015760139061507772469973L }; + 0.8306046932331322568306997975099527L, 0.9662347571015760139061507772469973L }; const RealScalar weights[] = { 0.0856622461895851725201480710863665L, 0.1803807865240693037849167569188581L, 0.2339569672863455236949351719947755L, 0.2339569672863455236949351719947755L, - 0.1803807865240693037849167569188581L, 0.0856622461895851725201480710863665L }; - assert(degree <= maxPadeDegree); + 0.1803807865240693037849167569188581L, 0.0856622461895851725201480710863665L }; + eigen_assert(degree <= maxPadeDegree); MatrixType TminusI = T - MatrixType::Identity(T.rows(), T.rows()); result.setZero(T.rows(), T.rows()); for (int k = 0; k < degree; ++k) @@ -318,7 +309,7 @@ void MatrixLogarithmAtomic<MatrixType>::computePade7(MatrixType& result, const M 0.1909150252525594724751848877444876L, 0.2089795918367346938775510204081633L, 0.1909150252525594724751848877444876L, 0.1398526957446383339507338857118898L, 0.0647424830844348466353057163395410L }; - assert(degree <= maxPadeDegree); + eigen_assert(degree <= maxPadeDegree); MatrixType TminusI = T - MatrixType::Identity(T.rows(), T.rows()); result.setZero(T.rows(), T.rows()); for (int k = 0; k < degree; ++k) @@ -338,7 +329,7 @@ void MatrixLogarithmAtomic<MatrixType>::computePade8(MatrixType& result, const M 0.1568533229389436436689811009933007L, 0.1813418916891809914825752246385978L, 0.1813418916891809914825752246385978L, 0.1568533229389436436689811009933007L, 0.1111905172266872352721779972131204L, 0.0506142681451881295762656771549811L }; - assert(degree <= maxPadeDegree); + eigen_assert(degree <= maxPadeDegree); MatrixType TminusI = T - MatrixType::Identity(T.rows(), T.rows()); result.setZero(T.rows(), T.rows()); for (int k = 0; k < degree; ++k) @@ -360,7 +351,7 @@ void MatrixLogarithmAtomic<MatrixType>::computePade9(MatrixType& result, const M 0.1651196775006298815822625346434870L, 0.1561735385200014200343152032922218L, 0.1303053482014677311593714347093164L, 0.0903240803474287020292360156214564L, 0.0406371941807872059859460790552618L }; - assert(degree <= maxPadeDegree); + eigen_assert(degree <= maxPadeDegree); MatrixType TminusI = T - MatrixType::Identity(T.rows(), T.rows()); result.setZero(T.rows(), T.rows()); for (int k = 0; k < degree; ++k) @@ -382,7 +373,7 @@ void MatrixLogarithmAtomic<MatrixType>::computePade10(MatrixType& result, const 0.1477621123573764350869464973256692L, 0.1477621123573764350869464973256692L, 0.1346333596549981775456134607847347L, 0.1095431812579910219977674671140816L, 0.0747256745752902965728881698288487L, 0.0333356721543440687967844049466659L }; - assert(degree <= maxPadeDegree); + eigen_assert(degree <= maxPadeDegree); MatrixType TminusI = T - MatrixType::Identity(T.rows(), T.rows()); result.setZero(T.rows(), T.rows()); for (int k = 0; k < degree; ++k) @@ -406,7 +397,7 @@ void MatrixLogarithmAtomic<MatrixType>::computePade11(MatrixType& result, const 0.1314022722551233310903444349452546L, 0.1165968822959952399592618524215876L, 0.0931451054638671257130488207158280L, 0.0627901847324523123173471496119701L, 0.0278342835580868332413768602212743L }; - assert(degree <= maxPadeDegree); + eigen_assert(degree <= maxPadeDegree); MatrixType TminusI = T - MatrixType::Identity(T.rows(), T.rows()); result.setZero(T.rows(), T.rows()); for (int k = 0; k < degree; ++k) @@ -423,8 +414,8 @@ void MatrixLogarithmAtomic<MatrixType>::computePade11(MatrixType& result, const * This class holds the argument to the matrix function until it is * assigned or evaluated for some other reason (so the argument * should not be changed in the meantime). It is the return type of - * matrixBase::matrixLogarithm() and most of the time this is the - * only way it is used. + * MatrixBase::log() and most of the time this is the only way it + * is used. */ template<typename Derived> class MatrixLogarithmReturnValue : public ReturnByValue<MatrixLogarithmReturnValue<Derived> > diff --git a/unsupported/Eigen/src/MatrixFunctions/MatrixPower.h b/unsupported/Eigen/src/MatrixFunctions/MatrixPower.h new file mode 100644 index 000000000..c32437281 --- /dev/null +++ b/unsupported/Eigen/src/MatrixFunctions/MatrixPower.h @@ -0,0 +1,509 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2012, 2013 Chen-Pang He <jdh8@ms63.hinet.net> +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + +#ifndef EIGEN_MATRIX_POWER +#define EIGEN_MATRIX_POWER + +namespace Eigen { + +template<typename MatrixType> class MatrixPower; + +template<typename MatrixType> +class MatrixPowerRetval : public ReturnByValue< MatrixPowerRetval<MatrixType> > +{ + public: + typedef typename MatrixType::RealScalar RealScalar; + typedef typename MatrixType::Index Index; + + MatrixPowerRetval(MatrixPower<MatrixType>& pow, RealScalar p) : m_pow(pow), m_p(p) + { } + + template<typename ResultType> + inline void evalTo(ResultType& res) const + { m_pow.compute(res, m_p); } + + Index rows() const { return m_pow.rows(); } + Index cols() const { return m_pow.cols(); } + + private: + MatrixPower<MatrixType>& m_pow; + const RealScalar m_p; + MatrixPowerRetval& operator=(const MatrixPowerRetval&); +}; + +template<typename MatrixType> +class MatrixPowerAtomic +{ + private: + enum { + RowsAtCompileTime = MatrixType::RowsAtCompileTime, + MaxRowsAtCompileTime = MatrixType::MaxRowsAtCompileTime + }; + typedef typename MatrixType::Scalar Scalar; + typedef typename MatrixType::RealScalar RealScalar; + typedef std::complex<RealScalar> ComplexScalar; + typedef typename MatrixType::Index Index; + typedef Array<Scalar, RowsAtCompileTime, 1, ColMajor, MaxRowsAtCompileTime> ArrayType; + + const MatrixType& m_A; + RealScalar m_p; + + void computePade(int degree, const MatrixType& IminusT, MatrixType& res) const; + void compute2x2(MatrixType& res, RealScalar p) const; + void computeBig(MatrixType& res) const; + static int getPadeDegree(float normIminusT); + static int getPadeDegree(double normIminusT); + static int getPadeDegree(long double normIminusT); + static ComplexScalar computeSuperDiag(const ComplexScalar&, const ComplexScalar&, RealScalar p); + static RealScalar computeSuperDiag(RealScalar, RealScalar, RealScalar p); + + public: + MatrixPowerAtomic(const MatrixType& T, RealScalar p); + void compute(MatrixType& res) const; +}; + +template<typename MatrixType> +MatrixPowerAtomic<MatrixType>::MatrixPowerAtomic(const MatrixType& T, RealScalar p) : + m_A(T), m_p(p) +{ eigen_assert(T.rows() == T.cols()); } + +template<typename MatrixType> +void MatrixPowerAtomic<MatrixType>::compute(MatrixType& res) const +{ + res.resizeLike(m_A); + switch (m_A.rows()) { + case 0: + break; + case 1: + res(0,0) = std::pow(m_A(0,0), m_p); + break; + case 2: + compute2x2(res, m_p); + break; + default: + computeBig(res); + } +} + +template<typename MatrixType> +void MatrixPowerAtomic<MatrixType>::computePade(int degree, const MatrixType& IminusT, MatrixType& res) const +{ + int i = degree<<1; + res = (m_p-degree) / ((i-1)<<1) * IminusT; + for (--i; i; --i) { + res = (MatrixType::Identity(IminusT.rows(), IminusT.cols()) + res).template triangularView<Upper>() + .solve((i==1 ? -m_p : i&1 ? (-m_p-(i>>1))/(i<<1) : (m_p-(i>>1))/((i-1)<<1)) * IminusT).eval(); + } + res += MatrixType::Identity(IminusT.rows(), IminusT.cols()); +} + +// This function assumes that res has the correct size (see bug 614) +template<typename MatrixType> +void MatrixPowerAtomic<MatrixType>::compute2x2(MatrixType& res, RealScalar p) const +{ + using std::abs; + using std::pow; + + ArrayType logTdiag = m_A.diagonal().array().log(); + res.coeffRef(0,0) = pow(m_A.coeff(0,0), p); + + for (Index i=1; i < m_A.cols(); ++i) { + res.coeffRef(i,i) = pow(m_A.coeff(i,i), p); + if (m_A.coeff(i-1,i-1) == m_A.coeff(i,i)) + res.coeffRef(i-1,i) = p * pow(m_A.coeff(i,i), p-1); + else if (2*abs(m_A.coeff(i-1,i-1)) < abs(m_A.coeff(i,i)) || 2*abs(m_A.coeff(i,i)) < abs(m_A.coeff(i-1,i-1))) + res.coeffRef(i-1,i) = (res.coeff(i,i)-res.coeff(i-1,i-1)) / (m_A.coeff(i,i)-m_A.coeff(i-1,i-1)); + else + res.coeffRef(i-1,i) = computeSuperDiag(m_A.coeff(i,i), m_A.coeff(i-1,i-1), p); + res.coeffRef(i-1,i) *= m_A.coeff(i-1,i); + } +} + +template<typename MatrixType> +void MatrixPowerAtomic<MatrixType>::computeBig(MatrixType& res) const +{ + const int digits = std::numeric_limits<RealScalar>::digits; + const RealScalar maxNormForPade = digits <= 24? 4.3386528e-1f: // sigle precision + digits <= 53? 2.789358995219730e-1: // double precision + digits <= 64? 2.4471944416607995472e-1L: // extended precision + digits <= 106? 1.1016843812851143391275867258512e-1L: // double-double + 9.134603732914548552537150753385375e-2L; // quadruple precision + MatrixType IminusT, sqrtT, T = m_A.template triangularView<Upper>(); + RealScalar normIminusT; + int degree, degree2, numberOfSquareRoots = 0; + bool hasExtraSquareRoot = false; + + /* FIXME + * For singular T, norm(I - T) >= 1 but maxNormForPade < 1, leads to infinite + * loop. We should move 0 eigenvalues to bottom right corner. We need not + * worry about tiny values (e.g. 1e-300) because they will reach 1 if + * repetitively sqrt'ed. + * + * If the 0 eigenvalues are semisimple, they can form a 0 matrix at the + * bottom right corner. + * + * [ T A ]^p [ T^p (T^-1 T^p A) ] + * [ ] = [ ] + * [ 0 0 ] [ 0 0 ] + */ + for (Index i=0; i < m_A.cols(); ++i) + eigen_assert(m_A(i,i) != RealScalar(0)); + + while (true) { + IminusT = MatrixType::Identity(m_A.rows(), m_A.cols()) - T; + normIminusT = IminusT.cwiseAbs().colwise().sum().maxCoeff(); + if (normIminusT < maxNormForPade) { + degree = getPadeDegree(normIminusT); + degree2 = getPadeDegree(normIminusT/2); + if (degree - degree2 <= 1 || hasExtraSquareRoot) + break; + hasExtraSquareRoot = true; + } + MatrixSquareRootTriangular<MatrixType>(T).compute(sqrtT); + T = sqrtT.template triangularView<Upper>(); + ++numberOfSquareRoots; + } + computePade(degree, IminusT, res); + + for (; numberOfSquareRoots; --numberOfSquareRoots) { + compute2x2(res, std::ldexp(m_p, -numberOfSquareRoots)); + res = res.template triangularView<Upper>() * res; + } + compute2x2(res, m_p); +} + +template<typename MatrixType> +inline int MatrixPowerAtomic<MatrixType>::getPadeDegree(float normIminusT) +{ + const float maxNormForPade[] = { 2.8064004e-1f /* degree = 3 */ , 4.3386528e-1f }; + int degree = 3; + for (; degree <= 4; ++degree) + if (normIminusT <= maxNormForPade[degree - 3]) + break; + return degree; +} + +template<typename MatrixType> +inline int MatrixPowerAtomic<MatrixType>::getPadeDegree(double normIminusT) +{ + const double maxNormForPade[] = { 1.884160592658218e-2 /* degree = 3 */ , 6.038881904059573e-2, 1.239917516308172e-1, + 1.999045567181744e-1, 2.789358995219730e-1 }; + int degree = 3; + for (; degree <= 7; ++degree) + if (normIminusT <= maxNormForPade[degree - 3]) + break; + return degree; +} + +template<typename MatrixType> +inline int MatrixPowerAtomic<MatrixType>::getPadeDegree(long double normIminusT) +{ +#if LDBL_MANT_DIG == 53 + const int maxPadeDegree = 7; + const double maxNormForPade[] = { 1.884160592658218e-2L /* degree = 3 */ , 6.038881904059573e-2L, 1.239917516308172e-1L, + 1.999045567181744e-1L, 2.789358995219730e-1L }; +#elif LDBL_MANT_DIG <= 64 + const int maxPadeDegree = 8; + const double maxNormForPade[] = { 6.3854693117491799460e-3L /* degree = 3 */ , 2.6394893435456973676e-2L, + 6.4216043030404063729e-2L, 1.1701165502926694307e-1L, 1.7904284231268670284e-1L, 2.4471944416607995472e-1L }; +#elif LDBL_MANT_DIG <= 106 + const int maxPadeDegree = 10; + const double maxNormForPade[] = { 1.0007161601787493236741409687186e-4L /* degree = 3 */ , + 1.0007161601787493236741409687186e-3L, 4.7069769360887572939882574746264e-3L, 1.3220386624169159689406653101695e-2L, + 2.8063482381631737920612944054906e-2L, 4.9625993951953473052385361085058e-2L, 7.7367040706027886224557538328171e-2L, + 1.1016843812851143391275867258512e-1L }; +#else + const int maxPadeDegree = 10; + const double maxNormForPade[] = { 5.524506147036624377378713555116378e-5L /* degree = 3 */ , + 6.640600568157479679823602193345995e-4L, 3.227716520106894279249709728084626e-3L, + 9.619593944683432960546978734646284e-3L, 2.134595382433742403911124458161147e-2L, + 3.908166513900489428442993794761185e-2L, 6.266780814639442865832535460550138e-2L, + 9.134603732914548552537150753385375e-2L }; +#endif + int degree = 3; + for (; degree <= maxPadeDegree; ++degree) + if (normIminusT <= maxNormForPade[degree - 3]) + break; + return degree; +} + +template<typename MatrixType> +inline typename MatrixPowerAtomic<MatrixType>::ComplexScalar +MatrixPowerAtomic<MatrixType>::computeSuperDiag(const ComplexScalar& curr, const ComplexScalar& prev, RealScalar p) +{ + ComplexScalar logCurr = std::log(curr); + ComplexScalar logPrev = std::log(prev); + int unwindingNumber = std::ceil((numext::imag(logCurr - logPrev) - M_PI) / (2*M_PI)); + ComplexScalar w = numext::atanh2(curr - prev, curr + prev) + ComplexScalar(0, M_PI*unwindingNumber); + return RealScalar(2) * std::exp(RealScalar(0.5) * p * (logCurr + logPrev)) * std::sinh(p * w) / (curr - prev); +} + +template<typename MatrixType> +inline typename MatrixPowerAtomic<MatrixType>::RealScalar +MatrixPowerAtomic<MatrixType>::computeSuperDiag(RealScalar curr, RealScalar prev, RealScalar p) +{ + RealScalar w = numext::atanh2(curr - prev, curr + prev); + return 2 * std::exp(p * (std::log(curr) + std::log(prev)) / 2) * std::sinh(p * w) / (curr - prev); +} + +/** + * \ingroup MatrixFunctions_Module + * + * \brief Class for computing matrix powers. + * + * \tparam MatrixType type of the base, expected to be an instantiation + * of the Matrix class template. + * + * This class is capable of computing real/complex matrices raised to + * an arbitrary real power. Meanwhile, it saves the result of Schur + * decomposition if an non-integral power has even been calculated. + * Therefore, if you want to compute multiple (>= 2) matrix powers + * for the same matrix, using the class directly is more efficient than + * calling MatrixBase::pow(). + * + * Example: + * \include MatrixPower_optimal.cpp + * Output: \verbinclude MatrixPower_optimal.out + */ +template<typename MatrixType> +class MatrixPower +{ + private: + enum { + RowsAtCompileTime = MatrixType::RowsAtCompileTime, + ColsAtCompileTime = MatrixType::ColsAtCompileTime, + MaxRowsAtCompileTime = MatrixType::MaxRowsAtCompileTime, + MaxColsAtCompileTime = MatrixType::MaxColsAtCompileTime + }; + typedef typename MatrixType::Scalar Scalar; + typedef typename MatrixType::RealScalar RealScalar; + typedef typename MatrixType::Index Index; + + public: + /** + * \brief Constructor. + * + * \param[in] A the base of the matrix power. + * + * The class stores a reference to A, so it should not be changed + * (or destroyed) before evaluation. + */ + explicit MatrixPower(const MatrixType& A) : m_A(A), m_conditionNumber(0) + { eigen_assert(A.rows() == A.cols()); } + + /** + * \brief Returns the matrix power. + * + * \param[in] p exponent, a real scalar. + * \return The expression \f$ A^p \f$, where A is specified in the + * constructor. + */ + const MatrixPowerRetval<MatrixType> operator()(RealScalar p) + { return MatrixPowerRetval<MatrixType>(*this, p); } + + /** + * \brief Compute the matrix power. + * + * \param[in] p exponent, a real scalar. + * \param[out] res \f$ A^p \f$ where A is specified in the + * constructor. + */ + template<typename ResultType> + void compute(ResultType& res, RealScalar p); + + Index rows() const { return m_A.rows(); } + Index cols() const { return m_A.cols(); } + + private: + typedef std::complex<RealScalar> ComplexScalar; + typedef Matrix<ComplexScalar, RowsAtCompileTime, ColsAtCompileTime, MatrixType::Options, + MaxRowsAtCompileTime, MaxColsAtCompileTime> ComplexMatrix; + + typename MatrixType::Nested m_A; + MatrixType m_tmp; + ComplexMatrix m_T, m_U, m_fT; + RealScalar m_conditionNumber; + + RealScalar modfAndInit(RealScalar, RealScalar*); + + template<typename ResultType> + void computeIntPower(ResultType&, RealScalar); + + template<typename ResultType> + void computeFracPower(ResultType&, RealScalar); + + template<int Rows, int Cols, int Options, int MaxRows, int MaxCols> + static void revertSchur( + Matrix<ComplexScalar, Rows, Cols, Options, MaxRows, MaxCols>& res, + const ComplexMatrix& T, + const ComplexMatrix& U); + + template<int Rows, int Cols, int Options, int MaxRows, int MaxCols> + static void revertSchur( + Matrix<RealScalar, Rows, Cols, Options, MaxRows, MaxCols>& res, + const ComplexMatrix& T, + const ComplexMatrix& U); +}; + +template<typename MatrixType> +template<typename ResultType> +void MatrixPower<MatrixType>::compute(ResultType& res, RealScalar p) +{ + switch (cols()) { + case 0: + break; + case 1: + res(0,0) = std::pow(m_A.coeff(0,0), p); + break; + default: + RealScalar intpart, x = modfAndInit(p, &intpart); + computeIntPower(res, intpart); + computeFracPower(res, x); + } +} + +template<typename MatrixType> +typename MatrixPower<MatrixType>::RealScalar +MatrixPower<MatrixType>::modfAndInit(RealScalar x, RealScalar* intpart) +{ + typedef Array<RealScalar, RowsAtCompileTime, 1, ColMajor, MaxRowsAtCompileTime> RealArray; + + *intpart = std::floor(x); + RealScalar res = x - *intpart; + + if (!m_conditionNumber && res) { + const ComplexSchur<MatrixType> schurOfA(m_A); + m_T = schurOfA.matrixT(); + m_U = schurOfA.matrixU(); + + const RealArray absTdiag = m_T.diagonal().array().abs(); + m_conditionNumber = absTdiag.maxCoeff() / absTdiag.minCoeff(); + } + + if (res>RealScalar(0.5) && res>(1-res)*std::pow(m_conditionNumber, res)) { + --res; + ++*intpart; + } + return res; +} + +template<typename MatrixType> +template<typename ResultType> +void MatrixPower<MatrixType>::computeIntPower(ResultType& res, RealScalar p) +{ + RealScalar pp = std::abs(p); + + if (p<0) m_tmp = m_A.inverse(); + else m_tmp = m_A; + + res = MatrixType::Identity(rows(), cols()); + while (pp >= 1) { + if (std::fmod(pp, 2) >= 1) + res = m_tmp * res; + m_tmp *= m_tmp; + pp /= 2; + } +} + +template<typename MatrixType> +template<typename ResultType> +void MatrixPower<MatrixType>::computeFracPower(ResultType& res, RealScalar p) +{ + if (p) { + eigen_assert(m_conditionNumber); + MatrixPowerAtomic<ComplexMatrix>(m_T, p).compute(m_fT); + revertSchur(m_tmp, m_fT, m_U); + res = m_tmp * res; + } +} + +template<typename MatrixType> +template<int Rows, int Cols, int Options, int MaxRows, int MaxCols> +inline void MatrixPower<MatrixType>::revertSchur( + Matrix<ComplexScalar, Rows, Cols, Options, MaxRows, MaxCols>& res, + const ComplexMatrix& T, + const ComplexMatrix& U) +{ res.noalias() = U * (T.template triangularView<Upper>() * U.adjoint()); } + +template<typename MatrixType> +template<int Rows, int Cols, int Options, int MaxRows, int MaxCols> +inline void MatrixPower<MatrixType>::revertSchur( + Matrix<RealScalar, Rows, Cols, Options, MaxRows, MaxCols>& res, + const ComplexMatrix& T, + const ComplexMatrix& U) +{ res.noalias() = (U * (T.template triangularView<Upper>() * U.adjoint())).real(); } + +/** + * \ingroup MatrixFunctions_Module + * + * \brief Proxy for the matrix power of some matrix (expression). + * + * \tparam Derived type of the base, a matrix (expression). + * + * This class holds the arguments to the matrix power until it is + * assigned or evaluated for some other reason (so the argument + * should not be changed in the meantime). It is the return type of + * MatrixBase::pow() and related functions and most of the + * time this is the only way it is used. + */ +template<typename Derived> +class MatrixPowerReturnValue : public ReturnByValue< MatrixPowerReturnValue<Derived> > +{ + public: + typedef typename Derived::PlainObject PlainObject; + typedef typename Derived::RealScalar RealScalar; + typedef typename Derived::Index Index; + + /** + * \brief Constructor. + * + * \param[in] A %Matrix (expression), the base of the matrix power. + * \param[in] p scalar, the exponent of the matrix power. + */ + MatrixPowerReturnValue(const Derived& A, RealScalar p) : m_A(A), m_p(p) + { } + + /** + * \brief Compute the matrix power. + * + * \param[out] result \f$ A^p \f$ where \p A and \p p are as in the + * constructor. + */ + template<typename ResultType> + inline void evalTo(ResultType& res) const + { MatrixPower<PlainObject>(m_A.eval()).compute(res, m_p); } + + Index rows() const { return m_A.rows(); } + Index cols() const { return m_A.cols(); } + + private: + const Derived& m_A; + const RealScalar m_p; + MatrixPowerReturnValue& operator=(const MatrixPowerReturnValue&); +}; + +namespace internal { + +template<typename MatrixPowerType> +struct traits< MatrixPowerRetval<MatrixPowerType> > +{ typedef typename MatrixPowerType::PlainObject ReturnType; }; + +template<typename Derived> +struct traits< MatrixPowerReturnValue<Derived> > +{ typedef typename Derived::PlainObject ReturnType; }; + +} + +template<typename Derived> +const MatrixPowerReturnValue<Derived> MatrixBase<Derived>::pow(const RealScalar& p) const +{ return MatrixPowerReturnValue<Derived>(derived(), p); } + +} // namespace Eigen + +#endif // EIGEN_MATRIX_POWER diff --git a/unsupported/Eigen/src/MatrixFunctions/MatrixSquareRoot.h b/unsupported/Eigen/src/MatrixFunctions/MatrixSquareRoot.h index 10319fa17..b48ea9d46 100644 --- a/unsupported/Eigen/src/MatrixFunctions/MatrixSquareRoot.h +++ b/unsupported/Eigen/src/MatrixFunctions/MatrixSquareRoot.h @@ -60,17 +60,17 @@ class MatrixSquareRootQuasiTriangular void computeOffDiagonalPartOfSqrt(MatrixType& sqrtT, const MatrixType& T); void compute2x2diagonalBlock(MatrixType& sqrtT, const MatrixType& T, typename MatrixType::Index i); void compute1x1offDiagonalBlock(MatrixType& sqrtT, const MatrixType& T, - typename MatrixType::Index i, typename MatrixType::Index j); + typename MatrixType::Index i, typename MatrixType::Index j); void compute1x2offDiagonalBlock(MatrixType& sqrtT, const MatrixType& T, - typename MatrixType::Index i, typename MatrixType::Index j); + typename MatrixType::Index i, typename MatrixType::Index j); void compute2x1offDiagonalBlock(MatrixType& sqrtT, const MatrixType& T, - typename MatrixType::Index i, typename MatrixType::Index j); + typename MatrixType::Index i, typename MatrixType::Index j); void compute2x2offDiagonalBlock(MatrixType& sqrtT, const MatrixType& T, - typename MatrixType::Index i, typename MatrixType::Index j); + typename MatrixType::Index i, typename MatrixType::Index j); template <typename SmallMatrixType> static void solveAuxiliaryEquation(SmallMatrixType& X, const SmallMatrixType& A, - const SmallMatrixType& B, const SmallMatrixType& C); + const SmallMatrixType& B, const SmallMatrixType& C); const MatrixType& m_A; }; @@ -79,18 +79,9 @@ template <typename MatrixType> template <typename ResultType> void MatrixSquareRootQuasiTriangular<MatrixType>::compute(ResultType &result) { - // Compute Schur decomposition of m_A - const RealSchur<MatrixType> schurOfA(m_A); - const MatrixType& T = schurOfA.matrixT(); - const MatrixType& U = schurOfA.matrixU(); - - // Compute square root of T - MatrixType sqrtT = MatrixType::Zero(m_A.rows(), m_A.rows()); - computeDiagonalPartOfSqrt(sqrtT, T); - computeOffDiagonalPartOfSqrt(sqrtT, T); - - // Compute square root of m_A - result = U * sqrtT * U.adjoint(); + result.resize(m_A.rows(), m_A.cols()); + computeDiagonalPartOfSqrt(result, m_A); + computeOffDiagonalPartOfSqrt(result, m_A); } // pre: T is quasi-upper-triangular and sqrtT is a zero matrix of the same size @@ -99,11 +90,12 @@ template <typename MatrixType> void MatrixSquareRootQuasiTriangular<MatrixType>::computeDiagonalPartOfSqrt(MatrixType& sqrtT, const MatrixType& T) { + using std::sqrt; const Index size = m_A.rows(); for (Index i = 0; i < size; i++) { if (i == size - 1 || T.coeff(i+1, i) == 0) { - eigen_assert(T(i,i) > 0); - sqrtT.coeffRef(i,i) = internal::sqrt(T.coeff(i,i)); + eigen_assert(T(i,i) >= 0); + sqrtT.coeffRef(i,i) = sqrt(T.coeff(i,i)); } else { compute2x2diagonalBlock(sqrtT, T, i); @@ -289,17 +281,14 @@ template <typename MatrixType> template <typename ResultType> void MatrixSquareRootTriangular<MatrixType>::compute(ResultType &result) { - // Compute Schur decomposition of m_A - const ComplexSchur<MatrixType> schurOfA(m_A); - const MatrixType& T = schurOfA.matrixT(); - const MatrixType& U = schurOfA.matrixU(); + using std::sqrt; - // Compute square root of T and store it in upper triangular part of result + // Compute square root of m_A and store it in upper triangular part of result // This uses that the square root of triangular matrices can be computed directly. result.resize(m_A.rows(), m_A.cols()); typedef typename MatrixType::Index Index; for (Index i = 0; i < m_A.rows(); i++) { - result.coeffRef(i,i) = internal::sqrt(T.coeff(i,i)); + result.coeffRef(i,i) = sqrt(m_A.coeff(i,i)); } for (Index j = 1; j < m_A.cols(); j++) { for (Index i = j-1; i >= 0; i--) { @@ -307,14 +296,9 @@ void MatrixSquareRootTriangular<MatrixType>::compute(ResultType &result) // if i = j-1, then segment has length 0 so tmp = 0 Scalar tmp = (result.row(i).segment(i+1,j-i-1) * result.col(j).segment(i+1,j-i-1)).value(); // denominator may be zero if original matrix is singular - result.coeffRef(i,j) = (T.coeff(i,j) - tmp) / (result.coeff(i,i) + result.coeff(j,j)); + result.coeffRef(i,j) = (m_A.coeff(i,j) - tmp) / (result.coeff(i,i) + result.coeff(j,j)); } } - - // Compute square root of m_A as U * result * U.adjoint() - MatrixType tmp; - tmp.noalias() = U * result.template triangularView<Upper>(); - result.noalias() = tmp * U.adjoint(); } @@ -371,9 +355,8 @@ class MatrixSquareRoot<MatrixType, 0> const MatrixType& U = schurOfA.matrixU(); // Compute square root of T - MatrixSquareRootQuasiTriangular<MatrixType> tmp(T); - MatrixType sqrtT = MatrixType::Zero(m_A.rows(), m_A.rows()); - tmp.compute(sqrtT); + MatrixType sqrtT = MatrixType::Zero(m_A.rows(), m_A.cols()); + MatrixSquareRootQuasiTriangular<MatrixType>(T).compute(sqrtT); // Compute square root of m_A result = U * sqrtT * U.adjoint(); @@ -405,12 +388,11 @@ class MatrixSquareRoot<MatrixType, 1> const MatrixType& U = schurOfA.matrixU(); // Compute square root of T - MatrixSquareRootTriangular<MatrixType> tmp(T); - MatrixType sqrtT = MatrixType::Zero(m_A.rows(), m_A.rows()); - tmp.compute(sqrtT); + MatrixType sqrtT; + MatrixSquareRootTriangular<MatrixType>(T).compute(sqrtT); // Compute square root of m_A - result = U * sqrtT * U.adjoint(); + result = U * (sqrtT.template triangularView<Upper>() * U.adjoint()); } private: diff --git a/unsupported/Eigen/src/NonLinearOptimization/HybridNonLinearSolver.h b/unsupported/Eigen/src/NonLinearOptimization/HybridNonLinearSolver.h index d9ce4eab6..b8ba6ddcb 100644 --- a/unsupported/Eigen/src/NonLinearOptimization/HybridNonLinearSolver.h +++ b/unsupported/Eigen/src/NonLinearOptimization/HybridNonLinearSolver.h @@ -52,7 +52,7 @@ public: Parameters() : factor(Scalar(100.)) , maxfev(1000) - , xtol(internal::sqrt(NumTraits<Scalar>::epsilon())) + , xtol(std::sqrt(NumTraits<Scalar>::epsilon())) , nb_of_subdiagonals(-1) , nb_of_superdiagonals(-1) , epsfcn(Scalar(0.)) {} @@ -70,7 +70,7 @@ public: HybridNonLinearSolverSpace::Status hybrj1( FVectorType &x, - const Scalar tol = internal::sqrt(NumTraits<Scalar>::epsilon()) + const Scalar tol = std::sqrt(NumTraits<Scalar>::epsilon()) ); HybridNonLinearSolverSpace::Status solveInit(FVectorType &x); @@ -79,7 +79,7 @@ public: HybridNonLinearSolverSpace::Status hybrd1( FVectorType &x, - const Scalar tol = internal::sqrt(NumTraits<Scalar>::epsilon()) + const Scalar tol = std::sqrt(NumTraits<Scalar>::epsilon()) ); HybridNonLinearSolverSpace::Status solveNumericalDiffInit(FVectorType &x); @@ -150,7 +150,7 @@ HybridNonLinearSolver<FunctorType,Scalar>::solveInit(FVectorType &x) fjac.resize(n, n); if (!useExternalScaling) diag.resize(n); - assert( (!useExternalScaling || diag.size()==n) || "When useExternalScaling is set, the caller must provide a valid 'diag'"); + eigen_assert( (!useExternalScaling || diag.size()==n) || "When useExternalScaling is set, the caller must provide a valid 'diag'"); /* Function Body */ nfev = 0; @@ -185,7 +185,9 @@ template<typename FunctorType, typename Scalar> HybridNonLinearSolverSpace::Status HybridNonLinearSolver<FunctorType,Scalar>::solveOneStep(FVectorType &x) { - assert(x.size()==n); // check the caller is not cheating us + using std::abs; + + eigen_assert(x.size()==n); // check the caller is not cheating us Index j; std::vector<JacobiRotation<Scalar> > v_givens(n), w_givens(n); @@ -252,14 +254,14 @@ HybridNonLinearSolver<FunctorType,Scalar>::solveOneStep(FVectorType &x) /* compute the scaled actual reduction. */ actred = -1.; if (fnorm1 < fnorm) /* Computing 2nd power */ - actred = 1. - internal::abs2(fnorm1 / fnorm); + actred = 1. - numext::abs2(fnorm1 / fnorm); /* compute the scaled predicted reduction. */ wa3 = R.template triangularView<Upper>()*wa1 + qtf; temp = wa3.stableNorm(); prered = 0.; if (temp < fnorm) /* Computing 2nd power */ - prered = 1. - internal::abs2(temp / fnorm); + prered = 1. - numext::abs2(temp / fnorm); /* compute the ratio of the actual to the predicted reduction. */ ratio = 0.; @@ -276,7 +278,7 @@ HybridNonLinearSolver<FunctorType,Scalar>::solveOneStep(FVectorType &x) ++ncsuc; if (ratio >= Scalar(.5) || ncsuc > 1) delta = (std::max)(delta, pnorm / Scalar(.5)); - if (internal::abs(ratio - 1.) <= Scalar(.1)) { + if (abs(ratio - 1.) <= Scalar(.1)) { delta = pnorm / Scalar(.5); } } @@ -388,7 +390,7 @@ HybridNonLinearSolver<FunctorType,Scalar>::solveNumericalDiffInit(FVectorType & fvec.resize(n); if (!useExternalScaling) diag.resize(n); - assert( (!useExternalScaling || diag.size()==n) || "When useExternalScaling is set, the caller must provide a valid 'diag'"); + eigen_assert( (!useExternalScaling || diag.size()==n) || "When useExternalScaling is set, the caller must provide a valid 'diag'"); /* Function Body */ nfev = 0; @@ -423,6 +425,9 @@ template<typename FunctorType, typename Scalar> HybridNonLinearSolverSpace::Status HybridNonLinearSolver<FunctorType,Scalar>::solveNumericalDiffOneStep(FVectorType &x) { + using std::sqrt; + using std::abs; + assert(x.size()==n); // check the caller is not cheating us Index j; @@ -492,14 +497,14 @@ HybridNonLinearSolver<FunctorType,Scalar>::solveNumericalDiffOneStep(FVectorType /* compute the scaled actual reduction. */ actred = -1.; if (fnorm1 < fnorm) /* Computing 2nd power */ - actred = 1. - internal::abs2(fnorm1 / fnorm); + actred = 1. - numext::abs2(fnorm1 / fnorm); /* compute the scaled predicted reduction. */ wa3 = R.template triangularView<Upper>()*wa1 + qtf; temp = wa3.stableNorm(); prered = 0.; if (temp < fnorm) /* Computing 2nd power */ - prered = 1. - internal::abs2(temp / fnorm); + prered = 1. - numext::abs2(temp / fnorm); /* compute the ratio of the actual to the predicted reduction. */ ratio = 0.; @@ -516,7 +521,7 @@ HybridNonLinearSolver<FunctorType,Scalar>::solveNumericalDiffOneStep(FVectorType ++ncsuc; if (ratio >= Scalar(.5) || ncsuc > 1) delta = (std::max)(delta, pnorm / Scalar(.5)); - if (internal::abs(ratio - 1.) <= Scalar(.1)) { + if (abs(ratio - 1.) <= Scalar(.1)) { delta = pnorm / Scalar(.5); } } diff --git a/unsupported/Eigen/src/NonLinearOptimization/LevenbergMarquardt.h b/unsupported/Eigen/src/NonLinearOptimization/LevenbergMarquardt.h index 075faeeb0..bfeb26fc9 100644 --- a/unsupported/Eigen/src/NonLinearOptimization/LevenbergMarquardt.h +++ b/unsupported/Eigen/src/NonLinearOptimization/LevenbergMarquardt.h @@ -55,8 +55,8 @@ public: Parameters() : factor(Scalar(100.)) , maxfev(400) - , ftol(internal::sqrt(NumTraits<Scalar>::epsilon())) - , xtol(internal::sqrt(NumTraits<Scalar>::epsilon())) + , ftol(std::sqrt(NumTraits<Scalar>::epsilon())) + , xtol(std::sqrt(NumTraits<Scalar>::epsilon())) , gtol(Scalar(0.)) , epsfcn(Scalar(0.)) {} Scalar factor; @@ -72,7 +72,7 @@ public: LevenbergMarquardtSpace::Status lmder1( FVectorType &x, - const Scalar tol = internal::sqrt(NumTraits<Scalar>::epsilon()) + const Scalar tol = std::sqrt(NumTraits<Scalar>::epsilon()) ); LevenbergMarquardtSpace::Status minimize(FVectorType &x); @@ -83,12 +83,12 @@ public: FunctorType &functor, FVectorType &x, Index *nfev, - const Scalar tol = internal::sqrt(NumTraits<Scalar>::epsilon()) + const Scalar tol = std::sqrt(NumTraits<Scalar>::epsilon()) ); LevenbergMarquardtSpace::Status lmstr1( FVectorType &x, - const Scalar tol = internal::sqrt(NumTraits<Scalar>::epsilon()) + const Scalar tol = std::sqrt(NumTraits<Scalar>::epsilon()) ); LevenbergMarquardtSpace::Status minimizeOptimumStorage(FVectorType &x); @@ -172,7 +172,7 @@ LevenbergMarquardt<FunctorType,Scalar>::minimizeInit(FVectorType &x) fjac.resize(m, n); if (!useExternalScaling) diag.resize(n); - assert( (!useExternalScaling || diag.size()==n) || "When useExternalScaling is set, the caller must provide a valid 'diag'"); + eigen_assert( (!useExternalScaling || diag.size()==n) || "When useExternalScaling is set, the caller must provide a valid 'diag'"); qtf.resize(n); /* Function Body */ @@ -206,7 +206,10 @@ template<typename FunctorType, typename Scalar> LevenbergMarquardtSpace::Status LevenbergMarquardt<FunctorType,Scalar>::minimizeOneStep(FVectorType &x) { - assert(x.size()==n); // check the caller is not cheating us + using std::abs; + using std::sqrt; + + eigen_assert(x.size()==n); // check the caller is not cheating us /* calculate the jacobian matrix. */ Index df_ret = functor.df(x, fjac); @@ -249,7 +252,7 @@ LevenbergMarquardt<FunctorType,Scalar>::minimizeOneStep(FVectorType &x) if (fnorm != 0.) for (Index j = 0; j < n; ++j) if (wa2[permutation.indices()[j]] != 0.) - gnorm = (std::max)(gnorm, internal::abs( fjac.col(j).head(j+1).dot(qtf.head(j+1)/fnorm) / wa2[permutation.indices()[j]])); + gnorm = (std::max)(gnorm, abs( fjac.col(j).head(j+1).dot(qtf.head(j+1)/fnorm) / wa2[permutation.indices()[j]])); /* test for convergence of the gradient norm. */ if (gnorm <= parameters.gtol) @@ -282,13 +285,13 @@ LevenbergMarquardt<FunctorType,Scalar>::minimizeOneStep(FVectorType &x) /* compute the scaled actual reduction. */ actred = -1.; if (Scalar(.1) * fnorm1 < fnorm) - actred = 1. - internal::abs2(fnorm1 / fnorm); + actred = 1. - numext::abs2(fnorm1 / fnorm); /* compute the scaled predicted reduction and */ /* the scaled directional derivative. */ wa3 = fjac.template triangularView<Upper>() * (qrfac.colsPermutation().inverse() *wa1); - temp1 = internal::abs2(wa3.stableNorm() / fnorm); - temp2 = internal::abs2(internal::sqrt(par) * pnorm / fnorm); + temp1 = numext::abs2(wa3.stableNorm() / fnorm); + temp2 = numext::abs2(sqrt(par) * pnorm / fnorm); prered = temp1 + temp2 / Scalar(.5); dirder = -(temp1 + temp2); @@ -326,9 +329,9 @@ LevenbergMarquardt<FunctorType,Scalar>::minimizeOneStep(FVectorType &x) } /* tests for convergence. */ - if (internal::abs(actred) <= parameters.ftol && prered <= parameters.ftol && Scalar(.5) * ratio <= 1. && delta <= parameters.xtol * xnorm) + if (abs(actred) <= parameters.ftol && prered <= parameters.ftol && Scalar(.5) * ratio <= 1. && delta <= parameters.xtol * xnorm) return LevenbergMarquardtSpace::RelativeErrorAndReductionTooSmall; - if (internal::abs(actred) <= parameters.ftol && prered <= parameters.ftol && Scalar(.5) * ratio <= 1.) + if (abs(actred) <= parameters.ftol && prered <= parameters.ftol && Scalar(.5) * ratio <= 1.) return LevenbergMarquardtSpace::RelativeReductionTooSmall; if (delta <= parameters.xtol * xnorm) return LevenbergMarquardtSpace::RelativeErrorTooSmall; @@ -336,7 +339,7 @@ LevenbergMarquardt<FunctorType,Scalar>::minimizeOneStep(FVectorType &x) /* tests for termination and stringent tolerances. */ if (nfev >= parameters.maxfev) return LevenbergMarquardtSpace::TooManyFunctionEvaluation; - if (internal::abs(actred) <= NumTraits<Scalar>::epsilon() && prered <= NumTraits<Scalar>::epsilon() && Scalar(.5) * ratio <= 1.) + if (abs(actred) <= NumTraits<Scalar>::epsilon() && prered <= NumTraits<Scalar>::epsilon() && Scalar(.5) * ratio <= 1.) return LevenbergMarquardtSpace::FtolTooSmall; if (delta <= NumTraits<Scalar>::epsilon() * xnorm) return LevenbergMarquardtSpace::XtolTooSmall; @@ -388,7 +391,7 @@ LevenbergMarquardt<FunctorType,Scalar>::minimizeOptimumStorageInit(FVectorType fjac.resize(n, n); if (!useExternalScaling) diag.resize(n); - assert( (!useExternalScaling || diag.size()==n) || "When useExternalScaling is set, the caller must provide a valid 'diag'"); + eigen_assert( (!useExternalScaling || diag.size()==n) || "When useExternalScaling is set, the caller must provide a valid 'diag'"); qtf.resize(n); /* Function Body */ @@ -423,7 +426,10 @@ template<typename FunctorType, typename Scalar> LevenbergMarquardtSpace::Status LevenbergMarquardt<FunctorType,Scalar>::minimizeOptimumStorageOneStep(FVectorType &x) { - assert(x.size()==n); // check the caller is not cheating us + using std::abs; + using std::sqrt; + + eigen_assert(x.size()==n); // check the caller is not cheating us Index i, j; bool sing; @@ -496,7 +502,7 @@ LevenbergMarquardt<FunctorType,Scalar>::minimizeOptimumStorageOneStep(FVectorTyp if (fnorm != 0.) for (j = 0; j < n; ++j) if (wa2[permutation.indices()[j]] != 0.) - gnorm = (std::max)(gnorm, internal::abs( fjac.col(j).head(j+1).dot(qtf.head(j+1)/fnorm) / wa2[permutation.indices()[j]])); + gnorm = (std::max)(gnorm, abs( fjac.col(j).head(j+1).dot(qtf.head(j+1)/fnorm) / wa2[permutation.indices()[j]])); /* test for convergence of the gradient norm. */ if (gnorm <= parameters.gtol) @@ -529,13 +535,13 @@ LevenbergMarquardt<FunctorType,Scalar>::minimizeOptimumStorageOneStep(FVectorTyp /* compute the scaled actual reduction. */ actred = -1.; if (Scalar(.1) * fnorm1 < fnorm) - actred = 1. - internal::abs2(fnorm1 / fnorm); + actred = 1. - numext::abs2(fnorm1 / fnorm); /* compute the scaled predicted reduction and */ /* the scaled directional derivative. */ wa3 = fjac.topLeftCorner(n,n).template triangularView<Upper>() * (permutation.inverse() * wa1); - temp1 = internal::abs2(wa3.stableNorm() / fnorm); - temp2 = internal::abs2(internal::sqrt(par) * pnorm / fnorm); + temp1 = numext::abs2(wa3.stableNorm() / fnorm); + temp2 = numext::abs2(sqrt(par) * pnorm / fnorm); prered = temp1 + temp2 / Scalar(.5); dirder = -(temp1 + temp2); @@ -573,9 +579,9 @@ LevenbergMarquardt<FunctorType,Scalar>::minimizeOptimumStorageOneStep(FVectorTyp } /* tests for convergence. */ - if (internal::abs(actred) <= parameters.ftol && prered <= parameters.ftol && Scalar(.5) * ratio <= 1. && delta <= parameters.xtol * xnorm) + if (abs(actred) <= parameters.ftol && prered <= parameters.ftol && Scalar(.5) * ratio <= 1. && delta <= parameters.xtol * xnorm) return LevenbergMarquardtSpace::RelativeErrorAndReductionTooSmall; - if (internal::abs(actred) <= parameters.ftol && prered <= parameters.ftol && Scalar(.5) * ratio <= 1.) + if (abs(actred) <= parameters.ftol && prered <= parameters.ftol && Scalar(.5) * ratio <= 1.) return LevenbergMarquardtSpace::RelativeReductionTooSmall; if (delta <= parameters.xtol * xnorm) return LevenbergMarquardtSpace::RelativeErrorTooSmall; @@ -583,7 +589,7 @@ LevenbergMarquardt<FunctorType,Scalar>::minimizeOptimumStorageOneStep(FVectorTyp /* tests for termination and stringent tolerances. */ if (nfev >= parameters.maxfev) return LevenbergMarquardtSpace::TooManyFunctionEvaluation; - if (internal::abs(actred) <= NumTraits<Scalar>::epsilon() && prered <= NumTraits<Scalar>::epsilon() && Scalar(.5) * ratio <= 1.) + if (abs(actred) <= NumTraits<Scalar>::epsilon() && prered <= NumTraits<Scalar>::epsilon() && Scalar(.5) * ratio <= 1.) return LevenbergMarquardtSpace::FtolTooSmall; if (delta <= NumTraits<Scalar>::epsilon() * xnorm) return LevenbergMarquardtSpace::XtolTooSmall; diff --git a/unsupported/Eigen/src/NonLinearOptimization/chkder.h b/unsupported/Eigen/src/NonLinearOptimization/chkder.h index ad37c5029..db8ff7d6e 100644 --- a/unsupported/Eigen/src/NonLinearOptimization/chkder.h +++ b/unsupported/Eigen/src/NonLinearOptimization/chkder.h @@ -16,6 +16,10 @@ void chkder( Matrix< Scalar, Dynamic, 1 > &err ) { + using std::sqrt; + using std::abs; + using std::log; + typedef DenseIndex Index; const Scalar eps = sqrt(NumTraits<Scalar>::epsilon()); diff --git a/unsupported/Eigen/src/NonLinearOptimization/covar.h b/unsupported/Eigen/src/NonLinearOptimization/covar.h index c73a09645..68260d191 100644 --- a/unsupported/Eigen/src/NonLinearOptimization/covar.h +++ b/unsupported/Eigen/src/NonLinearOptimization/covar.h @@ -6,8 +6,9 @@ template <typename Scalar> void covar( Matrix< Scalar, Dynamic, Dynamic > &r, const VectorXi &ipvt, - Scalar tol = sqrt(NumTraits<Scalar>::epsilon()) ) + Scalar tol = std::sqrt(NumTraits<Scalar>::epsilon()) ) { + using std::abs; typedef DenseIndex Index; /* Local variables */ @@ -19,7 +20,7 @@ void covar( const Index n = r.cols(); const Scalar tolr = tol * abs(r(0,0)); Matrix< Scalar, Dynamic, 1 > wa(n); - assert(ipvt.size()==n); + eigen_assert(ipvt.size()==n); /* form the inverse of r in the full upper triangle of r. */ l = -1; diff --git a/unsupported/Eigen/src/NonLinearOptimization/dogleg.h b/unsupported/Eigen/src/NonLinearOptimization/dogleg.h index 4fbc98bfc..80c5d277b 100644 --- a/unsupported/Eigen/src/NonLinearOptimization/dogleg.h +++ b/unsupported/Eigen/src/NonLinearOptimization/dogleg.h @@ -10,6 +10,9 @@ void dogleg( Scalar delta, Matrix< Scalar, Dynamic, 1 > &x) { + using std::abs; + using std::sqrt; + typedef DenseIndex Index; /* Local variables */ @@ -21,9 +24,9 @@ void dogleg( /* Function Body */ const Scalar epsmch = NumTraits<Scalar>::epsilon(); const Index n = qrfac.cols(); - assert(n==qtb.size()); - assert(n==x.size()); - assert(n==diag.size()); + eigen_assert(n==qtb.size()); + eigen_assert(n==x.size()); + eigen_assert(n==diag.size()); Matrix< Scalar, Dynamic, 1 > wa1(n), wa2(n); /* first, calculate the gauss-newton direction. */ @@ -89,8 +92,8 @@ void dogleg( /* at which the quadratic is minimized. */ bnorm = qtb.stableNorm(); temp = bnorm / gnorm * (bnorm / qnorm) * (sgnorm / delta); - temp = temp - delta / qnorm * abs2(sgnorm / delta) + sqrt(abs2(temp - delta / qnorm) + (1.-abs2(delta / qnorm)) * (1.-abs2(sgnorm / delta))); - alpha = delta / qnorm * (1. - abs2(sgnorm / delta)) / temp; + temp = temp - delta / qnorm * numext::abs2(sgnorm / delta) + sqrt(numext::abs2(temp - delta / qnorm) + (1.-numext::abs2(delta / qnorm)) * (1.-numext::abs2(sgnorm / delta))); + alpha = delta / qnorm * (1. - numext::abs2(sgnorm / delta)) / temp; algo_end: /* form appropriate convex combination of the gauss-newton */ diff --git a/unsupported/Eigen/src/NonLinearOptimization/fdjac1.h b/unsupported/Eigen/src/NonLinearOptimization/fdjac1.h index 1cabe69ae..bb7cf267b 100644 --- a/unsupported/Eigen/src/NonLinearOptimization/fdjac1.h +++ b/unsupported/Eigen/src/NonLinearOptimization/fdjac1.h @@ -11,6 +11,9 @@ DenseIndex fdjac1( DenseIndex ml, DenseIndex mu, Scalar epsfcn) { + using std::sqrt; + using std::abs; + typedef DenseIndex Index; /* Local variables */ @@ -24,7 +27,7 @@ DenseIndex fdjac1( /* Function Body */ const Scalar epsmch = NumTraits<Scalar>::epsilon(); const Index n = x.size(); - assert(fvec.size()==n); + eigen_assert(fvec.size()==n); Matrix< Scalar, Dynamic, 1 > wa1(n); Matrix< Scalar, Dynamic, 1 > wa2(n); diff --git a/unsupported/Eigen/src/NonLinearOptimization/lmpar.h b/unsupported/Eigen/src/NonLinearOptimization/lmpar.h index cc1ca530f..4c17d4cdf 100644 --- a/unsupported/Eigen/src/NonLinearOptimization/lmpar.h +++ b/unsupported/Eigen/src/NonLinearOptimization/lmpar.h @@ -12,6 +12,8 @@ void lmpar( Scalar &par, Matrix< Scalar, Dynamic, 1 > &x) { + using std::abs; + using std::sqrt; typedef DenseIndex Index; /* Local variables */ @@ -25,11 +27,11 @@ void lmpar( /* Function Body */ - const Scalar dwarf = std::numeric_limits<Scalar>::min(); + const Scalar dwarf = (std::numeric_limits<Scalar>::min)(); const Index n = r.cols(); - assert(n==diag.size()); - assert(n==qtb.size()); - assert(n==x.size()); + eigen_assert(n==diag.size()); + eigen_assert(n==qtb.size()); + eigen_assert(n==x.size()); Matrix< Scalar, Dynamic, 1 > wa1, wa2; @@ -168,6 +170,8 @@ void lmpar2( Matrix< Scalar, Dynamic, 1 > &x) { + using std::sqrt; + using std::abs; typedef DenseIndex Index; /* Local variables */ @@ -181,10 +185,10 @@ void lmpar2( /* Function Body */ - const Scalar dwarf = std::numeric_limits<Scalar>::min(); + const Scalar dwarf = (std::numeric_limits<Scalar>::min)(); const Index n = qr.matrixQR().cols(); - assert(n==diag.size()); - assert(n==qtb.size()); + eigen_assert(n==diag.size()); + eigen_assert(n==qtb.size()); Matrix< Scalar, Dynamic, 1 > wa1, wa2; diff --git a/unsupported/Eigen/src/NonLinearOptimization/r1updt.h b/unsupported/Eigen/src/NonLinearOptimization/r1updt.h index 55fae5ae8..f28766061 100644 --- a/unsupported/Eigen/src/NonLinearOptimization/r1updt.h +++ b/unsupported/Eigen/src/NonLinearOptimization/r1updt.h @@ -24,10 +24,10 @@ void r1updt( // r1updt had a broader usecase, but we dont use it here. And, more // importantly, we can not test it. - assert(m==n); - assert(u.size()==m); - assert(v.size()==n); - assert(w.size()==n); + eigen_assert(m==n); + eigen_assert(u.size()==m); + eigen_assert(v.size()==n); + eigen_assert(w.size()==n); /* move the nontrivial part of the last column of s into w. */ w[n-1] = s(n-1,n-1); diff --git a/unsupported/Eigen/src/NonLinearOptimization/rwupdt.h b/unsupported/Eigen/src/NonLinearOptimization/rwupdt.h index 9ce079e22..6ebf8563f 100644 --- a/unsupported/Eigen/src/NonLinearOptimization/rwupdt.h +++ b/unsupported/Eigen/src/NonLinearOptimization/rwupdt.h @@ -12,7 +12,7 @@ void rwupdt( typedef DenseIndex Index; const Index n = r.cols(); - assert(r.rows()>=n); + eigen_assert(r.rows()>=n); std::vector<JacobiRotation<Scalar> > givens(n); /* Local variables */ diff --git a/unsupported/Eigen/src/NumericalDiff/NumericalDiff.h b/unsupported/Eigen/src/NumericalDiff/NumericalDiff.h index d848cb407..ea5d8bc27 100644 --- a/unsupported/Eigen/src/NumericalDiff/NumericalDiff.h +++ b/unsupported/Eigen/src/NumericalDiff/NumericalDiff.h @@ -63,11 +63,13 @@ public: */ int df(const InputType& _x, JacobianType &jac) const { + using std::sqrt; + using std::abs; /* Local variables */ Scalar h; int nfev=0; const typename InputType::Index n = _x.size(); - const Scalar eps = internal::sqrt(((std::max)(epsfcn,NumTraits<Scalar>::epsilon() ))); + const Scalar eps = sqrt(((std::max)(epsfcn,NumTraits<Scalar>::epsilon() ))); ValueType val1, val2; InputType x = _x; // TODO : we should do this only if the size is not already known @@ -84,12 +86,12 @@ public: // do nothing break; default: - assert(false); + eigen_assert(false); }; // Function Body for (int j = 0; j < n; ++j) { - h = eps * internal::abs(x[j]); + h = eps * abs(x[j]); if (h == 0.) { h = eps; } @@ -110,7 +112,7 @@ public: jac.col(j) = (val2-val1)/(2*h); break; default: - assert(false); + eigen_assert(false); }; } return nfev; diff --git a/unsupported/Eigen/src/Polynomials/Companion.h b/unsupported/Eigen/src/Polynomials/Companion.h index 4badd9d58..b515c2920 100644 --- a/unsupported/Eigen/src/Polynomials/Companion.h +++ b/unsupported/Eigen/src/Polynomials/Companion.h @@ -210,6 +210,7 @@ bool companion<_Scalar,_Deg>::balancedR( Scalar colNorm, Scalar rowNorm, template< typename _Scalar, int _Deg > void companion<_Scalar,_Deg>::balance() { + using std::abs; EIGEN_STATIC_ASSERT( Deg == Dynamic || 1 < Deg, YOU_MADE_A_PROGRAMMING_MISTAKE ); const Index deg = m_monic.size(); const Index deg_1 = deg-1; diff --git a/unsupported/Eigen/src/Polynomials/PolynomialSolver.h b/unsupported/Eigen/src/Polynomials/PolynomialSolver.h index 70b873dbc..cd5c04bbf 100644 --- a/unsupported/Eigen/src/Polynomials/PolynomialSolver.h +++ b/unsupported/Eigen/src/Polynomials/PolynomialSolver.h @@ -69,10 +69,11 @@ class PolynomialSolverBase inline void realRoots( Stl_back_insertion_sequence& bi_seq, const RealScalar& absImaginaryThreshold = NumTraits<Scalar>::dummy_precision() ) const { + using std::abs; bi_seq.clear(); for(Index i=0; i<m_roots.size(); ++i ) { - if( internal::abs( m_roots[i].imag() ) < absImaginaryThreshold ){ + if( abs( m_roots[i].imag() ) < absImaginaryThreshold ){ bi_seq.push_back( m_roots[i].real() ); } } } @@ -82,10 +83,10 @@ class PolynomialSolverBase inline const RootType& selectComplexRoot_withRespectToNorm( squaredNormBinaryPredicate& pred ) const { Index res=0; - RealScalar norm2 = internal::abs2( m_roots[0] ); + RealScalar norm2 = numext::abs2( m_roots[0] ); for( Index i=1; i<m_roots.size(); ++i ) { - const RealScalar currNorm2 = internal::abs2( m_roots[i] ); + const RealScalar currNorm2 = numext::abs2( m_roots[i] ); if( pred( currNorm2, norm2 ) ){ res=i; norm2=currNorm2; } } @@ -118,13 +119,14 @@ class PolynomialSolverBase bool& hasArealRoot, const RealScalar& absImaginaryThreshold = NumTraits<Scalar>::dummy_precision() ) const { + using std::abs; hasArealRoot = false; Index res=0; RealScalar abs2(0); for( Index i=0; i<m_roots.size(); ++i ) { - if( internal::abs( m_roots[i].imag() ) < absImaginaryThreshold ) + if( abs( m_roots[i].imag() ) < absImaginaryThreshold ) { if( !hasArealRoot ) { @@ -144,11 +146,11 @@ class PolynomialSolverBase } else { - if( internal::abs( m_roots[i].imag() ) < internal::abs( m_roots[res].imag() ) ){ + if( abs( m_roots[i].imag() ) < abs( m_roots[res].imag() ) ){ res = i; } } } - return internal::real_ref(m_roots[res]); + return numext::real_ref(m_roots[res]); } @@ -158,13 +160,14 @@ class PolynomialSolverBase bool& hasArealRoot, const RealScalar& absImaginaryThreshold = NumTraits<Scalar>::dummy_precision() ) const { + using std::abs; hasArealRoot = false; Index res=0; RealScalar val(0); for( Index i=0; i<m_roots.size(); ++i ) { - if( internal::abs( m_roots[i].imag() ) < absImaginaryThreshold ) + if( abs( m_roots[i].imag() ) < absImaginaryThreshold ) { if( !hasArealRoot ) { @@ -184,11 +187,11 @@ class PolynomialSolverBase } else { - if( internal::abs( m_roots[i].imag() ) < internal::abs( m_roots[res].imag() ) ){ + if( abs( m_roots[i].imag() ) < abs( m_roots[res].imag() ) ){ res = i; } } } - return internal::real_ref(m_roots[res]); + return numext::real_ref(m_roots[res]); } public: @@ -341,7 +344,7 @@ class PolynomialSolver : public PolynomialSolverBase<_Scalar,_Deg> template< typename OtherPolynomial > void compute( const OtherPolynomial& poly ) { - assert( Scalar(0) != poly[poly.size()-1] ); + eigen_assert( Scalar(0) != poly[poly.size()-1] ); internal::companion<Scalar,_Deg> companion( poly ); companion.balance(); m_eigenSolver.compute( companion.denseMatrix() ); @@ -373,7 +376,7 @@ class PolynomialSolver<_Scalar,1> : public PolynomialSolverBase<_Scalar,1> template< typename OtherPolynomial > void compute( const OtherPolynomial& poly ) { - assert( Scalar(0) != poly[poly.size()-1] ); + eigen_assert( Scalar(0) != poly[poly.size()-1] ); m_roots[0] = -poly[0]/poly[poly.size()-1]; } diff --git a/unsupported/Eigen/src/Polynomials/PolynomialUtils.h b/unsupported/Eigen/src/Polynomials/PolynomialUtils.h index c23204c65..2bb8bc84a 100644 --- a/unsupported/Eigen/src/Polynomials/PolynomialUtils.h +++ b/unsupported/Eigen/src/Polynomials/PolynomialUtils.h @@ -47,7 +47,7 @@ T poly_eval( const Polynomials& poly, const T& x ) { typedef typename NumTraits<T>::Real Real; - if( internal::abs2( x ) <= Real(1) ){ + if( numext::abs2( x ) <= Real(1) ){ return poly_eval_horner( poly, x ); } else { @@ -74,15 +74,16 @@ template <typename Polynomial> inline typename NumTraits<typename Polynomial::Scalar>::Real cauchy_max_bound( const Polynomial& poly ) { + using std::abs; typedef typename Polynomial::Scalar Scalar; typedef typename NumTraits<Scalar>::Real Real; - assert( Scalar(0) != poly[poly.size()-1] ); + eigen_assert( Scalar(0) != poly[poly.size()-1] ); const Scalar inv_leading_coeff = Scalar(1)/poly[poly.size()-1]; Real cb(0); for( DenseIndex i=0; i<poly.size()-1; ++i ){ - cb += internal::abs(poly[i]*inv_leading_coeff); } + cb += abs(poly[i]*inv_leading_coeff); } return cb + Real(1); } @@ -96,6 +97,7 @@ template <typename Polynomial> inline typename NumTraits<typename Polynomial::Scalar>::Real cauchy_min_bound( const Polynomial& poly ) { + using std::abs; typedef typename Polynomial::Scalar Scalar; typedef typename NumTraits<Scalar>::Real Real; @@ -107,7 +109,7 @@ typename NumTraits<typename Polynomial::Scalar>::Real cauchy_min_bound( const Po const Scalar inv_min_coeff = Scalar(1)/poly[i]; Real cb(1); for( DenseIndex j=i+1; j<poly.size(); ++j ){ - cb += internal::abs(poly[j]*inv_min_coeff); } + cb += abs(poly[j]*inv_min_coeff); } return Real(1)/cb; } diff --git a/unsupported/Eigen/src/SVD/BDCSVD.h b/unsupported/Eigen/src/SVD/BDCSVD.h new file mode 100644 index 000000000..11d4882e4 --- /dev/null +++ b/unsupported/Eigen/src/SVD/BDCSVD.h @@ -0,0 +1,748 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// We used the "A Divide-And-Conquer Algorithm for the Bidiagonal SVD" +// research report written by Ming Gu and Stanley C.Eisenstat +// The code variable names correspond to the names they used in their +// report +// +// Copyright (C) 2013 Gauthier Brun <brun.gauthier@gmail.com> +// Copyright (C) 2013 Nicolas Carre <nicolas.carre@ensimag.fr> +// Copyright (C) 2013 Jean Ceccato <jean.ceccato@ensimag.fr> +// Copyright (C) 2013 Pierre Zoppitelli <pierre.zoppitelli@ensimag.fr> +// +// Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + +#ifndef EIGEN_BDCSVD_H +#define EIGEN_BDCSVD_H + +#define EPSILON 0.0000000000000001 + +#define ALGOSWAP 32 + +namespace Eigen { +/** \ingroup SVD_Module + * + * + * \class BDCSVD + * + * \brief class Bidiagonal Divide and Conquer SVD + * + * \param MatrixType the type of the matrix of which we are computing the SVD decomposition + * We plan to have a very similar interface to JacobiSVD on this class. + * It should be used to speed up the calcul of SVD for big matrices. + */ +template<typename _MatrixType> +class BDCSVD : public SVDBase<_MatrixType> +{ + typedef SVDBase<_MatrixType> Base; + +public: + using Base::rows; + using Base::cols; + + typedef _MatrixType MatrixType; + typedef typename MatrixType::Scalar Scalar; + typedef typename NumTraits<typename MatrixType::Scalar>::Real RealScalar; + typedef typename MatrixType::Index Index; + enum { + RowsAtCompileTime = MatrixType::RowsAtCompileTime, + ColsAtCompileTime = MatrixType::ColsAtCompileTime, + DiagSizeAtCompileTime = EIGEN_SIZE_MIN_PREFER_DYNAMIC(RowsAtCompileTime, ColsAtCompileTime), + MaxRowsAtCompileTime = MatrixType::MaxRowsAtCompileTime, + MaxColsAtCompileTime = MatrixType::MaxColsAtCompileTime, + MaxDiagSizeAtCompileTime = EIGEN_SIZE_MIN_PREFER_FIXED(MaxRowsAtCompileTime, MaxColsAtCompileTime), + MatrixOptions = MatrixType::Options + }; + + typedef Matrix<Scalar, RowsAtCompileTime, RowsAtCompileTime, + MatrixOptions, MaxRowsAtCompileTime, MaxRowsAtCompileTime> + MatrixUType; + typedef Matrix<Scalar, ColsAtCompileTime, ColsAtCompileTime, + MatrixOptions, MaxColsAtCompileTime, MaxColsAtCompileTime> + MatrixVType; + typedef typename internal::plain_diag_type<MatrixType, RealScalar>::type SingularValuesType; + typedef typename internal::plain_row_type<MatrixType>::type RowType; + typedef typename internal::plain_col_type<MatrixType>::type ColType; + typedef Matrix<Scalar, Dynamic, Dynamic> MatrixX; + typedef Matrix<RealScalar, Dynamic, Dynamic> MatrixXr; + typedef Matrix<RealScalar, Dynamic, 1> VectorType; + + /** \brief Default Constructor. + * + * The default constructor is useful in cases in which the user intends to + * perform decompositions via BDCSVD::compute(const MatrixType&). + */ + BDCSVD() + : SVDBase<_MatrixType>::SVDBase(), + algoswap(ALGOSWAP) + {} + + + /** \brief Default Constructor with memory preallocation + * + * Like the default constructor but with preallocation of the internal data + * according to the specified problem size. + * \sa BDCSVD() + */ + BDCSVD(Index rows, Index cols, unsigned int computationOptions = 0) + : SVDBase<_MatrixType>::SVDBase(), + algoswap(ALGOSWAP) + { + allocate(rows, cols, computationOptions); + } + + /** \brief Constructor performing the decomposition of given matrix. + * + * \param matrix the matrix to decompose + * \param computationOptions optional parameter allowing to specify if you want full or thin U or V unitaries to be computed. + * By default, none is computed. This is a bit - field, the possible bits are #ComputeFullU, #ComputeThinU, + * #ComputeFullV, #ComputeThinV. + * + * Thin unitaries are only available if your matrix type has a Dynamic number of columns (for example MatrixXf). They also are not + * available with the (non - default) FullPivHouseholderQR preconditioner. + */ + BDCSVD(const MatrixType& matrix, unsigned int computationOptions = 0) + : SVDBase<_MatrixType>::SVDBase(), + algoswap(ALGOSWAP) + { + compute(matrix, computationOptions); + } + + ~BDCSVD() + { + } + /** \brief Method performing the decomposition of given matrix using custom options. + * + * \param matrix the matrix to decompose + * \param computationOptions optional parameter allowing to specify if you want full or thin U or V unitaries to be computed. + * By default, none is computed. This is a bit - field, the possible bits are #ComputeFullU, #ComputeThinU, + * #ComputeFullV, #ComputeThinV. + * + * Thin unitaries are only available if your matrix type has a Dynamic number of columns (for example MatrixXf). They also are not + * available with the (non - default) FullPivHouseholderQR preconditioner. + */ + SVDBase<MatrixType>& compute(const MatrixType& matrix, unsigned int computationOptions); + + /** \brief Method performing the decomposition of given matrix using current options. + * + * \param matrix the matrix to decompose + * + * This method uses the current \a computationOptions, as already passed to the constructor or to compute(const MatrixType&, unsigned int). + */ + SVDBase<MatrixType>& compute(const MatrixType& matrix) + { + return compute(matrix, this->m_computationOptions); + } + + void setSwitchSize(int s) + { + eigen_assert(s>3 && "BDCSVD the size of the algo switch has to be greater than 4"); + algoswap = s; + } + + + /** \returns a (least squares) solution of \f$ A x = b \f$ using the current SVD decomposition of A. + * + * \param b the right - hand - side of the equation to solve. + * + * \note Solving requires both U and V to be computed. Thin U and V are enough, there is no need for full U or V. + * + * \note SVD solving is implicitly least - squares. Thus, this method serves both purposes of exact solving and least - squares solving. + * In other words, the returned solution is guaranteed to minimize the Euclidean norm \f$ \Vert A x - b \Vert \f$. + */ + template<typename Rhs> + inline const internal::solve_retval<BDCSVD, Rhs> + solve(const MatrixBase<Rhs>& b) const + { + eigen_assert(this->m_isInitialized && "BDCSVD is not initialized."); + eigen_assert(SVDBase<_MatrixType>::computeU() && SVDBase<_MatrixType>::computeV() && + "BDCSVD::solve() requires both unitaries U and V to be computed (thin unitaries suffice)."); + return internal::solve_retval<BDCSVD, Rhs>(*this, b.derived()); + } + + + const MatrixUType& matrixU() const + { + eigen_assert(this->m_isInitialized && "SVD is not initialized."); + if (isTranspose){ + eigen_assert(this->computeV() && "This SVD decomposition didn't compute U. Did you ask for it?"); + return this->m_matrixV; + } + else + { + eigen_assert(this->computeU() && "This SVD decomposition didn't compute U. Did you ask for it?"); + return this->m_matrixU; + } + + } + + + const MatrixVType& matrixV() const + { + eigen_assert(this->m_isInitialized && "SVD is not initialized."); + if (isTranspose){ + eigen_assert(this->computeU() && "This SVD decomposition didn't compute V. Did you ask for it?"); + return this->m_matrixU; + } + else + { + eigen_assert(this->computeV() && "This SVD decomposition didn't compute V. Did you ask for it?"); + return this->m_matrixV; + } + } + +private: + void allocate(Index rows, Index cols, unsigned int computationOptions); + void divide (Index firstCol, Index lastCol, Index firstRowW, + Index firstColW, Index shift); + void deflation43(Index firstCol, Index shift, Index i, Index size); + void deflation44(Index firstColu , Index firstColm, Index firstRowW, Index firstColW, Index i, Index j, Index size); + void deflation(Index firstCol, Index lastCol, Index k, Index firstRowW, Index firstColW, Index shift); + void copyUV(MatrixXr naiveU, MatrixXr naiveV, MatrixX householderU, MatrixX houseHolderV); + +protected: + MatrixXr m_naiveU, m_naiveV; + MatrixXr m_computed; + Index nRec; + int algoswap; + bool isTranspose, compU, compV; + +}; //end class BDCSVD + + +// Methode to allocate ans initialize matrix and attributs +template<typename MatrixType> +void BDCSVD<MatrixType>::allocate(Index rows, Index cols, unsigned int computationOptions) +{ + isTranspose = (cols > rows); + if (SVDBase<MatrixType>::allocate(rows, cols, computationOptions)) return; + m_computed = MatrixXr::Zero(this->m_diagSize + 1, this->m_diagSize ); + if (isTranspose){ + compU = this->computeU(); + compV = this->computeV(); + } + else + { + compV = this->computeU(); + compU = this->computeV(); + } + if (compU) m_naiveU = MatrixXr::Zero(this->m_diagSize + 1, this->m_diagSize + 1 ); + else m_naiveU = MatrixXr::Zero(2, this->m_diagSize + 1 ); + + if (compV) m_naiveV = MatrixXr::Zero(this->m_diagSize, this->m_diagSize); + + + //should be changed for a cleaner implementation + if (isTranspose){ + bool aux; + if (this->computeU()||this->computeV()){ + aux = this->m_computeFullU; + this->m_computeFullU = this->m_computeFullV; + this->m_computeFullV = aux; + aux = this->m_computeThinU; + this->m_computeThinU = this->m_computeThinV; + this->m_computeThinV = aux; + } + } +}// end allocate + +// Methode which compute the BDCSVD for the int +template<> +SVDBase<Matrix<int, Dynamic, Dynamic> >& +BDCSVD<Matrix<int, Dynamic, Dynamic> >::compute(const MatrixType& matrix, unsigned int computationOptions) { + allocate(matrix.rows(), matrix.cols(), computationOptions); + this->m_nonzeroSingularValues = 0; + m_computed = Matrix<int, Dynamic, Dynamic>::Zero(rows(), cols()); + for (int i=0; i<this->m_diagSize; i++) { + this->m_singularValues.coeffRef(i) = 0; + } + if (this->m_computeFullU) this->m_matrixU = Matrix<int, Dynamic, Dynamic>::Zero(rows(), rows()); + if (this->m_computeFullV) this->m_matrixV = Matrix<int, Dynamic, Dynamic>::Zero(cols(), cols()); + this->m_isInitialized = true; + return *this; +} + + +// Methode which compute the BDCSVD +template<typename MatrixType> +SVDBase<MatrixType>& +BDCSVD<MatrixType>::compute(const MatrixType& matrix, unsigned int computationOptions) +{ + allocate(matrix.rows(), matrix.cols(), computationOptions); + using std::abs; + + //**** step 1 Bidiagonalization isTranspose = (matrix.cols()>matrix.rows()) ; + MatrixType copy; + if (isTranspose) copy = matrix.adjoint(); + else copy = matrix; + + internal::UpperBidiagonalization<MatrixX > bid(copy); + + //**** step 2 Divide + // this is ugly and has to be redone (care of complex cast) + MatrixXr temp; + temp = bid.bidiagonal().toDenseMatrix().transpose(); + m_computed.setZero(); + for (int i=0; i<this->m_diagSize - 1; i++) { + m_computed(i, i) = temp(i, i); + m_computed(i + 1, i) = temp(i + 1, i); + } + m_computed(this->m_diagSize - 1, this->m_diagSize - 1) = temp(this->m_diagSize - 1, this->m_diagSize - 1); + divide(0, this->m_diagSize - 1, 0, 0, 0); + + //**** step 3 copy + for (int i=0; i<this->m_diagSize; i++) { + RealScalar a = abs(m_computed.coeff(i, i)); + this->m_singularValues.coeffRef(i) = a; + if (a == 0){ + this->m_nonzeroSingularValues = i; + break; + } + else if (i == this->m_diagSize - 1) + { + this->m_nonzeroSingularValues = i + 1; + break; + } + } + copyUV(m_naiveV, m_naiveU, bid.householderU(), bid.householderV()); + this->m_isInitialized = true; + return *this; +}// end compute + + +template<typename MatrixType> +void BDCSVD<MatrixType>::copyUV(MatrixXr naiveU, MatrixXr naiveV, MatrixX householderU, MatrixX householderV){ + if (this->computeU()){ + MatrixX temp = MatrixX::Zero(naiveU.rows(), naiveU.cols()); + temp.real() = naiveU; + if (this->m_computeThinU){ + this->m_matrixU = MatrixX::Identity(householderU.cols(), this->m_nonzeroSingularValues ); + this->m_matrixU.block(0, 0, this->m_diagSize, this->m_nonzeroSingularValues) = + temp.block(0, 0, this->m_diagSize, this->m_nonzeroSingularValues); + this->m_matrixU = householderU * this->m_matrixU ; + } + else + { + this->m_matrixU = MatrixX::Identity(householderU.cols(), householderU.cols()); + this->m_matrixU.block(0, 0, this->m_diagSize, this->m_diagSize) = temp.block(0, 0, this->m_diagSize, this->m_diagSize); + this->m_matrixU = householderU * this->m_matrixU ; + } + } + if (this->computeV()){ + MatrixX temp = MatrixX::Zero(naiveV.rows(), naiveV.cols()); + temp.real() = naiveV; + if (this->m_computeThinV){ + this->m_matrixV = MatrixX::Identity(householderV.cols(),this->m_nonzeroSingularValues ); + this->m_matrixV.block(0, 0, this->m_nonzeroSingularValues, this->m_nonzeroSingularValues) = + temp.block(0, 0, this->m_nonzeroSingularValues, this->m_nonzeroSingularValues); + this->m_matrixV = householderV * this->m_matrixV ; + } + else + { + this->m_matrixV = MatrixX::Identity(householderV.cols(), householderV.cols()); + this->m_matrixV.block(0, 0, this->m_diagSize, this->m_diagSize) = temp.block(0, 0, this->m_diagSize, this->m_diagSize); + this->m_matrixV = householderV * this->m_matrixV; + } + } +} + +// The divide algorithm is done "in place", we are always working on subsets of the same matrix. The divide methods takes as argument the +// place of the submatrix we are currently working on. + +//@param firstCol : The Index of the first column of the submatrix of m_computed and for m_naiveU; +//@param lastCol : The Index of the last column of the submatrix of m_computed and for m_naiveU; +// lastCol + 1 - firstCol is the size of the submatrix. +//@param firstRowW : The Index of the first row of the matrix W that we are to change. (see the reference paper section 1 for more information on W) +//@param firstRowW : Same as firstRowW with the column. +//@param shift : Each time one takes the left submatrix, one must add 1 to the shift. Why? Because! We actually want the last column of the U submatrix +// to become the first column (*coeff) and to shift all the other columns to the right. There are more details on the reference paper. +template<typename MatrixType> +void BDCSVD<MatrixType>::divide (Index firstCol, Index lastCol, Index firstRowW, + Index firstColW, Index shift) +{ + // requires nbRows = nbCols + 1; + using std::pow; + using std::sqrt; + using std::abs; + const Index n = lastCol - firstCol + 1; + const Index k = n/2; + RealScalar alphaK; + RealScalar betaK; + RealScalar r0; + RealScalar lambda, phi, c0, s0; + MatrixXr l, f; + // We use the other algorithm which is more efficient for small + // matrices. + if (n < algoswap){ + JacobiSVD<MatrixXr> b(m_computed.block(firstCol, firstCol, n + 1, n), + ComputeFullU | (ComputeFullV * compV)) ; + if (compU) m_naiveU.block(firstCol, firstCol, n + 1, n + 1).real() << b.matrixU(); + else + { + m_naiveU.row(0).segment(firstCol, n + 1).real() << b.matrixU().row(0); + m_naiveU.row(1).segment(firstCol, n + 1).real() << b.matrixU().row(n); + } + if (compV) m_naiveV.block(firstRowW, firstColW, n, n).real() << b.matrixV(); + m_computed.block(firstCol + shift, firstCol + shift, n + 1, n).setZero(); + for (int i=0; i<n; i++) + { + m_computed(firstCol + shift + i, firstCol + shift +i) = b.singularValues().coeffRef(i); + } + return; + } + // We use the divide and conquer algorithm + alphaK = m_computed(firstCol + k, firstCol + k); + betaK = m_computed(firstCol + k + 1, firstCol + k); + // The divide must be done in that order in order to have good results. Divide change the data inside the submatrices + // and the divide of the right submatrice reads one column of the left submatrice. That's why we need to treat the + // right submatrix before the left one. + divide(k + 1 + firstCol, lastCol, k + 1 + firstRowW, k + 1 + firstColW, shift); + divide(firstCol, k - 1 + firstCol, firstRowW, firstColW + 1, shift + 1); + if (compU) + { + lambda = m_naiveU(firstCol + k, firstCol + k); + phi = m_naiveU(firstCol + k + 1, lastCol + 1); + } + else + { + lambda = m_naiveU(1, firstCol + k); + phi = m_naiveU(0, lastCol + 1); + } + r0 = sqrt((abs(alphaK * lambda) * abs(alphaK * lambda)) + + abs(betaK * phi) * abs(betaK * phi)); + if (compU) + { + l = m_naiveU.row(firstCol + k).segment(firstCol, k); + f = m_naiveU.row(firstCol + k + 1).segment(firstCol + k + 1, n - k - 1); + } + else + { + l = m_naiveU.row(1).segment(firstCol, k); + f = m_naiveU.row(0).segment(firstCol + k + 1, n - k - 1); + } + if (compV) m_naiveV(firstRowW+k, firstColW) = 1; + if (r0 == 0) + { + c0 = 1; + s0 = 0; + } + else + { + c0 = alphaK * lambda / r0; + s0 = betaK * phi / r0; + } + if (compU) + { + MatrixXr q1 (m_naiveU.col(firstCol + k).segment(firstCol, k + 1)); + // we shiftW Q1 to the right + for (Index i = firstCol + k - 1; i >= firstCol; i--) + { + m_naiveU.col(i + 1).segment(firstCol, k + 1) << m_naiveU.col(i).segment(firstCol, k + 1); + } + // we shift q1 at the left with a factor c0 + m_naiveU.col(firstCol).segment( firstCol, k + 1) << (q1 * c0); + // last column = q1 * - s0 + m_naiveU.col(lastCol + 1).segment(firstCol, k + 1) << (q1 * ( - s0)); + // first column = q2 * s0 + m_naiveU.col(firstCol).segment(firstCol + k + 1, n - k) << + m_naiveU.col(lastCol + 1).segment(firstCol + k + 1, n - k) *s0; + // q2 *= c0 + m_naiveU.col(lastCol + 1).segment(firstCol + k + 1, n - k) *= c0; + } + else + { + RealScalar q1 = (m_naiveU(0, firstCol + k)); + // we shift Q1 to the right + for (Index i = firstCol + k - 1; i >= firstCol; i--) + { + m_naiveU(0, i + 1) = m_naiveU(0, i); + } + // we shift q1 at the left with a factor c0 + m_naiveU(0, firstCol) = (q1 * c0); + // last column = q1 * - s0 + m_naiveU(0, lastCol + 1) = (q1 * ( - s0)); + // first column = q2 * s0 + m_naiveU(1, firstCol) = m_naiveU(1, lastCol + 1) *s0; + // q2 *= c0 + m_naiveU(1, lastCol + 1) *= c0; + m_naiveU.row(1).segment(firstCol + 1, k).setZero(); + m_naiveU.row(0).segment(firstCol + k + 1, n - k - 1).setZero(); + } + m_computed(firstCol + shift, firstCol + shift) = r0; + m_computed.col(firstCol + shift).segment(firstCol + shift + 1, k) << alphaK * l.transpose().real(); + m_computed.col(firstCol + shift).segment(firstCol + shift + k + 1, n - k - 1) << betaK * f.transpose().real(); + + + // the line below do the deflation of the matrix for the third part of the algorithm + // Here the deflation is commented because the third part of the algorithm is not implemented + // the third part of the algorithm is a fast SVD on the matrix m_computed which works thanks to the deflation + + deflation(firstCol, lastCol, k, firstRowW, firstColW, shift); + + // Third part of the algorithm, since the real third part of the algorithm is not implemeted we use a JacobiSVD + JacobiSVD<MatrixXr> res= JacobiSVD<MatrixXr>(m_computed.block(firstCol + shift, firstCol +shift, n + 1, n), + ComputeFullU | (ComputeFullV * compV)) ; + if (compU) m_naiveU.block(firstCol, firstCol, n + 1, n + 1) *= res.matrixU(); + else m_naiveU.block(0, firstCol, 2, n + 1) *= res.matrixU(); + + if (compV) m_naiveV.block(firstRowW, firstColW, n, n) *= res.matrixV(); + m_computed.block(firstCol + shift, firstCol + shift, n, n) << MatrixXr::Zero(n, n); + for (int i=0; i<n; i++) + m_computed(firstCol + shift + i, firstCol + shift +i) = res.singularValues().coeffRef(i); + // end of the third part + + +}// end divide + + +// page 12_13 +// i >= 1, di almost null and zi non null. +// We use a rotation to zero out zi applied to the left of M +template <typename MatrixType> +void BDCSVD<MatrixType>::deflation43(Index firstCol, Index shift, Index i, Index size){ + using std::abs; + using std::sqrt; + using std::pow; + RealScalar c = m_computed(firstCol + shift, firstCol + shift); + RealScalar s = m_computed(i, firstCol + shift); + RealScalar r = sqrt(pow(abs(c), 2) + pow(abs(s), 2)); + if (r == 0){ + m_computed(i, i)=0; + return; + } + c/=r; + s/=r; + m_computed(firstCol + shift, firstCol + shift) = r; + m_computed(i, firstCol + shift) = 0; + m_computed(i, i) = 0; + if (compU){ + m_naiveU.col(firstCol).segment(firstCol,size) = + c * m_naiveU.col(firstCol).segment(firstCol, size) - + s * m_naiveU.col(i).segment(firstCol, size) ; + + m_naiveU.col(i).segment(firstCol, size) = + (c + s*s/c) * m_naiveU.col(i).segment(firstCol, size) + + (s/c) * m_naiveU.col(firstCol).segment(firstCol,size); + } +}// end deflation 43 + + +// page 13 +// i,j >= 1, i != j and |di - dj| < epsilon * norm2(M) +// We apply two rotations to have zj = 0; +template <typename MatrixType> +void BDCSVD<MatrixType>::deflation44(Index firstColu , Index firstColm, Index firstRowW, Index firstColW, Index i, Index j, Index size){ + using std::abs; + using std::sqrt; + using std::conj; + using std::pow; + RealScalar c = m_computed(firstColm, firstColm + j - 1); + RealScalar s = m_computed(firstColm, firstColm + i - 1); + RealScalar r = sqrt(pow(abs(c), 2) + pow(abs(s), 2)); + if (r==0){ + m_computed(firstColm + i, firstColm + i) = m_computed(firstColm + j, firstColm + j); + return; + } + c/=r; + s/=r; + m_computed(firstColm + i, firstColm) = r; + m_computed(firstColm + i, firstColm + i) = m_computed(firstColm + j, firstColm + j); + m_computed(firstColm + j, firstColm) = 0; + if (compU){ + m_naiveU.col(firstColu + i).segment(firstColu, size) = + c * m_naiveU.col(firstColu + i).segment(firstColu, size) - + s * m_naiveU.col(firstColu + j).segment(firstColu, size) ; + + m_naiveU.col(firstColu + j).segment(firstColu, size) = + (c + s*s/c) * m_naiveU.col(firstColu + j).segment(firstColu, size) + + (s/c) * m_naiveU.col(firstColu + i).segment(firstColu, size); + } + if (compV){ + m_naiveV.col(firstColW + i).segment(firstRowW, size - 1) = + c * m_naiveV.col(firstColW + i).segment(firstRowW, size - 1) + + s * m_naiveV.col(firstColW + j).segment(firstRowW, size - 1) ; + + m_naiveV.col(firstColW + j).segment(firstRowW, size - 1) = + (c + s*s/c) * m_naiveV.col(firstColW + j).segment(firstRowW, size - 1) - + (s/c) * m_naiveV.col(firstColW + i).segment(firstRowW, size - 1); + } +}// end deflation 44 + + + +template <typename MatrixType> +void BDCSVD<MatrixType>::deflation(Index firstCol, Index lastCol, Index k, Index firstRowW, Index firstColW, Index shift){ + //condition 4.1 + RealScalar EPS = EPSILON * (std::max<RealScalar>(m_computed(firstCol + shift + 1, firstCol + shift + 1), m_computed(firstCol + k, firstCol + k))); + const Index length = lastCol + 1 - firstCol; + if (m_computed(firstCol + shift, firstCol + shift) < EPS){ + m_computed(firstCol + shift, firstCol + shift) = EPS; + } + //condition 4.2 + for (Index i=firstCol + shift + 1;i<=lastCol + shift;i++){ + if (std::abs(m_computed(i, firstCol + shift)) < EPS){ + m_computed(i, firstCol + shift) = 0; + } + } + + //condition 4.3 + for (Index i=firstCol + shift + 1;i<=lastCol + shift; i++){ + if (m_computed(i, i) < EPS){ + deflation43(firstCol, shift, i, length); + } + } + + //condition 4.4 + + Index i=firstCol + shift + 1, j=firstCol + shift + k + 1; + //we stock the final place of each line + Index *permutation = new Index[length]; + + for (Index p =1; p < length; p++) { + if (i> firstCol + shift + k){ + permutation[p] = j; + j++; + } else if (j> lastCol + shift) + { + permutation[p] = i; + i++; + } + else + { + if (m_computed(i, i) < m_computed(j, j)){ + permutation[p] = j; + j++; + } + else + { + permutation[p] = i; + i++; + } + } + } + //we do the permutation + RealScalar aux; + //we stock the current index of each col + //and the column of each index + Index *realInd = new Index[length]; + Index *realCol = new Index[length]; + for (int pos = 0; pos< length; pos++){ + realCol[pos] = pos + firstCol + shift; + realInd[pos] = pos; + } + const Index Zero = firstCol + shift; + VectorType temp; + for (int i = 1; i < length - 1; i++){ + const Index I = i + Zero; + const Index realI = realInd[i]; + const Index j = permutation[length - i] - Zero; + const Index J = realCol[j]; + + //diag displace + aux = m_computed(I, I); + m_computed(I, I) = m_computed(J, J); + m_computed(J, J) = aux; + + //firstrow displace + aux = m_computed(I, Zero); + m_computed(I, Zero) = m_computed(J, Zero); + m_computed(J, Zero) = aux; + + // change columns + if (compU) { + temp = m_naiveU.col(I - shift).segment(firstCol, length + 1); + m_naiveU.col(I - shift).segment(firstCol, length + 1) << + m_naiveU.col(J - shift).segment(firstCol, length + 1); + m_naiveU.col(J - shift).segment(firstCol, length + 1) << temp; + } + else + { + temp = m_naiveU.col(I - shift).segment(0, 2); + m_naiveU.col(I - shift).segment(0, 2) << + m_naiveU.col(J - shift).segment(0, 2); + m_naiveU.col(J - shift).segment(0, 2) << temp; + } + if (compV) { + const Index CWI = I + firstColW - Zero; + const Index CWJ = J + firstColW - Zero; + temp = m_naiveV.col(CWI).segment(firstRowW, length); + m_naiveV.col(CWI).segment(firstRowW, length) << m_naiveV.col(CWJ).segment(firstRowW, length); + m_naiveV.col(CWJ).segment(firstRowW, length) << temp; + } + + //update real pos + realCol[realI] = J; + realCol[j] = I; + realInd[J - Zero] = realI; + realInd[I - Zero] = j; + } + for (Index i = firstCol + shift + 1; i<lastCol + shift;i++){ + if ((m_computed(i + 1, i + 1) - m_computed(i, i)) < EPS){ + deflation44(firstCol , + firstCol + shift, + firstRowW, + firstColW, + i - Zero, + i + 1 - Zero, + length); + } + } + delete [] permutation; + delete [] realInd; + delete [] realCol; + +}//end deflation + + +namespace internal{ + +template<typename _MatrixType, typename Rhs> +struct solve_retval<BDCSVD<_MatrixType>, Rhs> + : solve_retval_base<BDCSVD<_MatrixType>, Rhs> +{ + typedef BDCSVD<_MatrixType> BDCSVDType; + EIGEN_MAKE_SOLVE_HELPERS(BDCSVDType, Rhs) + + template<typename Dest> void evalTo(Dest& dst) const + { + eigen_assert(rhs().rows() == dec().rows()); + // A = U S V^* + // So A^{ - 1} = V S^{ - 1} U^* + Index diagSize = (std::min)(dec().rows(), dec().cols()); + typename BDCSVDType::SingularValuesType invertedSingVals(diagSize); + Index nonzeroSingVals = dec().nonzeroSingularValues(); + invertedSingVals.head(nonzeroSingVals) = dec().singularValues().head(nonzeroSingVals).array().inverse(); + invertedSingVals.tail(diagSize - nonzeroSingVals).setZero(); + + dst = dec().matrixV().leftCols(diagSize) + * invertedSingVals.asDiagonal() + * dec().matrixU().leftCols(diagSize).adjoint() + * rhs(); + return; + } +}; + +} //end namespace internal + + /** \svd_module + * + * \return the singular value decomposition of \c *this computed by + * BDC Algorithm + * + * \sa class BDCSVD + */ +/* +template<typename Derived> +BDCSVD<typename MatrixBase<Derived>::PlainObject> +MatrixBase<Derived>::bdcSvd(unsigned int computationOptions) const +{ + return BDCSVD<PlainObject>(*this, computationOptions); +} +*/ + +} // end namespace Eigen + +#endif diff --git a/unsupported/Eigen/src/SVD/CMakeLists.txt b/unsupported/Eigen/src/SVD/CMakeLists.txt new file mode 100644 index 000000000..b40baf092 --- /dev/null +++ b/unsupported/Eigen/src/SVD/CMakeLists.txt @@ -0,0 +1,6 @@ +FILE(GLOB Eigen_SVD_SRCS "*.h") + +INSTALL(FILES + ${Eigen_SVD_SRCS} + DESTINATION ${INCLUDE_INSTALL_DIR}unsupported/Eigen/src/SVD COMPONENT Devel + ) diff --git a/unsupported/Eigen/src/SVD/JacobiSVD.h b/unsupported/Eigen/src/SVD/JacobiSVD.h new file mode 100644 index 000000000..02fac409e --- /dev/null +++ b/unsupported/Eigen/src/SVD/JacobiSVD.h @@ -0,0 +1,782 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2009-2010 Benoit Jacob <jacob.benoit.1@gmail.com> +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + +#ifndef EIGEN_JACOBISVD_H +#define EIGEN_JACOBISVD_H + +namespace Eigen { + +namespace internal { +// forward declaration (needed by ICC) +// the empty body is required by MSVC +template<typename MatrixType, int QRPreconditioner, + bool IsComplex = NumTraits<typename MatrixType::Scalar>::IsComplex> +struct svd_precondition_2x2_block_to_be_real {}; + +/*** QR preconditioners (R-SVD) + *** + *** Their role is to reduce the problem of computing the SVD to the case of a square matrix. + *** This approach, known as R-SVD, is an optimization for rectangular-enough matrices, and is a requirement for + *** JacobiSVD which by itself is only able to work on square matrices. + ***/ + +enum { PreconditionIfMoreColsThanRows, PreconditionIfMoreRowsThanCols }; + +template<typename MatrixType, int QRPreconditioner, int Case> +struct qr_preconditioner_should_do_anything +{ + enum { a = MatrixType::RowsAtCompileTime != Dynamic && + MatrixType::ColsAtCompileTime != Dynamic && + MatrixType::ColsAtCompileTime <= MatrixType::RowsAtCompileTime, + b = MatrixType::RowsAtCompileTime != Dynamic && + MatrixType::ColsAtCompileTime != Dynamic && + MatrixType::RowsAtCompileTime <= MatrixType::ColsAtCompileTime, + ret = !( (QRPreconditioner == NoQRPreconditioner) || + (Case == PreconditionIfMoreColsThanRows && bool(a)) || + (Case == PreconditionIfMoreRowsThanCols && bool(b)) ) + }; +}; + +template<typename MatrixType, int QRPreconditioner, int Case, + bool DoAnything = qr_preconditioner_should_do_anything<MatrixType, QRPreconditioner, Case>::ret +> struct qr_preconditioner_impl {}; + +template<typename MatrixType, int QRPreconditioner, int Case> +class qr_preconditioner_impl<MatrixType, QRPreconditioner, Case, false> +{ +public: + typedef typename MatrixType::Index Index; + void allocate(const JacobiSVD<MatrixType, QRPreconditioner>&) {} + bool run(JacobiSVD<MatrixType, QRPreconditioner>&, const MatrixType&) + { + return false; + } +}; + +/*** preconditioner using FullPivHouseholderQR ***/ + +template<typename MatrixType> +class qr_preconditioner_impl<MatrixType, FullPivHouseholderQRPreconditioner, PreconditionIfMoreRowsThanCols, true> +{ +public: + typedef typename MatrixType::Index Index; + typedef typename MatrixType::Scalar Scalar; + enum + { + RowsAtCompileTime = MatrixType::RowsAtCompileTime, + MaxRowsAtCompileTime = MatrixType::MaxRowsAtCompileTime + }; + typedef Matrix<Scalar, 1, RowsAtCompileTime, RowMajor, 1, MaxRowsAtCompileTime> WorkspaceType; + + void allocate(const JacobiSVD<MatrixType, FullPivHouseholderQRPreconditioner>& svd) + { + if (svd.rows() != m_qr.rows() || svd.cols() != m_qr.cols()) + { + m_qr.~QRType(); + ::new (&m_qr) QRType(svd.rows(), svd.cols()); + } + if (svd.m_computeFullU) m_workspace.resize(svd.rows()); + } + + bool run(JacobiSVD<MatrixType, FullPivHouseholderQRPreconditioner>& svd, const MatrixType& matrix) + { + if(matrix.rows() > matrix.cols()) + { + m_qr.compute(matrix); + svd.m_workMatrix = m_qr.matrixQR().block(0,0,matrix.cols(),matrix.cols()).template triangularView<Upper>(); + if(svd.m_computeFullU) m_qr.matrixQ().evalTo(svd.m_matrixU, m_workspace); + if(svd.computeV()) svd.m_matrixV = m_qr.colsPermutation(); + return true; + } + return false; + } +private: + typedef FullPivHouseholderQR<MatrixType> QRType; + QRType m_qr; + WorkspaceType m_workspace; +}; + +template<typename MatrixType> +class qr_preconditioner_impl<MatrixType, FullPivHouseholderQRPreconditioner, PreconditionIfMoreColsThanRows, true> +{ +public: + typedef typename MatrixType::Index Index; + typedef typename MatrixType::Scalar Scalar; + enum + { + RowsAtCompileTime = MatrixType::RowsAtCompileTime, + ColsAtCompileTime = MatrixType::ColsAtCompileTime, + MaxRowsAtCompileTime = MatrixType::MaxRowsAtCompileTime, + MaxColsAtCompileTime = MatrixType::MaxColsAtCompileTime, + Options = MatrixType::Options + }; + typedef Matrix<Scalar, ColsAtCompileTime, RowsAtCompileTime, Options, MaxColsAtCompileTime, MaxRowsAtCompileTime> + TransposeTypeWithSameStorageOrder; + + void allocate(const JacobiSVD<MatrixType, FullPivHouseholderQRPreconditioner>& svd) + { + if (svd.cols() != m_qr.rows() || svd.rows() != m_qr.cols()) + { + m_qr.~QRType(); + ::new (&m_qr) QRType(svd.cols(), svd.rows()); + } + m_adjoint.resize(svd.cols(), svd.rows()); + if (svd.m_computeFullV) m_workspace.resize(svd.cols()); + } + + bool run(JacobiSVD<MatrixType, FullPivHouseholderQRPreconditioner>& svd, const MatrixType& matrix) + { + if(matrix.cols() > matrix.rows()) + { + m_adjoint = matrix.adjoint(); + m_qr.compute(m_adjoint); + svd.m_workMatrix = m_qr.matrixQR().block(0,0,matrix.rows(),matrix.rows()).template triangularView<Upper>().adjoint(); + if(svd.m_computeFullV) m_qr.matrixQ().evalTo(svd.m_matrixV, m_workspace); + if(svd.computeU()) svd.m_matrixU = m_qr.colsPermutation(); + return true; + } + else return false; + } +private: + typedef FullPivHouseholderQR<TransposeTypeWithSameStorageOrder> QRType; + QRType m_qr; + TransposeTypeWithSameStorageOrder m_adjoint; + typename internal::plain_row_type<MatrixType>::type m_workspace; +}; + +/*** preconditioner using ColPivHouseholderQR ***/ + +template<typename MatrixType> +class qr_preconditioner_impl<MatrixType, ColPivHouseholderQRPreconditioner, PreconditionIfMoreRowsThanCols, true> +{ +public: + typedef typename MatrixType::Index Index; + + void allocate(const JacobiSVD<MatrixType, ColPivHouseholderQRPreconditioner>& svd) + { + if (svd.rows() != m_qr.rows() || svd.cols() != m_qr.cols()) + { + m_qr.~QRType(); + ::new (&m_qr) QRType(svd.rows(), svd.cols()); + } + if (svd.m_computeFullU) m_workspace.resize(svd.rows()); + else if (svd.m_computeThinU) m_workspace.resize(svd.cols()); + } + + bool run(JacobiSVD<MatrixType, ColPivHouseholderQRPreconditioner>& svd, const MatrixType& matrix) + { + if(matrix.rows() > matrix.cols()) + { + m_qr.compute(matrix); + svd.m_workMatrix = m_qr.matrixQR().block(0,0,matrix.cols(),matrix.cols()).template triangularView<Upper>(); + if(svd.m_computeFullU) m_qr.householderQ().evalTo(svd.m_matrixU, m_workspace); + else if(svd.m_computeThinU) + { + svd.m_matrixU.setIdentity(matrix.rows(), matrix.cols()); + m_qr.householderQ().applyThisOnTheLeft(svd.m_matrixU, m_workspace); + } + if(svd.computeV()) svd.m_matrixV = m_qr.colsPermutation(); + return true; + } + return false; + } + +private: + typedef ColPivHouseholderQR<MatrixType> QRType; + QRType m_qr; + typename internal::plain_col_type<MatrixType>::type m_workspace; +}; + +template<typename MatrixType> +class qr_preconditioner_impl<MatrixType, ColPivHouseholderQRPreconditioner, PreconditionIfMoreColsThanRows, true> +{ +public: + typedef typename MatrixType::Index Index; + typedef typename MatrixType::Scalar Scalar; + enum + { + RowsAtCompileTime = MatrixType::RowsAtCompileTime, + ColsAtCompileTime = MatrixType::ColsAtCompileTime, + MaxRowsAtCompileTime = MatrixType::MaxRowsAtCompileTime, + MaxColsAtCompileTime = MatrixType::MaxColsAtCompileTime, + Options = MatrixType::Options + }; + + typedef Matrix<Scalar, ColsAtCompileTime, RowsAtCompileTime, Options, MaxColsAtCompileTime, MaxRowsAtCompileTime> + TransposeTypeWithSameStorageOrder; + + void allocate(const JacobiSVD<MatrixType, ColPivHouseholderQRPreconditioner>& svd) + { + if (svd.cols() != m_qr.rows() || svd.rows() != m_qr.cols()) + { + m_qr.~QRType(); + ::new (&m_qr) QRType(svd.cols(), svd.rows()); + } + if (svd.m_computeFullV) m_workspace.resize(svd.cols()); + else if (svd.m_computeThinV) m_workspace.resize(svd.rows()); + m_adjoint.resize(svd.cols(), svd.rows()); + } + + bool run(JacobiSVD<MatrixType, ColPivHouseholderQRPreconditioner>& svd, const MatrixType& matrix) + { + if(matrix.cols() > matrix.rows()) + { + m_adjoint = matrix.adjoint(); + m_qr.compute(m_adjoint); + + svd.m_workMatrix = m_qr.matrixQR().block(0,0,matrix.rows(),matrix.rows()).template triangularView<Upper>().adjoint(); + if(svd.m_computeFullV) m_qr.householderQ().evalTo(svd.m_matrixV, m_workspace); + else if(svd.m_computeThinV) + { + svd.m_matrixV.setIdentity(matrix.cols(), matrix.rows()); + m_qr.householderQ().applyThisOnTheLeft(svd.m_matrixV, m_workspace); + } + if(svd.computeU()) svd.m_matrixU = m_qr.colsPermutation(); + return true; + } + else return false; + } + +private: + typedef ColPivHouseholderQR<TransposeTypeWithSameStorageOrder> QRType; + QRType m_qr; + TransposeTypeWithSameStorageOrder m_adjoint; + typename internal::plain_row_type<MatrixType>::type m_workspace; +}; + +/*** preconditioner using HouseholderQR ***/ + +template<typename MatrixType> +class qr_preconditioner_impl<MatrixType, HouseholderQRPreconditioner, PreconditionIfMoreRowsThanCols, true> +{ +public: + typedef typename MatrixType::Index Index; + + void allocate(const JacobiSVD<MatrixType, HouseholderQRPreconditioner>& svd) + { + if (svd.rows() != m_qr.rows() || svd.cols() != m_qr.cols()) + { + m_qr.~QRType(); + ::new (&m_qr) QRType(svd.rows(), svd.cols()); + } + if (svd.m_computeFullU) m_workspace.resize(svd.rows()); + else if (svd.m_computeThinU) m_workspace.resize(svd.cols()); + } + + bool run(JacobiSVD<MatrixType, HouseholderQRPreconditioner>& svd, const MatrixType& matrix) + { + if(matrix.rows() > matrix.cols()) + { + m_qr.compute(matrix); + svd.m_workMatrix = m_qr.matrixQR().block(0,0,matrix.cols(),matrix.cols()).template triangularView<Upper>(); + if(svd.m_computeFullU) m_qr.householderQ().evalTo(svd.m_matrixU, m_workspace); + else if(svd.m_computeThinU) + { + svd.m_matrixU.setIdentity(matrix.rows(), matrix.cols()); + m_qr.householderQ().applyThisOnTheLeft(svd.m_matrixU, m_workspace); + } + if(svd.computeV()) svd.m_matrixV.setIdentity(matrix.cols(), matrix.cols()); + return true; + } + return false; + } +private: + typedef HouseholderQR<MatrixType> QRType; + QRType m_qr; + typename internal::plain_col_type<MatrixType>::type m_workspace; +}; + +template<typename MatrixType> +class qr_preconditioner_impl<MatrixType, HouseholderQRPreconditioner, PreconditionIfMoreColsThanRows, true> +{ +public: + typedef typename MatrixType::Index Index; + typedef typename MatrixType::Scalar Scalar; + enum + { + RowsAtCompileTime = MatrixType::RowsAtCompileTime, + ColsAtCompileTime = MatrixType::ColsAtCompileTime, + MaxRowsAtCompileTime = MatrixType::MaxRowsAtCompileTime, + MaxColsAtCompileTime = MatrixType::MaxColsAtCompileTime, + Options = MatrixType::Options + }; + + typedef Matrix<Scalar, ColsAtCompileTime, RowsAtCompileTime, Options, MaxColsAtCompileTime, MaxRowsAtCompileTime> + TransposeTypeWithSameStorageOrder; + + void allocate(const JacobiSVD<MatrixType, HouseholderQRPreconditioner>& svd) + { + if (svd.cols() != m_qr.rows() || svd.rows() != m_qr.cols()) + { + m_qr.~QRType(); + ::new (&m_qr) QRType(svd.cols(), svd.rows()); + } + if (svd.m_computeFullV) m_workspace.resize(svd.cols()); + else if (svd.m_computeThinV) m_workspace.resize(svd.rows()); + m_adjoint.resize(svd.cols(), svd.rows()); + } + + bool run(JacobiSVD<MatrixType, HouseholderQRPreconditioner>& svd, const MatrixType& matrix) + { + if(matrix.cols() > matrix.rows()) + { + m_adjoint = matrix.adjoint(); + m_qr.compute(m_adjoint); + + svd.m_workMatrix = m_qr.matrixQR().block(0,0,matrix.rows(),matrix.rows()).template triangularView<Upper>().adjoint(); + if(svd.m_computeFullV) m_qr.householderQ().evalTo(svd.m_matrixV, m_workspace); + else if(svd.m_computeThinV) + { + svd.m_matrixV.setIdentity(matrix.cols(), matrix.rows()); + m_qr.householderQ().applyThisOnTheLeft(svd.m_matrixV, m_workspace); + } + if(svd.computeU()) svd.m_matrixU.setIdentity(matrix.rows(), matrix.rows()); + return true; + } + else return false; + } + +private: + typedef HouseholderQR<TransposeTypeWithSameStorageOrder> QRType; + QRType m_qr; + TransposeTypeWithSameStorageOrder m_adjoint; + typename internal::plain_row_type<MatrixType>::type m_workspace; +}; + +/*** 2x2 SVD implementation + *** + *** JacobiSVD consists in performing a series of 2x2 SVD subproblems + ***/ + +template<typename MatrixType, int QRPreconditioner> +struct svd_precondition_2x2_block_to_be_real<MatrixType, QRPreconditioner, false> +{ + typedef JacobiSVD<MatrixType, QRPreconditioner> SVD; + typedef typename SVD::Index Index; + static void run(typename SVD::WorkMatrixType&, SVD&, Index, Index) {} +}; + +template<typename MatrixType, int QRPreconditioner> +struct svd_precondition_2x2_block_to_be_real<MatrixType, QRPreconditioner, true> +{ + typedef JacobiSVD<MatrixType, QRPreconditioner> SVD; + typedef typename MatrixType::Scalar Scalar; + typedef typename MatrixType::RealScalar RealScalar; + typedef typename SVD::Index Index; + static void run(typename SVD::WorkMatrixType& work_matrix, SVD& svd, Index p, Index q) + { + using std::sqrt; + Scalar z; + JacobiRotation<Scalar> rot; + RealScalar n = sqrt(numext::abs2(work_matrix.coeff(p,p)) + numext::abs2(work_matrix.coeff(q,p))); + if(n==0) + { + z = abs(work_matrix.coeff(p,q)) / work_matrix.coeff(p,q); + work_matrix.row(p) *= z; + if(svd.computeU()) svd.m_matrixU.col(p) *= conj(z); + z = abs(work_matrix.coeff(q,q)) / work_matrix.coeff(q,q); + work_matrix.row(q) *= z; + if(svd.computeU()) svd.m_matrixU.col(q) *= conj(z); + } + else + { + rot.c() = conj(work_matrix.coeff(p,p)) / n; + rot.s() = work_matrix.coeff(q,p) / n; + work_matrix.applyOnTheLeft(p,q,rot); + if(svd.computeU()) svd.m_matrixU.applyOnTheRight(p,q,rot.adjoint()); + if(work_matrix.coeff(p,q) != Scalar(0)) + { + Scalar z = abs(work_matrix.coeff(p,q)) / work_matrix.coeff(p,q); + work_matrix.col(q) *= z; + if(svd.computeV()) svd.m_matrixV.col(q) *= z; + } + if(work_matrix.coeff(q,q) != Scalar(0)) + { + z = abs(work_matrix.coeff(q,q)) / work_matrix.coeff(q,q); + work_matrix.row(q) *= z; + if(svd.computeU()) svd.m_matrixU.col(q) *= conj(z); + } + } + } +}; + +template<typename MatrixType, typename RealScalar, typename Index> +void real_2x2_jacobi_svd(const MatrixType& matrix, Index p, Index q, + JacobiRotation<RealScalar> *j_left, + JacobiRotation<RealScalar> *j_right) +{ + using std::sqrt; + Matrix<RealScalar,2,2> m; + m << numext::real(matrix.coeff(p,p)), numext::real(matrix.coeff(p,q)), + numext::real(matrix.coeff(q,p)), numext::real(matrix.coeff(q,q)); + JacobiRotation<RealScalar> rot1; + RealScalar t = m.coeff(0,0) + m.coeff(1,1); + RealScalar d = m.coeff(1,0) - m.coeff(0,1); + if(t == RealScalar(0)) + { + rot1.c() = RealScalar(0); + rot1.s() = d > RealScalar(0) ? RealScalar(1) : RealScalar(-1); + } + else + { + RealScalar u = d / t; + rot1.c() = RealScalar(1) / sqrt(RealScalar(1) + numext::abs2(u)); + rot1.s() = rot1.c() * u; + } + m.applyOnTheLeft(0,1,rot1); + j_right->makeJacobi(m,0,1); + *j_left = rot1 * j_right->transpose(); +} + +} // end namespace internal + +/** \ingroup SVD_Module + * + * + * \class JacobiSVD + * + * \brief Two-sided Jacobi SVD decomposition of a rectangular matrix + * + * \param MatrixType the type of the matrix of which we are computing the SVD decomposition + * \param QRPreconditioner this optional parameter allows to specify the type of QR decomposition that will be used internally + * for the R-SVD step for non-square matrices. See discussion of possible values below. + * + * SVD decomposition consists in decomposing any n-by-p matrix \a A as a product + * \f[ A = U S V^* \f] + * where \a U is a n-by-n unitary, \a V is a p-by-p unitary, and \a S is a n-by-p real positive matrix which is zero outside of its main diagonal; + * the diagonal entries of S are known as the \em singular \em values of \a A and the columns of \a U and \a V are known as the left + * and right \em singular \em vectors of \a A respectively. + * + * Singular values are always sorted in decreasing order. + * + * This JacobiSVD decomposition computes only the singular values by default. If you want \a U or \a V, you need to ask for them explicitly. + * + * You can ask for only \em thin \a U or \a V to be computed, meaning the following. In case of a rectangular n-by-p matrix, letting \a m be the + * smaller value among \a n and \a p, there are only \a m singular vectors; the remaining columns of \a U and \a V do not correspond to actual + * singular vectors. Asking for \em thin \a U or \a V means asking for only their \a m first columns to be formed. So \a U is then a n-by-m matrix, + * and \a V is then a p-by-m matrix. Notice that thin \a U and \a V are all you need for (least squares) solving. + * + * Here's an example demonstrating basic usage: + * \include JacobiSVD_basic.cpp + * Output: \verbinclude JacobiSVD_basic.out + * + * This JacobiSVD class is a two-sided Jacobi R-SVD decomposition, ensuring optimal reliability and accuracy. The downside is that it's slower than + * bidiagonalizing SVD algorithms for large square matrices; however its complexity is still \f$ O(n^2p) \f$ where \a n is the smaller dimension and + * \a p is the greater dimension, meaning that it is still of the same order of complexity as the faster bidiagonalizing R-SVD algorithms. + * In particular, like any R-SVD, it takes advantage of non-squareness in that its complexity is only linear in the greater dimension. + * + * If the input matrix has inf or nan coefficients, the result of the computation is undefined, but the computation is guaranteed to + * terminate in finite (and reasonable) time. + * + * The possible values for QRPreconditioner are: + * \li ColPivHouseholderQRPreconditioner is the default. In practice it's very safe. It uses column-pivoting QR. + * \li FullPivHouseholderQRPreconditioner, is the safest and slowest. It uses full-pivoting QR. + * Contrary to other QRs, it doesn't allow computing thin unitaries. + * \li HouseholderQRPreconditioner is the fastest, and less safe and accurate than the pivoting variants. It uses non-pivoting QR. + * This is very similar in safety and accuracy to the bidiagonalization process used by bidiagonalizing SVD algorithms (since bidiagonalization + * is inherently non-pivoting). However the resulting SVD is still more reliable than bidiagonalizing SVDs because the Jacobi-based iterarive + * process is more reliable than the optimized bidiagonal SVD iterations. + * \li NoQRPreconditioner allows not to use a QR preconditioner at all. This is useful if you know that you will only be computing + * JacobiSVD decompositions of square matrices. Non-square matrices require a QR preconditioner. Using this option will result in + * faster compilation and smaller executable code. It won't significantly speed up computation, since JacobiSVD is always checking + * if QR preconditioning is needed before applying it anyway. + * + * \sa MatrixBase::jacobiSvd() + */ +template<typename _MatrixType, int QRPreconditioner> +class JacobiSVD : public SVDBase<_MatrixType> +{ + public: + + typedef _MatrixType MatrixType; + typedef typename MatrixType::Scalar Scalar; + typedef typename NumTraits<typename MatrixType::Scalar>::Real RealScalar; + typedef typename MatrixType::Index Index; + enum { + RowsAtCompileTime = MatrixType::RowsAtCompileTime, + ColsAtCompileTime = MatrixType::ColsAtCompileTime, + DiagSizeAtCompileTime = EIGEN_SIZE_MIN_PREFER_DYNAMIC(RowsAtCompileTime,ColsAtCompileTime), + MaxRowsAtCompileTime = MatrixType::MaxRowsAtCompileTime, + MaxColsAtCompileTime = MatrixType::MaxColsAtCompileTime, + MaxDiagSizeAtCompileTime = EIGEN_SIZE_MIN_PREFER_FIXED(MaxRowsAtCompileTime,MaxColsAtCompileTime), + MatrixOptions = MatrixType::Options + }; + + typedef Matrix<Scalar, RowsAtCompileTime, RowsAtCompileTime, + MatrixOptions, MaxRowsAtCompileTime, MaxRowsAtCompileTime> + MatrixUType; + typedef Matrix<Scalar, ColsAtCompileTime, ColsAtCompileTime, + MatrixOptions, MaxColsAtCompileTime, MaxColsAtCompileTime> + MatrixVType; + typedef typename internal::plain_diag_type<MatrixType, RealScalar>::type SingularValuesType; + typedef typename internal::plain_row_type<MatrixType>::type RowType; + typedef typename internal::plain_col_type<MatrixType>::type ColType; + typedef Matrix<Scalar, DiagSizeAtCompileTime, DiagSizeAtCompileTime, + MatrixOptions, MaxDiagSizeAtCompileTime, MaxDiagSizeAtCompileTime> + WorkMatrixType; + + /** \brief Default Constructor. + * + * The default constructor is useful in cases in which the user intends to + * perform decompositions via JacobiSVD::compute(const MatrixType&). + */ + JacobiSVD() + : SVDBase<_MatrixType>::SVDBase() + {} + + + /** \brief Default Constructor with memory preallocation + * + * Like the default constructor but with preallocation of the internal data + * according to the specified problem size. + * \sa JacobiSVD() + */ + JacobiSVD(Index rows, Index cols, unsigned int computationOptions = 0) + : SVDBase<_MatrixType>::SVDBase() + { + allocate(rows, cols, computationOptions); + } + + /** \brief Constructor performing the decomposition of given matrix. + * + * \param matrix the matrix to decompose + * \param computationOptions optional parameter allowing to specify if you want full or thin U or V unitaries to be computed. + * By default, none is computed. This is a bit-field, the possible bits are #ComputeFullU, #ComputeThinU, + * #ComputeFullV, #ComputeThinV. + * + * Thin unitaries are only available if your matrix type has a Dynamic number of columns (for example MatrixXf). They also are not + * available with the (non-default) FullPivHouseholderQR preconditioner. + */ + JacobiSVD(const MatrixType& matrix, unsigned int computationOptions = 0) + : SVDBase<_MatrixType>::SVDBase() + { + compute(matrix, computationOptions); + } + + /** \brief Method performing the decomposition of given matrix using custom options. + * + * \param matrix the matrix to decompose + * \param computationOptions optional parameter allowing to specify if you want full or thin U or V unitaries to be computed. + * By default, none is computed. This is a bit-field, the possible bits are #ComputeFullU, #ComputeThinU, + * #ComputeFullV, #ComputeThinV. + * + * Thin unitaries are only available if your matrix type has a Dynamic number of columns (for example MatrixXf). They also are not + * available with the (non-default) FullPivHouseholderQR preconditioner. + */ + SVDBase<MatrixType>& compute(const MatrixType& matrix, unsigned int computationOptions); + + /** \brief Method performing the decomposition of given matrix using current options. + * + * \param matrix the matrix to decompose + * + * This method uses the current \a computationOptions, as already passed to the constructor or to compute(const MatrixType&, unsigned int). + */ + SVDBase<MatrixType>& compute(const MatrixType& matrix) + { + return compute(matrix, this->m_computationOptions); + } + + /** \returns a (least squares) solution of \f$ A x = b \f$ using the current SVD decomposition of A. + * + * \param b the right-hand-side of the equation to solve. + * + * \note Solving requires both U and V to be computed. Thin U and V are enough, there is no need for full U or V. + * + * \note SVD solving is implicitly least-squares. Thus, this method serves both purposes of exact solving and least-squares solving. + * In other words, the returned solution is guaranteed to minimize the Euclidean norm \f$ \Vert A x - b \Vert \f$. + */ + template<typename Rhs> + inline const internal::solve_retval<JacobiSVD, Rhs> + solve(const MatrixBase<Rhs>& b) const + { + eigen_assert(this->m_isInitialized && "JacobiSVD is not initialized."); + eigen_assert(SVDBase<MatrixType>::computeU() && SVDBase<MatrixType>::computeV() && "JacobiSVD::solve() requires both unitaries U and V to be computed (thin unitaries suffice)."); + return internal::solve_retval<JacobiSVD, Rhs>(*this, b.derived()); + } + + + + private: + void allocate(Index rows, Index cols, unsigned int computationOptions); + + protected: + WorkMatrixType m_workMatrix; + + template<typename __MatrixType, int _QRPreconditioner, bool _IsComplex> + friend struct internal::svd_precondition_2x2_block_to_be_real; + template<typename __MatrixType, int _QRPreconditioner, int _Case, bool _DoAnything> + friend struct internal::qr_preconditioner_impl; + + internal::qr_preconditioner_impl<MatrixType, QRPreconditioner, internal::PreconditionIfMoreColsThanRows> m_qr_precond_morecols; + internal::qr_preconditioner_impl<MatrixType, QRPreconditioner, internal::PreconditionIfMoreRowsThanCols> m_qr_precond_morerows; +}; + +template<typename MatrixType, int QRPreconditioner> +void JacobiSVD<MatrixType, QRPreconditioner>::allocate(Index rows, Index cols, unsigned int computationOptions) +{ + if (SVDBase<MatrixType>::allocate(rows, cols, computationOptions)) return; + + if (QRPreconditioner == FullPivHouseholderQRPreconditioner) + { + eigen_assert(!(this->m_computeThinU || this->m_computeThinV) && + "JacobiSVD: can't compute thin U or thin V with the FullPivHouseholderQR preconditioner. " + "Use the ColPivHouseholderQR preconditioner instead."); + } + + m_workMatrix.resize(this->m_diagSize, this->m_diagSize); + + if(this->m_cols>this->m_rows) m_qr_precond_morecols.allocate(*this); + if(this->m_rows>this->m_cols) m_qr_precond_morerows.allocate(*this); +} + +template<typename MatrixType, int QRPreconditioner> +SVDBase<MatrixType>& +JacobiSVD<MatrixType, QRPreconditioner>::compute(const MatrixType& matrix, unsigned int computationOptions) +{ + using std::abs; + allocate(matrix.rows(), matrix.cols(), computationOptions); + + // currently we stop when we reach precision 2*epsilon as the last bit of precision can require an unreasonable number of iterations, + // only worsening the precision of U and V as we accumulate more rotations + const RealScalar precision = RealScalar(2) * NumTraits<Scalar>::epsilon(); + + // limit for very small denormal numbers to be considered zero in order to avoid infinite loops (see bug 286) + const RealScalar considerAsZero = RealScalar(2) * std::numeric_limits<RealScalar>::denorm_min(); + + /*** step 1. The R-SVD step: we use a QR decomposition to reduce to the case of a square matrix */ + + if(!m_qr_precond_morecols.run(*this, matrix) && !m_qr_precond_morerows.run(*this, matrix)) + { + m_workMatrix = matrix.block(0,0,this->m_diagSize,this->m_diagSize); + if(this->m_computeFullU) this->m_matrixU.setIdentity(this->m_rows,this->m_rows); + if(this->m_computeThinU) this->m_matrixU.setIdentity(this->m_rows,this->m_diagSize); + if(this->m_computeFullV) this->m_matrixV.setIdentity(this->m_cols,this->m_cols); + if(this->m_computeThinV) this->m_matrixV.setIdentity(this->m_cols, this->m_diagSize); + } + + /*** step 2. The main Jacobi SVD iteration. ***/ + + bool finished = false; + while(!finished) + { + finished = true; + + // do a sweep: for all index pairs (p,q), perform SVD of the corresponding 2x2 sub-matrix + + for(Index p = 1; p < this->m_diagSize; ++p) + { + for(Index q = 0; q < p; ++q) + { + // if this 2x2 sub-matrix is not diagonal already... + // notice that this comparison will evaluate to false if any NaN is involved, ensuring that NaN's don't + // keep us iterating forever. Similarly, small denormal numbers are considered zero. + using std::max; + RealScalar threshold = (max)(considerAsZero, precision * (max)(abs(m_workMatrix.coeff(p,p)), + abs(m_workMatrix.coeff(q,q)))); + if((max)(abs(m_workMatrix.coeff(p,q)),abs(m_workMatrix.coeff(q,p))) > threshold) + { + finished = false; + + // perform SVD decomposition of 2x2 sub-matrix corresponding to indices p,q to make it diagonal + internal::svd_precondition_2x2_block_to_be_real<MatrixType, QRPreconditioner>::run(m_workMatrix, *this, p, q); + JacobiRotation<RealScalar> j_left, j_right; + internal::real_2x2_jacobi_svd(m_workMatrix, p, q, &j_left, &j_right); + + // accumulate resulting Jacobi rotations + m_workMatrix.applyOnTheLeft(p,q,j_left); + if(SVDBase<MatrixType>::computeU()) this->m_matrixU.applyOnTheRight(p,q,j_left.transpose()); + + m_workMatrix.applyOnTheRight(p,q,j_right); + if(SVDBase<MatrixType>::computeV()) this->m_matrixV.applyOnTheRight(p,q,j_right); + } + } + } + } + + /*** step 3. The work matrix is now diagonal, so ensure it's positive so its diagonal entries are the singular values ***/ + + for(Index i = 0; i < this->m_diagSize; ++i) + { + RealScalar a = abs(m_workMatrix.coeff(i,i)); + this->m_singularValues.coeffRef(i) = a; + if(SVDBase<MatrixType>::computeU() && (a!=RealScalar(0))) this->m_matrixU.col(i) *= this->m_workMatrix.coeff(i,i)/a; + } + + /*** step 4. Sort singular values in descending order and compute the number of nonzero singular values ***/ + + this->m_nonzeroSingularValues = this->m_diagSize; + for(Index i = 0; i < this->m_diagSize; i++) + { + Index pos; + RealScalar maxRemainingSingularValue = this->m_singularValues.tail(this->m_diagSize-i).maxCoeff(&pos); + if(maxRemainingSingularValue == RealScalar(0)) + { + this->m_nonzeroSingularValues = i; + break; + } + if(pos) + { + pos += i; + std::swap(this->m_singularValues.coeffRef(i), this->m_singularValues.coeffRef(pos)); + if(SVDBase<MatrixType>::computeU()) this->m_matrixU.col(pos).swap(this->m_matrixU.col(i)); + if(SVDBase<MatrixType>::computeV()) this->m_matrixV.col(pos).swap(this->m_matrixV.col(i)); + } + } + + this->m_isInitialized = true; + return *this; +} + +namespace internal { +template<typename _MatrixType, int QRPreconditioner, typename Rhs> +struct solve_retval<JacobiSVD<_MatrixType, QRPreconditioner>, Rhs> + : solve_retval_base<JacobiSVD<_MatrixType, QRPreconditioner>, Rhs> +{ + typedef JacobiSVD<_MatrixType, QRPreconditioner> JacobiSVDType; + EIGEN_MAKE_SOLVE_HELPERS(JacobiSVDType,Rhs) + + template<typename Dest> void evalTo(Dest& dst) const + { + eigen_assert(rhs().rows() == dec().rows()); + + // A = U S V^* + // So A^{-1} = V S^{-1} U^* + + Index diagSize = (std::min)(dec().rows(), dec().cols()); + typename JacobiSVDType::SingularValuesType invertedSingVals(diagSize); + + Index nonzeroSingVals = dec().nonzeroSingularValues(); + invertedSingVals.head(nonzeroSingVals) = dec().singularValues().head(nonzeroSingVals).array().inverse(); + invertedSingVals.tail(diagSize - nonzeroSingVals).setZero(); + + dst = dec().matrixV().leftCols(diagSize) + * invertedSingVals.asDiagonal() + * dec().matrixU().leftCols(diagSize).adjoint() + * rhs(); + } +}; +} // end namespace internal + +/** \svd_module + * + * \return the singular value decomposition of \c *this computed by two-sided + * Jacobi transformations. + * + * \sa class JacobiSVD + */ +template<typename Derived> +JacobiSVD<typename MatrixBase<Derived>::PlainObject> +MatrixBase<Derived>::jacobiSvd(unsigned int computationOptions) const +{ + return JacobiSVD<PlainObject>(*this, computationOptions); +} + +} // end namespace Eigen + +#endif // EIGEN_JACOBISVD_H diff --git a/unsupported/Eigen/src/SVD/SVDBase.h b/unsupported/Eigen/src/SVD/SVDBase.h new file mode 100644 index 000000000..fd8af3b8c --- /dev/null +++ b/unsupported/Eigen/src/SVD/SVDBase.h @@ -0,0 +1,236 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2009-2010 Benoit Jacob <jacob.benoit.1@gmail.com> +// +// Copyright (C) 2013 Gauthier Brun <brun.gauthier@gmail.com> +// Copyright (C) 2013 Nicolas Carre <nicolas.carre@ensimag.fr> +// Copyright (C) 2013 Jean Ceccato <jean.ceccato@ensimag.fr> +// Copyright (C) 2013 Pierre Zoppitelli <pierre.zoppitelli@ensimag.fr> +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + +#ifndef EIGEN_SVD_H +#define EIGEN_SVD_H + +namespace Eigen { +/** \ingroup SVD_Module + * + * + * \class SVDBase + * + * \brief Mother class of SVD classes algorithms + * + * \param MatrixType the type of the matrix of which we are computing the SVD decomposition + * SVD decomposition consists in decomposing any n-by-p matrix \a A as a product + * \f[ A = U S V^* \f] + * where \a U is a n-by-n unitary, \a V is a p-by-p unitary, and \a S is a n-by-p real positive matrix which is zero outside of its main diagonal; + * the diagonal entries of S are known as the \em singular \em values of \a A and the columns of \a U and \a V are known as the left + * and right \em singular \em vectors of \a A respectively. + * + * Singular values are always sorted in decreasing order. + * + * + * You can ask for only \em thin \a U or \a V to be computed, meaning the following. In case of a rectangular n-by-p matrix, letting \a m be the + * smaller value among \a n and \a p, there are only \a m singular vectors; the remaining columns of \a U and \a V do not correspond to actual + * singular vectors. Asking for \em thin \a U or \a V means asking for only their \a m first columns to be formed. So \a U is then a n-by-m matrix, + * and \a V is then a p-by-m matrix. Notice that thin \a U and \a V are all you need for (least squares) solving. + * + * If the input matrix has inf or nan coefficients, the result of the computation is undefined, but the computation is guaranteed to + * terminate in finite (and reasonable) time. + * \sa MatrixBase::genericSvd() + */ +template<typename _MatrixType> +class SVDBase +{ + +public: + typedef _MatrixType MatrixType; + typedef typename MatrixType::Scalar Scalar; + typedef typename NumTraits<typename MatrixType::Scalar>::Real RealScalar; + typedef typename MatrixType::Index Index; + enum { + RowsAtCompileTime = MatrixType::RowsAtCompileTime, + ColsAtCompileTime = MatrixType::ColsAtCompileTime, + DiagSizeAtCompileTime = EIGEN_SIZE_MIN_PREFER_DYNAMIC(RowsAtCompileTime,ColsAtCompileTime), + MaxRowsAtCompileTime = MatrixType::MaxRowsAtCompileTime, + MaxColsAtCompileTime = MatrixType::MaxColsAtCompileTime, + MaxDiagSizeAtCompileTime = EIGEN_SIZE_MIN_PREFER_FIXED(MaxRowsAtCompileTime,MaxColsAtCompileTime), + MatrixOptions = MatrixType::Options + }; + + typedef Matrix<Scalar, RowsAtCompileTime, RowsAtCompileTime, + MatrixOptions, MaxRowsAtCompileTime, MaxRowsAtCompileTime> + MatrixUType; + typedef Matrix<Scalar, ColsAtCompileTime, ColsAtCompileTime, + MatrixOptions, MaxColsAtCompileTime, MaxColsAtCompileTime> + MatrixVType; + typedef typename internal::plain_diag_type<MatrixType, RealScalar>::type SingularValuesType; + typedef typename internal::plain_row_type<MatrixType>::type RowType; + typedef typename internal::plain_col_type<MatrixType>::type ColType; + typedef Matrix<Scalar, DiagSizeAtCompileTime, DiagSizeAtCompileTime, + MatrixOptions, MaxDiagSizeAtCompileTime, MaxDiagSizeAtCompileTime> + WorkMatrixType; + + + + + /** \brief Method performing the decomposition of given matrix using custom options. + * + * \param matrix the matrix to decompose + * \param computationOptions optional parameter allowing to specify if you want full or thin U or V unitaries to be computed. + * By default, none is computed. This is a bit-field, the possible bits are #ComputeFullU, #ComputeThinU, + * #ComputeFullV, #ComputeThinV. + * + * Thin unitaries are only available if your matrix type has a Dynamic number of columns (for example MatrixXf). They also are not + * available with the (non-default) FullPivHouseholderQR preconditioner. + */ + SVDBase& compute(const MatrixType& matrix, unsigned int computationOptions); + + /** \brief Method performing the decomposition of given matrix using current options. + * + * \param matrix the matrix to decompose + * + * This method uses the current \a computationOptions, as already passed to the constructor or to compute(const MatrixType&, unsigned int). + */ + //virtual SVDBase& compute(const MatrixType& matrix) = 0; + SVDBase& compute(const MatrixType& matrix); + + /** \returns the \a U matrix. + * + * For the SVDBase decomposition of a n-by-p matrix, letting \a m be the minimum of \a n and \a p, + * the U matrix is n-by-n if you asked for #ComputeFullU, and is n-by-m if you asked for #ComputeThinU. + * + * The \a m first columns of \a U are the left singular vectors of the matrix being decomposed. + * + * This method asserts that you asked for \a U to be computed. + */ + const MatrixUType& matrixU() const + { + eigen_assert(m_isInitialized && "SVD is not initialized."); + eigen_assert(computeU() && "This SVD decomposition didn't compute U. Did you ask for it?"); + return m_matrixU; + } + + /** \returns the \a V matrix. + * + * For the SVD decomposition of a n-by-p matrix, letting \a m be the minimum of \a n and \a p, + * the V matrix is p-by-p if you asked for #ComputeFullV, and is p-by-m if you asked for ComputeThinV. + * + * The \a m first columns of \a V are the right singular vectors of the matrix being decomposed. + * + * This method asserts that you asked for \a V to be computed. + */ + const MatrixVType& matrixV() const + { + eigen_assert(m_isInitialized && "SVD is not initialized."); + eigen_assert(computeV() && "This SVD decomposition didn't compute V. Did you ask for it?"); + return m_matrixV; + } + + /** \returns the vector of singular values. + * + * For the SVD decomposition of a n-by-p matrix, letting \a m be the minimum of \a n and \a p, the + * returned vector has size \a m. Singular values are always sorted in decreasing order. + */ + const SingularValuesType& singularValues() const + { + eigen_assert(m_isInitialized && "SVD is not initialized."); + return m_singularValues; + } + + + + /** \returns the number of singular values that are not exactly 0 */ + Index nonzeroSingularValues() const + { + eigen_assert(m_isInitialized && "SVD is not initialized."); + return m_nonzeroSingularValues; + } + + + /** \returns true if \a U (full or thin) is asked for in this SVD decomposition */ + inline bool computeU() const { return m_computeFullU || m_computeThinU; } + /** \returns true if \a V (full or thin) is asked for in this SVD decomposition */ + inline bool computeV() const { return m_computeFullV || m_computeThinV; } + + + inline Index rows() const { return m_rows; } + inline Index cols() const { return m_cols; } + + +protected: + // return true if already allocated + bool allocate(Index rows, Index cols, unsigned int computationOptions) ; + + MatrixUType m_matrixU; + MatrixVType m_matrixV; + SingularValuesType m_singularValues; + bool m_isInitialized, m_isAllocated; + bool m_computeFullU, m_computeThinU; + bool m_computeFullV, m_computeThinV; + unsigned int m_computationOptions; + Index m_nonzeroSingularValues, m_rows, m_cols, m_diagSize; + + + /** \brief Default Constructor. + * + * Default constructor of SVDBase + */ + SVDBase() + : m_isInitialized(false), + m_isAllocated(false), + m_computationOptions(0), + m_rows(-1), m_cols(-1) + {} + + +}; + + +template<typename MatrixType> +bool SVDBase<MatrixType>::allocate(Index rows, Index cols, unsigned int computationOptions) +{ + eigen_assert(rows >= 0 && cols >= 0); + + if (m_isAllocated && + rows == m_rows && + cols == m_cols && + computationOptions == m_computationOptions) + { + return true; + } + + m_rows = rows; + m_cols = cols; + m_isInitialized = false; + m_isAllocated = true; + m_computationOptions = computationOptions; + m_computeFullU = (computationOptions & ComputeFullU) != 0; + m_computeThinU = (computationOptions & ComputeThinU) != 0; + m_computeFullV = (computationOptions & ComputeFullV) != 0; + m_computeThinV = (computationOptions & ComputeThinV) != 0; + eigen_assert(!(m_computeFullU && m_computeThinU) && "SVDBase: you can't ask for both full and thin U"); + eigen_assert(!(m_computeFullV && m_computeThinV) && "SVDBase: you can't ask for both full and thin V"); + eigen_assert(EIGEN_IMPLIES(m_computeThinU || m_computeThinV, MatrixType::ColsAtCompileTime==Dynamic) && + "SVDBase: thin U and V are only available when your matrix has a dynamic number of columns."); + + m_diagSize = (std::min)(m_rows, m_cols); + m_singularValues.resize(m_diagSize); + if(RowsAtCompileTime==Dynamic) + m_matrixU.resize(m_rows, m_computeFullU ? m_rows + : m_computeThinU ? m_diagSize + : 0); + if(ColsAtCompileTime==Dynamic) + m_matrixV.resize(m_cols, m_computeFullV ? m_cols + : m_computeThinV ? m_diagSize + : 0); + + return false; +} + +}// end namespace + +#endif // EIGEN_SVD_H diff --git a/unsupported/Eigen/src/SVD/TODOBdcsvd.txt b/unsupported/Eigen/src/SVD/TODOBdcsvd.txt new file mode 100644 index 000000000..0bc9a46e6 --- /dev/null +++ b/unsupported/Eigen/src/SVD/TODOBdcsvd.txt @@ -0,0 +1,29 @@ +TO DO LIST + + + +(optional optimization) - do all the allocations in the allocate part + - support static matrices + - return a error at compilation time when using integer matrices (int, long, std::complex<int>, ...) + +to finish the algorithm : + -implement the last part of the algorithm as described on the reference paper. + You may find more information on that part on this paper + + -to replace the call to JacobiSVD at the end of the divide algorithm, just after the call to + deflation. + +(suggested step by step resolution) + 0) comment the call to Jacobi in the last part of the divide method and everything right after + until the end of the method. What is commented can be a guideline to steps 3) 4) and 6) + 1) solve the secular equation (Characteristic equation) on the values that are not null (zi!=0 and di!=0), after the deflation + wich should be uncommented in the divide method + 2) remember the values of the singular values that are already computed (zi=0) + 3) assign the singular values found in m_computed at the right places (with the ones found in step 2) ) + in decreasing order + 4) set the firstcol to zero (except the first element) in m_computed + 5) compute all the singular vectors when CompV is set to true and only the left vectors when + CompV is set to false + 6) multiply naiveU and naiveV to the right by the matrices found, only naiveU when CompV is set to + false, /!\ if CompU is false NaiveU has only 2 rows + 7) delete everything commented in step 0) diff --git a/unsupported/Eigen/src/SVD/doneInBDCSVD.txt b/unsupported/Eigen/src/SVD/doneInBDCSVD.txt new file mode 100644 index 000000000..8563ddab8 --- /dev/null +++ b/unsupported/Eigen/src/SVD/doneInBDCSVD.txt @@ -0,0 +1,21 @@ +This unsupported package is about a divide and conquer algorithm to compute SVD. + +The implementation follows as closely as possible the following reference paper : +http://www.cs.yale.edu/publications/techreports/tr933.pdf + +The code documentation uses the same names for variables as the reference paper. The code, deflation included, is +working but there are a few things that could be optimised as explained in the TODOBdsvd. + +In the code comments were put at the line where would be the third step of the algorithm so one could simply add the call +of a function doing the last part of the algorithm and that would not require any knowledge of the part we implemented. + +In the TODOBdcsvd we explain what is the main difficulty of the last part and suggest a reference paper to help solve it. + +The implemented has trouble with fixed size matrices. + +In the actual implementation, it returns matrices of zero when ask to do a svd on an int matrix. + + +Paper for the third part: +http://www.stat.uchicago.edu/~lekheng/courses/302/classics/greengard-rokhlin.pdf + diff --git a/unsupported/Eigen/src/SparseExtra/BlockOfDynamicSparseMatrix.h b/unsupported/Eigen/src/SparseExtra/BlockOfDynamicSparseMatrix.h index fd24a732d..e9ec746e3 100644 --- a/unsupported/Eigen/src/SparseExtra/BlockOfDynamicSparseMatrix.h +++ b/unsupported/Eigen/src/SparseExtra/BlockOfDynamicSparseMatrix.h @@ -12,6 +12,12 @@ namespace Eigen { +#if 0 + +// NOTE Have to be reimplemented as a specialization of BlockImpl< DynamicSparseMatrix<_Scalar, _Options, _Index>, ... > +// See SparseBlock.h for an example + + /*************************************************************************** * specialisation for DynamicSparseMatrix ***************************************************************************/ @@ -109,6 +115,8 @@ class SparseInnerVectorSet<DynamicSparseMatrix<_Scalar, _Options, _Index>, Size> }; +#endif + } // end namespace Eigen #endif // EIGEN_SPARSE_BLOCKFORDYNAMICMATRIX_H diff --git a/unsupported/Eigen/src/SparseExtra/MarketIO.h b/unsupported/Eigen/src/SparseExtra/MarketIO.h index de958de9f..7aafce928 100644 --- a/unsupported/Eigen/src/SparseExtra/MarketIO.h +++ b/unsupported/Eigen/src/SparseExtra/MarketIO.h @@ -116,7 +116,7 @@ inline bool getMarketHeader(const std::string& filename, int& sym, bool& iscompl std::string line; // The matrix header is always the first line in the file - std::getline(in, line); assert(in.good()); + std::getline(in, line); eigen_assert(in.good()); std::stringstream fmtline(line); std::string substr[5]; @@ -200,11 +200,11 @@ bool loadMarketVector(VectorType& vec, const std::string& filename) int n(0), col(0); do { // Skip comments - std::getline(in, line); assert(in.good()); + std::getline(in, line); eigen_assert(in.good()); } while (line[0] == '%'); std::istringstream newline(line); newline >> n >> col; - assert(n>0 && col>0); + eigen_assert(n>0 && col>0); vec.resize(n); int i = 0; Scalar value; diff --git a/unsupported/Eigen/src/SparseExtra/MatrixMarketIterator.h b/unsupported/Eigen/src/SparseExtra/MatrixMarketIterator.h index 4716b68e7..bf13cf21f 100644 --- a/unsupported/Eigen/src/SparseExtra/MatrixMarketIterator.h +++ b/unsupported/Eigen/src/SparseExtra/MatrixMarketIterator.h @@ -184,9 +184,20 @@ class MatrixMarketIterator // if (S_ISDIR(st_buf.st_mode)) continue; // Determine from the header if it is a matrix or a right hand side - bool isvector,iscomplex; + bool isvector,iscomplex=false; if(!getMarketHeader(curfile,m_sym,iscomplex,isvector)) continue; if(isvector) continue; + if (!iscomplex) + { + if(internal::is_same<Scalar, std::complex<float> >::value || internal::is_same<Scalar, std::complex<double> >::value) + continue; + } + if (iscomplex) + { + if(internal::is_same<Scalar, float>::value || internal::is_same<Scalar, double>::value) + continue; + } + // Get the matrix name std::string filename = m_curs_id->d_name; diff --git a/unsupported/Eigen/src/Splines/Spline.h b/unsupported/Eigen/src/Splines/Spline.h index 3680f013a..771f10432 100644 --- a/unsupported/Eigen/src/Splines/Spline.h +++ b/unsupported/Eigen/src/Splines/Spline.h @@ -16,7 +16,7 @@ namespace Eigen { /** * \ingroup Splines_Module - * \class Spline class + * \class Spline * \brief A class representing multi-dimensional spline curves. * * The class represents B-splines with non-uniform knot vectors. Each control @@ -50,6 +50,21 @@ namespace Eigen /** \brief The data type representing the spline's control points. */ typedef typename SplineTraits<Spline>::ControlPointVectorType ControlPointVectorType; + + /** + * \brief Creates a (constant) zero spline. + * For Splines with dynamic degree, the resulting degree will be 0. + **/ + Spline() + : m_knots(1, (Degree==Dynamic ? 2 : 2*Degree+2)) + , m_ctrls(ControlPointVectorType::Zero(2,(Degree==Dynamic ? 1 : Degree+1))) + { + // in theory this code can go to the initializer list but it will get pretty + // much unreadable ... + enum { MinDegree = (Degree==Dynamic ? 0 : Degree) }; + m_knots.template segment<MinDegree+1>(0) = Array<Scalar,1,MinDegree+1>::Zero(); + m_knots.template segment<MinDegree+1>(MinDegree+1) = Array<Scalar,1,MinDegree+1>::Ones(); + } /** * \brief Creates a spline from a knot vector and control points. @@ -280,11 +295,7 @@ namespace Eigen enum { Order = SplineTraits<SplineType>::OrderAtCompileTime }; enum { DerivativeOrder = DerivativeType::ColsAtCompileTime }; - typedef typename SplineTraits<SplineType>::Scalar Scalar; - - typedef typename SplineTraits<SplineType>::BasisVectorType BasisVectorType; typedef typename SplineTraits<SplineType>::ControlPointVectorType ControlPointVectorType; - typedef typename SplineTraits<SplineType,DerivativeOrder>::BasisDerivativeType BasisDerivativeType; typedef typename BasisDerivativeType::ConstRowXpr BasisDerivativeRowXpr; @@ -343,7 +354,6 @@ namespace Eigen typedef typename SplineTraits<SplineType>::Scalar Scalar; typedef typename SplineTraits<SplineType>::BasisVectorType BasisVectorType; typedef typename SplineTraits<SplineType>::KnotVectorType KnotVectorType; - typedef typename SplineTraits<SplineType>::ControlPointVectorType ControlPointVectorType; const KnotVectorType& U = spline.knots(); diff --git a/unsupported/Eigen/src/Splines/SplineFitting.h b/unsupported/Eigen/src/Splines/SplineFitting.h index 1b566332f..0265d532c 100644 --- a/unsupported/Eigen/src/Splines/SplineFitting.h +++ b/unsupported/Eigen/src/Splines/SplineFitting.h @@ -40,8 +40,6 @@ namespace Eigen template <typename KnotVectorType> void KnotAveraging(const KnotVectorType& parameters, DenseIndex degree, KnotVectorType& knots) { - typedef typename KnotVectorType::Scalar Scalar; - knots.resize(parameters.size()+degree+1); for (DenseIndex j=1; j<parameters.size()-degree; ++j) @@ -118,7 +116,6 @@ namespace Eigen SplineType SplineFitting<SplineType>::Interpolate(const PointArrayType& pts, DenseIndex degree, const KnotVectorType& knot_parameters) { typedef typename SplineType::KnotVectorType::Scalar Scalar; - typedef typename SplineType::BasisVectorType BasisVectorType; typedef typename SplineType::ControlPointVectorType ControlPointVectorType; typedef Matrix<Scalar,Dynamic,Dynamic> MatrixType; diff --git a/unsupported/bench/bench_svd.cpp b/unsupported/bench/bench_svd.cpp new file mode 100644 index 000000000..01d8231ae --- /dev/null +++ b/unsupported/bench/bench_svd.cpp @@ -0,0 +1,123 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2013 Gauthier Brun <brun.gauthier@gmail.com> +// Copyright (C) 2013 Nicolas Carre <nicolas.carre@ensimag.fr> +// Copyright (C) 2013 Jean Ceccato <jean.ceccato@ensimag.fr> +// Copyright (C) 2013 Pierre Zoppitelli <pierre.zoppitelli@ensimag.fr> +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/ + +// Bench to compare the efficiency of SVD algorithms + +#include <iostream> +#include <bench/BenchTimer.h> +#include <unsupported/Eigen/SVD> + + +using namespace Eigen; +using namespace std; + +// number of computations of each algorithm before the print of the time +#ifndef REPEAT +#define REPEAT 10 +#endif + +// number of tests of the same type +#ifndef NUMBER_SAMPLE +#define NUMBER_SAMPLE 2 +#endif + +template<typename MatrixType> +void bench_svd(const MatrixType& a = MatrixType()) +{ + MatrixType m = MatrixType::Random(a.rows(), a.cols()); + BenchTimer timerJacobi; + BenchTimer timerBDC; + timerJacobi.reset(); + timerBDC.reset(); + + cout << " Only compute Singular Values" <<endl; + for (int k=1; k<=NUMBER_SAMPLE; ++k) + { + timerBDC.start(); + for (int i=0; i<REPEAT; ++i) + { + BDCSVD<MatrixType> bdc_matrix(m); + } + timerBDC.stop(); + + timerJacobi.start(); + for (int i=0; i<REPEAT; ++i) + { + JacobiSVD<MatrixType> jacobi_matrix(m); + } + timerJacobi.stop(); + + + cout << "Sample " << k << " : " << REPEAT << " computations : Jacobi : " << fixed << timerJacobi.value() << "s "; + cout << " || " << " BDC : " << timerBDC.value() << "s " <<endl <<endl; + + if (timerBDC.value() >= timerJacobi.value()) + cout << "KO : BDC is " << timerJacobi.value() / timerBDC.value() << " times faster than Jacobi" <<endl; + else + cout << "OK : BDC is " << timerJacobi.value() / timerBDC.value() << " times faster than Jacobi" <<endl; + + } + cout << " =================" <<endl; + std::cout<< std::endl; + timerJacobi.reset(); + timerBDC.reset(); + cout << " Computes rotaion matrix" <<endl; + for (int k=1; k<=NUMBER_SAMPLE; ++k) + { + timerBDC.start(); + for (int i=0; i<REPEAT; ++i) + { + BDCSVD<MatrixType> bdc_matrix(m, ComputeFullU|ComputeFullV); + } + timerBDC.stop(); + + timerJacobi.start(); + for (int i=0; i<REPEAT; ++i) + { + JacobiSVD<MatrixType> jacobi_matrix(m, ComputeFullU|ComputeFullV); + } + timerJacobi.stop(); + + + cout << "Sample " << k << " : " << REPEAT << " computations : Jacobi : " << fixed << timerJacobi.value() << "s "; + cout << " || " << " BDC : " << timerBDC.value() << "s " <<endl <<endl; + + if (timerBDC.value() >= timerJacobi.value()) + cout << "KO : BDC is " << timerJacobi.value() / timerBDC.value() << " times faster than Jacobi" <<endl; + else + cout << "OK : BDC is " << timerJacobi.value() / timerBDC.value() << " times faster than Jacobi" <<endl; + + } + std::cout<< std::endl; +} + + + +int main(int argc, char* argv[]) +{ + std::cout<< std::endl; + + std::cout<<"On a (Dynamic, Dynamic) (6, 6) Matrix" <<std::endl; + bench_svd<Matrix<double,Dynamic,Dynamic> >(Matrix<double,Dynamic,Dynamic>(6, 6)); + + std::cout<<"On a (Dynamic, Dynamic) (32, 32) Matrix" <<std::endl; + bench_svd<Matrix<double,Dynamic,Dynamic> >(Matrix<double,Dynamic,Dynamic>(32, 32)); + + //std::cout<<"On a (Dynamic, Dynamic) (128, 128) Matrix" <<std::endl; + //bench_svd<Matrix<double,Dynamic,Dynamic> >(Matrix<double,Dynamic,Dynamic>(128, 128)); + + std::cout<<"On a (Dynamic, Dynamic) (160, 160) Matrix" <<std::endl; + bench_svd<Matrix<double,Dynamic,Dynamic> >(Matrix<double,Dynamic,Dynamic>(160, 160)); + + std::cout<< "--------------------------------------------------------------------"<< std::endl; + +} diff --git a/unsupported/doc/Doxyfile.in b/unsupported/doc/Doxyfile.in deleted file mode 100644 index 1facf2985..000000000 --- a/unsupported/doc/Doxyfile.in +++ /dev/null @@ -1,1460 +0,0 @@ -# This file describes the settings to be used by the documentation system -# doxygen (www.doxygen.org) for a project -# -# All text after a hash (#) is considered a comment and will be ignored -# The format is: -# TAG = value [value, ...] -# For lists items can also be appended using: -# TAG += value [value, ...] -# Values that contain spaces should be placed between quotes (" ") - -#--------------------------------------------------------------------------- -# Project related configuration options -#--------------------------------------------------------------------------- - -# This tag specifies the encoding used for all characters in the config file -# that follow. The default is UTF-8 which is also the encoding used for all -# text before the first occurrence of this tag. Doxygen uses libiconv (or the -# iconv built into libc) for the transcoding. See -# http://www.gnu.org/software/libiconv for the list of possible encodings. - -DOXYFILE_ENCODING = UTF-8 - -# The PROJECT_NAME tag is a single word (or a sequence of words surrounded -# by quotes) that should identify the project. - -PROJECT_NAME = Eigen - unsupported modules - -# The PROJECT_NUMBER tag can be used to enter a project or revision number. -# This could be handy for archiving the generated documentation or -# if some version control system is used. - -#EIGEN_VERSION is set in the root CMakeLists.txt -PROJECT_NUMBER = "${EIGEN_VERSION}" - -# The OUTPUT_DIRECTORY tag is used to specify the (relative or absolute) -# base path where the generated documentation will be put. -# If a relative path is entered, it will be relative to the location -# where doxygen was started. If left blank the current directory will be used. - -OUTPUT_DIRECTORY = "${Eigen_BINARY_DIR}/doc/unsupported" - -# If the CREATE_SUBDIRS tag is set to YES, then doxygen will create -# 4096 sub-directories (in 2 levels) under the output directory of each output -# format and will distribute the generated files over these directories. -# Enabling this option can be useful when feeding doxygen a huge amount of -# source files, where putting all generated files in the same directory would -# otherwise cause performance problems for the file system. - -CREATE_SUBDIRS = NO - -# The OUTPUT_LANGUAGE tag is used to specify the language in which all -# documentation generated by doxygen is written. Doxygen will use this -# information to generate all constant output in the proper language. -# The default language is English, other supported languages are: -# Afrikaans, Arabic, Brazilian, Catalan, Chinese, Chinese-Traditional, -# Croatian, Czech, Danish, Dutch, Farsi, Finnish, French, German, Greek, -# Hungarian, Italian, Japanese, Japanese-en (Japanese with English messages), -# Korean, Korean-en, Lithuanian, Norwegian, Macedonian, Persian, Polish, -# Portuguese, Romanian, Russian, Serbian, Slovak, Slovene, Spanish, Swedish, -# and Ukrainian. - -OUTPUT_LANGUAGE = English - -# If the BRIEF_MEMBER_DESC tag is set to YES (the default) Doxygen will -# include brief member descriptions after the members that are listed in -# the file and class documentation (similar to JavaDoc). -# Set to NO to disable this. - -BRIEF_MEMBER_DESC = YES - -# If the REPEAT_BRIEF tag is set to YES (the default) Doxygen will prepend -# the brief description of a member or function before the detailed description. -# Note: if both HIDE_UNDOC_MEMBERS and BRIEF_MEMBER_DESC are set to NO, the -# brief descriptions will be completely suppressed. - -REPEAT_BRIEF = YES - -# This tag implements a quasi-intelligent brief description abbreviator -# that is used to form the text in various listings. Each string -# in this list, if found as the leading text of the brief description, will be -# stripped from the text and the result after processing the whole list, is -# used as the annotated text. Otherwise, the brief description is used as-is. -# If left blank, the following values are used ("$name" is automatically -# replaced with the name of the entity): "The $name class" "The $name widget" -# "The $name file" "is" "provides" "specifies" "contains" -# "represents" "a" "an" "the" - -ABBREVIATE_BRIEF = "The $name class" \ - "The $name widget" \ - "The $name file" \ - is \ - provides \ - specifies \ - contains \ - represents \ - a \ - an \ - the - -# If the ALWAYS_DETAILED_SEC and REPEAT_BRIEF tags are both set to YES then -# Doxygen will generate a detailed section even if there is only a brief -# description. - -ALWAYS_DETAILED_SEC = NO - -# If the INLINE_INHERITED_MEMB tag is set to YES, doxygen will show all -# inherited members of a class in the documentation of that class as if those -# members were ordinary class members. Constructors, destructors and assignment -# operators of the base classes will not be shown. - -INLINE_INHERITED_MEMB = NO - -# If the FULL_PATH_NAMES tag is set to YES then Doxygen will prepend the full -# path before files name in the file list and in the header files. If set -# to NO the shortest path that makes the file name unique will be used. - -FULL_PATH_NAMES = NO - -# If the FULL_PATH_NAMES tag is set to YES then the STRIP_FROM_PATH tag -# can be used to strip a user-defined part of the path. Stripping is -# only done if one of the specified strings matches the left-hand part of -# the path. The tag can be used to show relative paths in the file list. -# If left blank the directory from which doxygen is run is used as the -# path to strip. - -STRIP_FROM_PATH = - -# The STRIP_FROM_INC_PATH tag can be used to strip a user-defined part of -# the path mentioned in the documentation of a class, which tells -# the reader which header file to include in order to use a class. -# If left blank only the name of the header file containing the class -# definition is used. Otherwise one should specify the include paths that -# are normally passed to the compiler using the -I flag. - -STRIP_FROM_INC_PATH = - -# If the SHORT_NAMES tag is set to YES, doxygen will generate much shorter -# (but less readable) file names. This can be useful is your file systems -# doesn't support long names like on DOS, Mac, or CD-ROM. - -SHORT_NAMES = NO - -# If the JAVADOC_AUTOBRIEF tag is set to YES then Doxygen -# will interpret the first line (until the first dot) of a JavaDoc-style -# comment as the brief description. If set to NO, the JavaDoc -# comments will behave just like regular Qt-style comments -# (thus requiring an explicit @brief command for a brief description.) - -JAVADOC_AUTOBRIEF = NO - -# If the QT_AUTOBRIEF tag is set to YES then Doxygen will -# interpret the first line (until the first dot) of a Qt-style -# comment as the brief description. If set to NO, the comments -# will behave just like regular Qt-style comments (thus requiring -# an explicit \brief command for a brief description.) - -QT_AUTOBRIEF = NO - -# The MULTILINE_CPP_IS_BRIEF tag can be set to YES to make Doxygen -# treat a multi-line C++ special comment block (i.e. a block of //! or /// -# comments) as a brief description. This used to be the default behaviour. -# The new default is to treat a multi-line C++ comment block as a detailed -# description. Set this tag to YES if you prefer the old behaviour instead. - -MULTILINE_CPP_IS_BRIEF = NO - -# If the DETAILS_AT_TOP tag is set to YES then Doxygen -# will output the detailed description near the top, like JavaDoc. -# If set to NO, the detailed description appears after the member -# documentation. - -DETAILS_AT_TOP = YES - -# If the INHERIT_DOCS tag is set to YES (the default) then an undocumented -# member inherits the documentation from any documented member that it -# re-implements. - -INHERIT_DOCS = YES - -# If the SEPARATE_MEMBER_PAGES tag is set to YES, then doxygen will produce -# a new page for each member. If set to NO, the documentation of a member will -# be part of the file/class/namespace that contains it. - -SEPARATE_MEMBER_PAGES = NO - -# The TAB_SIZE tag can be used to set the number of spaces in a tab. -# Doxygen uses this value to replace tabs by spaces in code fragments. - -TAB_SIZE = 8 - -# This tag can be used to specify a number of aliases that acts -# as commands in the documentation. An alias has the form "name=value". -# For example adding "sideeffect=\par Side Effects:\n" will allow you to -# put the command \sideeffect (or @sideeffect) in the documentation, which -# will result in a user-defined paragraph with heading "Side Effects:". -# You can put \n's in the value part of an alias to insert newlines. - -ALIASES = "only_for_vectors=This is only for vectors (either row-vectors or column-vectors), i.e. matrices which are known at compile-time to have either one row or one column." \ - "array_module=This is defined in the %Array module. \code #include <Eigen/Array> \endcode" \ - "lu_module=This is defined in the %LU module. \code #include <Eigen/LU> \endcode" \ - "cholesky_module=This is defined in the %Cholesky module. \code #include <Eigen/Cholesky> \endcode" \ - "qr_module=This is defined in the %QR module. \code #include <Eigen/QR> \endcode" \ - "svd_module=This is defined in the %SVD module. \code #include <Eigen/SVD> \endcode" \ - "geometry_module=This is defined in the %Geometry module. \code #include <Eigen/Geometry> \endcode" \ - "label=\bug" \ - "nonstableyet=\warning This is not considered to be part of the stable public API yet. Changes may happen in future releases. See \ref Experimental \"Experimental parts of Eigen\"" - -# Set the OPTIMIZE_OUTPUT_FOR_C tag to YES if your project consists of C -# sources only. Doxygen will then generate output that is more tailored for C. -# For instance, some of the names that are used will be different. The list -# of all members will be omitted, etc. - -OPTIMIZE_OUTPUT_FOR_C = NO - -# Set the OPTIMIZE_OUTPUT_JAVA tag to YES if your project consists of Java -# sources only. Doxygen will then generate output that is more tailored for -# Java. For instance, namespaces will be presented as packages, qualified -# scopes will look different, etc. - -OPTIMIZE_OUTPUT_JAVA = NO - -# Set the OPTIMIZE_FOR_FORTRAN tag to YES if your project consists of Fortran -# sources only. Doxygen will then generate output that is more tailored for -# Fortran. - -OPTIMIZE_FOR_FORTRAN = NO - -# Set the OPTIMIZE_OUTPUT_VHDL tag to YES if your project consists of VHDL -# sources. Doxygen will then generate output that is tailored for -# VHDL. - -OPTIMIZE_OUTPUT_VHDL = NO - -# If you use STL classes (i.e. std::string, std::vector, etc.) but do not want -# to include (a tag file for) the STL sources as input, then you should -# set this tag to YES in order to let doxygen match functions declarations and -# definitions whose arguments contain STL classes (e.g. func(std::string); v.s. -# func(std::string) {}). This also make the inheritance and collaboration -# diagrams that involve STL classes more complete and accurate. - -BUILTIN_STL_SUPPORT = NO - -# If you use Microsoft's C++/CLI language, you should set this option to YES to -# enable parsing support. - -CPP_CLI_SUPPORT = NO - -# Set the SIP_SUPPORT tag to YES if your project consists of sip sources only. -# Doxygen will parse them like normal C++ but will assume all classes use public -# instead of private inheritance when no explicit protection keyword is present. - -SIP_SUPPORT = NO - -# For Microsoft's IDL there are propget and propput attributes to indicate getter -# and setter methods for a property. Setting this option to YES (the default) -# will make doxygen to replace the get and set methods by a property in the -# documentation. This will only work if the methods are indeed getting or -# setting a simple type. If this is not the case, or you want to show the -# methods anyway, you should set this option to NO. - -IDL_PROPERTY_SUPPORT = YES - -# If member grouping is used in the documentation and the DISTRIBUTE_GROUP_DOC -# tag is set to YES, then doxygen will reuse the documentation of the first -# member in the group (if any) for the other members of the group. By default -# all members of a group must be documented explicitly. - -DISTRIBUTE_GROUP_DOC = NO - -# Set the SUBGROUPING tag to YES (the default) to allow class member groups of -# the same type (for instance a group of public functions) to be put as a -# subgroup of that type (e.g. under the Public Functions section). Set it to -# NO to prevent subgrouping. Alternatively, this can be done per class using -# the \nosubgrouping command. - -SUBGROUPING = YES - -# When TYPEDEF_HIDES_STRUCT is enabled, a typedef of a struct, union, or enum -# is documented as struct, union, or enum with the name of the typedef. So -# typedef struct TypeS {} TypeT, will appear in the documentation as a struct -# with name TypeT. When disabled the typedef will appear as a member of a file, -# namespace, or class. And the struct will be named TypeS. This can typically -# be useful for C code in case the coding convention dictates that all compound -# types are typedef'ed and only the typedef is referenced, never the tag name. - -TYPEDEF_HIDES_STRUCT = NO - -#--------------------------------------------------------------------------- -# Build related configuration options -#--------------------------------------------------------------------------- - -# If the EXTRACT_ALL tag is set to YES doxygen will assume all entities in -# documentation are documented, even if no documentation was available. -# Private class members and static file members will be hidden unless -# the EXTRACT_PRIVATE and EXTRACT_STATIC tags are set to YES - -EXTRACT_ALL = NO - -# If the EXTRACT_PRIVATE tag is set to YES all private members of a class -# will be included in the documentation. - -EXTRACT_PRIVATE = NO - -# If the EXTRACT_STATIC tag is set to YES all static members of a file -# will be included in the documentation. - -EXTRACT_STATIC = NO - -# If the EXTRACT_LOCAL_CLASSES tag is set to YES classes (and structs) -# defined locally in source files will be included in the documentation. -# If set to NO only classes defined in header files are included. - -EXTRACT_LOCAL_CLASSES = NO - -# This flag is only useful for Objective-C code. When set to YES local -# methods, which are defined in the implementation section but not in -# the interface are included in the documentation. -# If set to NO (the default) only methods in the interface are included. - -EXTRACT_LOCAL_METHODS = NO - -# If this flag is set to YES, the members of anonymous namespaces will be -# extracted and appear in the documentation as a namespace called -# 'anonymous_namespace{file}', where file will be replaced with the base -# name of the file that contains the anonymous namespace. By default -# anonymous namespace are hidden. - -EXTRACT_ANON_NSPACES = NO - -# If the HIDE_UNDOC_MEMBERS tag is set to YES, Doxygen will hide all -# undocumented members of documented classes, files or namespaces. -# If set to NO (the default) these members will be included in the -# various overviews, but no documentation section is generated. -# This option has no effect if EXTRACT_ALL is enabled. - -HIDE_UNDOC_MEMBERS = NO - -# If the HIDE_UNDOC_CLASSES tag is set to YES, Doxygen will hide all -# undocumented classes that are normally visible in the class hierarchy. -# If set to NO (the default) these classes will be included in the various -# overviews. This option has no effect if EXTRACT_ALL is enabled. - -HIDE_UNDOC_CLASSES = YES - -# If the HIDE_FRIEND_COMPOUNDS tag is set to YES, Doxygen will hide all -# friend (class|struct|union) declarations. -# If set to NO (the default) these declarations will be included in the -# documentation. - -HIDE_FRIEND_COMPOUNDS = YES - -# If the HIDE_IN_BODY_DOCS tag is set to YES, Doxygen will hide any -# documentation blocks found inside the body of a function. -# If set to NO (the default) these blocks will be appended to the -# function's detailed documentation block. - -HIDE_IN_BODY_DOCS = NO - -# The INTERNAL_DOCS tag determines if documentation -# that is typed after a \internal command is included. If the tag is set -# to NO (the default) then the documentation will be excluded. -# Set it to YES to include the internal documentation. - -INTERNAL_DOCS = NO - -# If the CASE_SENSE_NAMES tag is set to NO then Doxygen will only generate -# file names in lower-case letters. If set to YES upper-case letters are also -# allowed. This is useful if you have classes or files whose names only differ -# in case and if your file system supports case sensitive file names. Windows -# and Mac users are advised to set this option to NO. - -CASE_SENSE_NAMES = YES - -# If the HIDE_SCOPE_NAMES tag is set to NO (the default) then Doxygen -# will show members with their full class and namespace scopes in the -# documentation. If set to YES the scope will be hidden. - -HIDE_SCOPE_NAMES = YES - -# If the SHOW_INCLUDE_FILES tag is set to YES (the default) then Doxygen -# will put a list of the files that are included by a file in the documentation -# of that file. - -SHOW_INCLUDE_FILES = YES - -# If the INLINE_INFO tag is set to YES (the default) then a tag [inline] -# is inserted in the documentation for inline members. - -INLINE_INFO = YES - -# If the SORT_MEMBER_DOCS tag is set to YES (the default) then doxygen -# will sort the (detailed) documentation of file and class members -# alphabetically by member name. If set to NO the members will appear in -# declaration order. - -SORT_MEMBER_DOCS = YES - -# If the SORT_BRIEF_DOCS tag is set to YES then doxygen will sort the -# brief documentation of file, namespace and class members alphabetically -# by member name. If set to NO (the default) the members will appear in -# declaration order. - -SORT_BRIEF_DOCS = YES - -# If the SORT_GROUP_NAMES tag is set to YES then doxygen will sort the -# hierarchy of group names into alphabetical order. If set to NO (the default) -# the group names will appear in their defined order. - -SORT_GROUP_NAMES = NO - -# If the SORT_BY_SCOPE_NAME tag is set to YES, the class list will be -# sorted by fully-qualified names, including namespaces. If set to -# NO (the default), the class list will be sorted only by class name, -# not including the namespace part. -# Note: This option is not very useful if HIDE_SCOPE_NAMES is set to YES. -# Note: This option applies only to the class list, not to the -# alphabetical list. - -SORT_BY_SCOPE_NAME = NO - -# The GENERATE_TODOLIST tag can be used to enable (YES) or -# disable (NO) the todo list. This list is created by putting \todo -# commands in the documentation. - -GENERATE_TODOLIST = NO - -# The GENERATE_TESTLIST tag can be used to enable (YES) or -# disable (NO) the test list. This list is created by putting \test -# commands in the documentation. - -GENERATE_TESTLIST = NO - -# The GENERATE_BUGLIST tag can be used to enable (YES) or -# disable (NO) the bug list. This list is created by putting \bug -# commands in the documentation. - -GENERATE_BUGLIST = NO - -# The GENERATE_DEPRECATEDLIST tag can be used to enable (YES) or -# disable (NO) the deprecated list. This list is created by putting -# \deprecated commands in the documentation. - -GENERATE_DEPRECATEDLIST= NO - -# The ENABLED_SECTIONS tag can be used to enable conditional -# documentation sections, marked by \if sectionname ... \endif. - -ENABLED_SECTIONS = - -# The MAX_INITIALIZER_LINES tag determines the maximum number of lines -# the initial value of a variable or define consists of for it to appear in -# the documentation. If the initializer consists of more lines than specified -# here it will be hidden. Use a value of 0 to hide initializers completely. -# The appearance of the initializer of individual variables and defines in the -# documentation can be controlled using \showinitializer or \hideinitializer -# command in the documentation regardless of this setting. - -MAX_INITIALIZER_LINES = 0 - -# Set the SHOW_USED_FILES tag to NO to disable the list of files generated -# at the bottom of the documentation of classes and structs. If set to YES the -# list will mention the files that were used to generate the documentation. - -SHOW_USED_FILES = YES - -# If the sources in your project are distributed over multiple directories -# then setting the SHOW_DIRECTORIES tag to YES will show the directory hierarchy -# in the documentation. The default is NO. - -SHOW_DIRECTORIES = NO - -# Set the SHOW_FILES tag to NO to disable the generation of the Files page. -# This will remove the Files entry from the Quick Index and from the -# Folder Tree View (if specified). The default is YES. - -SHOW_FILES = YES - -# Set the SHOW_NAMESPACES tag to NO to disable the generation of the -# Namespaces page. This will remove the Namespaces entry from the Quick Index -# and from the Folder Tree View (if specified). The default is YES. - -SHOW_NAMESPACES = NO - -# The FILE_VERSION_FILTER tag can be used to specify a program or script that -# doxygen should invoke to get the current version for each file (typically from -# the version control system). Doxygen will invoke the program by executing (via -# popen()) the command <command> <input-file>, where <command> is the value of -# the FILE_VERSION_FILTER tag, and <input-file> is the name of an input file -# provided by doxygen. Whatever the program writes to standard output -# is used as the file version. See the manual for examples. - -FILE_VERSION_FILTER = - -#--------------------------------------------------------------------------- -# configuration options related to warning and progress messages -#--------------------------------------------------------------------------- - -# The QUIET tag can be used to turn on/off the messages that are generated -# by doxygen. Possible values are YES and NO. If left blank NO is used. - -QUIET = NO - -# The WARNINGS tag can be used to turn on/off the warning messages that are -# generated by doxygen. Possible values are YES and NO. If left blank -# NO is used. - -WARNINGS = YES - -# If WARN_IF_UNDOCUMENTED is set to YES, then doxygen will generate warnings -# for undocumented members. If EXTRACT_ALL is set to YES then this flag will -# automatically be disabled. - -WARN_IF_UNDOCUMENTED = NO - -# If WARN_IF_DOC_ERROR is set to YES, doxygen will generate warnings for -# potential errors in the documentation, such as not documenting some -# parameters in a documented function, or documenting parameters that -# don't exist or using markup commands wrongly. - -WARN_IF_DOC_ERROR = YES - -# This WARN_NO_PARAMDOC option can be abled to get warnings for -# functions that are documented, but have no documentation for their parameters -# or return value. If set to NO (the default) doxygen will only warn about -# wrong or incomplete parameter documentation, but not about the absence of -# documentation. - -WARN_NO_PARAMDOC = NO - -# The WARN_FORMAT tag determines the format of the warning messages that -# doxygen can produce. The string should contain the $file, $line, and $text -# tags, which will be replaced by the file and line number from which the -# warning originated and the warning text. Optionally the format may contain -# $version, which will be replaced by the version of the file (if it could -# be obtained via FILE_VERSION_FILTER) - -WARN_FORMAT = "$file:$line: $text" - -# The WARN_LOGFILE tag can be used to specify a file to which warning -# and error messages should be written. If left blank the output is written -# to stderr. - -WARN_LOGFILE = - -#--------------------------------------------------------------------------- -# configuration options related to the input files -#--------------------------------------------------------------------------- - -# The INPUT tag can be used to specify the files and/or directories that contain -# documented source files. You may enter file names like "myfile.cpp" or -# directories like "/usr/src/myproject". Separate the files or directories -# with spaces. - -INPUT = "${Eigen_SOURCE_DIR}/unsupported/Eigen" \ - "${Eigen_SOURCE_DIR}/unsupported/doc" - -# This tag can be used to specify the character encoding of the source files -# that doxygen parses. Internally doxygen uses the UTF-8 encoding, which is -# also the default input encoding. Doxygen uses libiconv (or the iconv built -# into libc) for the transcoding. See http://www.gnu.org/software/libiconv for -# the list of possible encodings. - -INPUT_ENCODING = UTF-8 - -# If the value of the INPUT tag contains directories, you can use the -# FILE_PATTERNS tag to specify one or more wildcard pattern (like *.cpp -# and *.h) to filter out the source-files in the directories. If left -# blank the following patterns are tested: -# *.c *.cc *.cxx *.cpp *.c++ *.java *.ii *.ixx *.ipp *.i++ *.inl *.h *.hh *.hxx -# *.hpp *.h++ *.idl *.odl *.cs *.php *.php3 *.inc *.m *.mm *.py *.f90 - -FILE_PATTERNS = * - -# The RECURSIVE tag can be used to turn specify whether or not subdirectories -# should be searched for input files as well. Possible values are YES and NO. -# If left blank NO is used. - -RECURSIVE = YES - -# The EXCLUDE tag can be used to specify files and/or directories that should -# excluded from the INPUT source files. This way you can easily exclude a -# subdirectory from a directory tree whose root is specified with the INPUT tag. - -EXCLUDE = "${Eigen_SOURCE_DIR}/unsupported/doc/examples" \ - "${Eigen_SOURCE_DIR}/unsupported/doc/snippets" - -# The EXCLUDE_SYMLINKS tag can be used select whether or not files or -# directories that are symbolic links (a Unix filesystem feature) are excluded -# from the input. - -EXCLUDE_SYMLINKS = NO - -# If the value of the INPUT tag contains directories, you can use the -# EXCLUDE_PATTERNS tag to specify one or more wildcard patterns to exclude -# certain files from those directories. Note that the wildcards are matched -# against the file with absolute path, so to exclude all test directories -# for example use the pattern */test/* - -EXCLUDE_PATTERNS = CMake* \ - *.txt \ - *.sh \ - *.diff \ - *.orig \ - diff \ - *~ - -# The EXCLUDE_SYMBOLS tag can be used to specify one or more symbol names -# (namespaces, classes, functions, etc.) that should be excluded from the -# output. The symbol name can be a fully qualified name, a word, or if the -# wildcard * is used, a substring. Examples: ANamespace, AClass, -# AClass::ANamespace, ANamespace::*Test - -EXCLUDE_SYMBOLS = MatrixBase<* MapBase<* RotationBase<* Matrix<* - -# The EXAMPLE_PATH tag can be used to specify one or more files or -# directories that contain example code fragments that are included (see -# the \include command). - -EXAMPLE_PATH = "${Eigen_SOURCE_DIR}/doc/snippets" \ - "${Eigen_BINARY_DIR}/doc/snippets" \ - "${Eigen_SOURCE_DIR}/doc/examples" \ - "${Eigen_BINARY_DIR}/doc/examples" \ - "${Eigen_SOURCE_DIR}/unsupported/doc/snippets" \ - "${Eigen_BINARY_DIR}/unsupported/doc/snippets" \ - "${Eigen_SOURCE_DIR}/unsupported/doc/examples" \ - "${Eigen_BINARY_DIR}/unsupported/doc/examples" - -# If the value of the EXAMPLE_PATH tag contains directories, you can use the -# EXAMPLE_PATTERNS tag to specify one or more wildcard pattern (like *.cpp -# and *.h) to filter out the source-files in the directories. If left -# blank all files are included. - -EXAMPLE_PATTERNS = * - -# If the EXAMPLE_RECURSIVE tag is set to YES then subdirectories will be -# searched for input files to be used with the \include or \dontinclude -# commands irrespective of the value of the RECURSIVE tag. -# Possible values are YES and NO. If left blank NO is used. - -EXAMPLE_RECURSIVE = NO - -# The IMAGE_PATH tag can be used to specify one or more files or -# directories that contain image that are included in the documentation (see -# the \image command). - -IMAGE_PATH = - -# The INPUT_FILTER tag can be used to specify a program that doxygen should -# invoke to filter for each input file. Doxygen will invoke the filter program -# by executing (via popen()) the command <filter> <input-file>, where <filter> -# is the value of the INPUT_FILTER tag, and <input-file> is the name of an -# input file. Doxygen will then use the output that the filter program writes -# to standard output. If FILTER_PATTERNS is specified, this tag will be -# ignored. - -INPUT_FILTER = - -# The FILTER_PATTERNS tag can be used to specify filters on a per file pattern -# basis. Doxygen will compare the file name with each pattern and apply the -# filter if there is a match. The filters are a list of the form: -# pattern=filter (like *.cpp=my_cpp_filter). See INPUT_FILTER for further -# info on how filters are used. If FILTER_PATTERNS is empty, INPUT_FILTER -# is applied to all files. - -FILTER_PATTERNS = - -# If the FILTER_SOURCE_FILES tag is set to YES, the input filter (if set using -# INPUT_FILTER) will be used to filter the input files when producing source -# files to browse (i.e. when SOURCE_BROWSER is set to YES). - -FILTER_SOURCE_FILES = NO - -#--------------------------------------------------------------------------- -# configuration options related to source browsing -#--------------------------------------------------------------------------- - -# If the SOURCE_BROWSER tag is set to YES then a list of source files will -# be generated. Documented entities will be cross-referenced with these sources. -# Note: To get rid of all source code in the generated output, make sure also -# VERBATIM_HEADERS is set to NO. - -SOURCE_BROWSER = NO - -# Setting the INLINE_SOURCES tag to YES will include the body -# of functions and classes directly in the documentation. - -INLINE_SOURCES = NO - -# Setting the STRIP_CODE_COMMENTS tag to YES (the default) will instruct -# doxygen to hide any special comment blocks from generated source code -# fragments. Normal C and C++ comments will always remain visible. - -STRIP_CODE_COMMENTS = YES - -# If the REFERENCED_BY_RELATION tag is set to YES -# then for each documented function all documented -# functions referencing it will be listed. - -REFERENCED_BY_RELATION = YES - -# If the REFERENCES_RELATION tag is set to YES -# then for each documented function all documented entities -# called/used by that function will be listed. - -REFERENCES_RELATION = YES - -# If the REFERENCES_LINK_SOURCE tag is set to YES (the default) -# and SOURCE_BROWSER tag is set to YES, then the hyperlinks from -# functions in REFERENCES_RELATION and REFERENCED_BY_RELATION lists will -# link to the source code. Otherwise they will link to the documentstion. - -REFERENCES_LINK_SOURCE = YES - -# If the USE_HTAGS tag is set to YES then the references to source code -# will point to the HTML generated by the htags(1) tool instead of doxygen -# built-in source browser. The htags tool is part of GNU's global source -# tagging system (see http://www.gnu.org/software/global/global.html). You -# will need version 4.8.6 or higher. - -USE_HTAGS = NO - -# If the VERBATIM_HEADERS tag is set to YES (the default) then Doxygen -# will generate a verbatim copy of the header file for each class for -# which an include is specified. Set to NO to disable this. - -VERBATIM_HEADERS = YES - -#--------------------------------------------------------------------------- -# configuration options related to the alphabetical class index -#--------------------------------------------------------------------------- - -# If the ALPHABETICAL_INDEX tag is set to YES, an alphabetical index -# of all compounds will be generated. Enable this if the project -# contains a lot of classes, structs, unions or interfaces. - -ALPHABETICAL_INDEX = NO - -# If the alphabetical index is enabled (see ALPHABETICAL_INDEX) then -# the COLS_IN_ALPHA_INDEX tag can be used to specify the number of columns -# in which this list will be split (can be a number in the range [1..20]) - -COLS_IN_ALPHA_INDEX = 5 - -# In case all classes in a project start with a common prefix, all -# classes will be put under the same header in the alphabetical index. -# The IGNORE_PREFIX tag can be used to specify one or more prefixes that -# should be ignored while generating the index headers. - -IGNORE_PREFIX = - -#--------------------------------------------------------------------------- -# configuration options related to the HTML output -#--------------------------------------------------------------------------- - -# If the GENERATE_HTML tag is set to YES (the default) Doxygen will -# generate HTML output. - -GENERATE_HTML = YES - -# The HTML_OUTPUT tag is used to specify where the HTML docs will be put. -# If a relative path is entered the value of OUTPUT_DIRECTORY will be -# put in front of it. If left blank `html' will be used as the default path. - -HTML_OUTPUT = "${Eigen_BINARY_DIR}/doc/html/unsupported" - -# The HTML_FILE_EXTENSION tag can be used to specify the file extension for -# each generated HTML page (for example: .htm,.php,.asp). If it is left blank -# doxygen will generate files with .html extension. - -HTML_FILE_EXTENSION = .html - -# The HTML_HEADER tag can be used to specify a personal HTML header for -# each generated HTML page. If it is left blank doxygen will generate a -# standard header. - -HTML_HEADER = "${Eigen_BINARY_DIR}/doc/eigendoxy_header.html" - -# The HTML_FOOTER tag can be used to specify a personal HTML footer for -# each generated HTML page. If it is left blank doxygen will generate a -# standard footer. - -# the footer has not been customized yet, so let's use the default one -# ${Eigen_BINARY_DIR}/doc/eigendoxy_footer.html -HTML_FOOTER = - -# The HTML_STYLESHEET tag can be used to specify a user-defined cascading -# style sheet that is used by each HTML page. It can be used to -# fine-tune the look of the HTML output. If the tag is left blank doxygen -# will generate a default style sheet. Note that doxygen will try to copy -# the style sheet file to the HTML output directory, so don't put your own -# stylesheet in the HTML output directory as well, or it will be erased! - -HTML_STYLESHEET = "${Eigen_SOURCE_DIR}/doc/eigendoxy.css" - -# If the HTML_ALIGN_MEMBERS tag is set to YES, the members of classes, -# files or namespaces will be aligned in HTML using tables. If set to -# NO a bullet list will be used. - -HTML_ALIGN_MEMBERS = YES - -# If the GENERATE_HTMLHELP tag is set to YES, additional index files -# will be generated that can be used as input for tools like the -# Microsoft HTML help workshop to generate a compiled HTML help file (.chm) -# of the generated HTML documentation. - -GENERATE_HTMLHELP = NO - -# If the GENERATE_DOCSET tag is set to YES, additional index files -# will be generated that can be used as input for Apple's Xcode 3 -# integrated development environment, introduced with OSX 10.5 (Leopard). -# To create a documentation set, doxygen will generate a Makefile in the -# HTML output directory. Running make will produce the docset in that -# directory and running "make install" will install the docset in -# ~/Library/Developer/Shared/Documentation/DocSets so that Xcode will find -# it at startup. - -GENERATE_DOCSET = NO - -# When GENERATE_DOCSET tag is set to YES, this tag determines the name of the -# feed. A documentation feed provides an umbrella under which multiple -# documentation sets from a single provider (such as a company or product suite) -# can be grouped. - -DOCSET_FEEDNAME = "Doxygen generated docs" - -# When GENERATE_DOCSET tag is set to YES, this tag specifies a string that -# should uniquely identify the documentation set bundle. This should be a -# reverse domain-name style string, e.g. com.mycompany.MyDocSet. Doxygen -# will append .docset to the name. - -DOCSET_BUNDLE_ID = org.doxygen.Project - -# If the HTML_DYNAMIC_SECTIONS tag is set to YES then the generated HTML -# documentation will contain sections that can be hidden and shown after the -# page has loaded. For this to work a browser that supports -# JavaScript and DHTML is required (for instance Mozilla 1.0+, Firefox -# Netscape 6.0+, Internet explorer 5.0+, Konqueror, or Safari). - -HTML_DYNAMIC_SECTIONS = NO - -# If the GENERATE_HTMLHELP tag is set to YES, the CHM_FILE tag can -# be used to specify the file name of the resulting .chm file. You -# can add a path in front of the file if the result should not be -# written to the html output directory. - -CHM_FILE = - -# If the GENERATE_HTMLHELP tag is set to YES, the HHC_LOCATION tag can -# be used to specify the location (absolute path including file name) of -# the HTML help compiler (hhc.exe). If non-empty doxygen will try to run -# the HTML help compiler on the generated index.hhp. - -HHC_LOCATION = - -# If the GENERATE_HTMLHELP tag is set to YES, the GENERATE_CHI flag -# controls if a separate .chi index file is generated (YES) or that -# it should be included in the master .chm file (NO). - -GENERATE_CHI = NO - -# If the GENERATE_HTMLHELP tag is set to YES, the CHM_INDEX_ENCODING -# is used to encode HtmlHelp index (hhk), content (hhc) and project file -# content. - -CHM_INDEX_ENCODING = - -# If the GENERATE_HTMLHELP tag is set to YES, the BINARY_TOC flag -# controls whether a binary table of contents is generated (YES) or a -# normal table of contents (NO) in the .chm file. - -BINARY_TOC = NO - -# The TOC_EXPAND flag can be set to YES to add extra items for group members -# to the contents of the HTML help documentation and to the tree view. - -TOC_EXPAND = NO - -# The DISABLE_INDEX tag can be used to turn on/off the condensed index at -# top of each HTML page. The value NO (the default) enables the index and -# the value YES disables it. - -DISABLE_INDEX = NO - -# This tag can be used to set the number of enum values (range [1..20]) -# that doxygen will group on one line in the generated HTML documentation. - -ENUM_VALUES_PER_LINE = 1 - -# The GENERATE_TREEVIEW tag is used to specify whether a tree-like index -# structure should be generated to display hierarchical information. -# If the tag value is set to FRAME, a side panel will be generated -# containing a tree-like index structure (just like the one that -# is generated for HTML Help). For this to work a browser that supports -# JavaScript, DHTML, CSS and frames is required (for instance Mozilla 1.0+, -# Netscape 6.0+, Internet explorer 5.0+, or Konqueror). Windows users are -# probably better off using the HTML help feature. Other possible values -# for this tag are: HIERARCHIES, which will generate the Groups, Directories, -# and Class Hiererachy pages using a tree view instead of an ordered list; -# ALL, which combines the behavior of FRAME and HIERARCHIES; and NONE, which -# disables this behavior completely. For backwards compatibility with previous -# releases of Doxygen, the values YES and NO are equivalent to FRAME and NONE -# respectively. - -GENERATE_TREEVIEW = NO - -# If the treeview is enabled (see GENERATE_TREEVIEW) then this tag can be -# used to set the initial width (in pixels) of the frame in which the tree -# is shown. - -TREEVIEW_WIDTH = 250 - -# Use this tag to change the font size of Latex formulas included -# as images in the HTML documentation. The default is 10. Note that -# when you change the font size after a successful doxygen run you need -# to manually remove any form_*.png images from the HTML output directory -# to force them to be regenerated. - -FORMULA_FONTSIZE = 12 - -#--------------------------------------------------------------------------- -# configuration options related to the LaTeX output -#--------------------------------------------------------------------------- - -# If the GENERATE_LATEX tag is set to YES (the default) Doxygen will -# generate Latex output. - -GENERATE_LATEX = NO - -# The LATEX_OUTPUT tag is used to specify where the LaTeX docs will be put. -# If a relative path is entered the value of OUTPUT_DIRECTORY will be -# put in front of it. If left blank `latex' will be used as the default path. - -LATEX_OUTPUT = latex - -# The LATEX_CMD_NAME tag can be used to specify the LaTeX command name to be -# invoked. If left blank `latex' will be used as the default command name. - -LATEX_CMD_NAME = latex - -# The MAKEINDEX_CMD_NAME tag can be used to specify the command name to -# generate index for LaTeX. If left blank `makeindex' will be used as the -# default command name. - -MAKEINDEX_CMD_NAME = makeindex - -# If the COMPACT_LATEX tag is set to YES Doxygen generates more compact -# LaTeX documents. This may be useful for small projects and may help to -# save some trees in general. - -COMPACT_LATEX = NO - -# The PAPER_TYPE tag can be used to set the paper type that is used -# by the printer. Possible values are: a4, a4wide, letter, legal and -# executive. If left blank a4wide will be used. - -PAPER_TYPE = a4wide - -# The EXTRA_PACKAGES tag can be to specify one or more names of LaTeX -# packages that should be included in the LaTeX output. - -EXTRA_PACKAGES = amssymb \ - amsmath - -# The LATEX_HEADER tag can be used to specify a personal LaTeX header for -# the generated latex document. The header should contain everything until -# the first chapter. If it is left blank doxygen will generate a -# standard header. Notice: only use this tag if you know what you are doing! - -LATEX_HEADER = - -# If the PDF_HYPERLINKS tag is set to YES, the LaTeX that is generated -# is prepared for conversion to pdf (using ps2pdf). The pdf file will -# contain links (just like the HTML output) instead of page references -# This makes the output suitable for online browsing using a pdf viewer. - -PDF_HYPERLINKS = NO - -# If the USE_PDFLATEX tag is set to YES, pdflatex will be used instead of -# plain latex in the generated Makefile. Set this option to YES to get a -# higher quality PDF documentation. - -USE_PDFLATEX = NO - -# If the LATEX_BATCHMODE tag is set to YES, doxygen will add the \\batchmode. -# command to the generated LaTeX files. This will instruct LaTeX to keep -# running if errors occur, instead of asking the user for help. -# This option is also used when generating formulas in HTML. - -LATEX_BATCHMODE = NO - -# If LATEX_HIDE_INDICES is set to YES then doxygen will not -# include the index chapters (such as File Index, Compound Index, etc.) -# in the output. - -LATEX_HIDE_INDICES = NO - -#--------------------------------------------------------------------------- -# configuration options related to the RTF output -#--------------------------------------------------------------------------- - -# If the GENERATE_RTF tag is set to YES Doxygen will generate RTF output -# The RTF output is optimized for Word 97 and may not look very pretty with -# other RTF readers or editors. - -GENERATE_RTF = NO - -# The RTF_OUTPUT tag is used to specify where the RTF docs will be put. -# If a relative path is entered the value of OUTPUT_DIRECTORY will be -# put in front of it. If left blank `rtf' will be used as the default path. - -RTF_OUTPUT = rtf - -# If the COMPACT_RTF tag is set to YES Doxygen generates more compact -# RTF documents. This may be useful for small projects and may help to -# save some trees in general. - -COMPACT_RTF = NO - -# If the RTF_HYPERLINKS tag is set to YES, the RTF that is generated -# will contain hyperlink fields. The RTF file will -# contain links (just like the HTML output) instead of page references. -# This makes the output suitable for online browsing using WORD or other -# programs which support those fields. -# Note: wordpad (write) and others do not support links. - -RTF_HYPERLINKS = NO - -# Load stylesheet definitions from file. Syntax is similar to doxygen's -# config file, i.e. a series of assignments. You only have to provide -# replacements, missing definitions are set to their default value. - -RTF_STYLESHEET_FILE = - -# Set optional variables used in the generation of an rtf document. -# Syntax is similar to doxygen's config file. - -RTF_EXTENSIONS_FILE = - -#--------------------------------------------------------------------------- -# configuration options related to the man page output -#--------------------------------------------------------------------------- - -# If the GENERATE_MAN tag is set to YES (the default) Doxygen will -# generate man pages - -GENERATE_MAN = NO - -# The MAN_OUTPUT tag is used to specify where the man pages will be put. -# If a relative path is entered the value of OUTPUT_DIRECTORY will be -# put in front of it. If left blank `man' will be used as the default path. - -MAN_OUTPUT = man - -# The MAN_EXTENSION tag determines the extension that is added to -# the generated man pages (default is the subroutine's section .3) - -MAN_EXTENSION = .3 - -# If the MAN_LINKS tag is set to YES and Doxygen generates man output, -# then it will generate one additional man file for each entity -# documented in the real man page(s). These additional files -# only source the real man page, but without them the man command -# would be unable to find the correct page. The default is NO. - -MAN_LINKS = NO - -#--------------------------------------------------------------------------- -# configuration options related to the XML output -#--------------------------------------------------------------------------- - -# If the GENERATE_XML tag is set to YES Doxygen will -# generate an XML file that captures the structure of -# the code including all documentation. - -GENERATE_XML = NO - -# The XML_OUTPUT tag is used to specify where the XML pages will be put. -# If a relative path is entered the value of OUTPUT_DIRECTORY will be -# put in front of it. If left blank `xml' will be used as the default path. - -XML_OUTPUT = xml - -# The XML_SCHEMA tag can be used to specify an XML schema, -# which can be used by a validating XML parser to check the -# syntax of the XML files. - -XML_SCHEMA = - -# The XML_DTD tag can be used to specify an XML DTD, -# which can be used by a validating XML parser to check the -# syntax of the XML files. - -XML_DTD = - -# If the XML_PROGRAMLISTING tag is set to YES Doxygen will -# dump the program listings (including syntax highlighting -# and cross-referencing information) to the XML output. Note that -# enabling this will significantly increase the size of the XML output. - -XML_PROGRAMLISTING = YES - -#--------------------------------------------------------------------------- -# configuration options for the AutoGen Definitions output -#--------------------------------------------------------------------------- - -# If the GENERATE_AUTOGEN_DEF tag is set to YES Doxygen will -# generate an AutoGen Definitions (see autogen.sf.net) file -# that captures the structure of the code including all -# documentation. Note that this feature is still experimental -# and incomplete at the moment. - -GENERATE_AUTOGEN_DEF = NO - -#--------------------------------------------------------------------------- -# configuration options related to the Perl module output -#--------------------------------------------------------------------------- - -# If the GENERATE_PERLMOD tag is set to YES Doxygen will -# generate a Perl module file that captures the structure of -# the code including all documentation. Note that this -# feature is still experimental and incomplete at the -# moment. - -GENERATE_PERLMOD = NO - -# If the PERLMOD_LATEX tag is set to YES Doxygen will generate -# the necessary Makefile rules, Perl scripts and LaTeX code to be able -# to generate PDF and DVI output from the Perl module output. - -PERLMOD_LATEX = NO - -# If the PERLMOD_PRETTY tag is set to YES the Perl module output will be -# nicely formatted so it can be parsed by a human reader. This is useful -# if you want to understand what is going on. On the other hand, if this -# tag is set to NO the size of the Perl module output will be much smaller -# and Perl will parse it just the same. - -PERLMOD_PRETTY = YES - -# The names of the make variables in the generated doxyrules.make file -# are prefixed with the string contained in PERLMOD_MAKEVAR_PREFIX. -# This is useful so different doxyrules.make files included by the same -# Makefile don't overwrite each other's variables. - -PERLMOD_MAKEVAR_PREFIX = - -#--------------------------------------------------------------------------- -# Configuration options related to the preprocessor -#--------------------------------------------------------------------------- - -# If the ENABLE_PREPROCESSING tag is set to YES (the default) Doxygen will -# evaluate all C-preprocessor directives found in the sources and include -# files. - -ENABLE_PREPROCESSING = YES - -# If the MACRO_EXPANSION tag is set to YES Doxygen will expand all macro -# names in the source code. If set to NO (the default) only conditional -# compilation will be performed. Macro expansion can be done in a controlled -# way by setting EXPAND_ONLY_PREDEF to YES. - -MACRO_EXPANSION = YES - -# If the EXPAND_ONLY_PREDEF and MACRO_EXPANSION tags are both set to YES -# then the macro expansion is limited to the macros specified with the -# PREDEFINED and EXPAND_AS_DEFINED tags. - -EXPAND_ONLY_PREDEF = YES - -# If the SEARCH_INCLUDES tag is set to YES (the default) the includes files -# in the INCLUDE_PATH (see below) will be search if a #include is found. - -SEARCH_INCLUDES = YES - -# The INCLUDE_PATH tag can be used to specify one or more directories that -# contain include files that are not input files but should be processed by -# the preprocessor. - -INCLUDE_PATH = - -# You can use the INCLUDE_FILE_PATTERNS tag to specify one or more wildcard -# patterns (like *.h and *.hpp) to filter out the header-files in the -# directories. If left blank, the patterns specified with FILE_PATTERNS will -# be used. - -INCLUDE_FILE_PATTERNS = - -# The PREDEFINED tag can be used to specify one or more macro names that -# are defined before the preprocessor is started (similar to the -D option of -# gcc). The argument of the tag is a list of macros of the form: name -# or name=definition (no spaces). If the definition and the = are -# omitted =1 is assumed. To prevent a macro definition from being -# undefined via #undef or recursively expanded use the := operator -# instead of the = operator. - -PREDEFINED = EIGEN_EMPTY_STRUCT \ - EIGEN_PARSED_BY_DOXYGEN \ - EIGEN_VECTORIZE \ - EIGEN_QT_SUPPORT \ - EIGEN_STRONG_INLINE=inline - -# If the MACRO_EXPANSION and EXPAND_ONLY_PREDEF tags are set to YES then -# this tag can be used to specify a list of macro names that should be expanded. -# The macro definition that is found in the sources will be used. -# Use the PREDEFINED tag if you want to use a different macro definition. - -EXPAND_AS_DEFINED = EIGEN_MAKE_SCALAR_OPS \ - EIGEN_MAKE_TYPEDEFS \ - EIGEN_MAKE_TYPEDEFS_ALL_SIZES \ - EIGEN_CWISE_UNOP_RETURN_TYPE \ - EIGEN_CWISE_BINOP_RETURN_TYPE - -# If the SKIP_FUNCTION_MACROS tag is set to YES (the default) then -# doxygen's preprocessor will remove all function-like macros that are alone -# on a line, have an all uppercase name, and do not end with a semicolon. Such -# function macros are typically used for boiler-plate code, and will confuse -# the parser if not removed. - -SKIP_FUNCTION_MACROS = YES - -#--------------------------------------------------------------------------- -# Configuration::additions related to external references -#--------------------------------------------------------------------------- - -# The TAGFILES option can be used to specify one or more tagfiles. -# Optionally an initial location of the external documentation -# can be added for each tagfile. The format of a tag file without -# this location is as follows: -# TAGFILES = file1 file2 ... -# Adding location for the tag files is done as follows: -# TAGFILES = file1=loc1 "file2 = loc2" ... -# where "loc1" and "loc2" can be relative or absolute paths or -# URLs. If a location is present for each tag, the installdox tool -# does not have to be run to correct the links. -# Note that each tag file must have a unique name -# (where the name does NOT include the path) -# If a tag file is not located in the directory in which doxygen -# is run, you must also specify the path to the tagfile here. - -TAGFILES = "${Eigen_BINARY_DIR}/doc/eigen.doxytags"=../ - -# When a file name is specified after GENERATE_TAGFILE, doxygen will create -# a tag file that is based on the input files it reads. - -GENERATE_TAGFILE = "${Eigen_BINARY_DIR}/doc/eigen-unsupported.doxytags" - -# If the ALLEXTERNALS tag is set to YES all external classes will be listed -# in the class index. If set to NO only the inherited external classes -# will be listed. - -ALLEXTERNALS = NO - -# If the EXTERNAL_GROUPS tag is set to YES all external groups will be listed -# in the modules index. If set to NO, only the current project's groups will -# be listed. - -EXTERNAL_GROUPS = YES - -# The PERL_PATH should be the absolute path and name of the perl script -# interpreter (i.e. the result of `which perl'). - -PERL_PATH = /usr/bin/perl - -#--------------------------------------------------------------------------- -# Configuration options related to the dot tool -#--------------------------------------------------------------------------- - -# If the CLASS_DIAGRAMS tag is set to YES (the default) Doxygen will -# generate a inheritance diagram (in HTML, RTF and LaTeX) for classes with base -# or super classes. Setting the tag to NO turns the diagrams off. Note that -# this option is superseded by the HAVE_DOT option below. This is only a -# fallback. It is recommended to install and use dot, since it yields more -# powerful graphs. - -CLASS_DIAGRAMS = NO - -# You can define message sequence charts within doxygen comments using the \msc -# command. Doxygen will then run the mscgen tool (see -# http://www.mcternan.me.uk/mscgen/) to produce the chart and insert it in the -# documentation. The MSCGEN_PATH tag allows you to specify the directory where -# the mscgen tool resides. If left empty the tool is assumed to be found in the -# default search path. - -MSCGEN_PATH = NO - -# If set to YES, the inheritance and collaboration graphs will hide -# inheritance and usage relations if the target is undocumented -# or is not a class. - -HIDE_UNDOC_RELATIONS = NO - -# If you set the HAVE_DOT tag to YES then doxygen will assume the dot tool is -# available from the path. This tool is part of Graphviz, a graph visualization -# toolkit from AT&T and Lucent Bell Labs. The other options in this section -# have no effect if this option is set to NO (the default) - -HAVE_DOT = NO - -# By default doxygen will write a font called FreeSans.ttf to the output -# directory and reference it in all dot files that doxygen generates. This -# font does not include all possible unicode characters however, so when you need -# these (or just want a differently looking font) you can specify the font name -# using DOT_FONTNAME. You need need to make sure dot is able to find the font, -# which can be done by putting it in a standard location or by setting the -# DOTFONTPATH environment variable or by setting DOT_FONTPATH to the directory -# containing the font. - -DOT_FONTNAME = FreeSans - -# By default doxygen will tell dot to use the output directory to look for the -# FreeSans.ttf font (which doxygen will put there itself). If you specify a -# different font using DOT_FONTNAME you can set the path where dot -# can find it using this tag. - -DOT_FONTPATH = - -# If the CLASS_GRAPH and HAVE_DOT tags are set to YES then doxygen -# will generate a graph for each documented class showing the direct and -# indirect inheritance relations. Setting this tag to YES will force the -# the CLASS_DIAGRAMS tag to NO. - -CLASS_GRAPH = NO - -# If the COLLABORATION_GRAPH and HAVE_DOT tags are set to YES then doxygen -# will generate a graph for each documented class showing the direct and -# indirect implementation dependencies (inheritance, containment, and -# class references variables) of the class with other documented classes. - -COLLABORATION_GRAPH = NO - -# If the GROUP_GRAPHS and HAVE_DOT tags are set to YES then doxygen -# will generate a graph for groups, showing the direct groups dependencies - -GROUP_GRAPHS = NO - -# If the UML_LOOK tag is set to YES doxygen will generate inheritance and -# collaboration diagrams in a style similar to the OMG's Unified Modeling -# Language. - -UML_LOOK = NO - -# If set to YES, the inheritance and collaboration graphs will show the -# relations between templates and their instances. - -TEMPLATE_RELATIONS = NO - -# If the ENABLE_PREPROCESSING, SEARCH_INCLUDES, INCLUDE_GRAPH, and HAVE_DOT -# tags are set to YES then doxygen will generate a graph for each documented -# file showing the direct and indirect include dependencies of the file with -# other documented files. - -INCLUDE_GRAPH = NO - -# If the ENABLE_PREPROCESSING, SEARCH_INCLUDES, INCLUDED_BY_GRAPH, and -# HAVE_DOT tags are set to YES then doxygen will generate a graph for each -# documented header file showing the documented files that directly or -# indirectly include this file. - -INCLUDED_BY_GRAPH = NO - -# If the CALL_GRAPH and HAVE_DOT options are set to YES then -# doxygen will generate a call dependency graph for every global function -# or class method. Note that enabling this option will significantly increase -# the time of a run. So in most cases it will be better to enable call graphs -# for selected functions only using the \callgraph command. - -CALL_GRAPH = NO - -# If the CALLER_GRAPH and HAVE_DOT tags are set to YES then -# doxygen will generate a caller dependency graph for every global function -# or class method. Note that enabling this option will significantly increase -# the time of a run. So in most cases it will be better to enable caller -# graphs for selected functions only using the \callergraph command. - -CALLER_GRAPH = NO - -# If the GRAPHICAL_HIERARCHY and HAVE_DOT tags are set to YES then doxygen -# will graphical hierarchy of all classes instead of a textual one. - -GRAPHICAL_HIERARCHY = NO - -# If the DIRECTORY_GRAPH, SHOW_DIRECTORIES and HAVE_DOT tags are set to YES -# then doxygen will show the dependencies a directory has on other directories -# in a graphical way. The dependency relations are determined by the #include -# relations between the files in the directories. - -DIRECTORY_GRAPH = NO - -# The DOT_IMAGE_FORMAT tag can be used to set the image format of the images -# generated by dot. Possible values are png, jpg, or gif -# If left blank png will be used. - -DOT_IMAGE_FORMAT = png - -# The tag DOT_PATH can be used to specify the path where the dot tool can be -# found. If left blank, it is assumed the dot tool can be found in the path. - -DOT_PATH = - -# The DOTFILE_DIRS tag can be used to specify one or more directories that -# contain dot files that are included in the documentation (see the -# \dotfile command). - -DOTFILE_DIRS = - -# The DOT_GRAPH_MAX_NODES tag can be used to set the maximum number of -# nodes that will be shown in the graph. If the number of nodes in a graph -# becomes larger than this value, doxygen will truncate the graph, which is -# visualized by representing a node as a red box. Note that doxygen if the -# number of direct children of the root node in a graph is already larger than -# DOT_GRAPH_MAX_NODES then the graph will not be shown at all. Also note -# that the size of a graph can be further restricted by MAX_DOT_GRAPH_DEPTH. - -DOT_GRAPH_MAX_NODES = 50 - -# The MAX_DOT_GRAPH_DEPTH tag can be used to set the maximum depth of the -# graphs generated by dot. A depth value of 3 means that only nodes reachable -# from the root by following a path via at most 3 edges will be shown. Nodes -# that lay further from the root node will be omitted. Note that setting this -# option to 1 or 2 may greatly reduce the computation time needed for large -# code bases. Also note that the size of a graph can be further restricted by -# DOT_GRAPH_MAX_NODES. Using a depth of 0 means no depth restriction. - -MAX_DOT_GRAPH_DEPTH = 1000 - -# Set the DOT_TRANSPARENT tag to YES to generate images with a transparent -# background. This is enabled by default, which results in a transparent -# background. Warning: Depending on the platform used, enabling this option -# may lead to badly anti-aliased labels on the edges of a graph (i.e. they -# become hard to read). - -DOT_TRANSPARENT = NO - -# Set the DOT_MULTI_TARGETS tag to YES allow dot to generate multiple output -# files in one run (i.e. multiple -o and -T options on the command line). This -# makes dot run faster, but since only newer versions of dot (>1.8.10) -# support this, this feature is disabled by default. - -DOT_MULTI_TARGETS = NO - -# If the GENERATE_LEGEND tag is set to YES (the default) Doxygen will -# generate a legend page explaining the meaning of the various boxes and -# arrows in the dot generated graphs. - -GENERATE_LEGEND = NO - -# If the DOT_CLEANUP tag is set to YES (the default) Doxygen will -# remove the intermediate dot files that are used to generate -# the various graphs. - -DOT_CLEANUP = NO - -#--------------------------------------------------------------------------- -# Configuration::additions related to the search engine -#--------------------------------------------------------------------------- - -# The SEARCHENGINE tag specifies whether or not a search engine should be -# used. If set to NO the values of all tags below this one will be ignored. - -SEARCHENGINE = NO diff --git a/unsupported/doc/Overview.dox b/unsupported/doc/Overview.dox index 458b507b5..d048377df 100644 --- a/unsupported/doc/Overview.dox +++ b/unsupported/doc/Overview.dox @@ -10,6 +10,9 @@ Click on the \e Modules tab at the top of this page to get a list of all unsuppo Don't miss the <a href="..//index.html">official Eigen documentation</a>. +*/ + +/* \defgroup Unsupported_modules Unsupported modules diff --git a/unsupported/doc/eigendoxy_layout.xml.in b/unsupported/doc/eigendoxy_layout.xml.in new file mode 100644 index 000000000..c93621ed3 --- /dev/null +++ b/unsupported/doc/eigendoxy_layout.xml.in @@ -0,0 +1,177 @@ +<?xml version="1.0"?> +<doxygenlayout version="1.0"> + <!-- Navigation index tabs for HTML output --> + <navindex> + <tab type="user" url="index.html" title="Overview" /> + <tab type="modules" visible="yes" title="Unsupported Modules" intro=""/> +<!-- <tab type="mainpage" visible="yes" title=""/> --> + <tab type="classlist" visible="yes" title="" intro=""/> +<!-- <tab type="classmembers" visible="yes" title="" intro=""/> --> + </navindex> + + <!-- Layout definition for a class page --> + <class> + <briefdescription visible="no"/> + <includes visible="$SHOW_INCLUDE_FILES"/> + <detaileddescription title=""/> + <inheritancegraph visible="$CLASS_GRAPH"/> + <collaborationgraph visible="$COLLABORATION_GRAPH"/> + <allmemberslink visible="yes"/> + <memberdecl> + <nestedclasses visible="yes" title=""/> + <publictypes title=""/> + <publicslots title=""/> + <signals title=""/> + <publicmethods title=""/> + <publicstaticmethods title=""/> + <publicattributes title=""/> + <publicstaticattributes title=""/> + <protectedtypes title=""/> + <protectedslots title=""/> + <protectedmethods title=""/> + <protectedstaticmethods title=""/> + <protectedattributes title=""/> + <protectedstaticattributes title=""/> + <packagetypes title=""/> + <packagemethods title=""/> + <packagestaticmethods title=""/> + <packageattributes title=""/> + <packagestaticattributes title=""/> + <properties title=""/> + <events title=""/> + <privatetypes title=""/> + <privateslots title=""/> + <privatemethods title=""/> + <privatestaticmethods title=""/> + <privateattributes title=""/> + <privatestaticattributes title=""/> + <friends title=""/> + <related title="" subtitle=""/> + <membergroups visible="yes"/> + </memberdecl> + + <memberdef> + <inlineclasses title=""/> + <typedefs title=""/> + <enums title=""/> + <constructors title=""/> + <functions title=""/> + <related title=""/> + <variables title=""/> + <properties title=""/> + <events title=""/> + </memberdef> + <usedfiles visible="$SHOW_USED_FILES"/> + <authorsection visible="yes"/> + </class> + + <!-- Layout definition for a namespace page --> + <namespace> + <briefdescription visible="yes"/> + <memberdecl> + <nestednamespaces visible="yes" title=""/> + <classes visible="yes" title=""/> + <typedefs title=""/> + <enums title=""/> + <functions title=""/> + <variables title=""/> + <membergroups visible="yes"/> + </memberdecl> + <detaileddescription title=""/> + <memberdef> + <inlineclasses title=""/> + <typedefs title=""/> + <enums title=""/> + <functions title=""/> + <variables title=""/> + </memberdef> + <authorsection visible="yes"/> + </namespace> + + <!-- Layout definition for a file page --> + <file> + <briefdescription visible="yes"/> + <includes visible="$SHOW_INCLUDE_FILES"/> + <includegraph visible="$INCLUDE_GRAPH"/> + <includedbygraph visible="$INCLUDED_BY_GRAPH"/> + <sourcelink visible="yes"/> + <memberdecl> + <classes visible="yes" title=""/> + <namespaces visible="yes" title=""/> + <defines title=""/> + <typedefs title=""/> + <enums title=""/> + <functions title=""/> + <variables title=""/> + <membergroups visible="yes"/> + </memberdecl> + <detaileddescription title=""/> + <memberdef> + <inlineclasses title=""/> + <defines title=""/> + <typedefs title=""/> + <enums title=""/> + <functions title=""/> + <variables title=""/> + </memberdef> + <authorsection/> + </file> + + <!-- Layout definition for a group page --> + <group> + <briefdescription visible="no"/> + <detaileddescription title=""/> + <groupgraph visible="$GROUP_GRAPHS"/> + <memberdecl> + <nestedgroups visible="yes" title=""/> + <dirs visible="yes" title=""/> + <files visible="yes" title=""/> + <namespaces visible="yes" title=""/> + <classes visible="yes" title=""/> + <defines title=""/> + <typedefs title=""/> + <enums title=""/> + <enumvalues title=""/> + <functions title=""/> + <variables title=""/> + <signals title=""/> + <publicslots title=""/> + <protectedslots title=""/> + <privateslots title=""/> + <events title=""/> + <properties title=""/> + <friends title=""/> + <membergroups visible="yes"/> + </memberdecl> + + <memberdef> + <pagedocs/> + <inlineclasses title=""/> + <defines title=""/> + <typedefs title=""/> + <enums title=""/> + <enumvalues title=""/> + <functions title=""/> + <variables title=""/> + <signals title=""/> + <publicslots title=""/> + <protectedslots title=""/> + <privateslots title=""/> + <events title=""/> + <properties title=""/> + <friends title=""/> + </memberdef> + <authorsection visible="yes"/> + </group> + + <!-- Layout definition for a directory page --> + <directory> + <briefdescription visible="yes"/> + <directorygraph visible="yes"/> + <memberdecl> + <dirs visible="yes"/> + <files visible="yes"/> + </memberdecl> + <detaileddescription title=""/> + </directory> +</doxygenlayout> diff --git a/unsupported/doc/examples/MatrixPower.cpp b/unsupported/doc/examples/MatrixPower.cpp new file mode 100644 index 000000000..222452476 --- /dev/null +++ b/unsupported/doc/examples/MatrixPower.cpp @@ -0,0 +1,16 @@ +#include <unsupported/Eigen/MatrixFunctions> +#include <iostream> + +using namespace Eigen; + +int main() +{ + const double pi = std::acos(-1.0); + Matrix3d A; + A << cos(1), -sin(1), 0, + sin(1), cos(1), 0, + 0 , 0 , 1; + std::cout << "The matrix A is:\n" << A << "\n\n" + "The matrix power A^(pi/4) is:\n" << A.pow(pi/4) << std::endl; + return 0; +} diff --git a/unsupported/doc/examples/MatrixPower_optimal.cpp b/unsupported/doc/examples/MatrixPower_optimal.cpp new file mode 100644 index 000000000..86470ba0a --- /dev/null +++ b/unsupported/doc/examples/MatrixPower_optimal.cpp @@ -0,0 +1,17 @@ +#include <unsupported/Eigen/MatrixFunctions> +#include <iostream> + +using namespace Eigen; + +int main() +{ + Matrix4cd A = Matrix4cd::Random(); + MatrixPower<Matrix4cd> Apow(A); + + std::cout << "The matrix A is:\n" << A << "\n\n" + "A^3.1 is:\n" << Apow(3.1) << "\n\n" + "A^3.3 is:\n" << Apow(3.3) << "\n\n" + "A^3.7 is:\n" << Apow(3.7) << "\n\n" + "A^3.9 is:\n" << Apow(3.9) << std::endl; + return 0; +} diff --git a/unsupported/doc/examples/PolynomialSolver1.cpp b/unsupported/doc/examples/PolynomialSolver1.cpp index 71e6b825f..cd777a4e2 100644 --- a/unsupported/doc/examples/PolynomialSolver1.cpp +++ b/unsupported/doc/examples/PolynomialSolver1.cpp @@ -49,5 +49,5 @@ int main() cout.precision(10); cout << "The last root in float then in double: " << psolvef.roots()[5] << "\t" << psolve6d.roots()[5] << endl; std::complex<float> castedRoot( psolve6d.roots()[5].real(), psolve6d.roots()[5].imag() ); - cout << "Norm of the difference: " << internal::abs( psolvef.roots()[5] - castedRoot ) << endl; + cout << "Norm of the difference: " << std::abs( psolvef.roots()[5] - castedRoot ) << endl; } diff --git a/unsupported/test/CMakeLists.txt b/unsupported/test/CMakeLists.txt index b34b151b1..a94a3b5e5 100644 --- a/unsupported/test/CMakeLists.txt +++ b/unsupported/test/CMakeLists.txt @@ -1,4 +1,7 @@ +set_property(GLOBAL PROPERTY EIGEN_CURRENT_SUBPROJECT "Unsupported") +add_custom_target(BuildUnsupported) + include_directories(../../test ../../unsupported ../../Eigen ${CMAKE_CURRENT_BINARY_DIR}/../../test) @@ -22,7 +25,7 @@ endif(ADOLC_FOUND) # this test seems to never have been successful on x87, so is considered to contain a FP-related bug. # see thread: "non-linear optimization test summary" -#ei_add_test(NonLinearOptimization) +ei_add_test(NonLinearOptimization) ei_add_test(NumericalDiff) ei_add_test(autodiff) @@ -33,6 +36,7 @@ endif() ei_add_test(matrix_exponential) ei_add_test(matrix_function) +ei_add_test(matrix_power) ei_add_test(matrix_square_root) ei_add_test(alignedvector3) ei_add_test(FFT) @@ -84,4 +88,6 @@ ei_add_test(polynomialutils) ei_add_test(kronecker_product) ei_add_test(splines) ei_add_test(gmres) - +ei_add_test(minres) +ei_add_test(levenberg_marquardt) +ei_add_test(bdcsvd) diff --git a/unsupported/test/FFTW.cpp b/unsupported/test/FFTW.cpp index a07bf274b..d3718e2d2 100644 --- a/unsupported/test/FFTW.cpp +++ b/unsupported/test/FFTW.cpp @@ -16,9 +16,6 @@ std::complex<T> RandomCpx() { return std::complex<T>( (T)(rand()/(T)RAND_MAX - . using namespace std; using namespace Eigen; -float norm(float x) {return x*x;} -double norm(double x) {return x*x;} -long double norm(long double x) {return x*x;} template < typename T> complex<long double> promote(complex<T> x) { return complex<long double>(x.real(),x.imag()); } @@ -40,11 +37,11 @@ complex<long double> promote(long double x) { return complex<long double>( x); for (size_t k1=0;k1<(size_t)timebuf.size();++k1) { acc += promote( timebuf[k1] ) * exp( complex<long double>(0,k1*phinc) ); } - totalpower += norm(acc); + totalpower += numext::abs2(acc); complex<long double> x = promote(fftbuf[k0]); complex<long double> dif = acc - x; - difpower += norm(dif); - //cerr << k0 << "\t" << acc << "\t" << x << "\t" << sqrt(norm(dif)) << endl; + difpower += numext::abs2(dif); + //cerr << k0 << "\t" << acc << "\t" << x << "\t" << sqrt(numext::abs2(dif)) << endl; } cerr << "rmse:" << sqrt(difpower/totalpower) << endl; return sqrt(difpower/totalpower); @@ -57,8 +54,8 @@ complex<long double> promote(long double x) { return complex<long double>( x); long double difpower=0; size_t n = (min)( buf1.size(),buf2.size() ); for (size_t k=0;k<n;++k) { - totalpower += (norm( buf1[k] ) + norm(buf2[k]) )/2.; - difpower += norm(buf1[k] - buf2[k]); + totalpower += (numext::abs2( buf1[k] ) + numext::abs2(buf2[k]) )/2.; + difpower += numext::abs2(buf1[k] - buf2[k]); } return sqrt(difpower/totalpower); } diff --git a/unsupported/test/NonLinearOptimization.cpp b/unsupported/test/NonLinearOptimization.cpp index 81b066897..d7376b0f5 100644 --- a/unsupported/test/NonLinearOptimization.cpp +++ b/unsupported/test/NonLinearOptimization.cpp @@ -12,6 +12,8 @@ // It is intended to be done for this test only. #include <Eigen/src/Core/util/DisableStupidWarnings.h> +using std::sqrt; + int fcn_chkder(const VectorXd &x, VectorXd &fvec, MatrixXd &fjac, int iflag) { /* subroutine fcn for chkder example. */ @@ -795,7 +797,9 @@ struct hahn1_functor : Functor<double> static const double m_x[236]; int operator()(const VectorXd &b, VectorXd &fvec) { - static const double m_y[236] = { .591E0 , 1.547E0 , 2.902E0 , 2.894E0 , 4.703E0 , 6.307E0 , 7.03E0 , 7.898E0 , 9.470E0 , 9.484E0 , 10.072E0 , 10.163E0 , 11.615E0 , 12.005E0 , 12.478E0 , 12.982E0 , 12.970E0 , 13.926E0 , 14.452E0 , 14.404E0 , 15.190E0 , 15.550E0 , 15.528E0 , 15.499E0 , 16.131E0 , 16.438E0 , 16.387E0 , 16.549E0 , 16.872E0 , 16.830E0 , 16.926E0 , 16.907E0 , 16.966E0 , 17.060E0 , 17.122E0 , 17.311E0 , 17.355E0 , 17.668E0 , 17.767E0 , 17.803E0 , 17.765E0 , 17.768E0 , 17.736E0 , 17.858E0 , 17.877E0 , 17.912E0 , 18.046E0 , 18.085E0 , 18.291E0 , 18.357E0 , 18.426E0 , 18.584E0 , 18.610E0 , 18.870E0 , 18.795E0 , 19.111E0 , .367E0 , .796E0 , 0.892E0 , 1.903E0 , 2.150E0 , 3.697E0 , 5.870E0 , 6.421E0 , 7.422E0 , 9.944E0 , 11.023E0 , 11.87E0 , 12.786E0 , 14.067E0 , 13.974E0 , 14.462E0 , 14.464E0 , 15.381E0 , 15.483E0 , 15.59E0 , 16.075E0 , 16.347E0 , 16.181E0 , 16.915E0 , 17.003E0 , 16.978E0 , 17.756E0 , 17.808E0 , 17.868E0 , 18.481E0 , 18.486E0 , 19.090E0 , 16.062E0 , 16.337E0 , 16.345E0 , 16.388E0 , 17.159E0 , 17.116E0 , 17.164E0 , 17.123E0 , 17.979E0 , 17.974E0 , 18.007E0 , 17.993E0 , 18.523E0 , 18.669E0 , 18.617E0 , 19.371E0 , 19.330E0 , 0.080E0 , 0.248E0 , 1.089E0 , 1.418E0 , 2.278E0 , 3.624E0 , 4.574E0 , 5.556E0 , 7.267E0 , 7.695E0 , 9.136E0 , 9.959E0 , 9.957E0 , 11.600E0 , 13.138E0 , 13.564E0 , 13.871E0 , 13.994E0 , 14.947E0 , 15.473E0 , 15.379E0 , 15.455E0 , 15.908E0 , 16.114E0 , 17.071E0 , 17.135E0 , 17.282E0 , 17.368E0 , 17.483E0 , 17.764E0 , 18.185E0 , 18.271E0 , 18.236E0 , 18.237E0 , 18.523E0 , 18.627E0 , 18.665E0 , 19.086E0 , 0.214E0 , 0.943E0 , 1.429E0 , 2.241E0 , 2.951E0 , 3.782E0 , 4.757E0 , 5.602E0 , 7.169E0 , 8.920E0 , 10.055E0 , 12.035E0 , 12.861E0 , 13.436E0 , 14.167E0 , 14.755E0 , 15.168E0 , 15.651E0 , 15.746E0 , 16.216E0 , 16.445E0 , 16.965E0 , 17.121E0 , 17.206E0 , 17.250E0 , 17.339E0 , 17.793E0 , 18.123E0 , 18.49E0 , 18.566E0 , 18.645E0 , 18.706E0 , 18.924E0 , 19.1E0 , 0.375E0 , 0.471E0 , 1.504E0 , 2.204E0 , 2.813E0 , 4.765E0 , 9.835E0 , 10.040E0 , 11.946E0 , 12.596E0 , 13.303E0 , 13.922E0 , 14.440E0 , 14.951E0 , 15.627E0 , 15.639E0 , 15.814E0 , 16.315E0 , 16.334E0 , 16.430E0 , 16.423E0 , 17.024E0 , 17.009E0 , 17.165E0 , 17.134E0 , 17.349E0 , 17.576E0 , 17.848E0 , 18.090E0 , 18.276E0 , 18.404E0 , 18.519E0 , 19.133E0 , 19.074E0 , 19.239E0 , 19.280E0 , 19.101E0 , 19.398E0 , 19.252E0 , 19.89E0 , 20.007E0 , 19.929E0 , 19.268E0 , 19.324E0 , 20.049E0 , 20.107E0 , 20.062E0 , 20.065E0 , 19.286E0 , 19.972E0 , 20.088E0 , 20.743E0 , 20.83E0 , 20.935E0 , 21.035E0 , 20.93E0 , 21.074E0 , 21.085E0 , 20.935E0 }; + static const double m_y[236] = { .591E0 , 1.547E0 , 2.902E0 , 2.894E0 , 4.703E0 , 6.307E0 , 7.03E0 , 7.898E0 , 9.470E0 , 9.484E0 , 10.072E0 , 10.163E0 , 11.615E0 , 12.005E0 , 12.478E0 , 12.982E0 , 12.970E0 , 13.926E0 , 14.452E0 , 14.404E0 , 15.190E0 , 15.550E0 , 15.528E0 , 15.499E0 , 16.131E0 , 16.438E0 , 16.387E0 , 16.549E0 , 16.872E0 , 16.830E0 , 16.926E0 , 16.907E0 , 16.966E0 , 17.060E0 , 17.122E0 , 17.311E0 , 17.355E0 , 17.668E0 , 17.767E0 , 17.803E0 , 17.765E0 , 17.768E0 , 17.736E0 , 17.858E0 , 17.877E0 , 17.912E0 , 18.046E0 , 18.085E0 , 18.291E0 , 18.357E0 , 18.426E0 , 18.584E0 , 18.610E0 , 18.870E0 , 18.795E0 , 19.111E0 , .367E0 , .796E0 , 0.892E0 , 1.903E0 , 2.150E0 , 3.697E0 , 5.870E0 , 6.421E0 , 7.422E0 , 9.944E0 , 11.023E0 , 11.87E0 , 12.786E0 , 14.067E0 , 13.974E0 , 14.462E0 , 14.464E0 , 15.381E0 , 15.483E0 , 15.59E0 , 16.075E0 , 16.347E0 , 16.181E0 , 16.915E0 , 17.003E0 , 16.978E0 , 17.756E0 , 17.808E0 , 17.868E0 , 18.481E0 , 18.486E0 , 19.090E0 , 16.062E0 , 16.337E0 , 16.345E0 , + 16.388E0 , 17.159E0 , 17.116E0 , 17.164E0 , 17.123E0 , 17.979E0 , 17.974E0 , 18.007E0 , 17.993E0 , 18.523E0 , 18.669E0 , 18.617E0 , 19.371E0 , 19.330E0 , 0.080E0 , 0.248E0 , 1.089E0 , 1.418E0 , 2.278E0 , 3.624E0 , 4.574E0 , 5.556E0 , 7.267E0 , 7.695E0 , 9.136E0 , 9.959E0 , 9.957E0 , 11.600E0 , 13.138E0 , 13.564E0 , 13.871E0 , 13.994E0 , 14.947E0 , 15.473E0 , 15.379E0 , 15.455E0 , 15.908E0 , 16.114E0 , 17.071E0 , 17.135E0 , 17.282E0 , 17.368E0 , 17.483E0 , 17.764E0 , 18.185E0 , 18.271E0 , 18.236E0 , 18.237E0 , 18.523E0 , 18.627E0 , 18.665E0 , 19.086E0 , 0.214E0 , 0.943E0 , 1.429E0 , 2.241E0 , 2.951E0 , 3.782E0 , 4.757E0 , 5.602E0 , 7.169E0 , 8.920E0 , 10.055E0 , 12.035E0 , 12.861E0 , 13.436E0 , 14.167E0 , 14.755E0 , 15.168E0 , 15.651E0 , 15.746E0 , 16.216E0 , 16.445E0 , 16.965E0 , 17.121E0 , 17.206E0 , 17.250E0 , 17.339E0 , 17.793E0 , 18.123E0 , 18.49E0 , 18.566E0 , 18.645E0 , 18.706E0 , 18.924E0 , 19.1E0 , 0.375E0 , 0.471E0 , 1.504E0 , 2.204E0 , 2.813E0 , 4.765E0 , 9.835E0 , 10.040E0 , 11.946E0 , 12.596E0 , +13.303E0 , 13.922E0 , 14.440E0 , 14.951E0 , 15.627E0 , 15.639E0 , 15.814E0 , 16.315E0 , 16.334E0 , 16.430E0 , 16.423E0 , 17.024E0 , 17.009E0 , 17.165E0 , 17.134E0 , 17.349E0 , 17.576E0 , 17.848E0 , 18.090E0 , 18.276E0 , 18.404E0 , 18.519E0 , 19.133E0 , 19.074E0 , 19.239E0 , 19.280E0 , 19.101E0 , 19.398E0 , 19.252E0 , 19.89E0 , 20.007E0 , 19.929E0 , 19.268E0 , 19.324E0 , 20.049E0 , 20.107E0 , 20.062E0 , 20.065E0 , 19.286E0 , 19.972E0 , 20.088E0 , 20.743E0 , 20.83E0 , 20.935E0 , 21.035E0 , 20.93E0 , 21.074E0 , 21.085E0 , 20.935E0 }; // int called=0; printf("call hahn1_functor with iflag=%d, called=%d\n", iflag, called); if (iflag==1) called++; @@ -828,7 +832,9 @@ struct hahn1_functor : Functor<double> return 0; } }; -const double hahn1_functor::m_x[236] = { 24.41E0 , 34.82E0 , 44.09E0 , 45.07E0 , 54.98E0 , 65.51E0 , 70.53E0 , 75.70E0 , 89.57E0 , 91.14E0 , 96.40E0 , 97.19E0 , 114.26E0 , 120.25E0 , 127.08E0 , 133.55E0 , 133.61E0 , 158.67E0 , 172.74E0 , 171.31E0 , 202.14E0 , 220.55E0 , 221.05E0 , 221.39E0 , 250.99E0 , 268.99E0 , 271.80E0 , 271.97E0 , 321.31E0 , 321.69E0 , 330.14E0 , 333.03E0 , 333.47E0 , 340.77E0 , 345.65E0 , 373.11E0 , 373.79E0 , 411.82E0 , 419.51E0 , 421.59E0 , 422.02E0 , 422.47E0 , 422.61E0 , 441.75E0 , 447.41E0 , 448.7E0 , 472.89E0 , 476.69E0 , 522.47E0 , 522.62E0 , 524.43E0 , 546.75E0 , 549.53E0 , 575.29E0 , 576.00E0 , 625.55E0 , 20.15E0 , 28.78E0 , 29.57E0 , 37.41E0 , 39.12E0 , 50.24E0 , 61.38E0 , 66.25E0 , 73.42E0 , 95.52E0 , 107.32E0 , 122.04E0 , 134.03E0 , 163.19E0 , 163.48E0 , 175.70E0 , 179.86E0 , 211.27E0 , 217.78E0 , 219.14E0 , 262.52E0 , 268.01E0 , 268.62E0 , 336.25E0 , 337.23E0 , 339.33E0 , 427.38E0 , 428.58E0 , 432.68E0 , 528.99E0 , 531.08E0 , 628.34E0 , 253.24E0 , 273.13E0 , 273.66E0 , 282.10E0 , 346.62E0 , 347.19E0 , 348.78E0 , 351.18E0 , 450.10E0 , 450.35E0 , 451.92E0 , 455.56E0 , 552.22E0 , 553.56E0 , 555.74E0 , 652.59E0 , 656.20E0 , 14.13E0 , 20.41E0 , 31.30E0 , 33.84E0 , 39.70E0 , 48.83E0 , 54.50E0 , 60.41E0 , 72.77E0 , 75.25E0 , 86.84E0 , 94.88E0 , 96.40E0 , 117.37E0 , 139.08E0 , 147.73E0 , 158.63E0 , 161.84E0 , 192.11E0 , 206.76E0 , 209.07E0 , 213.32E0 , 226.44E0 , 237.12E0 , 330.90E0 , 358.72E0 , 370.77E0 , 372.72E0 , 396.24E0 , 416.59E0 , 484.02E0 , 495.47E0 , 514.78E0 , 515.65E0 , 519.47E0 , 544.47E0 , 560.11E0 , 620.77E0 , 18.97E0 , 28.93E0 , 33.91E0 , 40.03E0 , 44.66E0 , 49.87E0 , 55.16E0 , 60.90E0 , 72.08E0 , 85.15E0 , 97.06E0 , 119.63E0 , 133.27E0 , 143.84E0 , 161.91E0 , 180.67E0 , 198.44E0 , 226.86E0 , 229.65E0 , 258.27E0 , 273.77E0 , 339.15E0 , 350.13E0 , 362.75E0 , 371.03E0 , 393.32E0 , 448.53E0 , 473.78E0 , 511.12E0 , 524.70E0 , 548.75E0 , 551.64E0 , 574.02E0 , 623.86E0 , 21.46E0 , 24.33E0 , 33.43E0 , 39.22E0 , 44.18E0 , 55.02E0 , 94.33E0 , 96.44E0 , 118.82E0 , 128.48E0 , 141.94E0 , 156.92E0 , 171.65E0 , 190.00E0 , 223.26E0 , 223.88E0 , 231.50E0 , 265.05E0 , 269.44E0 , 271.78E0 , 273.46E0 , 334.61E0 , 339.79E0 , 349.52E0 , 358.18E0 , 377.98E0 , 394.77E0 , 429.66E0 , 468.22E0 , 487.27E0 , 519.54E0 , 523.03E0 , 612.99E0 , 638.59E0 , 641.36E0 , 622.05E0 , 631.50E0 , 663.97E0 , 646.9E0 , 748.29E0 , 749.21E0 , 750.14E0 , 647.04E0 , 646.89E0 , 746.9E0 , 748.43E0 , 747.35E0 , 749.27E0 , 647.61E0 , 747.78E0 , 750.51E0 , 851.37E0 , 845.97E0 , 847.54E0 , 849.93E0 , 851.61E0 , 849.75E0 , 850.98E0 , 848.23E0}; +const double hahn1_functor::m_x[236] = { 24.41E0 , 34.82E0 , 44.09E0 , 45.07E0 , 54.98E0 , 65.51E0 , 70.53E0 , 75.70E0 , 89.57E0 , 91.14E0 , 96.40E0 , 97.19E0 , 114.26E0 , 120.25E0 , 127.08E0 , 133.55E0 , 133.61E0 , 158.67E0 , 172.74E0 , 171.31E0 , 202.14E0 , 220.55E0 , 221.05E0 , 221.39E0 , 250.99E0 , 268.99E0 , 271.80E0 , 271.97E0 , 321.31E0 , 321.69E0 , 330.14E0 , 333.03E0 , 333.47E0 , 340.77E0 , 345.65E0 , 373.11E0 , 373.79E0 , 411.82E0 , 419.51E0 , 421.59E0 , 422.02E0 , 422.47E0 , 422.61E0 , 441.75E0 , 447.41E0 , 448.7E0 , 472.89E0 , 476.69E0 , 522.47E0 , 522.62E0 , 524.43E0 , 546.75E0 , 549.53E0 , 575.29E0 , 576.00E0 , 625.55E0 , 20.15E0 , 28.78E0 , 29.57E0 , 37.41E0 , 39.12E0 , 50.24E0 , 61.38E0 , 66.25E0 , 73.42E0 , 95.52E0 , 107.32E0 , 122.04E0 , 134.03E0 , 163.19E0 , 163.48E0 , 175.70E0 , 179.86E0 , 211.27E0 , 217.78E0 , 219.14E0 , 262.52E0 , 268.01E0 , 268.62E0 , 336.25E0 , 337.23E0 , 339.33E0 , 427.38E0 , 428.58E0 , 432.68E0 , 528.99E0 , 531.08E0 , 628.34E0 , 253.24E0 , 273.13E0 , 273.66E0 , +282.10E0 , 346.62E0 , 347.19E0 , 348.78E0 , 351.18E0 , 450.10E0 , 450.35E0 , 451.92E0 , 455.56E0 , 552.22E0 , 553.56E0 , 555.74E0 , 652.59E0 , 656.20E0 , 14.13E0 , 20.41E0 , 31.30E0 , 33.84E0 , 39.70E0 , 48.83E0 , 54.50E0 , 60.41E0 , 72.77E0 , 75.25E0 , 86.84E0 , 94.88E0 , 96.40E0 , 117.37E0 , 139.08E0 , 147.73E0 , 158.63E0 , 161.84E0 , 192.11E0 , 206.76E0 , 209.07E0 , 213.32E0 , 226.44E0 , 237.12E0 , 330.90E0 , 358.72E0 , 370.77E0 , 372.72E0 , 396.24E0 , 416.59E0 , 484.02E0 , 495.47E0 , 514.78E0 , 515.65E0 , 519.47E0 , 544.47E0 , 560.11E0 , 620.77E0 , 18.97E0 , 28.93E0 , 33.91E0 , 40.03E0 , 44.66E0 , 49.87E0 , 55.16E0 , 60.90E0 , 72.08E0 , 85.15E0 , 97.06E0 , 119.63E0 , 133.27E0 , 143.84E0 , 161.91E0 , 180.67E0 , 198.44E0 , 226.86E0 , 229.65E0 , 258.27E0 , 273.77E0 , 339.15E0 , 350.13E0 , 362.75E0 , 371.03E0 , 393.32E0 , 448.53E0 , 473.78E0 , 511.12E0 , 524.70E0 , 548.75E0 , 551.64E0 , 574.02E0 , 623.86E0 , 21.46E0 , 24.33E0 , 33.43E0 , 39.22E0 , 44.18E0 , 55.02E0 , 94.33E0 , 96.44E0 , 118.82E0 , 128.48E0 , +141.94E0 , 156.92E0 , 171.65E0 , 190.00E0 , 223.26E0 , 223.88E0 , 231.50E0 , 265.05E0 , 269.44E0 , 271.78E0 , 273.46E0 , 334.61E0 , 339.79E0 , 349.52E0 , 358.18E0 , 377.98E0 , 394.77E0 , 429.66E0 , 468.22E0 , 487.27E0 , 519.54E0 , 523.03E0 , 612.99E0 , 638.59E0 , 641.36E0 , 622.05E0 , 631.50E0 , 663.97E0 , 646.9E0 , 748.29E0 , 749.21E0 , 750.14E0 , 647.04E0 , 646.89E0 , 746.9E0 , 748.43E0 , 747.35E0 , 749.27E0 , 647.61E0 , 747.78E0 , 750.51E0 , 851.37E0 , 845.97E0 , 847.54E0 , 849.93E0 , 851.61E0 , 849.75E0 , 850.98E0 , 848.23E0}; // http://www.itl.nist.gov/div898/strd/nls/data/hahn1.shtml void testNistHahn1(void) @@ -1485,8 +1491,11 @@ struct Bennett5_functor : Functor<double> return 0; } }; -const double Bennett5_functor::x[154] = { 7.447168E0, 8.102586E0, 8.452547E0, 8.711278E0, 8.916774E0, 9.087155E0, 9.232590E0, 9.359535E0, 9.472166E0, 9.573384E0, 9.665293E0, 9.749461E0, 9.827092E0, 9.899128E0, 9.966321E0, 10.029280E0, 10.088510E0, 10.144430E0, 10.197380E0, 10.247670E0, 10.295560E0, 10.341250E0, 10.384950E0, 10.426820E0, 10.467000E0, 10.505640E0, 10.542830E0, 10.578690E0, 10.613310E0, 10.646780E0, 10.679150E0, 10.710520E0, 10.740920E0, 10.770440E0, 10.799100E0, 10.826970E0, 10.854080E0, 10.880470E0, 10.906190E0, 10.931260E0, 10.955720E0, 10.979590E0, 11.002910E0, 11.025700E0, 11.047980E0, 11.069770E0, 11.091100E0, 11.111980E0, 11.132440E0, 11.152480E0, 11.172130E0, 11.191410E0, 11.210310E0, 11.228870E0, 11.247090E0, 11.264980E0, 11.282560E0, 11.299840E0, 11.316820E0, 11.333520E0, 11.349940E0, 11.366100E0, 11.382000E0, 11.397660E0, 11.413070E0, 11.428240E0, 11.443200E0, 11.457930E0, 11.472440E0, 11.486750E0, 11.500860E0, 11.514770E0, 11.528490E0, 11.542020E0, 11.555380E0, 11.568550E0, 11.581560E0, 11.594420E0, 11.607121E0, 11.619640E0, 11.632000E0, 11.644210E0, 11.656280E0, 11.668200E0, 11.679980E0, 11.691620E0, 11.703130E0, 11.714510E0, 11.725760E0, 11.736880E0, 11.747890E0, 11.758780E0, 11.769550E0, 11.780200E0, 11.790730E0, 11.801160E0, 11.811480E0, 11.821700E0, 11.831810E0, 11.841820E0, 11.851730E0, 11.861550E0, 11.871270E0, 11.880890E0, 11.890420E0, 11.899870E0, 11.909220E0, 11.918490E0, 11.927680E0, 11.936780E0, 11.945790E0, 11.954730E0, 11.963590E0, 11.972370E0, 11.981070E0, 11.989700E0, 11.998260E0, 12.006740E0, 12.015150E0, 12.023490E0, 12.031760E0, 12.039970E0, 12.048100E0, 12.056170E0, 12.064180E0, 12.072120E0, 12.080010E0, 12.087820E0, 12.095580E0, 12.103280E0, 12.110920E0, 12.118500E0, 12.126030E0, 12.133500E0, 12.140910E0, 12.148270E0, 12.155570E0, 12.162830E0, 12.170030E0, 12.177170E0, 12.184270E0, 12.191320E0, 12.198320E0, 12.205270E0, 12.212170E0, 12.219030E0, 12.225840E0, 12.232600E0, 12.239320E0, 12.245990E0, 12.252620E0, 12.259200E0, 12.265750E0, 12.272240E0 }; -const double Bennett5_functor::y[154] = { -34.834702E0 ,-34.393200E0 ,-34.152901E0 ,-33.979099E0 ,-33.845901E0 ,-33.732899E0 ,-33.640301E0 ,-33.559200E0 ,-33.486801E0 ,-33.423100E0 ,-33.365101E0 ,-33.313000E0 ,-33.260899E0 ,-33.217400E0 ,-33.176899E0 ,-33.139198E0 ,-33.101601E0 ,-33.066799E0 ,-33.035000E0 ,-33.003101E0 ,-32.971298E0 ,-32.942299E0 ,-32.916302E0 ,-32.890202E0 ,-32.864101E0 ,-32.841000E0 ,-32.817799E0 ,-32.797501E0 ,-32.774300E0 ,-32.757000E0 ,-32.733799E0 ,-32.716400E0 ,-32.699100E0 ,-32.678799E0 ,-32.661400E0 ,-32.644001E0 ,-32.626701E0 ,-32.612202E0 ,-32.597698E0 ,-32.583199E0 ,-32.568699E0 ,-32.554298E0 ,-32.539799E0 ,-32.525299E0 ,-32.510799E0 ,-32.499199E0 ,-32.487598E0 ,-32.473202E0 ,-32.461601E0 ,-32.435501E0 ,-32.435501E0 ,-32.426800E0 ,-32.412300E0 ,-32.400799E0 ,-32.392101E0 ,-32.380501E0 ,-32.366001E0 ,-32.357300E0 ,-32.348598E0 ,-32.339901E0 ,-32.328400E0 ,-32.319698E0 ,-32.311001E0 ,-32.299400E0 ,-32.290699E0 ,-32.282001E0 ,-32.273300E0 ,-32.264599E0 ,-32.256001E0 ,-32.247299E0 ,-32.238602E0 ,-32.229900E0 ,-32.224098E0 ,-32.215401E0 ,-32.203800E0 ,-32.198002E0 ,-32.189400E0 ,-32.183601E0 ,-32.174900E0 ,-32.169102E0 ,-32.163300E0 ,-32.154598E0 ,-32.145901E0 ,-32.140099E0 ,-32.131401E0 ,-32.125599E0 ,-32.119801E0 ,-32.111198E0 ,-32.105400E0 ,-32.096699E0 ,-32.090900E0 ,-32.088001E0 ,-32.079300E0 ,-32.073502E0 ,-32.067699E0 ,-32.061901E0 ,-32.056099E0 ,-32.050301E0 ,-32.044498E0 ,-32.038799E0 ,-32.033001E0 ,-32.027199E0 ,-32.024300E0 ,-32.018501E0 ,-32.012699E0 ,-32.004002E0 ,-32.001099E0 ,-31.995300E0 ,-31.989500E0 ,-31.983700E0 ,-31.977900E0 ,-31.972099E0 ,-31.969299E0 ,-31.963501E0 ,-31.957701E0 ,-31.951900E0 ,-31.946100E0 ,-31.940300E0 ,-31.937401E0 ,-31.931601E0 ,-31.925800E0 ,-31.922899E0 ,-31.917101E0 ,-31.911301E0 ,-31.908400E0 ,-31.902599E0 ,-31.896900E0 ,-31.893999E0 ,-31.888201E0 ,-31.885300E0 ,-31.882401E0 ,-31.876600E0 ,-31.873699E0 ,-31.867901E0 ,-31.862101E0 ,-31.859200E0 ,-31.856300E0 ,-31.850500E0 ,-31.844700E0 ,-31.841801E0 ,-31.838900E0 ,-31.833099E0 ,-31.830200E0 ,-31.827299E0 ,-31.821600E0 ,-31.818701E0 ,-31.812901E0 ,-31.809999E0 ,-31.807100E0 ,-31.801300E0 ,-31.798401E0 ,-31.795500E0 ,-31.789700E0 ,-31.786800E0 }; +const double Bennett5_functor::x[154] = { 7.447168E0, 8.102586E0, 8.452547E0, 8.711278E0, 8.916774E0, 9.087155E0, 9.232590E0, 9.359535E0, 9.472166E0, 9.573384E0, 9.665293E0, 9.749461E0, 9.827092E0, 9.899128E0, 9.966321E0, 10.029280E0, 10.088510E0, 10.144430E0, 10.197380E0, 10.247670E0, 10.295560E0, 10.341250E0, 10.384950E0, 10.426820E0, 10.467000E0, 10.505640E0, 10.542830E0, 10.578690E0, 10.613310E0, 10.646780E0, 10.679150E0, 10.710520E0, 10.740920E0, 10.770440E0, 10.799100E0, 10.826970E0, 10.854080E0, 10.880470E0, 10.906190E0, 10.931260E0, 10.955720E0, 10.979590E0, 11.002910E0, 11.025700E0, 11.047980E0, 11.069770E0, 11.091100E0, 11.111980E0, 11.132440E0, 11.152480E0, 11.172130E0, 11.191410E0, 11.210310E0, 11.228870E0, 11.247090E0, 11.264980E0, 11.282560E0, 11.299840E0, 11.316820E0, 11.333520E0, 11.349940E0, 11.366100E0, 11.382000E0, 11.397660E0, 11.413070E0, 11.428240E0, 11.443200E0, 11.457930E0, 11.472440E0, 11.486750E0, 11.500860E0, 11.514770E0, 11.528490E0, 11.542020E0, 11.555380E0, 11.568550E0, +11.581560E0, 11.594420E0, 11.607121E0, 11.619640E0, 11.632000E0, 11.644210E0, 11.656280E0, 11.668200E0, 11.679980E0, 11.691620E0, 11.703130E0, 11.714510E0, 11.725760E0, 11.736880E0, 11.747890E0, 11.758780E0, 11.769550E0, 11.780200E0, 11.790730E0, 11.801160E0, 11.811480E0, 11.821700E0, 11.831810E0, 11.841820E0, 11.851730E0, 11.861550E0, 11.871270E0, 11.880890E0, 11.890420E0, 11.899870E0, 11.909220E0, 11.918490E0, 11.927680E0, 11.936780E0, 11.945790E0, 11.954730E0, 11.963590E0, 11.972370E0, 11.981070E0, 11.989700E0, 11.998260E0, 12.006740E0, 12.015150E0, 12.023490E0, 12.031760E0, 12.039970E0, 12.048100E0, 12.056170E0, 12.064180E0, 12.072120E0, 12.080010E0, 12.087820E0, 12.095580E0, 12.103280E0, 12.110920E0, 12.118500E0, 12.126030E0, 12.133500E0, 12.140910E0, 12.148270E0, 12.155570E0, 12.162830E0, 12.170030E0, 12.177170E0, 12.184270E0, 12.191320E0, 12.198320E0, 12.205270E0, 12.212170E0, 12.219030E0, 12.225840E0, 12.232600E0, 12.239320E0, 12.245990E0, 12.252620E0, 12.259200E0, 12.265750E0, 12.272240E0 }; +const double Bennett5_functor::y[154] = { -34.834702E0 ,-34.393200E0 ,-34.152901E0 ,-33.979099E0 ,-33.845901E0 ,-33.732899E0 ,-33.640301E0 ,-33.559200E0 ,-33.486801E0 ,-33.423100E0 ,-33.365101E0 ,-33.313000E0 ,-33.260899E0 ,-33.217400E0 ,-33.176899E0 ,-33.139198E0 ,-33.101601E0 ,-33.066799E0 ,-33.035000E0 ,-33.003101E0 ,-32.971298E0 ,-32.942299E0 ,-32.916302E0 ,-32.890202E0 ,-32.864101E0 ,-32.841000E0 ,-32.817799E0 ,-32.797501E0 ,-32.774300E0 ,-32.757000E0 ,-32.733799E0 ,-32.716400E0 ,-32.699100E0 ,-32.678799E0 ,-32.661400E0 ,-32.644001E0 ,-32.626701E0 ,-32.612202E0 ,-32.597698E0 ,-32.583199E0 ,-32.568699E0 ,-32.554298E0 ,-32.539799E0 ,-32.525299E0 ,-32.510799E0 ,-32.499199E0 ,-32.487598E0 ,-32.473202E0 ,-32.461601E0 ,-32.435501E0 ,-32.435501E0 ,-32.426800E0 ,-32.412300E0 ,-32.400799E0 ,-32.392101E0 ,-32.380501E0 ,-32.366001E0 ,-32.357300E0 ,-32.348598E0 ,-32.339901E0 ,-32.328400E0 ,-32.319698E0 ,-32.311001E0 ,-32.299400E0 ,-32.290699E0 ,-32.282001E0 ,-32.273300E0 ,-32.264599E0 ,-32.256001E0 ,-32.247299E0 +,-32.238602E0 ,-32.229900E0 ,-32.224098E0 ,-32.215401E0 ,-32.203800E0 ,-32.198002E0 ,-32.189400E0 ,-32.183601E0 ,-32.174900E0 ,-32.169102E0 ,-32.163300E0 ,-32.154598E0 ,-32.145901E0 ,-32.140099E0 ,-32.131401E0 ,-32.125599E0 ,-32.119801E0 ,-32.111198E0 ,-32.105400E0 ,-32.096699E0 ,-32.090900E0 ,-32.088001E0 ,-32.079300E0 ,-32.073502E0 ,-32.067699E0 ,-32.061901E0 ,-32.056099E0 ,-32.050301E0 ,-32.044498E0 ,-32.038799E0 ,-32.033001E0 ,-32.027199E0 ,-32.024300E0 ,-32.018501E0 ,-32.012699E0 ,-32.004002E0 ,-32.001099E0 ,-31.995300E0 ,-31.989500E0 ,-31.983700E0 ,-31.977900E0 ,-31.972099E0 ,-31.969299E0 ,-31.963501E0 ,-31.957701E0 ,-31.951900E0 ,-31.946100E0 ,-31.940300E0 ,-31.937401E0 ,-31.931601E0 ,-31.925800E0 ,-31.922899E0 ,-31.917101E0 ,-31.911301E0 ,-31.908400E0 ,-31.902599E0 ,-31.896900E0 ,-31.893999E0 ,-31.888201E0 ,-31.885300E0 ,-31.882401E0 ,-31.876600E0 ,-31.873699E0 ,-31.867901E0 ,-31.862101E0 ,-31.859200E0 ,-31.856300E0 ,-31.850500E0 ,-31.844700E0 ,-31.841801E0 ,-31.838900E0 ,-31.833099E0 ,-31.830200E0 , +-31.827299E0 ,-31.821600E0 ,-31.818701E0 ,-31.812901E0 ,-31.809999E0 ,-31.807100E0 ,-31.801300E0 ,-31.798401E0 ,-31.795500E0 ,-31.789700E0 ,-31.786800E0 }; // http://www.itl.nist.gov/div898/strd/nls/data/bennett5.shtml void testNistBennett5(void) @@ -1823,14 +1832,14 @@ void test_NonLinearOptimization() // NIST tests, level of difficulty = "Average" CALL_SUBTEST/*_5*/(testNistHahn1()); CALL_SUBTEST/*_6*/(testNistMisra1d()); - CALL_SUBTEST/*_7*/(testNistMGH17()); - CALL_SUBTEST/*_8*/(testNistLanczos1()); +// CALL_SUBTEST/*_7*/(testNistMGH17()); +// CALL_SUBTEST/*_8*/(testNistLanczos1()); - // NIST tests, level of difficulty = "Higher" +// // NIST tests, level of difficulty = "Higher" CALL_SUBTEST/*_9*/(testNistRat42()); - CALL_SUBTEST/*_10*/(testNistMGH10()); +// CALL_SUBTEST/*_10*/(testNistMGH10()); CALL_SUBTEST/*_11*/(testNistBoxBOD()); - CALL_SUBTEST/*_12*/(testNistMGH09()); +// CALL_SUBTEST/*_12*/(testNistMGH09()); CALL_SUBTEST/*_13*/(testNistBennett5()); CALL_SUBTEST/*_14*/(testNistThurber()); CALL_SUBTEST/*_15*/(testNistRat43()); diff --git a/unsupported/test/autodiff.cpp b/unsupported/test/autodiff.cpp index 6eb417e8d..087e7c542 100644 --- a/unsupported/test/autodiff.cpp +++ b/unsupported/test/autodiff.cpp @@ -127,46 +127,47 @@ template<typename Func> void forward_jacobian(const Func& f) VERIFY_IS_APPROX(j, jref); } + +// TODO also check actual derivatives! void test_autodiff_scalar() { - std::cerr << foo<float>(1,2) << "\n"; + Vector2f p = Vector2f::Random(); typedef AutoDiffScalar<Vector2f> AD; - AD ax(1,Vector2f::UnitX()); - AD ay(2,Vector2f::UnitY()); + AD ax(p.x(),Vector2f::UnitX()); + AD ay(p.y(),Vector2f::UnitY()); AD res = foo<AD>(ax,ay); - std::cerr << res.value() << " <> " - << res.derivatives().transpose() << "\n\n"; + VERIFY_IS_APPROX(res.value(), foo(p.x(),p.y())); } +// TODO also check actual derivatives! void test_autodiff_vector() { - std::cerr << foo<Vector2f>(Vector2f(1,2)) << "\n"; + Vector2f p = Vector2f::Random(); typedef AutoDiffScalar<Vector2f> AD; typedef Matrix<AD,2,1> VectorAD; - VectorAD p(AD(1),AD(-1)); - p.x().derivatives() = Vector2f::UnitX(); - p.y().derivatives() = Vector2f::UnitY(); + VectorAD ap = p.cast<AD>(); + ap.x().derivatives() = Vector2f::UnitX(); + ap.y().derivatives() = Vector2f::UnitY(); - AD res = foo<VectorAD>(p); - std::cerr << res.value() << " <> " - << res.derivatives().transpose() << "\n\n"; + AD res = foo<VectorAD>(ap); + VERIFY_IS_APPROX(res.value(), foo(p)); } void test_autodiff_jacobian() { - for(int i = 0; i < g_repeat; i++) { - CALL_SUBTEST(( forward_jacobian(TestFunc1<double,2,2>()) )); - CALL_SUBTEST(( forward_jacobian(TestFunc1<double,2,3>()) )); - CALL_SUBTEST(( forward_jacobian(TestFunc1<double,3,2>()) )); - CALL_SUBTEST(( forward_jacobian(TestFunc1<double,3,3>()) )); - CALL_SUBTEST(( forward_jacobian(TestFunc1<double>(3,3)) )); - } + CALL_SUBTEST(( forward_jacobian(TestFunc1<double,2,2>()) )); + CALL_SUBTEST(( forward_jacobian(TestFunc1<double,2,3>()) )); + CALL_SUBTEST(( forward_jacobian(TestFunc1<double,3,2>()) )); + CALL_SUBTEST(( forward_jacobian(TestFunc1<double,3,3>()) )); + CALL_SUBTEST(( forward_jacobian(TestFunc1<double>(3,3)) )); } void test_autodiff() { - test_autodiff_scalar(); - test_autodiff_vector(); -// test_autodiff_jacobian(); + for(int i = 0; i < g_repeat; i++) { + CALL_SUBTEST_1( test_autodiff_scalar() ); + CALL_SUBTEST_2( test_autodiff_vector() ); + CALL_SUBTEST_3( test_autodiff_jacobian() ); + } } diff --git a/unsupported/test/bdcsvd.cpp b/unsupported/test/bdcsvd.cpp new file mode 100644 index 000000000..115a649b0 --- /dev/null +++ b/unsupported/test/bdcsvd.cpp @@ -0,0 +1,213 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2013 Gauthier Brun <brun.gauthier@gmail.com> +// Copyright (C) 2013 Nicolas Carre <nicolas.carre@ensimag.fr> +// Copyright (C) 2013 Jean Ceccato <jean.ceccato@ensimag.fr> +// Copyright (C) 2013 Pierre Zoppitelli <pierre.zoppitelli@ensimag.fr> +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/ + +#include "svd_common.h" +#include <iostream> +#include <Eigen/LU> + +// check if "svd" is the good image of "m" +template<typename MatrixType> +void bdcsvd_check_full(const MatrixType& m, const BDCSVD<MatrixType>& svd) +{ + svd_check_full< MatrixType, BDCSVD< MatrixType > >(m, svd); +} + +// Compare to a reference value +template<typename MatrixType> +void bdcsvd_compare_to_full(const MatrixType& m, + unsigned int computationOptions, + const BDCSVD<MatrixType>& referenceSvd) +{ + svd_compare_to_full< MatrixType, BDCSVD< MatrixType > >(m, computationOptions, referenceSvd); +} // end bdcsvd_compare_to_full + + +template<typename MatrixType> +void bdcsvd_solve(const MatrixType& m, unsigned int computationOptions) +{ + svd_solve< MatrixType, BDCSVD< MatrixType > >(m, computationOptions); +} // end template bdcsvd_solve + + +// test the computations options +template<typename MatrixType> +void bdcsvd_test_all_computation_options(const MatrixType& m) +{ + BDCSVD<MatrixType> fullSvd(m, ComputeFullU|ComputeFullV); + svd_test_computation_options_1< MatrixType, BDCSVD< MatrixType > >(m, fullSvd); + svd_test_computation_options_2< MatrixType, BDCSVD< MatrixType > >(m, fullSvd); +} // end bdcsvd_test_all_computation_options + + +// Call a test with all the computations options +template<typename MatrixType> +void bdcsvd(const MatrixType& a = MatrixType(), bool pickrandom = true) +{ + MatrixType m = pickrandom ? MatrixType::Random(a.rows(), a.cols()) : a; + bdcsvd_test_all_computation_options<MatrixType>(m); +} // end template bdcsvd + + +// verify assert +template<typename MatrixType> +void bdcsvd_verify_assert(const MatrixType& m) +{ + svd_verify_assert< MatrixType, BDCSVD< MatrixType > >(m); +}// end template bdcsvd_verify_assert + + +// test weird values +template<typename MatrixType> +void bdcsvd_inf_nan() +{ + svd_inf_nan< MatrixType, BDCSVD< MatrixType > >(); +}// end template bdcsvd_inf_nan + + + +void bdcsvd_preallocate() +{ + svd_preallocate< BDCSVD< MatrixXf > >(); +} // end bdcsvd_preallocate + + +// compare the Singular values returned with Jacobi and Bdc +template<typename MatrixType> +void compare_bdc_jacobi(const MatrixType& a = MatrixType(), unsigned int computationOptions = 0) +{ + std::cout << "debut compare" << std::endl; + MatrixType m = MatrixType::Random(a.rows(), a.cols()); + BDCSVD<MatrixType> bdc_svd(m); + JacobiSVD<MatrixType> jacobi_svd(m); + VERIFY_IS_APPROX(bdc_svd.singularValues(), jacobi_svd.singularValues()); + if(computationOptions & ComputeFullU) + VERIFY_IS_APPROX(bdc_svd.matrixU(), jacobi_svd.matrixU()); + if(computationOptions & ComputeThinU) + VERIFY_IS_APPROX(bdc_svd.matrixU(), jacobi_svd.matrixU()); + if(computationOptions & ComputeFullV) + VERIFY_IS_APPROX(bdc_svd.matrixV(), jacobi_svd.matrixV()); + if(computationOptions & ComputeThinV) + VERIFY_IS_APPROX(bdc_svd.matrixV(), jacobi_svd.matrixV()); + std::cout << "fin compare" << std::endl; +} // end template compare_bdc_jacobi + + +// call the tests +void test_bdcsvd() +{ + // test of Dynamic defined Matrix (42, 42) of float + CALL_SUBTEST_11(( bdcsvd_verify_assert<Matrix<float,Dynamic,Dynamic> > + (Matrix<float,Dynamic,Dynamic>(42,42)) )); + CALL_SUBTEST_11(( compare_bdc_jacobi<Matrix<float,Dynamic,Dynamic> > + (Matrix<float,Dynamic,Dynamic>(42,42), 0) )); + CALL_SUBTEST_11(( bdcsvd<Matrix<float,Dynamic,Dynamic> > + (Matrix<float,Dynamic,Dynamic>(42,42)) )); + + // test of Dynamic defined Matrix (50, 50) of double + CALL_SUBTEST_13(( bdcsvd_verify_assert<Matrix<double,Dynamic,Dynamic> > + (Matrix<double,Dynamic,Dynamic>(50,50)) )); + CALL_SUBTEST_13(( compare_bdc_jacobi<Matrix<double,Dynamic,Dynamic> > + (Matrix<double,Dynamic,Dynamic>(50,50), 0) )); + CALL_SUBTEST_13(( bdcsvd<Matrix<double,Dynamic,Dynamic> > + (Matrix<double,Dynamic,Dynamic>(50, 50)) )); + + // test of Dynamic defined Matrix (22, 22) of complex double + CALL_SUBTEST_14(( bdcsvd_verify_assert<Matrix<std::complex<double>,Dynamic,Dynamic> > + (Matrix<std::complex<double>,Dynamic,Dynamic>(22,22)) )); + CALL_SUBTEST_14(( compare_bdc_jacobi<Matrix<std::complex<double>,Dynamic,Dynamic> > + (Matrix<std::complex<double>, Dynamic, Dynamic> (22,22), 0) )); + CALL_SUBTEST_14(( bdcsvd<Matrix<std::complex<double>,Dynamic,Dynamic> > + (Matrix<std::complex<double>,Dynamic,Dynamic>(22, 22)) )); + + // test of Dynamic defined Matrix (10, 10) of int + //CALL_SUBTEST_15(( bdcsvd_verify_assert<Matrix<int,Dynamic,Dynamic> > + // (Matrix<int,Dynamic,Dynamic>(10,10)) )); + //CALL_SUBTEST_15(( compare_bdc_jacobi<Matrix<int,Dynamic,Dynamic> > + // (Matrix<int,Dynamic,Dynamic>(10,10), 0) )); + //CALL_SUBTEST_15(( bdcsvd<Matrix<int,Dynamic,Dynamic> > + // (Matrix<int,Dynamic,Dynamic>(10, 10)) )); + + + // test of Dynamic defined Matrix (8, 6) of double + + CALL_SUBTEST_16(( bdcsvd_verify_assert<Matrix<double,Dynamic,Dynamic> > + (Matrix<double,Dynamic,Dynamic>(8,6)) )); + CALL_SUBTEST_16(( compare_bdc_jacobi<Matrix<double,Dynamic,Dynamic> > + (Matrix<double,Dynamic,Dynamic>(8, 6), 0) )); + CALL_SUBTEST_16(( bdcsvd<Matrix<double,Dynamic,Dynamic> > + (Matrix<double,Dynamic,Dynamic>(8, 6)) )); + + + + // test of Dynamic defined Matrix (36, 12) of float + CALL_SUBTEST_17(( compare_bdc_jacobi<Matrix<float,Dynamic,Dynamic> > + (Matrix<float,Dynamic,Dynamic>(36, 12), 0) )); + CALL_SUBTEST_17(( bdcsvd<Matrix<float,Dynamic,Dynamic> > + (Matrix<float,Dynamic,Dynamic>(36, 12)) )); + + // test of Dynamic defined Matrix (5, 8) of double + CALL_SUBTEST_18(( compare_bdc_jacobi<Matrix<double,Dynamic,Dynamic> > + (Matrix<double,Dynamic,Dynamic>(5, 8), 0) )); + CALL_SUBTEST_18(( bdcsvd<Matrix<double,Dynamic,Dynamic> > + (Matrix<double,Dynamic,Dynamic>(5, 8)) )); + + + // non regression tests + CALL_SUBTEST_3(( bdcsvd_verify_assert(Matrix3f()) )); + CALL_SUBTEST_4(( bdcsvd_verify_assert(Matrix4d()) )); + CALL_SUBTEST_7(( bdcsvd_verify_assert(MatrixXf(10,12)) )); + CALL_SUBTEST_8(( bdcsvd_verify_assert(MatrixXcd(7,5)) )); + + // SUBTESTS 1 and 2 on specifics matrix + for(int i = 0; i < g_repeat; i++) { + Matrix2cd m; + m << 0, 1, + 0, 1; + CALL_SUBTEST_1(( bdcsvd(m, false) )); + m << 1, 0, + 1, 0; + CALL_SUBTEST_1(( bdcsvd(m, false) )); + + Matrix2d n; + n << 0, 0, + 0, 0; + CALL_SUBTEST_2(( bdcsvd(n, false) )); + n << 0, 0, + 0, 1; + CALL_SUBTEST_2(( bdcsvd(n, false) )); + + // Statics matrix don't work with BDSVD yet + // bdc algo on a random 3x3 float matrix + // CALL_SUBTEST_3(( bdcsvd<Matrix3f>() )); + // bdc algo on a random 4x4 double matrix + // CALL_SUBTEST_4(( bdcsvd<Matrix4d>() )); + // bdc algo on a random 3x5 float matrix + // CALL_SUBTEST_5(( bdcsvd<Matrix<float,3,5> >() )); + + int r = internal::random<int>(1, 30), + c = internal::random<int>(1, 30); + CALL_SUBTEST_7(( bdcsvd<MatrixXf>(MatrixXf(r,c)) )); + CALL_SUBTEST_8(( bdcsvd<MatrixXcd>(MatrixXcd(r,c)) )); + (void) r; + (void) c; + + // Test on inf/nan matrix + CALL_SUBTEST_7( bdcsvd_inf_nan<MatrixXf>() ); + } + + CALL_SUBTEST_7(( bdcsvd<MatrixXf>(MatrixXf(internal::random<int>(EIGEN_TEST_MAX_SIZE/4, EIGEN_TEST_MAX_SIZE/2), internal::random<int>(EIGEN_TEST_MAX_SIZE/4, EIGEN_TEST_MAX_SIZE/2))) )); + CALL_SUBTEST_8(( bdcsvd<MatrixXcd>(MatrixXcd(internal::random<int>(EIGEN_TEST_MAX_SIZE/4, EIGEN_TEST_MAX_SIZE/3), internal::random<int>(EIGEN_TEST_MAX_SIZE/4, EIGEN_TEST_MAX_SIZE/3))) )); + + // Test problem size constructors + CALL_SUBTEST_7( BDCSVD<MatrixXf>(10,10) ); + +} // end test_bdcsvd diff --git a/unsupported/test/dgmres.cpp b/unsupported/test/dgmres.cpp new file mode 100644 index 000000000..2b11807c8 --- /dev/null +++ b/unsupported/test/dgmres.cpp @@ -0,0 +1,31 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2011 Gael Guennebaud <g.gael@free.fr> +// Copyright (C) 2012 desire Nuentsa <desire.nuentsa_wakam@inria.fr +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + +#include "../../test/sparse_solver.h" +#include <Eigen/src/IterativeSolvers/DGMRES.h> + +template<typename T> void test_dgmres_T() +{ + DGMRES<SparseMatrix<T>, DiagonalPreconditioner<T> > dgmres_colmajor_diag; + DGMRES<SparseMatrix<T>, IdentityPreconditioner > dgmres_colmajor_I; + DGMRES<SparseMatrix<T>, IncompleteLUT<T> > dgmres_colmajor_ilut; + //GMRES<SparseMatrix<T>, SSORPreconditioner<T> > dgmres_colmajor_ssor; + + CALL_SUBTEST( check_sparse_square_solving(dgmres_colmajor_diag) ); +// CALL_SUBTEST( check_sparse_square_solving(dgmres_colmajor_I) ); + CALL_SUBTEST( check_sparse_square_solving(dgmres_colmajor_ilut) ); + //CALL_SUBTEST( check_sparse_square_solving(dgmres_colmajor_ssor) ); +} + +void test_dgmres() +{ + CALL_SUBTEST_1(test_dgmres_T<double>()); + CALL_SUBTEST_2(test_dgmres_T<std::complex<double> >()); +} diff --git a/unsupported/test/gmres.cpp b/unsupported/test/gmres.cpp index 647c16927..f2969116b 100644 --- a/unsupported/test/gmres.cpp +++ b/unsupported/test/gmres.cpp @@ -26,8 +26,6 @@ template<typename T> void test_gmres_T() void test_gmres() { - for(int i = 0; i < g_repeat; i++) { - CALL_SUBTEST_1(test_gmres_T<double>()); - CALL_SUBTEST_2(test_gmres_T<std::complex<double> >()); - } + CALL_SUBTEST_1(test_gmres_T<double>()); + CALL_SUBTEST_2(test_gmres_T<std::complex<double> >()); } diff --git a/unsupported/test/jacobisvd.cpp b/unsupported/test/jacobisvd.cpp new file mode 100644 index 000000000..b4e884eee --- /dev/null +++ b/unsupported/test/jacobisvd.cpp @@ -0,0 +1,198 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2008 Gael Guennebaud <gael.guennebaud@inria.fr> +// Copyright (C) 2009 Benoit Jacob <jacob.benoit.1@gmail.com> +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + +#include "svd_common.h" + +template<typename MatrixType, int QRPreconditioner> +void jacobisvd_check_full(const MatrixType& m, const JacobiSVD<MatrixType, QRPreconditioner>& svd) +{ + svd_check_full<MatrixType, JacobiSVD<MatrixType, QRPreconditioner > >(m, svd); +} + +template<typename MatrixType, int QRPreconditioner> +void jacobisvd_compare_to_full(const MatrixType& m, + unsigned int computationOptions, + const JacobiSVD<MatrixType, QRPreconditioner>& referenceSvd) +{ + svd_compare_to_full<MatrixType, JacobiSVD<MatrixType, QRPreconditioner> >(m, computationOptions, referenceSvd); +} + + +template<typename MatrixType, int QRPreconditioner> +void jacobisvd_solve(const MatrixType& m, unsigned int computationOptions) +{ + svd_solve< MatrixType, JacobiSVD< MatrixType, QRPreconditioner > >(m, computationOptions); +} + + + +template<typename MatrixType, int QRPreconditioner> +void jacobisvd_test_all_computation_options(const MatrixType& m) +{ + + if (QRPreconditioner == NoQRPreconditioner && m.rows() != m.cols()) + return; + + JacobiSVD< MatrixType, QRPreconditioner > fullSvd(m, ComputeFullU|ComputeFullV); + svd_test_computation_options_1< MatrixType, JacobiSVD< MatrixType, QRPreconditioner > >(m, fullSvd); + + if(QRPreconditioner == FullPivHouseholderQRPreconditioner) + return; + svd_test_computation_options_2< MatrixType, JacobiSVD< MatrixType, QRPreconditioner > >(m, fullSvd); + +} + +template<typename MatrixType> +void jacobisvd(const MatrixType& a = MatrixType(), bool pickrandom = true) +{ + MatrixType m = pickrandom ? MatrixType::Random(a.rows(), a.cols()) : a; + + jacobisvd_test_all_computation_options<MatrixType, FullPivHouseholderQRPreconditioner>(m); + jacobisvd_test_all_computation_options<MatrixType, ColPivHouseholderQRPreconditioner>(m); + jacobisvd_test_all_computation_options<MatrixType, HouseholderQRPreconditioner>(m); + jacobisvd_test_all_computation_options<MatrixType, NoQRPreconditioner>(m); +} + + +template<typename MatrixType> +void jacobisvd_verify_assert(const MatrixType& m) +{ + + svd_verify_assert<MatrixType, JacobiSVD< MatrixType > >(m); + + typedef typename MatrixType::Index Index; + Index rows = m.rows(); + Index cols = m.cols(); + + enum { + RowsAtCompileTime = MatrixType::RowsAtCompileTime, + ColsAtCompileTime = MatrixType::ColsAtCompileTime + }; + + MatrixType a = MatrixType::Zero(rows, cols); + a.setZero(); + + if (ColsAtCompileTime == Dynamic) + { + JacobiSVD<MatrixType, FullPivHouseholderQRPreconditioner> svd_fullqr; + VERIFY_RAISES_ASSERT(svd_fullqr.compute(a, ComputeFullU|ComputeThinV)) + VERIFY_RAISES_ASSERT(svd_fullqr.compute(a, ComputeThinU|ComputeThinV)) + VERIFY_RAISES_ASSERT(svd_fullqr.compute(a, ComputeThinU|ComputeFullV)) + } +} + +template<typename MatrixType> +void jacobisvd_method() +{ + enum { Size = MatrixType::RowsAtCompileTime }; + typedef typename MatrixType::RealScalar RealScalar; + typedef Matrix<RealScalar, Size, 1> RealVecType; + MatrixType m = MatrixType::Identity(); + VERIFY_IS_APPROX(m.jacobiSvd().singularValues(), RealVecType::Ones()); + VERIFY_RAISES_ASSERT(m.jacobiSvd().matrixU()); + VERIFY_RAISES_ASSERT(m.jacobiSvd().matrixV()); + VERIFY_IS_APPROX(m.jacobiSvd(ComputeFullU|ComputeFullV).solve(m), m); +} + + + +template<typename MatrixType> +void jacobisvd_inf_nan() +{ + svd_inf_nan<MatrixType, JacobiSVD< MatrixType > >(); +} + + +// Regression test for bug 286: JacobiSVD loops indefinitely with some +// matrices containing denormal numbers. +void jacobisvd_bug286() +{ +#if defined __INTEL_COMPILER +// shut up warning #239: floating point underflow +#pragma warning push +#pragma warning disable 239 +#endif + Matrix2d M; + M << -7.90884e-313, -4.94e-324, + 0, 5.60844e-313; +#if defined __INTEL_COMPILER +#pragma warning pop +#endif + JacobiSVD<Matrix2d> svd; + svd.compute(M); // just check we don't loop indefinitely +} + + +void jacobisvd_preallocate() +{ + svd_preallocate< JacobiSVD <MatrixXf> >(); +} + +void test_jacobisvd() +{ + CALL_SUBTEST_11(( jacobisvd<Matrix<double,Dynamic,Dynamic> > + (Matrix<double,Dynamic,Dynamic>(16, 6)) )); + + CALL_SUBTEST_3(( jacobisvd_verify_assert(Matrix3f()) )); + CALL_SUBTEST_4(( jacobisvd_verify_assert(Matrix4d()) )); + CALL_SUBTEST_7(( jacobisvd_verify_assert(MatrixXf(10,12)) )); + CALL_SUBTEST_8(( jacobisvd_verify_assert(MatrixXcd(7,5)) )); + + for(int i = 0; i < g_repeat; i++) { + Matrix2cd m; + m << 0, 1, + 0, 1; + CALL_SUBTEST_1(( jacobisvd(m, false) )); + m << 1, 0, + 1, 0; + CALL_SUBTEST_1(( jacobisvd(m, false) )); + + Matrix2d n; + n << 0, 0, + 0, 0; + CALL_SUBTEST_2(( jacobisvd(n, false) )); + n << 0, 0, + 0, 1; + CALL_SUBTEST_2(( jacobisvd(n, false) )); + + CALL_SUBTEST_3(( jacobisvd<Matrix3f>() )); + CALL_SUBTEST_4(( jacobisvd<Matrix4d>() )); + CALL_SUBTEST_5(( jacobisvd<Matrix<float,3,5> >() )); + CALL_SUBTEST_6(( jacobisvd<Matrix<double,Dynamic,2> >(Matrix<double,Dynamic,2>(10,2)) )); + + int r = internal::random<int>(1, 30), + c = internal::random<int>(1, 30); + CALL_SUBTEST_7(( jacobisvd<MatrixXf>(MatrixXf(r,c)) )); + CALL_SUBTEST_8(( jacobisvd<MatrixXcd>(MatrixXcd(r,c)) )); + (void) r; + (void) c; + + // Test on inf/nan matrix + CALL_SUBTEST_7( jacobisvd_inf_nan<MatrixXf>() ); + } + + CALL_SUBTEST_7(( jacobisvd<MatrixXf>(MatrixXf(internal::random<int>(EIGEN_TEST_MAX_SIZE/4, EIGEN_TEST_MAX_SIZE/2), internal::random<int>(EIGEN_TEST_MAX_SIZE/4, EIGEN_TEST_MAX_SIZE/2))) )); + CALL_SUBTEST_8(( jacobisvd<MatrixXcd>(MatrixXcd(internal::random<int>(EIGEN_TEST_MAX_SIZE/4, EIGEN_TEST_MAX_SIZE/3), internal::random<int>(EIGEN_TEST_MAX_SIZE/4, EIGEN_TEST_MAX_SIZE/3))) )); + + + // test matrixbase method + CALL_SUBTEST_1(( jacobisvd_method<Matrix2cd>() )); + CALL_SUBTEST_3(( jacobisvd_method<Matrix3f>() )); + + + // Test problem size constructors + CALL_SUBTEST_7( JacobiSVD<MatrixXf>(10,10) ); + + // Check that preallocation avoids subsequent mallocs + CALL_SUBTEST_9( jacobisvd_preallocate() ); + + // Regression check for bug 286 + CALL_SUBTEST_2( jacobisvd_bug286() ); +} diff --git a/unsupported/test/kronecker_product.cpp b/unsupported/test/kronecker_product.cpp index a60bd3022..8ddc6ec28 100644 --- a/unsupported/test/kronecker_product.cpp +++ b/unsupported/test/kronecker_product.cpp @@ -3,6 +3,7 @@ // // Copyright (C) 2011 Kolja Brix <brix@igpm.rwth-aachen.de> // Copyright (C) 2011 Andreas Platen <andiplaten@gmx.de> +// Copyright (C) 2012 Chen-Pang He <jdh8@ms63.hinet.net> // // This Source Code Form is subject to the terms of the Mozilla // Public License v. 2.0. If a copy of the MPL was not distributed @@ -15,7 +16,7 @@ template<typename MatrixType> -void check_dimension(const MatrixType& ab, const unsigned int rows, const unsigned int cols) +void check_dimension(const MatrixType& ab, const int rows, const int cols) { VERIFY_IS_EQUAL(ab.rows(), rows); VERIFY_IS_EQUAL(ab.cols(), cols); @@ -85,68 +86,69 @@ void check_sparse_kronecker_product(const MatrixType& ab) void test_kronecker_product() { // DM = dense matrix; SM = sparse matrix + Matrix<double, 2, 3> DM_a; - MatrixXd DM_b(3,2); SparseMatrix<double> SM_a(2,3); + SM_a.insert(0,0) = DM_a.coeffRef(0,0) = -0.4461540300782201; + SM_a.insert(0,1) = DM_a.coeffRef(0,1) = -0.8057364375283049; + SM_a.insert(0,2) = DM_a.coeffRef(0,2) = 0.3896572459516341; + SM_a.insert(1,0) = DM_a.coeffRef(1,0) = -0.9076572187376921; + SM_a.insert(1,1) = DM_a.coeffRef(1,1) = 0.6469156566545853; + SM_a.insert(1,2) = DM_a.coeffRef(1,2) = -0.3658010398782789; + + MatrixXd DM_b(3,2); SparseMatrix<double> SM_b(3,2); - SM_a.insert(0,0) = DM_a(0,0) = -0.4461540300782201; - SM_a.insert(0,1) = DM_a(0,1) = -0.8057364375283049; - SM_a.insert(0,2) = DM_a(0,2) = 0.3896572459516341; - SM_a.insert(1,0) = DM_a(1,0) = -0.9076572187376921; - SM_a.insert(1,1) = DM_a(1,1) = 0.6469156566545853; - SM_a.insert(1,2) = DM_a(1,2) = -0.3658010398782789; - SM_b.insert(0,0) = DM_b(0,0) = 0.9004440976767099; - SM_b.insert(0,1) = DM_b(0,1) = -0.2368830858139832; - SM_b.insert(1,0) = DM_b(1,0) = -0.9311078389941825; - SM_b.insert(1,1) = DM_b(1,1) = 0.5310335762980047; - SM_b.insert(2,0) = DM_b(2,0) = -0.1225112806872035; - SM_b.insert(2,1) = DM_b(2,1) = 0.5903998022741264; + SM_b.insert(0,0) = DM_b.coeffRef(0,0) = 0.9004440976767099; + SM_b.insert(0,1) = DM_b.coeffRef(0,1) = -0.2368830858139832; + SM_b.insert(1,0) = DM_b.coeffRef(1,0) = -0.9311078389941825; + SM_b.insert(1,1) = DM_b.coeffRef(1,1) = 0.5310335762980047; + SM_b.insert(2,0) = DM_b.coeffRef(2,0) = -0.1225112806872035; + SM_b.insert(2,1) = DM_b.coeffRef(2,1) = 0.5903998022741264; + SparseMatrix<double,RowMajor> SM_row_a(SM_a), SM_row_b(SM_b); // test kroneckerProduct(DM_block,DM,DM_fixedSize) - Matrix<double, 6, 6> DM_fix_ab; - DM_fix_ab(0,0)=37.0; - kroneckerProduct(DM_a.block(0,0,2,3),DM_b,DM_fix_ab); + Matrix<double, 6, 6> DM_fix_ab = kroneckerProduct(DM_a.topLeftCorner<2,3>(),DM_b); + CALL_SUBTEST(check_kronecker_product(DM_fix_ab)); + for(int i=0;i<DM_fix_ab.rows();++i) + for(int j=0;j<DM_fix_ab.cols();++j) + VERIFY_IS_APPROX(kroneckerProduct(DM_a,DM_b).coeff(i,j), DM_fix_ab(i,j)); + // test kroneckerProduct(DM,DM,DM_block) MatrixXd DM_block_ab(10,15); - DM_block_ab(0,0)=37.0; - kroneckerProduct(DM_a,DM_b,DM_block_ab.block(2,5,6,6)); - CALL_SUBTEST(check_kronecker_product(DM_block_ab.block(2,5,6,6))); + DM_block_ab.block<6,6>(2,5) = kroneckerProduct(DM_a,DM_b); + CALL_SUBTEST(check_kronecker_product(DM_block_ab.block<6,6>(2,5))); // test kroneckerProduct(DM,DM,DM) - MatrixXd DM_ab(1,5); - DM_ab(0,0)=37.0; - kroneckerProduct(DM_a,DM_b,DM_ab); + MatrixXd DM_ab = kroneckerProduct(DM_a,DM_b); CALL_SUBTEST(check_kronecker_product(DM_ab)); // test kroneckerProduct(SM,DM,SM) - SparseMatrix<double> SM_ab(1,20); - SM_ab.insert(0,0)=37.0; - kroneckerProduct(SM_a,DM_b,SM_ab); + SparseMatrix<double> SM_ab = kroneckerProduct(SM_a,DM_b); CALL_SUBTEST(check_kronecker_product(SM_ab)); - SparseMatrix<double,RowMajor> SM_ab2(10,3); - SM_ab2.insert(0,0)=37.0; - kroneckerProduct(SM_a,DM_b,SM_ab2); + SparseMatrix<double,RowMajor> SM_ab2 = kroneckerProduct(SM_a,DM_b); CALL_SUBTEST(check_kronecker_product(SM_ab2)); // test kroneckerProduct(DM,SM,SM) + SM_ab.setZero(); SM_ab.insert(0,0)=37.0; - kroneckerProduct(DM_a,SM_b,SM_ab); + SM_ab = kroneckerProduct(DM_a,SM_b); CALL_SUBTEST(check_kronecker_product(SM_ab)); + SM_ab2.setZero(); SM_ab2.insert(0,0)=37.0; - kroneckerProduct(DM_a,SM_b,SM_ab2); + SM_ab2 = kroneckerProduct(DM_a,SM_b); CALL_SUBTEST(check_kronecker_product(SM_ab2)); // test kroneckerProduct(SM,SM,SM) SM_ab.resize(2,33); SM_ab.insert(0,0)=37.0; - kroneckerProduct(SM_a,SM_b,SM_ab); + SM_ab = kroneckerProduct(SM_a,SM_b); CALL_SUBTEST(check_kronecker_product(SM_ab)); SM_ab2.resize(5,11); SM_ab2.insert(0,0)=37.0; - kroneckerProduct(SM_a,SM_b,SM_ab2); + SM_ab2 = kroneckerProduct(SM_a,SM_b); CALL_SUBTEST(check_kronecker_product(SM_ab2)); // test kroneckerProduct(SM,SM,SM) with sparse pattern @@ -158,22 +160,22 @@ void test_kronecker_product() SM_a.insert(0,3) = -0.2; SM_a.insert(2,4) = 0.3; SM_a.finalize(); + SM_b.insert(0,0) = 0.4; SM_b.insert(2,1) = -0.5; SM_b.finalize(); SM_ab.resize(1,1); SM_ab.insert(0,0)=37.0; - kroneckerProduct(SM_a,SM_b,SM_ab); + SM_ab = kroneckerProduct(SM_a,SM_b); CALL_SUBTEST(check_sparse_kronecker_product(SM_ab)); // test dimension of result of kroneckerProduct(DM,DM,DM) MatrixXd DM_a2(2,1); MatrixXd DM_b2(5,4); - MatrixXd DM_ab2; - kroneckerProduct(DM_a2,DM_b2,DM_ab2); + MatrixXd DM_ab2 = kroneckerProduct(DM_a2,DM_b2); CALL_SUBTEST(check_dimension(DM_ab2,2*5,1*4)); DM_a2.resize(10,9); DM_b2.resize(4,8); - kroneckerProduct(DM_a2,DM_b2,DM_ab2); + DM_ab2 = kroneckerProduct(DM_a2,DM_b2); CALL_SUBTEST(check_dimension(DM_ab2,10*4,9*8)); } diff --git a/unsupported/test/levenberg_marquardt.cpp b/unsupported/test/levenberg_marquardt.cpp new file mode 100644 index 000000000..04464727d --- /dev/null +++ b/unsupported/test/levenberg_marquardt.cpp @@ -0,0 +1,1448 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2009 Thomas Capricelli <orzel@freehackers.org> +// Copyright (C) 2012 desire Nuentsa <desire.nuentsa_wakam@inria.fr +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + + +#include <stdio.h> + +#include "main.h" +#include <unsupported/Eigen/LevenbergMarquardt> + +// This disables some useless Warnings on MSVC. +// It is intended to be done for this test only. +#include <Eigen/src/Core/util/DisableStupidWarnings.h> + +using std::sqrt; + +struct lmder_functor : DenseFunctor<double> +{ + lmder_functor(void): DenseFunctor<double>(3,15) {} + int operator()(const VectorXd &x, VectorXd &fvec) const + { + double tmp1, tmp2, tmp3; + static const double y[15] = {1.4e-1, 1.8e-1, 2.2e-1, 2.5e-1, 2.9e-1, 3.2e-1, 3.5e-1, + 3.9e-1, 3.7e-1, 5.8e-1, 7.3e-1, 9.6e-1, 1.34, 2.1, 4.39}; + + for (int i = 0; i < values(); i++) + { + tmp1 = i+1; + tmp2 = 16 - i - 1; + tmp3 = (i>=8)? tmp2 : tmp1; + fvec[i] = y[i] - (x[0] + tmp1/(x[1]*tmp2 + x[2]*tmp3)); + } + return 0; + } + + int df(const VectorXd &x, MatrixXd &fjac) const + { + double tmp1, tmp2, tmp3, tmp4; + for (int i = 0; i < values(); i++) + { + tmp1 = i+1; + tmp2 = 16 - i - 1; + tmp3 = (i>=8)? tmp2 : tmp1; + tmp4 = (x[1]*tmp2 + x[2]*tmp3); tmp4 = tmp4*tmp4; + fjac(i,0) = -1; + fjac(i,1) = tmp1*tmp2/tmp4; + fjac(i,2) = tmp1*tmp3/tmp4; + } + return 0; + } +}; + +void testLmder1() +{ + int n=3, info; + + VectorXd x; + + /* the following starting values provide a rough fit. */ + x.setConstant(n, 1.); + + // do the computation + lmder_functor functor; + LevenbergMarquardt<lmder_functor> lm(functor); + info = lm.lmder1(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 6); + VERIFY_IS_EQUAL(lm.njev(), 5); + + // check norm + VERIFY_IS_APPROX(lm.fvec().blueNorm(), 0.09063596); + + // check x + VectorXd x_ref(n); + x_ref << 0.08241058, 1.133037, 2.343695; + VERIFY_IS_APPROX(x, x_ref); +} + +void testLmder() +{ + const int m=15, n=3; + int info; + double fnorm, covfac; + VectorXd x; + + /* the following starting values provide a rough fit. */ + x.setConstant(n, 1.); + + // do the computation + lmder_functor functor; + LevenbergMarquardt<lmder_functor> lm(functor); + info = lm.minimize(x); + + // check return values + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 6); + VERIFY_IS_EQUAL(lm.njev(), 5); + + // check norm + fnorm = lm.fvec().blueNorm(); + VERIFY_IS_APPROX(fnorm, 0.09063596); + + // check x + VectorXd x_ref(n); + x_ref << 0.08241058, 1.133037, 2.343695; + VERIFY_IS_APPROX(x, x_ref); + + // check covariance + covfac = fnorm*fnorm/(m-n); + internal::covar(lm.matrixR(), lm.permutation().indices()); // TODO : move this as a function of lm + + MatrixXd cov_ref(n,n); + cov_ref << + 0.0001531202, 0.002869941, -0.002656662, + 0.002869941, 0.09480935, -0.09098995, + -0.002656662, -0.09098995, 0.08778727; + +// std::cout << fjac*covfac << std::endl; + + MatrixXd cov; + cov = covfac*lm.matrixR().topLeftCorner<n,n>(); + VERIFY_IS_APPROX( cov, cov_ref); + // TODO: why isn't this allowed ? : + // VERIFY_IS_APPROX( covfac*fjac.topLeftCorner<n,n>() , cov_ref); +} + +struct lmdif_functor : DenseFunctor<double> +{ + lmdif_functor(void) : DenseFunctor<double>(3,15) {} + int operator()(const VectorXd &x, VectorXd &fvec) const + { + int i; + double tmp1,tmp2,tmp3; + static const double y[15]={1.4e-1,1.8e-1,2.2e-1,2.5e-1,2.9e-1,3.2e-1,3.5e-1,3.9e-1, + 3.7e-1,5.8e-1,7.3e-1,9.6e-1,1.34e0,2.1e0,4.39e0}; + + assert(x.size()==3); + assert(fvec.size()==15); + for (i=0; i<15; i++) + { + tmp1 = i+1; + tmp2 = 15 - i; + tmp3 = tmp1; + + if (i >= 8) tmp3 = tmp2; + fvec[i] = y[i] - (x[0] + tmp1/(x[1]*tmp2 + x[2]*tmp3)); + } + return 0; + } +}; + +void testLmdif1() +{ + const int n=3; + int info; + + VectorXd x(n), fvec(15); + + /* the following starting values provide a rough fit. */ + x.setConstant(n, 1.); + + // do the computation + lmdif_functor functor; + DenseIndex nfev; + info = LevenbergMarquardt<lmdif_functor>::lmdif1(functor, x, &nfev); + + // check return value + VERIFY_IS_EQUAL(info, 1); +// VERIFY_IS_EQUAL(nfev, 26); + + // check norm + functor(x, fvec); + VERIFY_IS_APPROX(fvec.blueNorm(), 0.09063596); + + // check x + VectorXd x_ref(n); + x_ref << 0.0824106, 1.1330366, 2.3436947; + VERIFY_IS_APPROX(x, x_ref); + +} + +void testLmdif() +{ + const int m=15, n=3; + int info; + double fnorm, covfac; + VectorXd x(n); + + /* the following starting values provide a rough fit. */ + x.setConstant(n, 1.); + + // do the computation + lmdif_functor functor; + NumericalDiff<lmdif_functor> numDiff(functor); + LevenbergMarquardt<NumericalDiff<lmdif_functor> > lm(numDiff); + info = lm.minimize(x); + + // check return values + VERIFY_IS_EQUAL(info, 1); +// VERIFY_IS_EQUAL(lm.nfev(), 26); + + // check norm + fnorm = lm.fvec().blueNorm(); + VERIFY_IS_APPROX(fnorm, 0.09063596); + + // check x + VectorXd x_ref(n); + x_ref << 0.08241058, 1.133037, 2.343695; + VERIFY_IS_APPROX(x, x_ref); + + // check covariance + covfac = fnorm*fnorm/(m-n); + internal::covar(lm.matrixR(), lm.permutation().indices()); // TODO : move this as a function of lm + + MatrixXd cov_ref(n,n); + cov_ref << + 0.0001531202, 0.002869942, -0.002656662, + 0.002869942, 0.09480937, -0.09098997, + -0.002656662, -0.09098997, 0.08778729; + +// std::cout << fjac*covfac << std::endl; + + MatrixXd cov; + cov = covfac*lm.matrixR().topLeftCorner<n,n>(); + VERIFY_IS_APPROX( cov, cov_ref); + // TODO: why isn't this allowed ? : + // VERIFY_IS_APPROX( covfac*fjac.topLeftCorner<n,n>() , cov_ref); +} + +struct chwirut2_functor : DenseFunctor<double> +{ + chwirut2_functor(void) : DenseFunctor<double>(3,54) {} + static const double m_x[54]; + static const double m_y[54]; + int operator()(const VectorXd &b, VectorXd &fvec) + { + int i; + + assert(b.size()==3); + assert(fvec.size()==54); + for(i=0; i<54; i++) { + double x = m_x[i]; + fvec[i] = exp(-b[0]*x)/(b[1]+b[2]*x) - m_y[i]; + } + return 0; + } + int df(const VectorXd &b, MatrixXd &fjac) + { + assert(b.size()==3); + assert(fjac.rows()==54); + assert(fjac.cols()==3); + for(int i=0; i<54; i++) { + double x = m_x[i]; + double factor = 1./(b[1]+b[2]*x); + double e = exp(-b[0]*x); + fjac(i,0) = -x*e*factor; + fjac(i,1) = -e*factor*factor; + fjac(i,2) = -x*e*factor*factor; + } + return 0; + } +}; +const double chwirut2_functor::m_x[54] = { 0.500E0, 1.000E0, 1.750E0, 3.750E0, 5.750E0, 0.875E0, 2.250E0, 3.250E0, 5.250E0, 0.750E0, 1.750E0, 2.750E0, 4.750E0, 0.625E0, 1.250E0, 2.250E0, 4.250E0, .500E0, 3.000E0, .750E0, 3.000E0, 1.500E0, 6.000E0, 3.000E0, 6.000E0, 1.500E0, 3.000E0, .500E0, 2.000E0, 4.000E0, .750E0, 2.000E0, 5.000E0, .750E0, 2.250E0, 3.750E0, 5.750E0, 3.000E0, .750E0, 2.500E0, 4.000E0, .750E0, 2.500E0, 4.000E0, .750E0, 2.500E0, 4.000E0, .500E0, 6.000E0, 3.000E0, .500E0, 2.750E0, .500E0, 1.750E0}; +const double chwirut2_functor::m_y[54] = { 92.9000E0 ,57.1000E0 ,31.0500E0 ,11.5875E0 ,8.0250E0 ,63.6000E0 ,21.4000E0 ,14.2500E0 ,8.4750E0 ,63.8000E0 ,26.8000E0 ,16.4625E0 ,7.1250E0 ,67.3000E0 ,41.0000E0 ,21.1500E0 ,8.1750E0 ,81.5000E0 ,13.1200E0 ,59.9000E0 ,14.6200E0 ,32.9000E0 ,5.4400E0 ,12.5600E0 ,5.4400E0 ,32.0000E0 ,13.9500E0 ,75.8000E0 ,20.0000E0 ,10.4200E0 ,59.5000E0 ,21.6700E0 ,8.5500E0 ,62.0000E0 ,20.2000E0 ,7.7600E0 ,3.7500E0 ,11.8100E0 ,54.7000E0 ,23.7000E0 ,11.5500E0 ,61.3000E0 ,17.7000E0 ,8.7400E0 ,59.2000E0 ,16.3000E0 ,8.6200E0 ,81.0000E0 ,4.8700E0 ,14.6200E0 ,81.7000E0 ,17.1700E0 ,81.3000E0 ,28.9000E0 }; + +// http://www.itl.nist.gov/div898/strd/nls/data/chwirut2.shtml +void testNistChwirut2(void) +{ + const int n=3; + int info; + + VectorXd x(n); + + /* + * First try + */ + x<< 0.1, 0.01, 0.02; + // do the computation + chwirut2_functor functor; + LevenbergMarquardt<chwirut2_functor> lm(functor); + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); +// VERIFY_IS_EQUAL(lm.nfev(), 10); + VERIFY_IS_EQUAL(lm.njev(), 8); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 5.1304802941E+02); + // check x + VERIFY_IS_APPROX(x[0], 1.6657666537E-01); + VERIFY_IS_APPROX(x[1], 5.1653291286E-03); + VERIFY_IS_APPROX(x[2], 1.2150007096E-02); + + /* + * Second try + */ + x<< 0.15, 0.008, 0.010; + // do the computation + lm.resetParameters(); + lm.setFtol(1.E6*NumTraits<double>::epsilon()); + lm.setXtol(1.E6*NumTraits<double>::epsilon()); + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); +// VERIFY_IS_EQUAL(lm.nfev(), 7); + VERIFY_IS_EQUAL(lm.njev(), 6); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 5.1304802941E+02); + // check x + VERIFY_IS_APPROX(x[0], 1.6657666537E-01); + VERIFY_IS_APPROX(x[1], 5.1653291286E-03); + VERIFY_IS_APPROX(x[2], 1.2150007096E-02); +} + + +struct misra1a_functor : DenseFunctor<double> +{ + misra1a_functor(void) : DenseFunctor<double>(2,14) {} + static const double m_x[14]; + static const double m_y[14]; + int operator()(const VectorXd &b, VectorXd &fvec) + { + assert(b.size()==2); + assert(fvec.size()==14); + for(int i=0; i<14; i++) { + fvec[i] = b[0]*(1.-exp(-b[1]*m_x[i])) - m_y[i] ; + } + return 0; + } + int df(const VectorXd &b, MatrixXd &fjac) + { + assert(b.size()==2); + assert(fjac.rows()==14); + assert(fjac.cols()==2); + for(int i=0; i<14; i++) { + fjac(i,0) = (1.-exp(-b[1]*m_x[i])); + fjac(i,1) = (b[0]*m_x[i]*exp(-b[1]*m_x[i])); + } + return 0; + } +}; +const double misra1a_functor::m_x[14] = { 77.6E0, 114.9E0, 141.1E0, 190.8E0, 239.9E0, 289.0E0, 332.8E0, 378.4E0, 434.8E0, 477.3E0, 536.8E0, 593.1E0, 689.1E0, 760.0E0}; +const double misra1a_functor::m_y[14] = { 10.07E0, 14.73E0, 17.94E0, 23.93E0, 29.61E0, 35.18E0, 40.02E0, 44.82E0, 50.76E0, 55.05E0, 61.01E0, 66.40E0, 75.47E0, 81.78E0}; + +// http://www.itl.nist.gov/div898/strd/nls/data/misra1a.shtml +void testNistMisra1a(void) +{ + const int n=2; + int info; + + VectorXd x(n); + + /* + * First try + */ + x<< 500., 0.0001; + // do the computation + misra1a_functor functor; + LevenbergMarquardt<misra1a_functor> lm(functor); + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 19); + VERIFY_IS_EQUAL(lm.njev(), 15); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 1.2455138894E-01); + // check x + VERIFY_IS_APPROX(x[0], 2.3894212918E+02); + VERIFY_IS_APPROX(x[1], 5.5015643181E-04); + + /* + * Second try + */ + x<< 250., 0.0005; + // do the computation + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 5); + VERIFY_IS_EQUAL(lm.njev(), 4); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 1.2455138894E-01); + // check x + VERIFY_IS_APPROX(x[0], 2.3894212918E+02); + VERIFY_IS_APPROX(x[1], 5.5015643181E-04); +} + +struct hahn1_functor : DenseFunctor<double> +{ + hahn1_functor(void) : DenseFunctor<double>(7,236) {} + static const double m_x[236]; + int operator()(const VectorXd &b, VectorXd &fvec) + { + static const double m_y[236] = { .591E0 , 1.547E0 , 2.902E0 , 2.894E0 , 4.703E0 , 6.307E0 , 7.03E0 , 7.898E0 , 9.470E0 , 9.484E0 , 10.072E0 , 10.163E0 , 11.615E0 , 12.005E0 , 12.478E0 , 12.982E0 , 12.970E0 , 13.926E0 , 14.452E0 , 14.404E0 , 15.190E0 , 15.550E0 , 15.528E0 , 15.499E0 , 16.131E0 , 16.438E0 , 16.387E0 , 16.549E0 , 16.872E0 , 16.830E0 , 16.926E0 , 16.907E0 , 16.966E0 , 17.060E0 , 17.122E0 , 17.311E0 , 17.355E0 , 17.668E0 , 17.767E0 , 17.803E0 , 17.765E0 , 17.768E0 , 17.736E0 , 17.858E0 , 17.877E0 , 17.912E0 , 18.046E0 , 18.085E0 , 18.291E0 , 18.357E0 , 18.426E0 , 18.584E0 , 18.610E0 , 18.870E0 , 18.795E0 , 19.111E0 , .367E0 , .796E0 , 0.892E0 , 1.903E0 , 2.150E0 , 3.697E0 , 5.870E0 , 6.421E0 , 7.422E0 , 9.944E0 , 11.023E0 , 11.87E0 , 12.786E0 , 14.067E0 , 13.974E0 , 14.462E0 , 14.464E0 , 15.381E0 , 15.483E0 , 15.59E0 , 16.075E0 , 16.347E0 , 16.181E0 , 16.915E0 , 17.003E0 , 16.978E0 , 17.756E0 , 17.808E0 , 17.868E0 , 18.481E0 , 18.486E0 , 19.090E0 , 16.062E0 , 16.337E0 , 16.345E0 , + 16.388E0 , 17.159E0 , 17.116E0 , 17.164E0 , 17.123E0 , 17.979E0 , 17.974E0 , 18.007E0 , 17.993E0 , 18.523E0 , 18.669E0 , 18.617E0 , 19.371E0 , 19.330E0 , 0.080E0 , 0.248E0 , 1.089E0 , 1.418E0 , 2.278E0 , 3.624E0 , 4.574E0 , 5.556E0 , 7.267E0 , 7.695E0 , 9.136E0 , 9.959E0 , 9.957E0 , 11.600E0 , 13.138E0 , 13.564E0 , 13.871E0 , 13.994E0 , 14.947E0 , 15.473E0 , 15.379E0 , 15.455E0 , 15.908E0 , 16.114E0 , 17.071E0 , 17.135E0 , 17.282E0 , 17.368E0 , 17.483E0 , 17.764E0 , 18.185E0 , 18.271E0 , 18.236E0 , 18.237E0 , 18.523E0 , 18.627E0 , 18.665E0 , 19.086E0 , 0.214E0 , 0.943E0 , 1.429E0 , 2.241E0 , 2.951E0 , 3.782E0 , 4.757E0 , 5.602E0 , 7.169E0 , 8.920E0 , 10.055E0 , 12.035E0 , 12.861E0 , 13.436E0 , 14.167E0 , 14.755E0 , 15.168E0 , 15.651E0 , 15.746E0 , 16.216E0 , 16.445E0 , 16.965E0 , 17.121E0 , 17.206E0 , 17.250E0 , 17.339E0 , 17.793E0 , 18.123E0 , 18.49E0 , 18.566E0 , 18.645E0 , 18.706E0 , 18.924E0 , 19.1E0 , 0.375E0 , 0.471E0 , 1.504E0 , 2.204E0 , 2.813E0 , 4.765E0 , 9.835E0 , 10.040E0 , 11.946E0 , +12.596E0 , +13.303E0 , 13.922E0 , 14.440E0 , 14.951E0 , 15.627E0 , 15.639E0 , 15.814E0 , 16.315E0 , 16.334E0 , 16.430E0 , 16.423E0 , 17.024E0 , 17.009E0 , 17.165E0 , 17.134E0 , 17.349E0 , 17.576E0 , 17.848E0 , 18.090E0 , 18.276E0 , 18.404E0 , 18.519E0 , 19.133E0 , 19.074E0 , 19.239E0 , 19.280E0 , 19.101E0 , 19.398E0 , 19.252E0 , 19.89E0 , 20.007E0 , 19.929E0 , 19.268E0 , 19.324E0 , 20.049E0 , 20.107E0 , 20.062E0 , 20.065E0 , 19.286E0 , 19.972E0 , 20.088E0 , 20.743E0 , 20.83E0 , 20.935E0 , 21.035E0 , 20.93E0 , 21.074E0 , 21.085E0 , 20.935E0 }; + + // int called=0; printf("call hahn1_functor with iflag=%d, called=%d\n", iflag, called); if (iflag==1) called++; + + assert(b.size()==7); + assert(fvec.size()==236); + for(int i=0; i<236; i++) { + double x=m_x[i], xx=x*x, xxx=xx*x; + fvec[i] = (b[0]+b[1]*x+b[2]*xx+b[3]*xxx) / (1.+b[4]*x+b[5]*xx+b[6]*xxx) - m_y[i]; + } + return 0; + } + + int df(const VectorXd &b, MatrixXd &fjac) + { + assert(b.size()==7); + assert(fjac.rows()==236); + assert(fjac.cols()==7); + for(int i=0; i<236; i++) { + double x=m_x[i], xx=x*x, xxx=xx*x; + double fact = 1./(1.+b[4]*x+b[5]*xx+b[6]*xxx); + fjac(i,0) = 1.*fact; + fjac(i,1) = x*fact; + fjac(i,2) = xx*fact; + fjac(i,3) = xxx*fact; + fact = - (b[0]+b[1]*x+b[2]*xx+b[3]*xxx) * fact * fact; + fjac(i,4) = x*fact; + fjac(i,5) = xx*fact; + fjac(i,6) = xxx*fact; + } + return 0; + } +}; +const double hahn1_functor::m_x[236] = { 24.41E0 , 34.82E0 , 44.09E0 , 45.07E0 , 54.98E0 , 65.51E0 , 70.53E0 , 75.70E0 , 89.57E0 , 91.14E0 , 96.40E0 , 97.19E0 , 114.26E0 , 120.25E0 , 127.08E0 , 133.55E0 , 133.61E0 , 158.67E0 , 172.74E0 , 171.31E0 , 202.14E0 , 220.55E0 , 221.05E0 , 221.39E0 , 250.99E0 , 268.99E0 , 271.80E0 , 271.97E0 , 321.31E0 , 321.69E0 , 330.14E0 , 333.03E0 , 333.47E0 , 340.77E0 , 345.65E0 , 373.11E0 , 373.79E0 , 411.82E0 , 419.51E0 , 421.59E0 , 422.02E0 , 422.47E0 , 422.61E0 , 441.75E0 , 447.41E0 , 448.7E0 , 472.89E0 , 476.69E0 , 522.47E0 , 522.62E0 , 524.43E0 , 546.75E0 , 549.53E0 , 575.29E0 , 576.00E0 , 625.55E0 , 20.15E0 , 28.78E0 , 29.57E0 , 37.41E0 , 39.12E0 , 50.24E0 , 61.38E0 , 66.25E0 , 73.42E0 , 95.52E0 , 107.32E0 , 122.04E0 , 134.03E0 , 163.19E0 , 163.48E0 , 175.70E0 , 179.86E0 , 211.27E0 , 217.78E0 , 219.14E0 , 262.52E0 , 268.01E0 , 268.62E0 , 336.25E0 , 337.23E0 , 339.33E0 , 427.38E0 , 428.58E0 , 432.68E0 , 528.99E0 , 531.08E0 , 628.34E0 , 253.24E0 , 273.13E0 , 273.66E0 , +282.10E0 , 346.62E0 , 347.19E0 , 348.78E0 , 351.18E0 , 450.10E0 , 450.35E0 , 451.92E0 , 455.56E0 , 552.22E0 , 553.56E0 , 555.74E0 , 652.59E0 , 656.20E0 , 14.13E0 , 20.41E0 , 31.30E0 , 33.84E0 , 39.70E0 , 48.83E0 , 54.50E0 , 60.41E0 , 72.77E0 , 75.25E0 , 86.84E0 , 94.88E0 , 96.40E0 , 117.37E0 , 139.08E0 , 147.73E0 , 158.63E0 , 161.84E0 , 192.11E0 , 206.76E0 , 209.07E0 , 213.32E0 , 226.44E0 , 237.12E0 , 330.90E0 , 358.72E0 , 370.77E0 , 372.72E0 , 396.24E0 , 416.59E0 , 484.02E0 , 495.47E0 , 514.78E0 , 515.65E0 , 519.47E0 , 544.47E0 , 560.11E0 , 620.77E0 , 18.97E0 , 28.93E0 , 33.91E0 , 40.03E0 , 44.66E0 , 49.87E0 , 55.16E0 , 60.90E0 , 72.08E0 , 85.15E0 , 97.06E0 , 119.63E0 , 133.27E0 , 143.84E0 , 161.91E0 , 180.67E0 , 198.44E0 , 226.86E0 , 229.65E0 , 258.27E0 , 273.77E0 , 339.15E0 , 350.13E0 , 362.75E0 , 371.03E0 , 393.32E0 , 448.53E0 , 473.78E0 , 511.12E0 , 524.70E0 , 548.75E0 , 551.64E0 , 574.02E0 , 623.86E0 , 21.46E0 , 24.33E0 , 33.43E0 , 39.22E0 , 44.18E0 , 55.02E0 , 94.33E0 , 96.44E0 , 118.82E0 , 128.48E0 , +141.94E0 , 156.92E0 , 171.65E0 , 190.00E0 , 223.26E0 , 223.88E0 , 231.50E0 , 265.05E0 , 269.44E0 , 271.78E0 , 273.46E0 , 334.61E0 , 339.79E0 , 349.52E0 , 358.18E0 , 377.98E0 , 394.77E0 , 429.66E0 , 468.22E0 , 487.27E0 , 519.54E0 , 523.03E0 , 612.99E0 , 638.59E0 , 641.36E0 , 622.05E0 , 631.50E0 , 663.97E0 , 646.9E0 , 748.29E0 , 749.21E0 , 750.14E0 , 647.04E0 , 646.89E0 , 746.9E0 , 748.43E0 , 747.35E0 , 749.27E0 , 647.61E0 , 747.78E0 , 750.51E0 , 851.37E0 , 845.97E0 , 847.54E0 , 849.93E0 , 851.61E0 , 849.75E0 , 850.98E0 , 848.23E0}; + +// http://www.itl.nist.gov/div898/strd/nls/data/hahn1.shtml +void testNistHahn1(void) +{ + const int n=7; + int info; + + VectorXd x(n); + + /* + * First try + */ + x<< 10., -1., .05, -.00001, -.05, .001, -.000001; + // do the computation + hahn1_functor functor; + LevenbergMarquardt<hahn1_functor> lm(functor); + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 11); + VERIFY_IS_EQUAL(lm.njev(), 10); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 1.5324382854E+00); + // check x + VERIFY_IS_APPROX(x[0], 1.0776351733E+00); + VERIFY_IS_APPROX(x[1],-1.2269296921E-01); + VERIFY_IS_APPROX(x[2], 4.0863750610E-03); + VERIFY_IS_APPROX(x[3],-1.426264e-06); // shoulde be : -1.4262662514E-06 + VERIFY_IS_APPROX(x[4],-5.7609940901E-03); + VERIFY_IS_APPROX(x[5], 2.4053735503E-04); + VERIFY_IS_APPROX(x[6],-1.2314450199E-07); + + /* + * Second try + */ + x<< .1, -.1, .005, -.000001, -.005, .0001, -.0000001; + // do the computation + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); +// VERIFY_IS_EQUAL(lm.nfev(), 11); + VERIFY_IS_EQUAL(lm.njev(), 10); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 1.5324382854E+00); + // check x + VERIFY_IS_APPROX(x[0], 1.077640); // should be : 1.0776351733E+00 + VERIFY_IS_APPROX(x[1], -0.1226933); // should be : -1.2269296921E-01 + VERIFY_IS_APPROX(x[2], 0.004086383); // should be : 4.0863750610E-03 + VERIFY_IS_APPROX(x[3], -1.426277e-06); // shoulde be : -1.4262662514E-06 + VERIFY_IS_APPROX(x[4],-5.7609940901E-03); + VERIFY_IS_APPROX(x[5], 0.00024053772); // should be : 2.4053735503E-04 + VERIFY_IS_APPROX(x[6], -1.231450e-07); // should be : -1.2314450199E-07 + +} + +struct misra1d_functor : DenseFunctor<double> +{ + misra1d_functor(void) : DenseFunctor<double>(2,14) {} + static const double x[14]; + static const double y[14]; + int operator()(const VectorXd &b, VectorXd &fvec) + { + assert(b.size()==2); + assert(fvec.size()==14); + for(int i=0; i<14; i++) { + fvec[i] = b[0]*b[1]*x[i]/(1.+b[1]*x[i]) - y[i]; + } + return 0; + } + int df(const VectorXd &b, MatrixXd &fjac) + { + assert(b.size()==2); + assert(fjac.rows()==14); + assert(fjac.cols()==2); + for(int i=0; i<14; i++) { + double den = 1.+b[1]*x[i]; + fjac(i,0) = b[1]*x[i] / den; + fjac(i,1) = b[0]*x[i]*(den-b[1]*x[i])/den/den; + } + return 0; + } +}; +const double misra1d_functor::x[14] = { 77.6E0, 114.9E0, 141.1E0, 190.8E0, 239.9E0, 289.0E0, 332.8E0, 378.4E0, 434.8E0, 477.3E0, 536.8E0, 593.1E0, 689.1E0, 760.0E0}; +const double misra1d_functor::y[14] = { 10.07E0, 14.73E0, 17.94E0, 23.93E0, 29.61E0, 35.18E0, 40.02E0, 44.82E0, 50.76E0, 55.05E0, 61.01E0, 66.40E0, 75.47E0, 81.78E0}; + +// http://www.itl.nist.gov/div898/strd/nls/data/misra1d.shtml +void testNistMisra1d(void) +{ + const int n=2; + int info; + + VectorXd x(n); + + /* + * First try + */ + x<< 500., 0.0001; + // do the computation + misra1d_functor functor; + LevenbergMarquardt<misra1d_functor> lm(functor); + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 9); + VERIFY_IS_EQUAL(lm.njev(), 7); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 5.6419295283E-02); + // check x + VERIFY_IS_APPROX(x[0], 4.3736970754E+02); + VERIFY_IS_APPROX(x[1], 3.0227324449E-04); + + /* + * Second try + */ + x<< 450., 0.0003; + // do the computation + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 4); + VERIFY_IS_EQUAL(lm.njev(), 3); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 5.6419295283E-02); + // check x + VERIFY_IS_APPROX(x[0], 4.3736970754E+02); + VERIFY_IS_APPROX(x[1], 3.0227324449E-04); +} + + +struct lanczos1_functor : DenseFunctor<double> +{ + lanczos1_functor(void) : DenseFunctor<double>(6,24) {} + static const double x[24]; + static const double y[24]; + int operator()(const VectorXd &b, VectorXd &fvec) + { + assert(b.size()==6); + assert(fvec.size()==24); + for(int i=0; i<24; i++) + fvec[i] = b[0]*exp(-b[1]*x[i]) + b[2]*exp(-b[3]*x[i]) + b[4]*exp(-b[5]*x[i]) - y[i]; + return 0; + } + int df(const VectorXd &b, MatrixXd &fjac) + { + assert(b.size()==6); + assert(fjac.rows()==24); + assert(fjac.cols()==6); + for(int i=0; i<24; i++) { + fjac(i,0) = exp(-b[1]*x[i]); + fjac(i,1) = -b[0]*x[i]*exp(-b[1]*x[i]); + fjac(i,2) = exp(-b[3]*x[i]); + fjac(i,3) = -b[2]*x[i]*exp(-b[3]*x[i]); + fjac(i,4) = exp(-b[5]*x[i]); + fjac(i,5) = -b[4]*x[i]*exp(-b[5]*x[i]); + } + return 0; + } +}; +const double lanczos1_functor::x[24] = { 0.000000000000E+00, 5.000000000000E-02, 1.000000000000E-01, 1.500000000000E-01, 2.000000000000E-01, 2.500000000000E-01, 3.000000000000E-01, 3.500000000000E-01, 4.000000000000E-01, 4.500000000000E-01, 5.000000000000E-01, 5.500000000000E-01, 6.000000000000E-01, 6.500000000000E-01, 7.000000000000E-01, 7.500000000000E-01, 8.000000000000E-01, 8.500000000000E-01, 9.000000000000E-01, 9.500000000000E-01, 1.000000000000E+00, 1.050000000000E+00, 1.100000000000E+00, 1.150000000000E+00 }; +const double lanczos1_functor::y[24] = { 2.513400000000E+00 ,2.044333373291E+00 ,1.668404436564E+00 ,1.366418021208E+00 ,1.123232487372E+00 ,9.268897180037E-01 ,7.679338563728E-01 ,6.388775523106E-01 ,5.337835317402E-01 ,4.479363617347E-01 ,3.775847884350E-01 ,3.197393199326E-01 ,2.720130773746E-01 ,2.324965529032E-01 ,1.996589546065E-01 ,1.722704126914E-01 ,1.493405660168E-01 ,1.300700206922E-01 ,1.138119324644E-01 ,1.000415587559E-01 ,8.833209084540E-02 ,7.833544019350E-02 ,6.976693743449E-02 ,6.239312536719E-02 }; + +// http://www.itl.nist.gov/div898/strd/nls/data/lanczos1.shtml +void testNistLanczos1(void) +{ + const int n=6; + int info; + + VectorXd x(n); + + /* + * First try + */ + x<< 1.2, 0.3, 5.6, 5.5, 6.5, 7.6; + // do the computation + lanczos1_functor functor; + LevenbergMarquardt<lanczos1_functor> lm(functor); + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 2); + VERIFY_IS_EQUAL(lm.nfev(), 79); + VERIFY_IS_EQUAL(lm.njev(), 72); + // check norm^2 +// VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 1.430899764097e-25); // should be 1.4307867721E-25, but nist results are on 128-bit floats + // check x + VERIFY_IS_APPROX(x[0], 9.5100000027E-02); + VERIFY_IS_APPROX(x[1], 1.0000000001E+00); + VERIFY_IS_APPROX(x[2], 8.6070000013E-01); + VERIFY_IS_APPROX(x[3], 3.0000000002E+00); + VERIFY_IS_APPROX(x[4], 1.5575999998E+00); + VERIFY_IS_APPROX(x[5], 5.0000000001E+00); + + /* + * Second try + */ + x<< 0.5, 0.7, 3.6, 4.2, 4., 6.3; + // do the computation + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 2); + VERIFY_IS_EQUAL(lm.nfev(), 9); + VERIFY_IS_EQUAL(lm.njev(), 8); + // check norm^2 +// VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 1.428595533845e-25); // should be 1.4307867721E-25, but nist results are on 128-bit floats + // check x + VERIFY_IS_APPROX(x[0], 9.5100000027E-02); + VERIFY_IS_APPROX(x[1], 1.0000000001E+00); + VERIFY_IS_APPROX(x[2], 8.6070000013E-01); + VERIFY_IS_APPROX(x[3], 3.0000000002E+00); + VERIFY_IS_APPROX(x[4], 1.5575999998E+00); + VERIFY_IS_APPROX(x[5], 5.0000000001E+00); + +} + +struct rat42_functor : DenseFunctor<double> +{ + rat42_functor(void) : DenseFunctor<double>(3,9) {} + static const double x[9]; + static const double y[9]; + int operator()(const VectorXd &b, VectorXd &fvec) + { + assert(b.size()==3); + assert(fvec.size()==9); + for(int i=0; i<9; i++) { + fvec[i] = b[0] / (1.+exp(b[1]-b[2]*x[i])) - y[i]; + } + return 0; + } + + int df(const VectorXd &b, MatrixXd &fjac) + { + assert(b.size()==3); + assert(fjac.rows()==9); + assert(fjac.cols()==3); + for(int i=0; i<9; i++) { + double e = exp(b[1]-b[2]*x[i]); + fjac(i,0) = 1./(1.+e); + fjac(i,1) = -b[0]*e/(1.+e)/(1.+e); + fjac(i,2) = +b[0]*e*x[i]/(1.+e)/(1.+e); + } + return 0; + } +}; +const double rat42_functor::x[9] = { 9.000E0, 14.000E0, 21.000E0, 28.000E0, 42.000E0, 57.000E0, 63.000E0, 70.000E0, 79.000E0 }; +const double rat42_functor::y[9] = { 8.930E0 ,10.800E0 ,18.590E0 ,22.330E0 ,39.350E0 ,56.110E0 ,61.730E0 ,64.620E0 ,67.080E0 }; + +// http://www.itl.nist.gov/div898/strd/nls/data/ratkowsky2.shtml +void testNistRat42(void) +{ + const int n=3; + int info; + + VectorXd x(n); + + /* + * First try + */ + x<< 100., 1., 0.1; + // do the computation + rat42_functor functor; + LevenbergMarquardt<rat42_functor> lm(functor); + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 10); + VERIFY_IS_EQUAL(lm.njev(), 8); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 8.0565229338E+00); + // check x + VERIFY_IS_APPROX(x[0], 7.2462237576E+01); + VERIFY_IS_APPROX(x[1], 2.6180768402E+00); + VERIFY_IS_APPROX(x[2], 6.7359200066E-02); + + /* + * Second try + */ + x<< 75., 2.5, 0.07; + // do the computation + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 6); + VERIFY_IS_EQUAL(lm.njev(), 5); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 8.0565229338E+00); + // check x + VERIFY_IS_APPROX(x[0], 7.2462237576E+01); + VERIFY_IS_APPROX(x[1], 2.6180768402E+00); + VERIFY_IS_APPROX(x[2], 6.7359200066E-02); +} + +struct MGH10_functor : DenseFunctor<double> +{ + MGH10_functor(void) : DenseFunctor<double>(3,16) {} + static const double x[16]; + static const double y[16]; + int operator()(const VectorXd &b, VectorXd &fvec) + { + assert(b.size()==3); + assert(fvec.size()==16); + for(int i=0; i<16; i++) + fvec[i] = b[0] * exp(b[1]/(x[i]+b[2])) - y[i]; + return 0; + } + int df(const VectorXd &b, MatrixXd &fjac) + { + assert(b.size()==3); + assert(fjac.rows()==16); + assert(fjac.cols()==3); + for(int i=0; i<16; i++) { + double factor = 1./(x[i]+b[2]); + double e = exp(b[1]*factor); + fjac(i,0) = e; + fjac(i,1) = b[0]*factor*e; + fjac(i,2) = -b[1]*b[0]*factor*factor*e; + } + return 0; + } +}; +const double MGH10_functor::x[16] = { 5.000000E+01, 5.500000E+01, 6.000000E+01, 6.500000E+01, 7.000000E+01, 7.500000E+01, 8.000000E+01, 8.500000E+01, 9.000000E+01, 9.500000E+01, 1.000000E+02, 1.050000E+02, 1.100000E+02, 1.150000E+02, 1.200000E+02, 1.250000E+02 }; +const double MGH10_functor::y[16] = { 3.478000E+04, 2.861000E+04, 2.365000E+04, 1.963000E+04, 1.637000E+04, 1.372000E+04, 1.154000E+04, 9.744000E+03, 8.261000E+03, 7.030000E+03, 6.005000E+03, 5.147000E+03, 4.427000E+03, 3.820000E+03, 3.307000E+03, 2.872000E+03 }; + +// http://www.itl.nist.gov/div898/strd/nls/data/mgh10.shtml +void testNistMGH10(void) +{ + const int n=3; + int info; + + VectorXd x(n); + + /* + * First try + */ + x<< 2., 400000., 25000.; + // do the computation + MGH10_functor functor; + LevenbergMarquardt<MGH10_functor> lm(functor); + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 284 ); + VERIFY_IS_EQUAL(lm.njev(), 249 ); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 8.7945855171E+01); + // check x + VERIFY_IS_APPROX(x[0], 5.6096364710E-03); + VERIFY_IS_APPROX(x[1], 6.1813463463E+03); + VERIFY_IS_APPROX(x[2], 3.4522363462E+02); + + /* + * Second try + */ + x<< 0.02, 4000., 250.; + // do the computation + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 126); + VERIFY_IS_EQUAL(lm.njev(), 116); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 8.7945855171E+01); + // check x + VERIFY_IS_APPROX(x[0], 5.6096364710E-03); + VERIFY_IS_APPROX(x[1], 6.1813463463E+03); + VERIFY_IS_APPROX(x[2], 3.4522363462E+02); +} + + +struct BoxBOD_functor : DenseFunctor<double> +{ + BoxBOD_functor(void) : DenseFunctor<double>(2,6) {} + static const double x[6]; + int operator()(const VectorXd &b, VectorXd &fvec) + { + static const double y[6] = { 109., 149., 149., 191., 213., 224. }; + assert(b.size()==2); + assert(fvec.size()==6); + for(int i=0; i<6; i++) + fvec[i] = b[0]*(1.-exp(-b[1]*x[i])) - y[i]; + return 0; + } + int df(const VectorXd &b, MatrixXd &fjac) + { + assert(b.size()==2); + assert(fjac.rows()==6); + assert(fjac.cols()==2); + for(int i=0; i<6; i++) { + double e = exp(-b[1]*x[i]); + fjac(i,0) = 1.-e; + fjac(i,1) = b[0]*x[i]*e; + } + return 0; + } +}; +const double BoxBOD_functor::x[6] = { 1., 2., 3., 5., 7., 10. }; + +// http://www.itl.nist.gov/div898/strd/nls/data/boxbod.shtml +void testNistBoxBOD(void) +{ + const int n=2; + int info; + + VectorXd x(n); + + /* + * First try + */ + x<< 1., 1.; + // do the computation + BoxBOD_functor functor; + LevenbergMarquardt<BoxBOD_functor> lm(functor); + lm.setFtol(1.E6*NumTraits<double>::epsilon()); + lm.setXtol(1.E6*NumTraits<double>::epsilon()); + lm.setFactor(10); + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 31); + VERIFY_IS_EQUAL(lm.njev(), 25); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 1.1680088766E+03); + // check x + VERIFY_IS_APPROX(x[0], 2.1380940889E+02); + VERIFY_IS_APPROX(x[1], 5.4723748542E-01); + + /* + * Second try + */ + x<< 100., 0.75; + // do the computation + lm.resetParameters(); + lm.setFtol(NumTraits<double>::epsilon()); + lm.setXtol( NumTraits<double>::epsilon()); + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 15 ); + VERIFY_IS_EQUAL(lm.njev(), 14 ); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 1.1680088766E+03); + // check x + VERIFY_IS_APPROX(x[0], 2.1380940889E+02); + VERIFY_IS_APPROX(x[1], 5.4723748542E-01); +} + +struct MGH17_functor : DenseFunctor<double> +{ + MGH17_functor(void) : DenseFunctor<double>(5,33) {} + static const double x[33]; + static const double y[33]; + int operator()(const VectorXd &b, VectorXd &fvec) + { + assert(b.size()==5); + assert(fvec.size()==33); + for(int i=0; i<33; i++) + fvec[i] = b[0] + b[1]*exp(-b[3]*x[i]) + b[2]*exp(-b[4]*x[i]) - y[i]; + return 0; + } + int df(const VectorXd &b, MatrixXd &fjac) + { + assert(b.size()==5); + assert(fjac.rows()==33); + assert(fjac.cols()==5); + for(int i=0; i<33; i++) { + fjac(i,0) = 1.; + fjac(i,1) = exp(-b[3]*x[i]); + fjac(i,2) = exp(-b[4]*x[i]); + fjac(i,3) = -x[i]*b[1]*exp(-b[3]*x[i]); + fjac(i,4) = -x[i]*b[2]*exp(-b[4]*x[i]); + } + return 0; + } +}; +const double MGH17_functor::x[33] = { 0.000000E+00, 1.000000E+01, 2.000000E+01, 3.000000E+01, 4.000000E+01, 5.000000E+01, 6.000000E+01, 7.000000E+01, 8.000000E+01, 9.000000E+01, 1.000000E+02, 1.100000E+02, 1.200000E+02, 1.300000E+02, 1.400000E+02, 1.500000E+02, 1.600000E+02, 1.700000E+02, 1.800000E+02, 1.900000E+02, 2.000000E+02, 2.100000E+02, 2.200000E+02, 2.300000E+02, 2.400000E+02, 2.500000E+02, 2.600000E+02, 2.700000E+02, 2.800000E+02, 2.900000E+02, 3.000000E+02, 3.100000E+02, 3.200000E+02 }; +const double MGH17_functor::y[33] = { 8.440000E-01, 9.080000E-01, 9.320000E-01, 9.360000E-01, 9.250000E-01, 9.080000E-01, 8.810000E-01, 8.500000E-01, 8.180000E-01, 7.840000E-01, 7.510000E-01, 7.180000E-01, 6.850000E-01, 6.580000E-01, 6.280000E-01, 6.030000E-01, 5.800000E-01, 5.580000E-01, 5.380000E-01, 5.220000E-01, 5.060000E-01, 4.900000E-01, 4.780000E-01, 4.670000E-01, 4.570000E-01, 4.480000E-01, 4.380000E-01, 4.310000E-01, 4.240000E-01, 4.200000E-01, 4.140000E-01, 4.110000E-01, 4.060000E-01 }; + +// http://www.itl.nist.gov/div898/strd/nls/data/mgh17.shtml +void testNistMGH17(void) +{ + const int n=5; + int info; + + VectorXd x(n); + + /* + * First try + */ + x<< 50., 150., -100., 1., 2.; + // do the computation + MGH17_functor functor; + LevenbergMarquardt<MGH17_functor> lm(functor); + lm.setFtol(NumTraits<double>::epsilon()); + lm.setXtol(NumTraits<double>::epsilon()); + lm.setMaxfev(1000); + info = lm.minimize(x); + + // check return value +// VERIFY_IS_EQUAL(info, 2); //FIXME Use (lm.info() == Success) +// VERIFY_IS_EQUAL(lm.nfev(), 602 ); + VERIFY_IS_EQUAL(lm.njev(), 545 ); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 5.4648946975E-05); + // check x + VERIFY_IS_APPROX(x[0], 3.7541005211E-01); + VERIFY_IS_APPROX(x[1], 1.9358469127E+00); + VERIFY_IS_APPROX(x[2], -1.4646871366E+00); + VERIFY_IS_APPROX(x[3], 1.2867534640E-02); + VERIFY_IS_APPROX(x[4], 2.2122699662E-02); + + /* + * Second try + */ + x<< 0.5 ,1.5 ,-1 ,0.01 ,0.02; + // do the computation + lm.resetParameters(); + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 18); + VERIFY_IS_EQUAL(lm.njev(), 15); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 5.4648946975E-05); + // check x + VERIFY_IS_APPROX(x[0], 3.7541005211E-01); + VERIFY_IS_APPROX(x[1], 1.9358469127E+00); + VERIFY_IS_APPROX(x[2], -1.4646871366E+00); + VERIFY_IS_APPROX(x[3], 1.2867534640E-02); + VERIFY_IS_APPROX(x[4], 2.2122699662E-02); +} + +struct MGH09_functor : DenseFunctor<double> +{ + MGH09_functor(void) : DenseFunctor<double>(4,11) {} + static const double _x[11]; + static const double y[11]; + int operator()(const VectorXd &b, VectorXd &fvec) + { + assert(b.size()==4); + assert(fvec.size()==11); + for(int i=0; i<11; i++) { + double x = _x[i], xx=x*x; + fvec[i] = b[0]*(xx+x*b[1])/(xx+x*b[2]+b[3]) - y[i]; + } + return 0; + } + int df(const VectorXd &b, MatrixXd &fjac) + { + assert(b.size()==4); + assert(fjac.rows()==11); + assert(fjac.cols()==4); + for(int i=0; i<11; i++) { + double x = _x[i], xx=x*x; + double factor = 1./(xx+x*b[2]+b[3]); + fjac(i,0) = (xx+x*b[1]) * factor; + fjac(i,1) = b[0]*x* factor; + fjac(i,2) = - b[0]*(xx+x*b[1]) * x * factor * factor; + fjac(i,3) = - b[0]*(xx+x*b[1]) * factor * factor; + } + return 0; + } +}; +const double MGH09_functor::_x[11] = { 4., 2., 1., 5.E-1 , 2.5E-01, 1.670000E-01, 1.250000E-01, 1.E-01, 8.330000E-02, 7.140000E-02, 6.250000E-02 }; +const double MGH09_functor::y[11] = { 1.957000E-01, 1.947000E-01, 1.735000E-01, 1.600000E-01, 8.440000E-02, 6.270000E-02, 4.560000E-02, 3.420000E-02, 3.230000E-02, 2.350000E-02, 2.460000E-02 }; + +// http://www.itl.nist.gov/div898/strd/nls/data/mgh09.shtml +void testNistMGH09(void) +{ + const int n=4; + int info; + + VectorXd x(n); + + /* + * First try + */ + x<< 25., 39, 41.5, 39.; + // do the computation + MGH09_functor functor; + LevenbergMarquardt<MGH09_functor> lm(functor); + lm.setMaxfev(1000); + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 490 ); + VERIFY_IS_EQUAL(lm.njev(), 376 ); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 3.0750560385E-04); + // check x + VERIFY_IS_APPROX(x[0], 0.1928077089); // should be 1.9280693458E-01 + VERIFY_IS_APPROX(x[1], 0.19126423573); // should be 1.9128232873E-01 + VERIFY_IS_APPROX(x[2], 0.12305309914); // should be 1.2305650693E-01 + VERIFY_IS_APPROX(x[3], 0.13605395375); // should be 1.3606233068E-01 + + /* + * Second try + */ + x<< 0.25, 0.39, 0.415, 0.39; + // do the computation + lm.resetParameters(); + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 18); + VERIFY_IS_EQUAL(lm.njev(), 16); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 3.0750560385E-04); + // check x + VERIFY_IS_APPROX(x[0], 0.19280781); // should be 1.9280693458E-01 + VERIFY_IS_APPROX(x[1], 0.19126265); // should be 1.9128232873E-01 + VERIFY_IS_APPROX(x[2], 0.12305280); // should be 1.2305650693E-01 + VERIFY_IS_APPROX(x[3], 0.13605322); // should be 1.3606233068E-01 +} + + + +struct Bennett5_functor : DenseFunctor<double> +{ + Bennett5_functor(void) : DenseFunctor<double>(3,154) {} + static const double x[154]; + static const double y[154]; + int operator()(const VectorXd &b, VectorXd &fvec) + { + assert(b.size()==3); + assert(fvec.size()==154); + for(int i=0; i<154; i++) + fvec[i] = b[0]* pow(b[1]+x[i],-1./b[2]) - y[i]; + return 0; + } + int df(const VectorXd &b, MatrixXd &fjac) + { + assert(b.size()==3); + assert(fjac.rows()==154); + assert(fjac.cols()==3); + for(int i=0; i<154; i++) { + double e = pow(b[1]+x[i],-1./b[2]); + fjac(i,0) = e; + fjac(i,1) = - b[0]*e/b[2]/(b[1]+x[i]); + fjac(i,2) = b[0]*e*log(b[1]+x[i])/b[2]/b[2]; + } + return 0; + } +}; +const double Bennett5_functor::x[154] = { 7.447168E0, 8.102586E0, 8.452547E0, 8.711278E0, 8.916774E0, 9.087155E0, 9.232590E0, 9.359535E0, 9.472166E0, 9.573384E0, 9.665293E0, 9.749461E0, 9.827092E0, 9.899128E0, 9.966321E0, 10.029280E0, 10.088510E0, 10.144430E0, 10.197380E0, 10.247670E0, 10.295560E0, 10.341250E0, 10.384950E0, 10.426820E0, 10.467000E0, 10.505640E0, 10.542830E0, 10.578690E0, 10.613310E0, 10.646780E0, 10.679150E0, 10.710520E0, 10.740920E0, 10.770440E0, 10.799100E0, 10.826970E0, 10.854080E0, 10.880470E0, 10.906190E0, 10.931260E0, 10.955720E0, 10.979590E0, 11.002910E0, 11.025700E0, 11.047980E0, 11.069770E0, 11.091100E0, 11.111980E0, 11.132440E0, 11.152480E0, 11.172130E0, 11.191410E0, 11.210310E0, 11.228870E0, 11.247090E0, 11.264980E0, 11.282560E0, 11.299840E0, 11.316820E0, 11.333520E0, 11.349940E0, 11.366100E0, 11.382000E0, 11.397660E0, 11.413070E0, 11.428240E0, 11.443200E0, 11.457930E0, 11.472440E0, 11.486750E0, 11.500860E0, 11.514770E0, 11.528490E0, 11.542020E0, 11.555380E0, 11.568550E0, +11.581560E0, 11.594420E0, 11.607121E0, 11.619640E0, 11.632000E0, 11.644210E0, 11.656280E0, 11.668200E0, 11.679980E0, 11.691620E0, 11.703130E0, 11.714510E0, 11.725760E0, 11.736880E0, 11.747890E0, 11.758780E0, 11.769550E0, 11.780200E0, 11.790730E0, 11.801160E0, 11.811480E0, 11.821700E0, 11.831810E0, 11.841820E0, 11.851730E0, 11.861550E0, 11.871270E0, 11.880890E0, 11.890420E0, 11.899870E0, 11.909220E0, 11.918490E0, 11.927680E0, 11.936780E0, 11.945790E0, 11.954730E0, 11.963590E0, 11.972370E0, 11.981070E0, 11.989700E0, 11.998260E0, 12.006740E0, 12.015150E0, 12.023490E0, 12.031760E0, 12.039970E0, 12.048100E0, 12.056170E0, 12.064180E0, 12.072120E0, 12.080010E0, 12.087820E0, 12.095580E0, 12.103280E0, 12.110920E0, 12.118500E0, 12.126030E0, 12.133500E0, 12.140910E0, 12.148270E0, 12.155570E0, 12.162830E0, 12.170030E0, 12.177170E0, 12.184270E0, 12.191320E0, 12.198320E0, 12.205270E0, 12.212170E0, 12.219030E0, 12.225840E0, 12.232600E0, 12.239320E0, 12.245990E0, 12.252620E0, 12.259200E0, 12.265750E0, 12.272240E0 }; +const double Bennett5_functor::y[154] = { -34.834702E0 ,-34.393200E0 ,-34.152901E0 ,-33.979099E0 ,-33.845901E0 ,-33.732899E0 ,-33.640301E0 ,-33.559200E0 ,-33.486801E0 ,-33.423100E0 ,-33.365101E0 ,-33.313000E0 ,-33.260899E0 ,-33.217400E0 ,-33.176899E0 ,-33.139198E0 ,-33.101601E0 ,-33.066799E0 ,-33.035000E0 ,-33.003101E0 ,-32.971298E0 ,-32.942299E0 ,-32.916302E0 ,-32.890202E0 ,-32.864101E0 ,-32.841000E0 ,-32.817799E0 ,-32.797501E0 ,-32.774300E0 ,-32.757000E0 ,-32.733799E0 ,-32.716400E0 ,-32.699100E0 ,-32.678799E0 ,-32.661400E0 ,-32.644001E0 ,-32.626701E0 ,-32.612202E0 ,-32.597698E0 ,-32.583199E0 ,-32.568699E0 ,-32.554298E0 ,-32.539799E0 ,-32.525299E0 ,-32.510799E0 ,-32.499199E0 ,-32.487598E0 ,-32.473202E0 ,-32.461601E0 ,-32.435501E0 ,-32.435501E0 ,-32.426800E0 ,-32.412300E0 ,-32.400799E0 ,-32.392101E0 ,-32.380501E0 ,-32.366001E0 ,-32.357300E0 ,-32.348598E0 ,-32.339901E0 ,-32.328400E0 ,-32.319698E0 ,-32.311001E0 ,-32.299400E0 ,-32.290699E0 ,-32.282001E0 ,-32.273300E0 ,-32.264599E0 ,-32.256001E0 ,-32.247299E0 +,-32.238602E0 ,-32.229900E0 ,-32.224098E0 ,-32.215401E0 ,-32.203800E0 ,-32.198002E0 ,-32.189400E0 ,-32.183601E0 ,-32.174900E0 ,-32.169102E0 ,-32.163300E0 ,-32.154598E0 ,-32.145901E0 ,-32.140099E0 ,-32.131401E0 ,-32.125599E0 ,-32.119801E0 ,-32.111198E0 ,-32.105400E0 ,-32.096699E0 ,-32.090900E0 ,-32.088001E0 ,-32.079300E0 ,-32.073502E0 ,-32.067699E0 ,-32.061901E0 ,-32.056099E0 ,-32.050301E0 ,-32.044498E0 ,-32.038799E0 ,-32.033001E0 ,-32.027199E0 ,-32.024300E0 ,-32.018501E0 ,-32.012699E0 ,-32.004002E0 ,-32.001099E0 ,-31.995300E0 ,-31.989500E0 ,-31.983700E0 ,-31.977900E0 ,-31.972099E0 ,-31.969299E0 ,-31.963501E0 ,-31.957701E0 ,-31.951900E0 ,-31.946100E0 ,-31.940300E0 ,-31.937401E0 ,-31.931601E0 ,-31.925800E0 ,-31.922899E0 ,-31.917101E0 ,-31.911301E0 ,-31.908400E0 ,-31.902599E0 ,-31.896900E0 ,-31.893999E0 ,-31.888201E0 ,-31.885300E0 ,-31.882401E0 ,-31.876600E0 ,-31.873699E0 ,-31.867901E0 ,-31.862101E0 ,-31.859200E0 ,-31.856300E0 ,-31.850500E0 ,-31.844700E0 ,-31.841801E0 ,-31.838900E0 ,-31.833099E0 ,-31.830200E0 , +-31.827299E0 ,-31.821600E0 ,-31.818701E0 ,-31.812901E0 ,-31.809999E0 ,-31.807100E0 ,-31.801300E0 ,-31.798401E0 ,-31.795500E0 ,-31.789700E0 ,-31.786800E0 }; + +// http://www.itl.nist.gov/div898/strd/nls/data/bennett5.shtml +void testNistBennett5(void) +{ + const int n=3; + int info; + + VectorXd x(n); + + /* + * First try + */ + x<< -2000., 50., 0.8; + // do the computation + Bennett5_functor functor; + LevenbergMarquardt<Bennett5_functor> lm(functor); + lm.setMaxfev(1000); + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 758); + VERIFY_IS_EQUAL(lm.njev(), 744); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 5.2404744073E-04); + // check x + VERIFY_IS_APPROX(x[0], -2.5235058043E+03); + VERIFY_IS_APPROX(x[1], 4.6736564644E+01); + VERIFY_IS_APPROX(x[2], 9.3218483193E-01); + /* + * Second try + */ + x<< -1500., 45., 0.85; + // do the computation + lm.resetParameters(); + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 203); + VERIFY_IS_EQUAL(lm.njev(), 192); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 5.2404744073E-04); + // check x + VERIFY_IS_APPROX(x[0], -2523.3007865); // should be -2.5235058043E+03 + VERIFY_IS_APPROX(x[1], 46.735705771); // should be 4.6736564644E+01); + VERIFY_IS_APPROX(x[2], 0.93219881891); // should be 9.3218483193E-01); +} + +struct thurber_functor : DenseFunctor<double> +{ + thurber_functor(void) : DenseFunctor<double>(7,37) {} + static const double _x[37]; + static const double _y[37]; + int operator()(const VectorXd &b, VectorXd &fvec) + { + // int called=0; printf("call hahn1_functor with iflag=%d, called=%d\n", iflag, called); if (iflag==1) called++; + assert(b.size()==7); + assert(fvec.size()==37); + for(int i=0; i<37; i++) { + double x=_x[i], xx=x*x, xxx=xx*x; + fvec[i] = (b[0]+b[1]*x+b[2]*xx+b[3]*xxx) / (1.+b[4]*x+b[5]*xx+b[6]*xxx) - _y[i]; + } + return 0; + } + int df(const VectorXd &b, MatrixXd &fjac) + { + assert(b.size()==7); + assert(fjac.rows()==37); + assert(fjac.cols()==7); + for(int i=0; i<37; i++) { + double x=_x[i], xx=x*x, xxx=xx*x; + double fact = 1./(1.+b[4]*x+b[5]*xx+b[6]*xxx); + fjac(i,0) = 1.*fact; + fjac(i,1) = x*fact; + fjac(i,2) = xx*fact; + fjac(i,3) = xxx*fact; + fact = - (b[0]+b[1]*x+b[2]*xx+b[3]*xxx) * fact * fact; + fjac(i,4) = x*fact; + fjac(i,5) = xx*fact; + fjac(i,6) = xxx*fact; + } + return 0; + } +}; +const double thurber_functor::_x[37] = { -3.067E0, -2.981E0, -2.921E0, -2.912E0, -2.840E0, -2.797E0, -2.702E0, -2.699E0, -2.633E0, -2.481E0, -2.363E0, -2.322E0, -1.501E0, -1.460E0, -1.274E0, -1.212E0, -1.100E0, -1.046E0, -0.915E0, -0.714E0, -0.566E0, -0.545E0, -0.400E0, -0.309E0, -0.109E0, -0.103E0, 0.010E0, 0.119E0, 0.377E0, 0.790E0, 0.963E0, 1.006E0, 1.115E0, 1.572E0, 1.841E0, 2.047E0, 2.200E0 }; +const double thurber_functor::_y[37] = { 80.574E0, 84.248E0, 87.264E0, 87.195E0, 89.076E0, 89.608E0, 89.868E0, 90.101E0, 92.405E0, 95.854E0, 100.696E0, 101.060E0, 401.672E0, 390.724E0, 567.534E0, 635.316E0, 733.054E0, 759.087E0, 894.206E0, 990.785E0, 1090.109E0, 1080.914E0, 1122.643E0, 1178.351E0, 1260.531E0, 1273.514E0, 1288.339E0, 1327.543E0, 1353.863E0, 1414.509E0, 1425.208E0, 1421.384E0, 1442.962E0, 1464.350E0, 1468.705E0, 1447.894E0, 1457.628E0}; + +// http://www.itl.nist.gov/div898/strd/nls/data/thurber.shtml +void testNistThurber(void) +{ + const int n=7; + int info; + + VectorXd x(n); + + /* + * First try + */ + x<< 1000 ,1000 ,400 ,40 ,0.7,0.3,0.0 ; + // do the computation + thurber_functor functor; + LevenbergMarquardt<thurber_functor> lm(functor); + lm.setFtol(1.E4*NumTraits<double>::epsilon()); + lm.setXtol(1.E4*NumTraits<double>::epsilon()); + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 39); + VERIFY_IS_EQUAL(lm.njev(), 36); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 5.6427082397E+03); + // check x + VERIFY_IS_APPROX(x[0], 1.2881396800E+03); + VERIFY_IS_APPROX(x[1], 1.4910792535E+03); + VERIFY_IS_APPROX(x[2], 5.8323836877E+02); + VERIFY_IS_APPROX(x[3], 7.5416644291E+01); + VERIFY_IS_APPROX(x[4], 9.6629502864E-01); + VERIFY_IS_APPROX(x[5], 3.9797285797E-01); + VERIFY_IS_APPROX(x[6], 4.9727297349E-02); + + /* + * Second try + */ + x<< 1300 ,1500 ,500 ,75 ,1 ,0.4 ,0.05 ; + // do the computation + lm.resetParameters(); + lm.setFtol(1.E4*NumTraits<double>::epsilon()); + lm.setXtol(1.E4*NumTraits<double>::epsilon()); + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 29); + VERIFY_IS_EQUAL(lm.njev(), 28); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 5.6427082397E+03); + // check x + VERIFY_IS_APPROX(x[0], 1.2881396800E+03); + VERIFY_IS_APPROX(x[1], 1.4910792535E+03); + VERIFY_IS_APPROX(x[2], 5.8323836877E+02); + VERIFY_IS_APPROX(x[3], 7.5416644291E+01); + VERIFY_IS_APPROX(x[4], 9.6629502864E-01); + VERIFY_IS_APPROX(x[5], 3.9797285797E-01); + VERIFY_IS_APPROX(x[6], 4.9727297349E-02); +} + +struct rat43_functor : DenseFunctor<double> +{ + rat43_functor(void) : DenseFunctor<double>(4,15) {} + static const double x[15]; + static const double y[15]; + int operator()(const VectorXd &b, VectorXd &fvec) + { + assert(b.size()==4); + assert(fvec.size()==15); + for(int i=0; i<15; i++) + fvec[i] = b[0] * pow(1.+exp(b[1]-b[2]*x[i]),-1./b[3]) - y[i]; + return 0; + } + int df(const VectorXd &b, MatrixXd &fjac) + { + assert(b.size()==4); + assert(fjac.rows()==15); + assert(fjac.cols()==4); + for(int i=0; i<15; i++) { + double e = exp(b[1]-b[2]*x[i]); + double power = -1./b[3]; + fjac(i,0) = pow(1.+e, power); + fjac(i,1) = power*b[0]*e*pow(1.+e, power-1.); + fjac(i,2) = -power*b[0]*e*x[i]*pow(1.+e, power-1.); + fjac(i,3) = b[0]*power*power*log(1.+e)*pow(1.+e, power); + } + return 0; + } +}; +const double rat43_functor::x[15] = { 1., 2., 3., 4., 5., 6., 7., 8., 9., 10., 11., 12., 13., 14., 15. }; +const double rat43_functor::y[15] = { 16.08, 33.83, 65.80, 97.20, 191.55, 326.20, 386.87, 520.53, 590.03, 651.92, 724.93, 699.56, 689.96, 637.56, 717.41 }; + +// http://www.itl.nist.gov/div898/strd/nls/data/ratkowsky3.shtml +void testNistRat43(void) +{ + const int n=4; + int info; + + VectorXd x(n); + + /* + * First try + */ + x<< 100., 10., 1., 1.; + // do the computation + rat43_functor functor; + LevenbergMarquardt<rat43_functor> lm(functor); + lm.setFtol(1.E6*NumTraits<double>::epsilon()); + lm.setXtol(1.E6*NumTraits<double>::epsilon()); + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 27); + VERIFY_IS_EQUAL(lm.njev(), 20); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 8.7864049080E+03); + // check x + VERIFY_IS_APPROX(x[0], 6.9964151270E+02); + VERIFY_IS_APPROX(x[1], 5.2771253025E+00); + VERIFY_IS_APPROX(x[2], 7.5962938329E-01); + VERIFY_IS_APPROX(x[3], 1.2792483859E+00); + + /* + * Second try + */ + x<< 700., 5., 0.75, 1.3; + // do the computation + lm.resetParameters(); + lm.setFtol(1.E5*NumTraits<double>::epsilon()); + lm.setXtol(1.E5*NumTraits<double>::epsilon()); + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 9); + VERIFY_IS_EQUAL(lm.njev(), 8); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 8.7864049080E+03); + // check x + VERIFY_IS_APPROX(x[0], 6.9964151270E+02); + VERIFY_IS_APPROX(x[1], 5.2771253025E+00); + VERIFY_IS_APPROX(x[2], 7.5962938329E-01); + VERIFY_IS_APPROX(x[3], 1.2792483859E+00); +} + + + +struct eckerle4_functor : DenseFunctor<double> +{ + eckerle4_functor(void) : DenseFunctor<double>(3,35) {} + static const double x[35]; + static const double y[35]; + int operator()(const VectorXd &b, VectorXd &fvec) + { + assert(b.size()==3); + assert(fvec.size()==35); + for(int i=0; i<35; i++) + fvec[i] = b[0]/b[1] * exp(-0.5*(x[i]-b[2])*(x[i]-b[2])/(b[1]*b[1])) - y[i]; + return 0; + } + int df(const VectorXd &b, MatrixXd &fjac) + { + assert(b.size()==3); + assert(fjac.rows()==35); + assert(fjac.cols()==3); + for(int i=0; i<35; i++) { + double b12 = b[1]*b[1]; + double e = exp(-0.5*(x[i]-b[2])*(x[i]-b[2])/b12); + fjac(i,0) = e / b[1]; + fjac(i,1) = ((x[i]-b[2])*(x[i]-b[2])/b12-1.) * b[0]*e/b12; + fjac(i,2) = (x[i]-b[2])*e*b[0]/b[1]/b12; + } + return 0; + } +}; +const double eckerle4_functor::x[35] = { 400.0, 405.0, 410.0, 415.0, 420.0, 425.0, 430.0, 435.0, 436.5, 438.0, 439.5, 441.0, 442.5, 444.0, 445.5, 447.0, 448.5, 450.0, 451.5, 453.0, 454.5, 456.0, 457.5, 459.0, 460.5, 462.0, 463.5, 465.0, 470.0, 475.0, 480.0, 485.0, 490.0, 495.0, 500.0}; +const double eckerle4_functor::y[35] = { 0.0001575, 0.0001699, 0.0002350, 0.0003102, 0.0004917, 0.0008710, 0.0017418, 0.0046400, 0.0065895, 0.0097302, 0.0149002, 0.0237310, 0.0401683, 0.0712559, 0.1264458, 0.2073413, 0.2902366, 0.3445623, 0.3698049, 0.3668534, 0.3106727, 0.2078154, 0.1164354, 0.0616764, 0.0337200, 0.0194023, 0.0117831, 0.0074357, 0.0022732, 0.0008800, 0.0004579, 0.0002345, 0.0001586, 0.0001143, 0.0000710 }; + +// http://www.itl.nist.gov/div898/strd/nls/data/eckerle4.shtml +void testNistEckerle4(void) +{ + const int n=3; + int info; + + VectorXd x(n); + + /* + * First try + */ + x<< 1., 10., 500.; + // do the computation + eckerle4_functor functor; + LevenbergMarquardt<eckerle4_functor> lm(functor); + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 18); + VERIFY_IS_EQUAL(lm.njev(), 15); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 1.4635887487E-03); + // check x + VERIFY_IS_APPROX(x[0], 1.5543827178); + VERIFY_IS_APPROX(x[1], 4.0888321754); + VERIFY_IS_APPROX(x[2], 4.5154121844E+02); + + /* + * Second try + */ + x<< 1.5, 5., 450.; + // do the computation + info = lm.minimize(x); + + // check return value + VERIFY_IS_EQUAL(info, 1); + VERIFY_IS_EQUAL(lm.nfev(), 7); + VERIFY_IS_EQUAL(lm.njev(), 6); + // check norm^2 + VERIFY_IS_APPROX(lm.fvec().squaredNorm(), 1.4635887487E-03); + // check x + VERIFY_IS_APPROX(x[0], 1.5543827178); + VERIFY_IS_APPROX(x[1], 4.0888321754); + VERIFY_IS_APPROX(x[2], 4.5154121844E+02); +} + +void test_levenberg_marquardt() +{ + // Tests using the examples provided by (c)minpack + CALL_SUBTEST(testLmder1()); + CALL_SUBTEST(testLmder()); + CALL_SUBTEST(testLmdif1()); +// CALL_SUBTEST(testLmstr1()); +// CALL_SUBTEST(testLmstr()); + CALL_SUBTEST(testLmdif()); + + // NIST tests, level of difficulty = "Lower" + CALL_SUBTEST(testNistMisra1a()); + CALL_SUBTEST(testNistChwirut2()); + + // NIST tests, level of difficulty = "Average" + CALL_SUBTEST(testNistHahn1()); + CALL_SUBTEST(testNistMisra1d()); + CALL_SUBTEST(testNistMGH17()); + CALL_SUBTEST(testNistLanczos1()); + +// // NIST tests, level of difficulty = "Higher" + CALL_SUBTEST(testNistRat42()); + CALL_SUBTEST(testNistMGH10()); + CALL_SUBTEST(testNistBoxBOD()); +// CALL_SUBTEST(testNistMGH09()); + CALL_SUBTEST(testNistBennett5()); + CALL_SUBTEST(testNistThurber()); + CALL_SUBTEST(testNistRat43()); + CALL_SUBTEST(testNistEckerle4()); +} diff --git a/unsupported/test/matrix_exponential.cpp b/unsupported/test/matrix_exponential.cpp index 695472f91..50dec083d 100644 --- a/unsupported/test/matrix_exponential.cpp +++ b/unsupported/test/matrix_exponential.cpp @@ -7,8 +7,7 @@ // Public License v. 2.0. If a copy of the MPL was not distributed // with this file, You can obtain one at http://mozilla.org/MPL/2.0/. -#include "main.h" -#include <unsupported/Eigen/MatrixFunctions> +#include "matrix_functions.h" double binom(int n, int k) { @@ -18,12 +17,6 @@ double binom(int n, int k) return res; } -template <typename Derived, typename OtherDerived> -double relerr(const MatrixBase<Derived>& A, const MatrixBase<OtherDerived>& B) -{ - return std::sqrt((A - B).cwiseAbs2().sum() / (std::min)(A.cwiseAbs2().sum(), B.cwiseAbs2().sum())); -} - template <typename T> T expfn(T x, int) { @@ -109,8 +102,7 @@ void randomTest(const MatrixType& m, double tol) */ typename MatrixType::Index rows = m.rows(); typename MatrixType::Index cols = m.cols(); - MatrixType m1(rows, cols), m2(rows, cols), m3(rows, cols), - identity = MatrixType::Identity(rows, rows); + MatrixType m1(rows, cols), m2(rows, cols), identity = MatrixType::Identity(rows, cols); typedef typename NumTraits<typename internal::traits<MatrixType>::Scalar>::Real RealScalar; diff --git a/unsupported/test/matrix_function.cpp b/unsupported/test/matrix_function.cpp index 0439c5a7d..3c76cfb65 100644 --- a/unsupported/test/matrix_function.cpp +++ b/unsupported/test/matrix_function.cpp @@ -110,7 +110,6 @@ void testMatrixLogarithm(const MatrixType& A) { typedef typename internal::traits<MatrixType>::Scalar Scalar; typedef typename NumTraits<Scalar>::Real RealScalar; - typedef std::complex<RealScalar> ComplexScalar; MatrixType scaledA; RealScalar maxImagPartOfSpectrum = A.eigenvalues().imag().cwiseAbs().maxCoeff(); diff --git a/unsupported/test/matrix_functions.h b/unsupported/test/matrix_functions.h new file mode 100644 index 000000000..5817caef6 --- /dev/null +++ b/unsupported/test/matrix_functions.h @@ -0,0 +1,47 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2009-2011 Jitse Niesen <jitse@maths.leeds.ac.uk> +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + +#include "main.h" +#include <unsupported/Eigen/MatrixFunctions> + +template <typename MatrixType, int IsComplex = NumTraits<typename internal::traits<MatrixType>::Scalar>::IsComplex> +struct generateTestMatrix; + +// for real matrices, make sure none of the eigenvalues are negative +template <typename MatrixType> +struct generateTestMatrix<MatrixType,0> +{ + static void run(MatrixType& result, typename MatrixType::Index size) + { + MatrixType mat = MatrixType::Random(size, size); + EigenSolver<MatrixType> es(mat); + typename EigenSolver<MatrixType>::EigenvalueType eivals = es.eigenvalues(); + for (typename MatrixType::Index i = 0; i < size; ++i) { + if (eivals(i).imag() == 0 && eivals(i).real() < 0) + eivals(i) = -eivals(i); + } + result = (es.eigenvectors() * eivals.asDiagonal() * es.eigenvectors().inverse()).real(); + } +}; + +// for complex matrices, any matrix is fine +template <typename MatrixType> +struct generateTestMatrix<MatrixType,1> +{ + static void run(MatrixType& result, typename MatrixType::Index size) + { + result = MatrixType::Random(size, size); + } +}; + +template <typename Derived, typename OtherDerived> +double relerr(const MatrixBase<Derived>& A, const MatrixBase<OtherDerived>& B) +{ + return std::sqrt((A - B).cwiseAbs2().sum() / (std::min)(A.cwiseAbs2().sum(), B.cwiseAbs2().sum())); +} diff --git a/unsupported/test/matrix_power.cpp b/unsupported/test/matrix_power.cpp new file mode 100644 index 000000000..b9d513b45 --- /dev/null +++ b/unsupported/test/matrix_power.cpp @@ -0,0 +1,133 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2012 Chen-Pang He <jdh8@ms63.hinet.net> +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + +#include "matrix_functions.h" + +template <typename MatrixType, int IsComplex = NumTraits<typename MatrixType::Scalar>::IsComplex> +struct generateTriangularMatrix; + +// for real matrices, make sure none of the eigenvalues are negative +template <typename MatrixType> +struct generateTriangularMatrix<MatrixType,0> +{ + static void run(MatrixType& result, typename MatrixType::Index size) + { + result.resize(size, size); + result.template triangularView<Upper>() = MatrixType::Random(size, size); + for (typename MatrixType::Index i = 0; i < size; ++i) + result.coeffRef(i,i) = std::abs(result.coeff(i,i)); + } +}; + +// for complex matrices, any matrix is fine +template <typename MatrixType> +struct generateTriangularMatrix<MatrixType,1> +{ + static void run(MatrixType& result, typename MatrixType::Index size) + { + result.resize(size, size); + result.template triangularView<Upper>() = MatrixType::Random(size, size); + } +}; + +template<typename T> +void test2dRotation(double tol) +{ + Matrix<T,2,2> A, B, C; + T angle, c, s; + + A << 0, 1, -1, 0; + MatrixPower<Matrix<T,2,2> > Apow(A); + + for (int i=0; i<=20; ++i) { + angle = pow(10, (i-10) / 5.); + c = std::cos(angle); + s = std::sin(angle); + B << c, s, -s, c; + + C = Apow(std::ldexp(angle,1) / M_PI); + std::cout << "test2dRotation: i = " << i << " error powerm = " << relerr(C,B) << '\n'; + VERIFY(C.isApprox(B, static_cast<T>(tol))); + } +} + +template<typename T> +void test2dHyperbolicRotation(double tol) +{ + Matrix<std::complex<T>,2,2> A, B, C; + T angle, ch = std::cosh((T)1); + std::complex<T> ish(0, std::sinh((T)1)); + + A << ch, ish, -ish, ch; + MatrixPower<Matrix<std::complex<T>,2,2> > Apow(A); + + for (int i=0; i<=20; ++i) { + angle = std::ldexp(static_cast<T>(i-10), -1); + ch = std::cosh(angle); + ish = std::complex<T>(0, std::sinh(angle)); + B << ch, ish, -ish, ch; + + C = Apow(angle); + std::cout << "test2dHyperbolicRotation: i = " << i << " error powerm = " << relerr(C,B) << '\n'; + VERIFY(C.isApprox(B, static_cast<T>(tol))); + } +} + +template<typename MatrixType> +void testExponentLaws(const MatrixType& m, double tol) +{ + typedef typename MatrixType::RealScalar RealScalar; + MatrixType m1, m2, m3, m4, m5; + RealScalar x, y; + + for (int i=0; i < g_repeat; ++i) { + generateTestMatrix<MatrixType>::run(m1, m.rows()); + MatrixPower<MatrixType> mpow(m1); + + x = internal::random<RealScalar>(); + y = internal::random<RealScalar>(); + m2 = mpow(x); + m3 = mpow(y); + + m4 = mpow(x+y); + m5.noalias() = m2 * m3; + VERIFY(m4.isApprox(m5, static_cast<RealScalar>(tol))); + + m4 = mpow(x*y); + m5 = m2.pow(y); + VERIFY(m4.isApprox(m5, static_cast<RealScalar>(tol))); + + m4 = (std::abs(x) * m1).pow(y); + m5 = std::pow(std::abs(x), y) * m3; + VERIFY(m4.isApprox(m5, static_cast<RealScalar>(tol))); + } +} + +typedef Matrix<double,3,3,RowMajor> Matrix3dRowMajor; +typedef Matrix<long double,Dynamic,Dynamic> MatrixXe; + +void test_matrix_power() +{ + CALL_SUBTEST_2(test2dRotation<double>(1e-13)); + CALL_SUBTEST_1(test2dRotation<float>(2e-5)); // was 1e-5, relaxed for clang 2.8 / linux / x86-64 + CALL_SUBTEST_9(test2dRotation<long double>(1e-13)); + CALL_SUBTEST_2(test2dHyperbolicRotation<double>(1e-14)); + CALL_SUBTEST_1(test2dHyperbolicRotation<float>(1e-5)); + CALL_SUBTEST_9(test2dHyperbolicRotation<long double>(1e-14)); + + CALL_SUBTEST_2(testExponentLaws(Matrix2d(), 1e-13)); + CALL_SUBTEST_7(testExponentLaws(Matrix3dRowMajor(), 1e-13)); + CALL_SUBTEST_3(testExponentLaws(Matrix4cd(), 1e-13)); + CALL_SUBTEST_4(testExponentLaws(MatrixXd(8,8), 2e-12)); + CALL_SUBTEST_1(testExponentLaws(Matrix2f(), 1e-4)); + CALL_SUBTEST_5(testExponentLaws(Matrix3cf(), 1e-4)); + CALL_SUBTEST_8(testExponentLaws(Matrix4f(), 1e-4)); + CALL_SUBTEST_6(testExponentLaws(MatrixXf(2,2), 1e-3)); // see bug 614 + CALL_SUBTEST_9(testExponentLaws(MatrixXe(7,7), 1e-13)); +} diff --git a/unsupported/test/matrix_square_root.cpp b/unsupported/test/matrix_square_root.cpp index 508619a7a..ea541e1ea 100644 --- a/unsupported/test/matrix_square_root.cpp +++ b/unsupported/test/matrix_square_root.cpp @@ -7,38 +7,7 @@ // Public License v. 2.0. If a copy of the MPL was not distributed // with this file, You can obtain one at http://mozilla.org/MPL/2.0/. -#include "main.h" -#include <unsupported/Eigen/MatrixFunctions> - -template <typename MatrixType, int IsComplex = NumTraits<typename internal::traits<MatrixType>::Scalar>::IsComplex> -struct generateTestMatrix; - -// for real matrices, make sure none of the eigenvalues are negative -template <typename MatrixType> -struct generateTestMatrix<MatrixType,0> -{ - static void run(MatrixType& result, typename MatrixType::Index size) - { - MatrixType mat = MatrixType::Random(size, size); - EigenSolver<MatrixType> es(mat); - typename EigenSolver<MatrixType>::EigenvalueType eivals = es.eigenvalues(); - for (typename MatrixType::Index i = 0; i < size; ++i) { - if (eivals(i).imag() == 0 && eivals(i).real() < 0) - eivals(i) = -eivals(i); - } - result = (es.eigenvectors() * eivals.asDiagonal() * es.eigenvectors().inverse()).real(); - } -}; - -// for complex matrices, any matrix is fine -template <typename MatrixType> -struct generateTestMatrix<MatrixType,1> -{ - static void run(MatrixType& result, typename MatrixType::Index size) - { - result = MatrixType::Random(size, size); - } -}; +#include "matrix_functions.h" template<typename MatrixType> void testMatrixSqrt(const MatrixType& m) diff --git a/unsupported/test/minres.cpp b/unsupported/test/minres.cpp new file mode 100644 index 000000000..fd12da548 --- /dev/null +++ b/unsupported/test/minres.cpp @@ -0,0 +1,32 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2011 Gael Guennebaud <g.gael@free.fr> +// Copyright (C) 2012 Giacomo Po <gpo@ucla.edu> +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. +#include <cmath> + +#include "../../test/sparse_solver.h" +#include <Eigen/IterativeSolvers> + +template<typename T> void test_minres_T() +{ + MINRES<SparseMatrix<T>, Lower, DiagonalPreconditioner<T> > minres_colmajor_diag; + MINRES<SparseMatrix<T>, Lower, IdentityPreconditioner > minres_colmajor_I; +// MINRES<SparseMatrix<T>, Lower, IncompleteLUT<T> > minres_colmajor_ilut; + //minres<SparseMatrix<T>, SSORPreconditioner<T> > minres_colmajor_ssor; + + CALL_SUBTEST( check_sparse_square_solving(minres_colmajor_diag) ); + CALL_SUBTEST( check_sparse_spd_solving(minres_colmajor_I) ); + // CALL_SUBTEST( check_sparse_square_solving(minres_colmajor_ilut) ); + //CALL_SUBTEST( check_sparse_square_solving(minres_colmajor_ssor) ); +} + +void test_minres() +{ + CALL_SUBTEST_1(test_minres_T<double>()); +// CALL_SUBTEST_2(test_minres_T<std::complex<double> >()); +} diff --git a/unsupported/test/mpreal/dlmalloc.c b/unsupported/test/mpreal/dlmalloc.c deleted file mode 100755 index 7ce8feb07..000000000 --- a/unsupported/test/mpreal/dlmalloc.c +++ /dev/null @@ -1,5703 +0,0 @@ -/*
- This is a version (aka dlmalloc) of malloc/free/realloc written by
- Doug Lea and released to the public domain, as explained at
- http://creativecommons.org/licenses/publicdomain. Send questions,
- comments, complaints, performance data, etc to dl@cs.oswego.edu
-
-* Version 2.8.4 Wed May 27 09:56:23 2009 Doug Lea (dl at gee)
-
- Note: There may be an updated version of this malloc obtainable at
- ftp://gee.cs.oswego.edu/pub/misc/malloc.c
- Check before installing!
-
-* Quickstart
-
- This library is all in one file to simplify the most common usage:
- ftp it, compile it (-O3), and link it into another program. All of
- the compile-time options default to reasonable values for use on
- most platforms. You might later want to step through various
- compile-time and dynamic tuning options.
-
- For convenience, an include file for code using this malloc is at:
- ftp://gee.cs.oswego.edu/pub/misc/malloc-2.8.4.h
- You don't really need this .h file unless you call functions not
- defined in your system include files. The .h file contains only the
- excerpts from this file needed for using this malloc on ANSI C/C++
- systems, so long as you haven't changed compile-time options about
- naming and tuning parameters. If you do, then you can create your
- own malloc.h that does include all settings by cutting at the point
- indicated below. Note that you may already by default be using a C
- library containing a malloc that is based on some version of this
- malloc (for example in linux). You might still want to use the one
- in this file to customize settings or to avoid overheads associated
- with library versions.
-
-* Vital statistics:
-
- Supported pointer/size_t representation: 4 or 8 bytes
- size_t MUST be an unsigned type of the same width as
- pointers. (If you are using an ancient system that declares
- size_t as a signed type, or need it to be a different width
- than pointers, you can use a previous release of this malloc
- (e.g. 2.7.2) supporting these.)
-
- Alignment: 8 bytes (default)
- This suffices for nearly all current machines and C compilers.
- However, you can define MALLOC_ALIGNMENT to be wider than this
- if necessary (up to 128bytes), at the expense of using more space.
-
- Minimum overhead per allocated chunk: 4 or 8 bytes (if 4byte sizes)
- 8 or 16 bytes (if 8byte sizes)
- Each malloced chunk has a hidden word of overhead holding size
- and status information, and additional cross-check word
- if FOOTERS is defined.
-
- Minimum allocated size: 4-byte ptrs: 16 bytes (including overhead)
- 8-byte ptrs: 32 bytes (including overhead)
-
- Even a request for zero bytes (i.e., malloc(0)) returns a
- pointer to something of the minimum allocatable size.
- The maximum overhead wastage (i.e., number of extra bytes
- allocated than were requested in malloc) is less than or equal
- to the minimum size, except for requests >= mmap_threshold that
- are serviced via mmap(), where the worst case wastage is about
- 32 bytes plus the remainder from a system page (the minimal
- mmap unit); typically 4096 or 8192 bytes.
-
- Security: static-safe; optionally more or less
- The "security" of malloc refers to the ability of malicious
- code to accentuate the effects of errors (for example, freeing
- space that is not currently malloc'ed or overwriting past the
- ends of chunks) in code that calls malloc. This malloc
- guarantees not to modify any memory locations below the base of
- heap, i.e., static variables, even in the presence of usage
- errors. The routines additionally detect most improper frees
- and reallocs. All this holds as long as the static bookkeeping
- for malloc itself is not corrupted by some other means. This
- is only one aspect of security -- these checks do not, and
- cannot, detect all possible programming errors.
-
- If FOOTERS is defined nonzero, then each allocated chunk
- carries an additional check word to verify that it was malloced
- from its space. These check words are the same within each
- execution of a program using malloc, but differ across
- executions, so externally crafted fake chunks cannot be
- freed. This improves security by rejecting frees/reallocs that
- could corrupt heap memory, in addition to the checks preventing
- writes to statics that are always on. This may further improve
- security at the expense of time and space overhead. (Note that
- FOOTERS may also be worth using with MSPACES.)
-
- By default detected errors cause the program to abort (calling
- "abort()"). You can override this to instead proceed past
- errors by defining PROCEED_ON_ERROR. In this case, a bad free
- has no effect, and a malloc that encounters a bad address
- caused by user overwrites will ignore the bad address by
- dropping pointers and indices to all known memory. This may
- be appropriate for programs that should continue if at all
- possible in the face of programming errors, although they may
- run out of memory because dropped memory is never reclaimed.
-
- If you don't like either of these options, you can define
- CORRUPTION_ERROR_ACTION and USAGE_ERROR_ACTION to do anything
- else. And if if you are sure that your program using malloc has
- no errors or vulnerabilities, you can define INSECURE to 1,
- which might (or might not) provide a small performance improvement.
-
- Thread-safety: NOT thread-safe unless USE_LOCKS defined
- When USE_LOCKS is defined, each public call to malloc, free,
- etc is surrounded with either a pthread mutex or a win32
- spinlock (depending on WIN32). This is not especially fast, and
- can be a major bottleneck. It is designed only to provide
- minimal protection in concurrent environments, and to provide a
- basis for extensions. If you are using malloc in a concurrent
- program, consider instead using nedmalloc
- (http://www.nedprod.com/programs/portable/nedmalloc/) or
- ptmalloc (See http://www.malloc.de), which are derived
- from versions of this malloc.
-
- System requirements: Any combination of MORECORE and/or MMAP/MUNMAP
- This malloc can use unix sbrk or any emulation (invoked using
- the CALL_MORECORE macro) and/or mmap/munmap or any emulation
- (invoked using CALL_MMAP/CALL_MUNMAP) to get and release system
- memory. On most unix systems, it tends to work best if both
- MORECORE and MMAP are enabled. On Win32, it uses emulations
- based on VirtualAlloc. It also uses common C library functions
- like memset.
-
- Compliance: I believe it is compliant with the Single Unix Specification
- (See http://www.unix.org). Also SVID/XPG, ANSI C, and probably
- others as well.
-
-* Overview of algorithms
-
- This is not the fastest, most space-conserving, most portable, or
- most tunable malloc ever written. However it is among the fastest
- while also being among the most space-conserving, portable and
- tunable. Consistent balance across these factors results in a good
- general-purpose allocator for malloc-intensive programs.
-
- In most ways, this malloc is a best-fit allocator. Generally, it
- chooses the best-fitting existing chunk for a request, with ties
- broken in approximately least-recently-used order. (This strategy
- normally maintains low fragmentation.) However, for requests less
- than 256bytes, it deviates from best-fit when there is not an
- exactly fitting available chunk by preferring to use space adjacent
- to that used for the previous small request, as well as by breaking
- ties in approximately most-recently-used order. (These enhance
- locality of series of small allocations.) And for very large requests
- (>= 256Kb by default), it relies on system memory mapping
- facilities, if supported. (This helps avoid carrying around and
- possibly fragmenting memory used only for large chunks.)
-
- All operations (except malloc_stats and mallinfo) have execution
- times that are bounded by a constant factor of the number of bits in
- a size_t, not counting any clearing in calloc or copying in realloc,
- or actions surrounding MORECORE and MMAP that have times
- proportional to the number of non-contiguous regions returned by
- system allocation routines, which is often just 1. In real-time
- applications, you can optionally suppress segment traversals using
- NO_SEGMENT_TRAVERSAL, which assures bounded execution even when
- system allocators return non-contiguous spaces, at the typical
- expense of carrying around more memory and increased fragmentation.
-
- The implementation is not very modular and seriously overuses
- macros. Perhaps someday all C compilers will do as good a job
- inlining modular code as can now be done by brute-force expansion,
- but now, enough of them seem not to.
-
- Some compilers issue a lot of warnings about code that is
- dead/unreachable only on some platforms, and also about intentional
- uses of negation on unsigned types. All known cases of each can be
- ignored.
-
- For a longer but out of date high-level description, see
- http://gee.cs.oswego.edu/dl/html/malloc.html
-
-* MSPACES
- If MSPACES is defined, then in addition to malloc, free, etc.,
- this file also defines mspace_malloc, mspace_free, etc. These
- are versions of malloc routines that take an "mspace" argument
- obtained using create_mspace, to control all internal bookkeeping.
- If ONLY_MSPACES is defined, only these versions are compiled.
- So if you would like to use this allocator for only some allocations,
- and your system malloc for others, you can compile with
- ONLY_MSPACES and then do something like...
- static mspace mymspace = create_mspace(0,0); // for example
- #define mymalloc(bytes) mspace_malloc(mymspace, bytes)
-
- (Note: If you only need one instance of an mspace, you can instead
- use "USE_DL_PREFIX" to relabel the global malloc.)
-
- You can similarly create thread-local allocators by storing
- mspaces as thread-locals. For example:
- static __thread mspace tlms = 0;
- void* tlmalloc(size_t bytes) {
- if (tlms == 0) tlms = create_mspace(0, 0);
- return mspace_malloc(tlms, bytes);
- }
- void tlfree(void* mem) { mspace_free(tlms, mem); }
-
- Unless FOOTERS is defined, each mspace is completely independent.
- You cannot allocate from one and free to another (although
- conformance is only weakly checked, so usage errors are not always
- caught). If FOOTERS is defined, then each chunk carries around a tag
- indicating its originating mspace, and frees are directed to their
- originating spaces.
-
- ------------------------- Compile-time options ---------------------------
-
-Be careful in setting #define values for numerical constants of type
-size_t. On some systems, literal values are not automatically extended
-to size_t precision unless they are explicitly casted. You can also
-use the symbolic values MAX_SIZE_T, SIZE_T_ONE, etc below.
-
-WIN32 default: defined if _WIN32 defined
- Defining WIN32 sets up defaults for MS environment and compilers.
- Otherwise defaults are for unix. Beware that there seem to be some
- cases where this malloc might not be a pure drop-in replacement for
- Win32 malloc: Random-looking failures from Win32 GDI API's (eg;
- SetDIBits()) may be due to bugs in some video driver implementations
- when pixel buffers are malloc()ed, and the region spans more than
- one VirtualAlloc()ed region. Because dlmalloc uses a small (64Kb)
- default granularity, pixel buffers may straddle virtual allocation
- regions more often than when using the Microsoft allocator. You can
- avoid this by using VirtualAlloc() and VirtualFree() for all pixel
- buffers rather than using malloc(). If this is not possible,
- recompile this malloc with a larger DEFAULT_GRANULARITY.
-
-MALLOC_ALIGNMENT default: (size_t)8
- Controls the minimum alignment for malloc'ed chunks. It must be a
- power of two and at least 8, even on machines for which smaller
- alignments would suffice. It may be defined as larger than this
- though. Note however that code and data structures are optimized for
- the case of 8-byte alignment.
-
-MSPACES default: 0 (false)
- If true, compile in support for independent allocation spaces.
- This is only supported if HAVE_MMAP is true.
-
-ONLY_MSPACES default: 0 (false)
- If true, only compile in mspace versions, not regular versions.
-
-USE_LOCKS default: 0 (false)
- Causes each call to each public routine to be surrounded with
- pthread or WIN32 mutex lock/unlock. (If set true, this can be
- overridden on a per-mspace basis for mspace versions.) If set to a
- non-zero value other than 1, locks are used, but their
- implementation is left out, so lock functions must be supplied manually,
- as described below.
-
-USE_SPIN_LOCKS default: 1 iff USE_LOCKS and on x86 using gcc or MSC
- If true, uses custom spin locks for locking. This is currently
- supported only for x86 platforms using gcc or recent MS compilers.
- Otherwise, posix locks or win32 critical sections are used.
-
-FOOTERS default: 0
- If true, provide extra checking and dispatching by placing
- information in the footers of allocated chunks. This adds
- space and time overhead.
-
-INSECURE default: 0
- If true, omit checks for usage errors and heap space overwrites.
-
-USE_DL_PREFIX default: NOT defined
- Causes compiler to prefix all public routines with the string 'dl'.
- This can be useful when you only want to use this malloc in one part
- of a program, using your regular system malloc elsewhere.
-
-ABORT default: defined as abort()
- Defines how to abort on failed checks. On most systems, a failed
- check cannot die with an "assert" or even print an informative
- message, because the underlying print routines in turn call malloc,
- which will fail again. Generally, the best policy is to simply call
- abort(). It's not very useful to do more than this because many
- errors due to overwriting will show up as address faults (null, odd
- addresses etc) rather than malloc-triggered checks, so will also
- abort. Also, most compilers know that abort() does not return, so
- can better optimize code conditionally calling it.
-
-PROCEED_ON_ERROR default: defined as 0 (false)
- Controls whether detected bad addresses cause them to bypassed
- rather than aborting. If set, detected bad arguments to free and
- realloc are ignored. And all bookkeeping information is zeroed out
- upon a detected overwrite of freed heap space, thus losing the
- ability to ever return it from malloc again, but enabling the
- application to proceed. If PROCEED_ON_ERROR is defined, the
- static variable malloc_corruption_error_count is compiled in
- and can be examined to see if errors have occurred. This option
- generates slower code than the default abort policy.
-
-DEBUG default: NOT defined
- The DEBUG setting is mainly intended for people trying to modify
- this code or diagnose problems when porting to new platforms.
- However, it may also be able to better isolate user errors than just
- using runtime checks. The assertions in the check routines spell
- out in more detail the assumptions and invariants underlying the
- algorithms. The checking is fairly extensive, and will slow down
- execution noticeably. Calling malloc_stats or mallinfo with DEBUG
- set will attempt to check every non-mmapped allocated and free chunk
- in the course of computing the summaries.
-
-ABORT_ON_ASSERT_FAILURE default: defined as 1 (true)
- Debugging assertion failures can be nearly impossible if your
- version of the assert macro causes malloc to be called, which will
- lead to a cascade of further failures, blowing the runtime stack.
- ABORT_ON_ASSERT_FAILURE cause assertions failures to call abort(),
- which will usually make debugging easier.
-
-MALLOC_FAILURE_ACTION default: sets errno to ENOMEM, or no-op on win32
- The action to take before "return 0" when malloc fails to be able to
- return memory because there is none available.
-
-HAVE_MORECORE default: 1 (true) unless win32 or ONLY_MSPACES
- True if this system supports sbrk or an emulation of it.
-
-MORECORE default: sbrk
- The name of the sbrk-style system routine to call to obtain more
- memory. See below for guidance on writing custom MORECORE
- functions. The type of the argument to sbrk/MORECORE varies across
- systems. It cannot be size_t, because it supports negative
- arguments, so it is normally the signed type of the same width as
- size_t (sometimes declared as "intptr_t"). It doesn't much matter
- though. Internally, we only call it with arguments less than half
- the max value of a size_t, which should work across all reasonable
- possibilities, although sometimes generating compiler warnings.
-
-MORECORE_CONTIGUOUS default: 1 (true) if HAVE_MORECORE
- If true, take advantage of fact that consecutive calls to MORECORE
- with positive arguments always return contiguous increasing
- addresses. This is true of unix sbrk. It does not hurt too much to
- set it true anyway, since malloc copes with non-contiguities.
- Setting it false when definitely non-contiguous saves time
- and possibly wasted space it would take to discover this though.
-
-MORECORE_CANNOT_TRIM default: NOT defined
- True if MORECORE cannot release space back to the system when given
- negative arguments. This is generally necessary only if you are
- using a hand-crafted MORECORE function that cannot handle negative
- arguments.
-
-NO_SEGMENT_TRAVERSAL default: 0
- If non-zero, suppresses traversals of memory segments
- returned by either MORECORE or CALL_MMAP. This disables
- merging of segments that are contiguous, and selectively
- releasing them to the OS if unused, but bounds execution times.
-
-HAVE_MMAP default: 1 (true)
- True if this system supports mmap or an emulation of it. If so, and
- HAVE_MORECORE is not true, MMAP is used for all system
- allocation. If set and HAVE_MORECORE is true as well, MMAP is
- primarily used to directly allocate very large blocks. It is also
- used as a backup strategy in cases where MORECORE fails to provide
- space from system. Note: A single call to MUNMAP is assumed to be
- able to unmap memory that may have be allocated using multiple calls
- to MMAP, so long as they are adjacent.
-
-HAVE_MREMAP default: 1 on linux, else 0
- If true realloc() uses mremap() to re-allocate large blocks and
- extend or shrink allocation spaces.
-
-MMAP_CLEARS default: 1 except on WINCE.
- True if mmap clears memory so calloc doesn't need to. This is true
- for standard unix mmap using /dev/zero and on WIN32 except for WINCE.
-
-USE_BUILTIN_FFS default: 0 (i.e., not used)
- Causes malloc to use the builtin ffs() function to compute indices.
- Some compilers may recognize and intrinsify ffs to be faster than the
- supplied C version. Also, the case of x86 using gcc is special-cased
- to an asm instruction, so is already as fast as it can be, and so
- this setting has no effect. Similarly for Win32 under recent MS compilers.
- (On most x86s, the asm version is only slightly faster than the C version.)
-
-malloc_getpagesize default: derive from system includes, or 4096.
- The system page size. To the extent possible, this malloc manages
- memory from the system in page-size units. This may be (and
- usually is) a function rather than a constant. This is ignored
- if WIN32, where page size is determined using getSystemInfo during
- initialization.
-
-USE_DEV_RANDOM default: 0 (i.e., not used)
- Causes malloc to use /dev/random to initialize secure magic seed for
- stamping footers. Otherwise, the current time is used.
-
-NO_MALLINFO default: 0
- If defined, don't compile "mallinfo". This can be a simple way
- of dealing with mismatches between system declarations and
- those in this file.
-
-MALLINFO_FIELD_TYPE default: size_t
- The type of the fields in the mallinfo struct. This was originally
- defined as "int" in SVID etc, but is more usefully defined as
- size_t. The value is used only if HAVE_USR_INCLUDE_MALLOC_H is not set
-
-REALLOC_ZERO_BYTES_FREES default: not defined
- This should be set if a call to realloc with zero bytes should
- be the same as a call to free. Some people think it should. Otherwise,
- since this malloc returns a unique pointer for malloc(0), so does
- realloc(p, 0).
-
-LACKS_UNISTD_H, LACKS_FCNTL_H, LACKS_SYS_PARAM_H, LACKS_SYS_MMAN_H
-LACKS_STRINGS_H, LACKS_STRING_H, LACKS_SYS_TYPES_H, LACKS_ERRNO_H
-LACKS_STDLIB_H default: NOT defined unless on WIN32
- Define these if your system does not have these header files.
- You might need to manually insert some of the declarations they provide.
-
-DEFAULT_GRANULARITY default: page size if MORECORE_CONTIGUOUS,
- system_info.dwAllocationGranularity in WIN32,
- otherwise 64K.
- Also settable using mallopt(M_GRANULARITY, x)
- The unit for allocating and deallocating memory from the system. On
- most systems with contiguous MORECORE, there is no reason to
- make this more than a page. However, systems with MMAP tend to
- either require or encourage larger granularities. You can increase
- this value to prevent system allocation functions to be called so
- often, especially if they are slow. The value must be at least one
- page and must be a power of two. Setting to 0 causes initialization
- to either page size or win32 region size. (Note: In previous
- versions of malloc, the equivalent of this option was called
- "TOP_PAD")
-
-DEFAULT_TRIM_THRESHOLD default: 2MB
- Also settable using mallopt(M_TRIM_THRESHOLD, x)
- The maximum amount of unused top-most memory to keep before
- releasing via malloc_trim in free(). Automatic trimming is mainly
- useful in long-lived programs using contiguous MORECORE. Because
- trimming via sbrk can be slow on some systems, and can sometimes be
- wasteful (in cases where programs immediately afterward allocate
- more large chunks) the value should be high enough so that your
- overall system performance would improve by releasing this much
- memory. As a rough guide, you might set to a value close to the
- average size of a process (program) running on your system.
- Releasing this much memory would allow such a process to run in
- memory. Generally, it is worth tuning trim thresholds when a
- program undergoes phases where several large chunks are allocated
- and released in ways that can reuse each other's storage, perhaps
- mixed with phases where there are no such chunks at all. The trim
- value must be greater than page size to have any useful effect. To
- disable trimming completely, you can set to MAX_SIZE_T. Note that the trick
- some people use of mallocing a huge space and then freeing it at
- program startup, in an attempt to reserve system memory, doesn't
- have the intended effect under automatic trimming, since that memory
- will immediately be returned to the system.
-
-DEFAULT_MMAP_THRESHOLD default: 256K
- Also settable using mallopt(M_MMAP_THRESHOLD, x)
- The request size threshold for using MMAP to directly service a
- request. Requests of at least this size that cannot be allocated
- using already-existing space will be serviced via mmap. (If enough
- normal freed space already exists it is used instead.) Using mmap
- segregates relatively large chunks of memory so that they can be
- individually obtained and released from the host system. A request
- serviced through mmap is never reused by any other request (at least
- not directly; the system may just so happen to remap successive
- requests to the same locations). Segregating space in this way has
- the benefits that: Mmapped space can always be individually released
- back to the system, which helps keep the system level memory demands
- of a long-lived program low. Also, mapped memory doesn't become
- `locked' between other chunks, as can happen with normally allocated
- chunks, which means that even trimming via malloc_trim would not
- release them. However, it has the disadvantage that the space
- cannot be reclaimed, consolidated, and then used to service later
- requests, as happens with normal chunks. The advantages of mmap
- nearly always outweigh disadvantages for "large" chunks, but the
- value of "large" may vary across systems. The default is an
- empirically derived value that works well in most systems. You can
- disable mmap by setting to MAX_SIZE_T.
-
-MAX_RELEASE_CHECK_RATE default: 4095 unless not HAVE_MMAP
- The number of consolidated frees between checks to release
- unused segments when freeing. When using non-contiguous segments,
- especially with multiple mspaces, checking only for topmost space
- doesn't always suffice to trigger trimming. To compensate for this,
- free() will, with a period of MAX_RELEASE_CHECK_RATE (or the
- current number of segments, if greater) try to release unused
- segments to the OS when freeing chunks that result in
- consolidation. The best value for this parameter is a compromise
- between slowing down frees with relatively costly checks that
- rarely trigger versus holding on to unused memory. To effectively
- disable, set to MAX_SIZE_T. This may lead to a very slight speed
- improvement at the expense of carrying around more memory.
-*/
-
-#define USE_DL_PREFIX
-//#define HAVE_USR_INCLUDE_MALLOC_H
-//#define MSPACES 1
-#define NO_SEGMENT_TRAVERSAL 1
-
-/* Version identifier to allow people to support multiple versions */
-#ifndef DLMALLOC_VERSION
-#define DLMALLOC_VERSION 20804
-#endif /* DLMALLOC_VERSION */
-
-#ifndef WIN32
-#ifdef _WIN32
-#define WIN32 1
-#endif /* _WIN32 */
-#ifdef _WIN32_WCE
-#define LACKS_FCNTL_H
-#define WIN32 1
-#endif /* _WIN32_WCE */
-#endif /* WIN32 */
-#ifdef WIN32
-#define WIN32_LEAN_AND_MEAN
-#include <windows.h>
-#define HAVE_MMAP 1
-#define HAVE_MORECORE 0
-#define LACKS_UNISTD_H
-#define LACKS_SYS_PARAM_H
-#define LACKS_SYS_MMAN_H
-#define LACKS_STRING_H
-#define LACKS_STRINGS_H
-#define LACKS_SYS_TYPES_H
-#define LACKS_ERRNO_H
-#ifndef MALLOC_FAILURE_ACTION
-#define MALLOC_FAILURE_ACTION
-#endif /* MALLOC_FAILURE_ACTION */
-#ifdef _WIN32_WCE /* WINCE reportedly does not clear */
-#define MMAP_CLEARS 0
-#else
-#define MMAP_CLEARS 1
-#endif /* _WIN32_WCE */
-#endif /* WIN32 */
-
-#if defined(DARWIN) || defined(_DARWIN)
-/* Mac OSX docs advise not to use sbrk; it seems better to use mmap */
-#ifndef HAVE_MORECORE
-#define HAVE_MORECORE 0
-#define HAVE_MMAP 1
-/* OSX allocators provide 16 byte alignment */
-#ifndef MALLOC_ALIGNMENT
-#define MALLOC_ALIGNMENT ((size_t)16U)
-#endif
-#endif /* HAVE_MORECORE */
-#endif /* DARWIN */
-
-#ifndef LACKS_SYS_TYPES_H
-#include <sys/types.h> /* For size_t */
-#endif /* LACKS_SYS_TYPES_H */
-
-#if (defined(__GNUC__) && ((defined(__i386__) || defined(__x86_64__)))) || (defined(_MSC_VER) && _MSC_VER>=1310)
-#define SPIN_LOCKS_AVAILABLE 1
-#else
-#define SPIN_LOCKS_AVAILABLE 0
-#endif
-
-/* The maximum possible size_t value has all bits set */
-#define MAX_SIZE_T (~(size_t)0)
-
-#ifndef ONLY_MSPACES
-#define ONLY_MSPACES 0 /* define to a value */
-#else
-#define ONLY_MSPACES 1
-#endif /* ONLY_MSPACES */
-#ifndef MSPACES
-#if ONLY_MSPACES
-#define MSPACES 1
-#else /* ONLY_MSPACES */
-#define MSPACES 0
-#endif /* ONLY_MSPACES */
-#endif /* MSPACES */
-#ifndef MALLOC_ALIGNMENT
-#define MALLOC_ALIGNMENT ((size_t)8U)
-#endif /* MALLOC_ALIGNMENT */
-#ifndef FOOTERS
-#define FOOTERS 0
-#endif /* FOOTERS */
-#ifndef ABORT
-#define ABORT abort()
-#endif /* ABORT */
-#ifndef ABORT_ON_ASSERT_FAILURE
-#define ABORT_ON_ASSERT_FAILURE 1
-#endif /* ABORT_ON_ASSERT_FAILURE */
-#ifndef PROCEED_ON_ERROR
-#define PROCEED_ON_ERROR 0
-#endif /* PROCEED_ON_ERROR */
-#ifndef USE_LOCKS
-#define USE_LOCKS 0
-#endif /* USE_LOCKS */
-#ifndef USE_SPIN_LOCKS
-#if USE_LOCKS && SPIN_LOCKS_AVAILABLE
-#define USE_SPIN_LOCKS 1
-#else
-#define USE_SPIN_LOCKS 0
-#endif /* USE_LOCKS && SPIN_LOCKS_AVAILABLE. */
-#endif /* USE_SPIN_LOCKS */
-#ifndef INSECURE
-#define INSECURE 0
-#endif /* INSECURE */
-#ifndef HAVE_MMAP
-#define HAVE_MMAP 1
-#endif /* HAVE_MMAP */
-#ifndef MMAP_CLEARS
-#define MMAP_CLEARS 1
-#endif /* MMAP_CLEARS */
-#ifndef HAVE_MREMAP
-#ifdef linux
-#define HAVE_MREMAP 1
-#else /* linux */
-#define HAVE_MREMAP 0
-#endif /* linux */
-#endif /* HAVE_MREMAP */
-#ifndef MALLOC_FAILURE_ACTION
-#define MALLOC_FAILURE_ACTION errno = ENOMEM;
-#endif /* MALLOC_FAILURE_ACTION */
-#ifndef HAVE_MORECORE
-#if ONLY_MSPACES
-#define HAVE_MORECORE 0
-#else /* ONLY_MSPACES */
-#define HAVE_MORECORE 1
-#endif /* ONLY_MSPACES */
-#endif /* HAVE_MORECORE */
-#if !HAVE_MORECORE
-#define MORECORE_CONTIGUOUS 0
-#else /* !HAVE_MORECORE */
-#define MORECORE_DEFAULT sbrk
-#ifndef MORECORE_CONTIGUOUS
-#define MORECORE_CONTIGUOUS 1
-#endif /* MORECORE_CONTIGUOUS */
-#endif /* HAVE_MORECORE */
-#ifndef DEFAULT_GRANULARITY
-#if (MORECORE_CONTIGUOUS || defined(WIN32))
-#define DEFAULT_GRANULARITY (0) /* 0 means to compute in init_mparams */
-#else /* MORECORE_CONTIGUOUS */
-#define DEFAULT_GRANULARITY ((size_t)64U * (size_t)1024U)
-#endif /* MORECORE_CONTIGUOUS */
-#endif /* DEFAULT_GRANULARITY */
-#ifndef DEFAULT_TRIM_THRESHOLD
-#ifndef MORECORE_CANNOT_TRIM
-#define DEFAULT_TRIM_THRESHOLD ((size_t)2U * (size_t)1024U * (size_t)1024U)
-#else /* MORECORE_CANNOT_TRIM */
-#define DEFAULT_TRIM_THRESHOLD MAX_SIZE_T
-#endif /* MORECORE_CANNOT_TRIM */
-#endif /* DEFAULT_TRIM_THRESHOLD */
-#ifndef DEFAULT_MMAP_THRESHOLD
-#if HAVE_MMAP
-#define DEFAULT_MMAP_THRESHOLD ((size_t)256U * (size_t)1024U)
-#else /* HAVE_MMAP */
-#define DEFAULT_MMAP_THRESHOLD MAX_SIZE_T
-#endif /* HAVE_MMAP */
-#endif /* DEFAULT_MMAP_THRESHOLD */
-#ifndef MAX_RELEASE_CHECK_RATE
-#if HAVE_MMAP
-#define MAX_RELEASE_CHECK_RATE 4095
-#else
-#define MAX_RELEASE_CHECK_RATE MAX_SIZE_T
-#endif /* HAVE_MMAP */
-#endif /* MAX_RELEASE_CHECK_RATE */
-#ifndef USE_BUILTIN_FFS
-#define USE_BUILTIN_FFS 0
-#endif /* USE_BUILTIN_FFS */
-#ifndef USE_DEV_RANDOM
-#define USE_DEV_RANDOM 0
-#endif /* USE_DEV_RANDOM */
-#ifndef NO_MALLINFO
-#define NO_MALLINFO 0
-#endif /* NO_MALLINFO */
-#ifndef MALLINFO_FIELD_TYPE
-#define MALLINFO_FIELD_TYPE size_t
-#endif /* MALLINFO_FIELD_TYPE */
-#ifndef NO_SEGMENT_TRAVERSAL
-#define NO_SEGMENT_TRAVERSAL 0
-#endif /* NO_SEGMENT_TRAVERSAL */
-
-/*
- mallopt tuning options. SVID/XPG defines four standard parameter
- numbers for mallopt, normally defined in malloc.h. None of these
- are used in this malloc, so setting them has no effect. But this
- malloc does support the following options.
-*/
-
-#define M_TRIM_THRESHOLD (-1)
-#define M_GRANULARITY (-2)
-#define M_MMAP_THRESHOLD (-3)
-
-/* ------------------------ Mallinfo declarations ------------------------ */
-
-#if !NO_MALLINFO
-/*
- This version of malloc supports the standard SVID/XPG mallinfo
- routine that returns a struct containing usage properties and
- statistics. It should work on any system that has a
- /usr/include/malloc.h defining struct mallinfo. The main
- declaration needed is the mallinfo struct that is returned (by-copy)
- by mallinfo(). The malloinfo struct contains a bunch of fields that
- are not even meaningful in this version of malloc. These fields are
- are instead filled by mallinfo() with other numbers that might be of
- interest.
-
- HAVE_USR_INCLUDE_MALLOC_H should be set if you have a
- /usr/include/malloc.h file that includes a declaration of struct
- mallinfo. If so, it is included; else a compliant version is
- declared below. These must be precisely the same for mallinfo() to
- work. The original SVID version of this struct, defined on most
- systems with mallinfo, declares all fields as ints. But some others
- define as unsigned long. If your system defines the fields using a
- type of different width than listed here, you MUST #include your
- system version and #define HAVE_USR_INCLUDE_MALLOC_H.
-*/
-
-/* #define HAVE_USR_INCLUDE_MALLOC_H */
-
-#ifdef HAVE_USR_INCLUDE_MALLOC_H
-#include "/usr/include/malloc.h"
-#else /* HAVE_USR_INCLUDE_MALLOC_H */
-#ifndef STRUCT_MALLINFO_DECLARED
-#define STRUCT_MALLINFO_DECLARED 1
-struct mallinfo {
- MALLINFO_FIELD_TYPE arena; /* non-mmapped space allocated from system */
- MALLINFO_FIELD_TYPE ordblks; /* number of free chunks */
- MALLINFO_FIELD_TYPE smblks; /* always 0 */
- MALLINFO_FIELD_TYPE hblks; /* always 0 */
- MALLINFO_FIELD_TYPE hblkhd; /* space in mmapped regions */
- MALLINFO_FIELD_TYPE usmblks; /* maximum total allocated space */
- MALLINFO_FIELD_TYPE fsmblks; /* always 0 */
- MALLINFO_FIELD_TYPE uordblks; /* total allocated space */
- MALLINFO_FIELD_TYPE fordblks; /* total free space */
- MALLINFO_FIELD_TYPE keepcost; /* releasable (via malloc_trim) space */
-};
-#endif /* STRUCT_MALLINFO_DECLARED */
-#endif /* HAVE_USR_INCLUDE_MALLOC_H */
-#endif /* NO_MALLINFO */
-
-/*
- Try to persuade compilers to inline. The most critical functions for
- inlining are defined as macros, so these aren't used for them.
-*/
-
-#ifndef FORCEINLINE
- #if defined(__GNUC__)
-#define FORCEINLINE __inline __attribute__ ((always_inline))
- #elif defined(_MSC_VER)
- #define FORCEINLINE __forceinline
- #endif
-#endif
-#ifndef NOINLINE
- #if defined(__GNUC__)
- #define NOINLINE __attribute__ ((noinline))
- #elif defined(_MSC_VER)
- #define NOINLINE __declspec(noinline)
- #else
- #define NOINLINE
- #endif
-#endif
-
-#ifdef __cplusplus
-extern "C" {
-#ifndef FORCEINLINE
- #define FORCEINLINE inline
-#endif
-#endif /* __cplusplus */
-#ifndef FORCEINLINE
- #define FORCEINLINE
-#endif
-
-#if !ONLY_MSPACES
-
-/* ------------------- Declarations of public routines ------------------- */
-
-#ifndef USE_DL_PREFIX
-#define dlcalloc calloc
-#define dlfree free
-#define dlmalloc malloc
-#define dlmemalign memalign
-#define dlrealloc realloc
-#define dlvalloc valloc
-#define dlpvalloc pvalloc
-#define dlmallinfo mallinfo
-#define dlmallopt mallopt
-#define dlmalloc_trim malloc_trim
-#define dlmalloc_stats malloc_stats
-#define dlmalloc_usable_size malloc_usable_size
-#define dlmalloc_footprint malloc_footprint
-#define dlmalloc_max_footprint malloc_max_footprint
-#define dlindependent_calloc independent_calloc
-#define dlindependent_comalloc independent_comalloc
-#endif /* USE_DL_PREFIX */
-
-
-/*
- malloc(size_t n)
- Returns a pointer to a newly allocated chunk of at least n bytes, or
- null if no space is available, in which case errno is set to ENOMEM
- on ANSI C systems.
-
- If n is zero, malloc returns a minimum-sized chunk. (The minimum
- size is 16 bytes on most 32bit systems, and 32 bytes on 64bit
- systems.) Note that size_t is an unsigned type, so calls with
- arguments that would be negative if signed are interpreted as
- requests for huge amounts of space, which will often fail. The
- maximum supported value of n differs across systems, but is in all
- cases less than the maximum representable value of a size_t.
-*/
-void* dlmalloc(size_t);
-
-/*
- free(void* p)
- Releases the chunk of memory pointed to by p, that had been previously
- allocated using malloc or a related routine such as realloc.
- It has no effect if p is null. If p was not malloced or already
- freed, free(p) will by default cause the current program to abort.
-*/
-void dlfree(void*);
-
-/*
- calloc(size_t n_elements, size_t element_size);
- Returns a pointer to n_elements * element_size bytes, with all locations
- set to zero.
-*/
-void* dlcalloc(size_t, size_t);
-
-/*
- realloc(void* p, size_t n)
- Returns a pointer to a chunk of size n that contains the same data
- as does chunk p up to the minimum of (n, p's size) bytes, or null
- if no space is available.
-
- The returned pointer may or may not be the same as p. The algorithm
- prefers extending p in most cases when possible, otherwise it
- employs the equivalent of a malloc-copy-free sequence.
-
- If p is null, realloc is equivalent to malloc.
-
- If space is not available, realloc returns null, errno is set (if on
- ANSI) and p is NOT freed.
-
- if n is for fewer bytes than already held by p, the newly unused
- space is lopped off and freed if possible. realloc with a size
- argument of zero (re)allocates a minimum-sized chunk.
-
- The old unix realloc convention of allowing the last-free'd chunk
- to be used as an argument to realloc is not supported.
-*/
-
-void* dlrealloc(void*, size_t);
-
-/*
- memalign(size_t alignment, size_t n);
- Returns a pointer to a newly allocated chunk of n bytes, aligned
- in accord with the alignment argument.
-
- The alignment argument should be a power of two. If the argument is
- not a power of two, the nearest greater power is used.
- 8-byte alignment is guaranteed by normal malloc calls, so don't
- bother calling memalign with an argument of 8 or less.
-
- Overreliance on memalign is a sure way to fragment space.
-*/
-void* dlmemalign(size_t, size_t);
-
-/*
- valloc(size_t n);
- Equivalent to memalign(pagesize, n), where pagesize is the page
- size of the system. If the pagesize is unknown, 4096 is used.
-*/
-void* dlvalloc(size_t);
-
-/*
- mallopt(int parameter_number, int parameter_value)
- Sets tunable parameters The format is to provide a
- (parameter-number, parameter-value) pair. mallopt then sets the
- corresponding parameter to the argument value if it can (i.e., so
- long as the value is meaningful), and returns 1 if successful else
- 0. To workaround the fact that mallopt is specified to use int,
- not size_t parameters, the value -1 is specially treated as the
- maximum unsigned size_t value.
-
- SVID/XPG/ANSI defines four standard param numbers for mallopt,
- normally defined in malloc.h. None of these are use in this malloc,
- so setting them has no effect. But this malloc also supports other
- options in mallopt. See below for details. Briefly, supported
- parameters are as follows (listed defaults are for "typical"
- configurations).
-
- Symbol param # default allowed param values
- M_TRIM_THRESHOLD -1 2*1024*1024 any (-1 disables)
- M_GRANULARITY -2 page size any power of 2 >= page size
- M_MMAP_THRESHOLD -3 256*1024 any (or 0 if no MMAP support)
-*/
-int dlmallopt(int, int);
-
-/*
- malloc_footprint();
- Returns the number of bytes obtained from the system. The total
- number of bytes allocated by malloc, realloc etc., is less than this
- value. Unlike mallinfo, this function returns only a precomputed
- result, so can be called frequently to monitor memory consumption.
- Even if locks are otherwise defined, this function does not use them,
- so results might not be up to date.
-*/
-size_t dlmalloc_footprint(void);
-
-/*
- malloc_max_footprint();
- Returns the maximum number of bytes obtained from the system. This
- value will be greater than current footprint if deallocated space
- has been reclaimed by the system. The peak number of bytes allocated
- by malloc, realloc etc., is less than this value. Unlike mallinfo,
- this function returns only a precomputed result, so can be called
- frequently to monitor memory consumption. Even if locks are
- otherwise defined, this function does not use them, so results might
- not be up to date.
-*/
-size_t dlmalloc_max_footprint(void);
-
-#if !NO_MALLINFO
-/*
- mallinfo()
- Returns (by copy) a struct containing various summary statistics:
-
- arena: current total non-mmapped bytes allocated from system
- ordblks: the number of free chunks
- smblks: always zero.
- hblks: current number of mmapped regions
- hblkhd: total bytes held in mmapped regions
- usmblks: the maximum total allocated space. This will be greater
- than current total if trimming has occurred.
- fsmblks: always zero
- uordblks: current total allocated space (normal or mmapped)
- fordblks: total free space
- keepcost: the maximum number of bytes that could ideally be released
- back to system via malloc_trim. ("ideally" means that
- it ignores page restrictions etc.)
-
- Because these fields are ints, but internal bookkeeping may
- be kept as longs, the reported values may wrap around zero and
- thus be inaccurate.
-*/
-struct mallinfo dlmallinfo(void);
-#endif /* NO_MALLINFO */
-
-/*
- independent_calloc(size_t n_elements, size_t element_size, void* chunks[]);
-
- independent_calloc is similar to calloc, but instead of returning a
- single cleared space, it returns an array of pointers to n_elements
- independent elements that can hold contents of size elem_size, each
- of which starts out cleared, and can be independently freed,
- realloc'ed etc. The elements are guaranteed to be adjacently
- allocated (this is not guaranteed to occur with multiple callocs or
- mallocs), which may also improve cache locality in some
- applications.
-
- The "chunks" argument is optional (i.e., may be null, which is
- probably the most typical usage). If it is null, the returned array
- is itself dynamically allocated and should also be freed when it is
- no longer needed. Otherwise, the chunks array must be of at least
- n_elements in length. It is filled in with the pointers to the
- chunks.
-
- In either case, independent_calloc returns this pointer array, or
- null if the allocation failed. If n_elements is zero and "chunks"
- is null, it returns a chunk representing an array with zero elements
- (which should be freed if not wanted).
-
- Each element must be individually freed when it is no longer
- needed. If you'd like to instead be able to free all at once, you
- should instead use regular calloc and assign pointers into this
- space to represent elements. (In this case though, you cannot
- independently free elements.)
-
- independent_calloc simplifies and speeds up implementations of many
- kinds of pools. It may also be useful when constructing large data
- structures that initially have a fixed number of fixed-sized nodes,
- but the number is not known at compile time, and some of the nodes
- may later need to be freed. For example:
-
- struct Node { int item; struct Node* next; };
-
- struct Node* build_list() {
- struct Node** pool;
- int n = read_number_of_nodes_needed();
- if (n <= 0) return 0;
- pool = (struct Node**)(independent_calloc(n, sizeof(struct Node), 0);
- if (pool == 0) die();
- // organize into a linked list...
- struct Node* first = pool[0];
- for (i = 0; i < n-1; ++i)
- pool[i]->next = pool[i+1];
- free(pool); // Can now free the array (or not, if it is needed later)
- return first;
- }
-*/
-void** dlindependent_calloc(size_t, size_t, void**);
-
-/*
- independent_comalloc(size_t n_elements, size_t sizes[], void* chunks[]);
-
- independent_comalloc allocates, all at once, a set of n_elements
- chunks with sizes indicated in the "sizes" array. It returns
- an array of pointers to these elements, each of which can be
- independently freed, realloc'ed etc. The elements are guaranteed to
- be adjacently allocated (this is not guaranteed to occur with
- multiple callocs or mallocs), which may also improve cache locality
- in some applications.
-
- The "chunks" argument is optional (i.e., may be null). If it is null
- the returned array is itself dynamically allocated and should also
- be freed when it is no longer needed. Otherwise, the chunks array
- must be of at least n_elements in length. It is filled in with the
- pointers to the chunks.
-
- In either case, independent_comalloc returns this pointer array, or
- null if the allocation failed. If n_elements is zero and chunks is
- null, it returns a chunk representing an array with zero elements
- (which should be freed if not wanted).
-
- Each element must be individually freed when it is no longer
- needed. If you'd like to instead be able to free all at once, you
- should instead use a single regular malloc, and assign pointers at
- particular offsets in the aggregate space. (In this case though, you
- cannot independently free elements.)
-
- independent_comallac differs from independent_calloc in that each
- element may have a different size, and also that it does not
- automatically clear elements.
-
- independent_comalloc can be used to speed up allocation in cases
- where several structs or objects must always be allocated at the
- same time. For example:
-
- struct Head { ... }
- struct Foot { ... }
-
- void send_message(char* msg) {
- int msglen = strlen(msg);
- size_t sizes[3] = { sizeof(struct Head), msglen, sizeof(struct Foot) };
- void* chunks[3];
- if (independent_comalloc(3, sizes, chunks) == 0)
- die();
- struct Head* head = (struct Head*)(chunks[0]);
- char* body = (char*)(chunks[1]);
- struct Foot* foot = (struct Foot*)(chunks[2]);
- // ...
- }
-
- In general though, independent_comalloc is worth using only for
- larger values of n_elements. For small values, you probably won't
- detect enough difference from series of malloc calls to bother.
-
- Overuse of independent_comalloc can increase overall memory usage,
- since it cannot reuse existing noncontiguous small chunks that
- might be available for some of the elements.
-*/
-void** dlindependent_comalloc(size_t, size_t*, void**);
-
-
-/*
- pvalloc(size_t n);
- Equivalent to valloc(minimum-page-that-holds(n)), that is,
- round up n to nearest pagesize.
- */
-void* dlpvalloc(size_t);
-
-/*
- malloc_trim(size_t pad);
-
- If possible, gives memory back to the system (via negative arguments
- to sbrk) if there is unused memory at the `high' end of the malloc
- pool or in unused MMAP segments. You can call this after freeing
- large blocks of memory to potentially reduce the system-level memory
- requirements of a program. However, it cannot guarantee to reduce
- memory. Under some allocation patterns, some large free blocks of
- memory will be locked between two used chunks, so they cannot be
- given back to the system.
-
- The `pad' argument to malloc_trim represents the amount of free
- trailing space to leave untrimmed. If this argument is zero, only
- the minimum amount of memory to maintain internal data structures
- will be left. Non-zero arguments can be supplied to maintain enough
- trailing space to service future expected allocations without having
- to re-obtain memory from the system.
-
- Malloc_trim returns 1 if it actually released any memory, else 0.
-*/
-int dlmalloc_trim(size_t);
-
-/*
- malloc_stats();
- Prints on stderr the amount of space obtained from the system (both
- via sbrk and mmap), the maximum amount (which may be more than
- current if malloc_trim and/or munmap got called), and the current
- number of bytes allocated via malloc (or realloc, etc) but not yet
- freed. Note that this is the number of bytes allocated, not the
- number requested. It will be larger than the number requested
- because of alignment and bookkeeping overhead. Because it includes
- alignment wastage as being in use, this figure may be greater than
- zero even when no user-level chunks are allocated.
-
- The reported current and maximum system memory can be inaccurate if
- a program makes other calls to system memory allocation functions
- (normally sbrk) outside of malloc.
-
- malloc_stats prints only the most commonly interesting statistics.
- More information can be obtained by calling mallinfo.
-*/
-void dlmalloc_stats(void);
-
-#endif /* ONLY_MSPACES */
-
-/*
- malloc_usable_size(void* p);
-
- Returns the number of bytes you can actually use in
- an allocated chunk, which may be more than you requested (although
- often not) due to alignment and minimum size constraints.
- You can use this many bytes without worrying about
- overwriting other allocated objects. This is not a particularly great
- programming practice. malloc_usable_size can be more useful in
- debugging and assertions, for example:
-
- p = malloc(n);
- assert(malloc_usable_size(p) >= 256);
-*/
-size_t dlmalloc_usable_size(void*);
-
-
-#if MSPACES
-
-/*
- mspace is an opaque type representing an independent
- region of space that supports mspace_malloc, etc.
-*/
-typedef void* mspace;
-
-/*
- create_mspace creates and returns a new independent space with the
- given initial capacity, or, if 0, the default granularity size. It
- returns null if there is no system memory available to create the
- space. If argument locked is non-zero, the space uses a separate
- lock to control access. The capacity of the space will grow
- dynamically as needed to service mspace_malloc requests. You can
- control the sizes of incremental increases of this space by
- compiling with a different DEFAULT_GRANULARITY or dynamically
- setting with mallopt(M_GRANULARITY, value).
-*/
-mspace create_mspace(size_t capacity, int locked);
-
-/*
- destroy_mspace destroys the given space, and attempts to return all
- of its memory back to the system, returning the total number of
- bytes freed. After destruction, the results of access to all memory
- used by the space become undefined.
-*/
-size_t destroy_mspace(mspace msp);
-
-/*
- create_mspace_with_base uses the memory supplied as the initial base
- of a new mspace. Part (less than 128*sizeof(size_t) bytes) of this
- space is used for bookkeeping, so the capacity must be at least this
- large. (Otherwise 0 is returned.) When this initial space is
- exhausted, additional memory will be obtained from the system.
- Destroying this space will deallocate all additionally allocated
- space (if possible) but not the initial base.
-*/
-mspace create_mspace_with_base(void* base, size_t capacity, int locked);
-
-/*
- mspace_track_large_chunks controls whether requests for large chunks
- are allocated in their own untracked mmapped regions, separate from
- others in this mspace. By default large chunks are not tracked,
- which reduces fragmentation. However, such chunks are not
- necessarily released to the system upon destroy_mspace. Enabling
- tracking by setting to true may increase fragmentation, but avoids
- leakage when relying on destroy_mspace to release all memory
- allocated using this space. The function returns the previous
- setting.
-*/
-int mspace_track_large_chunks(mspace msp, int enable);
-
-
-/*
- mspace_malloc behaves as malloc, but operates within
- the given space.
-*/
-void* mspace_malloc(mspace msp, size_t bytes);
-
-/*
- mspace_free behaves as free, but operates within
- the given space.
-
- If compiled with FOOTERS==1, mspace_free is not actually needed.
- free may be called instead of mspace_free because freed chunks from
- any space are handled by their originating spaces.
-*/
-void mspace_free(mspace msp, void* mem);
-
-/*
- mspace_realloc behaves as realloc, but operates within
- the given space.
-
- If compiled with FOOTERS==1, mspace_realloc is not actually
- needed. realloc may be called instead of mspace_realloc because
- realloced chunks from any space are handled by their originating
- spaces.
-*/
-void* mspace_realloc(mspace msp, void* mem, size_t newsize);
-
-/*
- mspace_calloc behaves as calloc, but operates within
- the given space.
-*/
-void* mspace_calloc(mspace msp, size_t n_elements, size_t elem_size);
-
-/*
- mspace_memalign behaves as memalign, but operates within
- the given space.
-*/
-void* mspace_memalign(mspace msp, size_t alignment, size_t bytes);
-
-/*
- mspace_independent_calloc behaves as independent_calloc, but
- operates within the given space.
-*/
-void** mspace_independent_calloc(mspace msp, size_t n_elements,
- size_t elem_size, void* chunks[]);
-
-/*
- mspace_independent_comalloc behaves as independent_comalloc, but
- operates within the given space.
-*/
-void** mspace_independent_comalloc(mspace msp, size_t n_elements,
- size_t sizes[], void* chunks[]);
-
-/*
- mspace_footprint() returns the number of bytes obtained from the
- system for this space.
-*/
-size_t mspace_footprint(mspace msp);
-
-/*
- mspace_max_footprint() returns the peak number of bytes obtained from the
- system for this space.
-*/
-size_t mspace_max_footprint(mspace msp);
-
-
-#if !NO_MALLINFO
-/*
- mspace_mallinfo behaves as mallinfo, but reports properties of
- the given space.
-*/
-struct mallinfo mspace_mallinfo(mspace msp);
-#endif /* NO_MALLINFO */
-
-/*
- malloc_usable_size(void* p) behaves the same as malloc_usable_size;
-*/
- size_t mspace_usable_size(void* mem);
-
-/*
- mspace_malloc_stats behaves as malloc_stats, but reports
- properties of the given space.
-*/
-void mspace_malloc_stats(mspace msp);
-
-/*
- mspace_trim behaves as malloc_trim, but
- operates within the given space.
-*/
-int mspace_trim(mspace msp, size_t pad);
-
-/*
- An alias for mallopt.
-*/
-int mspace_mallopt(int, int);
-
-#endif /* MSPACES */
-
-#ifdef __cplusplus
-} /* end of extern "C" */
-#endif /* __cplusplus */
-
-/*
- ========================================================================
- To make a fully customizable malloc.h header file, cut everything
- above this line, put into file malloc.h, edit to suit, and #include it
- on the next line, as well as in programs that use this malloc.
- ========================================================================
-*/
-
-/* #include "malloc.h" */
-
-/*------------------------------ internal #includes ---------------------- */
-
-#ifdef WIN32
-#pragma warning( disable : 4146 ) /* no "unsigned" warnings */
-#endif /* WIN32 */
-
-#include <stdio.h> /* for printing in malloc_stats */
-
-#ifndef LACKS_ERRNO_H
-#include <errno.h> /* for MALLOC_FAILURE_ACTION */
-#endif /* LACKS_ERRNO_H */
-/*#if FOOTERS || DEBUG
-*/
-#include <time.h> /* for magic initialization */
-/*#endif*/ /* FOOTERS */
-#ifndef LACKS_STDLIB_H
-#include <stdlib.h> /* for abort() */
-#endif /* LACKS_STDLIB_H */
-#ifdef DEBUG
-#if ABORT_ON_ASSERT_FAILURE
-#undef assert
-#define assert(x) if(!(x)) ABORT
-#else /* ABORT_ON_ASSERT_FAILURE */
-#include <assert.h>
-#endif /* ABORT_ON_ASSERT_FAILURE */
-#else /* DEBUG */
-#ifndef assert
-#define assert(x)
-#endif
-#define DEBUG 0
-#endif /* DEBUG */
-#ifndef LACKS_STRING_H
-#include <string.h> /* for memset etc */
-#endif /* LACKS_STRING_H */
-#if USE_BUILTIN_FFS
-#ifndef LACKS_STRINGS_H
-#include <strings.h> /* for ffs */
-#endif /* LACKS_STRINGS_H */
-#endif /* USE_BUILTIN_FFS */
-#if HAVE_MMAP
-#ifndef LACKS_SYS_MMAN_H
-/* On some versions of linux, mremap decl in mman.h needs __USE_GNU set */
-#if (defined(linux) && !defined(__USE_GNU))
-#define __USE_GNU 1
-#include <sys/mman.h> /* for mmap */
-#undef __USE_GNU
-#else
-#include <sys/mman.h> /* for mmap */
-#endif /* linux */
-#endif /* LACKS_SYS_MMAN_H */
-#ifndef LACKS_FCNTL_H
-#include <fcntl.h>
-#endif /* LACKS_FCNTL_H */
-#endif /* HAVE_MMAP */
-#ifndef LACKS_UNISTD_H
-#include <unistd.h> /* for sbrk, sysconf */
-#else /* LACKS_UNISTD_H */
-#if !defined(__FreeBSD__) && !defined(__OpenBSD__) && !defined(__NetBSD__)
-extern void* sbrk(ptrdiff_t);
-#endif /* FreeBSD etc */
-#endif /* LACKS_UNISTD_H */
-
-/* Declarations for locking */
-#if USE_LOCKS
-#ifndef WIN32
-#include <pthread.h>
-#if defined (__SVR4) && defined (__sun) /* solaris */
-#include <thread.h>
-#endif /* solaris */
-#else
-#ifndef _M_AMD64
-/* These are already defined on AMD64 builds */
-#ifdef __cplusplus
-extern "C" {
-#endif /* __cplusplus */
-LONG __cdecl _InterlockedCompareExchange(LONG volatile *Dest, LONG Exchange, LONG Comp);
-LONG __cdecl _InterlockedExchange(LONG volatile *Target, LONG Value);
-#ifdef __cplusplus
-}
-#endif /* __cplusplus */
-#endif /* _M_AMD64 */
-#pragma intrinsic (_InterlockedCompareExchange)
-#pragma intrinsic (_InterlockedExchange)
-#define interlockedcompareexchange _InterlockedCompareExchange
-#define interlockedexchange _InterlockedExchange
-#endif /* Win32 */
-#endif /* USE_LOCKS */
-
-/* Declarations for bit scanning on win32 */
-#if defined(_MSC_VER) && _MSC_VER>=1300
-#ifndef BitScanForward /* Try to avoid pulling in WinNT.h */
-#ifdef __cplusplus
-extern "C" {
-#endif /* __cplusplus */
-unsigned char _BitScanForward(unsigned long *index, unsigned long mask);
-unsigned char _BitScanReverse(unsigned long *index, unsigned long mask);
-#ifdef __cplusplus
-}
-#endif /* __cplusplus */
-
-#define BitScanForward _BitScanForward
-#define BitScanReverse _BitScanReverse
-#pragma intrinsic(_BitScanForward)
-#pragma intrinsic(_BitScanReverse)
-#endif /* BitScanForward */
-#endif /* defined(_MSC_VER) && _MSC_VER>=1300 */
-
-#ifndef WIN32
-#ifndef malloc_getpagesize
-# ifdef _SC_PAGESIZE /* some SVR4 systems omit an underscore */
-# ifndef _SC_PAGE_SIZE
-# define _SC_PAGE_SIZE _SC_PAGESIZE
-# endif
-# endif
-# ifdef _SC_PAGE_SIZE
-# define malloc_getpagesize sysconf(_SC_PAGE_SIZE)
-# else
-# if defined(BSD) || defined(DGUX) || defined(HAVE_GETPAGESIZE)
- extern size_t getpagesize();
-# define malloc_getpagesize getpagesize()
-# else
-# ifdef WIN32 /* use supplied emulation of getpagesize */
-# define malloc_getpagesize getpagesize()
-# else
-# ifndef LACKS_SYS_PARAM_H
-# include <sys/param.h>
-# endif
-# ifdef EXEC_PAGESIZE
-# define malloc_getpagesize EXEC_PAGESIZE
-# else
-# ifdef NBPG
-# ifndef CLSIZE
-# define malloc_getpagesize NBPG
-# else
-# define malloc_getpagesize (NBPG * CLSIZE)
-# endif
-# else
-# ifdef NBPC
-# define malloc_getpagesize NBPC
-# else
-# ifdef PAGESIZE
-# define malloc_getpagesize PAGESIZE
-# else /* just guess */
-# define malloc_getpagesize ((size_t)4096U)
-# endif
-# endif
-# endif
-# endif
-# endif
-# endif
-# endif
-#endif
-#endif
-
-
-
-/* ------------------- size_t and alignment properties -------------------- */
-
-/* The byte and bit size of a size_t */
-#define SIZE_T_SIZE (sizeof(size_t))
-#define SIZE_T_BITSIZE (sizeof(size_t) << 3)
-
-/* Some constants coerced to size_t */
-/* Annoying but necessary to avoid errors on some platforms */
-#define SIZE_T_ZERO ((size_t)0)
-#define SIZE_T_ONE ((size_t)1)
-#define SIZE_T_TWO ((size_t)2)
-#define SIZE_T_FOUR ((size_t)4)
-#define TWO_SIZE_T_SIZES (SIZE_T_SIZE<<1)
-#define FOUR_SIZE_T_SIZES (SIZE_T_SIZE<<2)
-#define SIX_SIZE_T_SIZES (FOUR_SIZE_T_SIZES+TWO_SIZE_T_SIZES)
-#define HALF_MAX_SIZE_T (MAX_SIZE_T / 2U)
-
-/* The bit mask value corresponding to MALLOC_ALIGNMENT */
-#define CHUNK_ALIGN_MASK (MALLOC_ALIGNMENT - SIZE_T_ONE)
-
-/* True if address a has acceptable alignment */
-#define is_aligned(A) (((size_t)((A)) & (CHUNK_ALIGN_MASK)) == 0)
-
-/* the number of bytes to offset an address to align it */
-#define align_offset(A)\
- ((((size_t)(A) & CHUNK_ALIGN_MASK) == 0)? 0 :\
- ((MALLOC_ALIGNMENT - ((size_t)(A) & CHUNK_ALIGN_MASK)) & CHUNK_ALIGN_MASK))
-
-/* -------------------------- MMAP preliminaries ------------------------- */
-
-/*
- If HAVE_MORECORE or HAVE_MMAP are false, we just define calls and
- checks to fail so compiler optimizer can delete code rather than
- using so many "#if"s.
-*/
-
-
-/* MORECORE and MMAP must return MFAIL on failure */
-#define MFAIL ((void*)(MAX_SIZE_T))
-#define CMFAIL ((char*)(MFAIL)) /* defined for convenience */
-
-#if HAVE_MMAP
-
-#ifndef WIN32
-#define MUNMAP_DEFAULT(a, s) munmap((a), (s))
-#define MMAP_PROT (PROT_READ|PROT_WRITE)
-#if !defined(MAP_ANONYMOUS) && defined(MAP_ANON)
-#define MAP_ANONYMOUS MAP_ANON
-#endif /* MAP_ANON */
-#ifdef MAP_ANONYMOUS
-#define MMAP_FLAGS (MAP_PRIVATE|MAP_ANONYMOUS)
-#define MMAP_DEFAULT(s) mmap(0, (s), MMAP_PROT, MMAP_FLAGS, -1, 0)
-#else /* MAP_ANONYMOUS */
-/*
- Nearly all versions of mmap support MAP_ANONYMOUS, so the following
- is unlikely to be needed, but is supplied just in case.
-*/
-#define MMAP_FLAGS (MAP_PRIVATE)
-static int dev_zero_fd = -1; /* Cached file descriptor for /dev/zero. */
-#define MMAP_DEFAULT(s) ((dev_zero_fd < 0) ? \
- (dev_zero_fd = open("/dev/zero", O_RDWR), \
- mmap(0, (s), MMAP_PROT, MMAP_FLAGS, dev_zero_fd, 0)) : \
- mmap(0, (s), MMAP_PROT, MMAP_FLAGS, dev_zero_fd, 0))
-#endif /* MAP_ANONYMOUS */
-
-#define DIRECT_MMAP_DEFAULT(s) MMAP_DEFAULT(s)
-
-#else /* WIN32 */
-
-/* Win32 MMAP via VirtualAlloc */
-static FORCEINLINE void* win32mmap(size_t size) {
- void* ptr = VirtualAlloc(0, size, MEM_RESERVE|MEM_COMMIT, PAGE_READWRITE);
- return (ptr != 0)? ptr: MFAIL;
-}
-
-/* For direct MMAP, use MEM_TOP_DOWN to minimize interference */
-static FORCEINLINE void* win32direct_mmap(size_t size) {
- void* ptr = VirtualAlloc(0, size, MEM_RESERVE|MEM_COMMIT|MEM_TOP_DOWN,
- PAGE_READWRITE);
- return (ptr != 0)? ptr: MFAIL;
-}
-
-/* This function supports releasing coalesed segments */
-static FORCEINLINE int win32munmap(void* ptr, size_t size) {
- MEMORY_BASIC_INFORMATION minfo;
- char* cptr = (char*)ptr;
- while (size) {
- if (VirtualQuery(cptr, &minfo, sizeof(minfo)) == 0)
- return -1;
- if (minfo.BaseAddress != cptr || minfo.AllocationBase != cptr ||
- minfo.State != MEM_COMMIT || minfo.RegionSize > size)
- return -1;
- if (VirtualFree(cptr, 0, MEM_RELEASE) == 0)
- return -1;
- cptr += minfo.RegionSize;
- size -= minfo.RegionSize;
- }
- return 0;
-}
-
-#define MMAP_DEFAULT(s) win32mmap(s)
-#define MUNMAP_DEFAULT(a, s) win32munmap((a), (s))
-#define DIRECT_MMAP_DEFAULT(s) win32direct_mmap(s)
-#endif /* WIN32 */
-#endif /* HAVE_MMAP */
-
-#if HAVE_MREMAP
-#ifndef WIN32
-#define MREMAP_DEFAULT(addr, osz, nsz, mv) mremap((addr), (osz), (nsz), (mv))
-#endif /* WIN32 */
-#endif /* HAVE_MREMAP */
-
-
-/**
- * Define CALL_MORECORE
- */
-#if HAVE_MORECORE
- #ifdef MORECORE
- #define CALL_MORECORE(S) MORECORE(S)
- #else /* MORECORE */
- #define CALL_MORECORE(S) MORECORE_DEFAULT(S)
- #endif /* MORECORE */
-#else /* HAVE_MORECORE */
- #define CALL_MORECORE(S) MFAIL
-#endif /* HAVE_MORECORE */
-
-/**
- * Define CALL_MMAP/CALL_MUNMAP/CALL_DIRECT_MMAP
- */
-#if HAVE_MMAP
- #define USE_MMAP_BIT (SIZE_T_ONE)
-
- #ifdef MMAP
- #define CALL_MMAP(s) MMAP(s)
- #else /* MMAP */
- #define CALL_MMAP(s) MMAP_DEFAULT(s)
- #endif /* MMAP */
- #ifdef MUNMAP
- #define CALL_MUNMAP(a, s) MUNMAP((a), (s))
- #else /* MUNMAP */
- #define CALL_MUNMAP(a, s) MUNMAP_DEFAULT((a), (s))
- #endif /* MUNMAP */
- #ifdef DIRECT_MMAP
- #define CALL_DIRECT_MMAP(s) DIRECT_MMAP(s)
- #else /* DIRECT_MMAP */
- #define CALL_DIRECT_MMAP(s) DIRECT_MMAP_DEFAULT(s)
- #endif /* DIRECT_MMAP */
-#else /* HAVE_MMAP */
- #define USE_MMAP_BIT (SIZE_T_ZERO)
-
- #define MMAP(s) MFAIL
- #define MUNMAP(a, s) (-1)
- #define DIRECT_MMAP(s) MFAIL
- #define CALL_DIRECT_MMAP(s) DIRECT_MMAP(s)
- #define CALL_MMAP(s) MMAP(s)
- #define CALL_MUNMAP(a, s) MUNMAP((a), (s))
-#endif /* HAVE_MMAP */
-
-/**
- * Define CALL_MREMAP
- */
-#if HAVE_MMAP && HAVE_MREMAP
- #ifdef MREMAP
- #define CALL_MREMAP(addr, osz, nsz, mv) MREMAP((addr), (osz), (nsz), (mv))
- #else /* MREMAP */
- #define CALL_MREMAP(addr, osz, nsz, mv) MREMAP_DEFAULT((addr), (osz), (nsz), (mv))
- #endif /* MREMAP */
-#else /* HAVE_MMAP && HAVE_MREMAP */
- #define CALL_MREMAP(addr, osz, nsz, mv) MFAIL
-#endif /* HAVE_MMAP && HAVE_MREMAP */
-
-/* mstate bit set if continguous morecore disabled or failed */
-#define USE_NONCONTIGUOUS_BIT (4U)
-
-/* segment bit set in create_mspace_with_base */
-#define EXTERN_BIT (8U)
-
-
-/* --------------------------- Lock preliminaries ------------------------ */
-
-/*
- When locks are defined, there is one global lock, plus
- one per-mspace lock.
-
- The global lock_ensures that mparams.magic and other unique
- mparams values are initialized only once. It also protects
- sequences of calls to MORECORE. In many cases sys_alloc requires
- two calls, that should not be interleaved with calls by other
- threads. This does not protect against direct calls to MORECORE
- by other threads not using this lock, so there is still code to
- cope the best we can on interference.
-
- Per-mspace locks surround calls to malloc, free, etc. To enable use
- in layered extensions, per-mspace locks are reentrant.
-
- Because lock-protected regions generally have bounded times, it is
- OK to use the supplied simple spinlocks in the custom versions for
- x86. Spinlocks are likely to improve performance for lightly
- contended applications, but worsen performance under heavy
- contention.
-
- If USE_LOCKS is > 1, the definitions of lock routines here are
- bypassed, in which case you will need to define the type MLOCK_T,
- and at least INITIAL_LOCK, ACQUIRE_LOCK, RELEASE_LOCK and possibly
- TRY_LOCK (which is not used in this malloc, but commonly needed in
- extensions.) You must also declare a
- static MLOCK_T malloc_global_mutex = { initialization values };.
-
-*/
-
-#if USE_LOCKS == 1
-
-#if USE_SPIN_LOCKS && SPIN_LOCKS_AVAILABLE
-#ifndef WIN32
-
-/* Custom pthread-style spin locks on x86 and x64 for gcc */
-struct pthread_mlock_t {
- volatile unsigned int l;
- unsigned int c;
- pthread_t threadid;
-};
-#define MLOCK_T struct pthread_mlock_t
-#define CURRENT_THREAD pthread_self()
-#define INITIAL_LOCK(sl) ((sl)->threadid = 0, (sl)->l = (sl)->c = 0, 0)
-#define ACQUIRE_LOCK(sl) pthread_acquire_lock(sl)
-#define RELEASE_LOCK(sl) pthread_release_lock(sl)
-#define TRY_LOCK(sl) pthread_try_lock(sl)
-#define SPINS_PER_YIELD 63
-
-static MLOCK_T malloc_global_mutex = { 0, 0, 0};
-
-static FORCEINLINE int pthread_acquire_lock (MLOCK_T *sl) {
- int spins = 0;
- volatile unsigned int* lp = &sl->l;
- for (;;) {
- if (*lp != 0) {
- if (sl->threadid == CURRENT_THREAD) {
- ++sl->c;
- return 0;
- }
- }
- else {
- /* place args to cmpxchgl in locals to evade oddities in some gccs */
- int cmp = 0;
- int val = 1;
- int ret;
- __asm__ __volatile__ ("lock; cmpxchgl %1, %2"
- : "=a" (ret)
- : "r" (val), "m" (*(lp)), "0"(cmp)
- : "memory", "cc");
- if (!ret) {
- assert(!sl->threadid);
- sl->threadid = CURRENT_THREAD;
- sl->c = 1;
- return 0;
- }
- }
- if ((++spins & SPINS_PER_YIELD) == 0) {
-#if defined (__SVR4) && defined (__sun) /* solaris */
- thr_yield();
-#else
-#if defined(__linux__) || defined(__FreeBSD__) || defined(__APPLE__)
- sched_yield();
-#else /* no-op yield on unknown systems */
- ;
-#endif /* __linux__ || __FreeBSD__ || __APPLE__ */
-#endif /* solaris */
- }
- }
-}
-
-static FORCEINLINE void pthread_release_lock (MLOCK_T *sl) {
- volatile unsigned int* lp = &sl->l;
- assert(*lp != 0);
- assert(sl->threadid == CURRENT_THREAD);
- if (--sl->c == 0) {
- sl->threadid = 0;
- int prev = 0;
- int ret;
- __asm__ __volatile__ ("lock; xchgl %0, %1"
- : "=r" (ret)
- : "m" (*(lp)), "0"(prev)
- : "memory");
- }
-}
-
-static FORCEINLINE int pthread_try_lock (MLOCK_T *sl) {
- volatile unsigned int* lp = &sl->l;
- if (*lp != 0) {
- if (sl->threadid == CURRENT_THREAD) {
- ++sl->c;
- return 1;
- }
- }
- else {
- int cmp = 0;
- int val = 1;
- int ret;
- __asm__ __volatile__ ("lock; cmpxchgl %1, %2"
- : "=a" (ret)
- : "r" (val), "m" (*(lp)), "0"(cmp)
- : "memory", "cc");
- if (!ret) {
- assert(!sl->threadid);
- sl->threadid = CURRENT_THREAD;
- sl->c = 1;
- return 1;
- }
- }
- return 0;
-}
-
-
-#else /* WIN32 */
-/* Custom win32-style spin locks on x86 and x64 for MSC */
-struct win32_mlock_t {
- volatile long l;
- unsigned int c;
- long threadid;
-};
-
-#define MLOCK_T struct win32_mlock_t
-#define CURRENT_THREAD GetCurrentThreadId()
-#define INITIAL_LOCK(sl) ((sl)->threadid = 0, (sl)->l = (sl)->c = 0, 0)
-#define ACQUIRE_LOCK(sl) win32_acquire_lock(sl)
-#define RELEASE_LOCK(sl) win32_release_lock(sl)
-#define TRY_LOCK(sl) win32_try_lock(sl)
-#define SPINS_PER_YIELD 63
-
-static MLOCK_T malloc_global_mutex = { 0, 0, 0};
-
-static FORCEINLINE int win32_acquire_lock (MLOCK_T *sl) {
- int spins = 0;
- for (;;) {
- if (sl->l != 0) {
- if (sl->threadid == CURRENT_THREAD) {
- ++sl->c;
- return 0;
- }
- }
- else {
- if (!interlockedexchange(&sl->l, 1)) {
- assert(!sl->threadid);
- sl->threadid = CURRENT_THREAD;
- sl->c = 1;
- return 0;
- }
- }
- if ((++spins & SPINS_PER_YIELD) == 0)
- SleepEx(0, FALSE);
- }
-}
-
-static FORCEINLINE void win32_release_lock (MLOCK_T *sl) {
- assert(sl->threadid == CURRENT_THREAD);
- assert(sl->l != 0);
- if (--sl->c == 0) {
- sl->threadid = 0;
- interlockedexchange (&sl->l, 0);
- }
-}
-
-static FORCEINLINE int win32_try_lock (MLOCK_T *sl) {
- if (sl->l != 0) {
- if (sl->threadid == CURRENT_THREAD) {
- ++sl->c;
- return 1;
- }
- }
- else {
- if (!interlockedexchange(&sl->l, 1)){
- assert(!sl->threadid);
- sl->threadid = CURRENT_THREAD;
- sl->c = 1;
- return 1;
- }
- }
- return 0;
-}
-
-#endif /* WIN32 */
-#else /* USE_SPIN_LOCKS */
-
-#ifndef WIN32
-/* pthreads-based locks */
-
-#define MLOCK_T pthread_mutex_t
-#define CURRENT_THREAD pthread_self()
-#define INITIAL_LOCK(sl) pthread_init_lock(sl)
-#define ACQUIRE_LOCK(sl) pthread_mutex_lock(sl)
-#define RELEASE_LOCK(sl) pthread_mutex_unlock(sl)
-#define TRY_LOCK(sl) (!pthread_mutex_trylock(sl))
-
-static MLOCK_T malloc_global_mutex = PTHREAD_MUTEX_INITIALIZER;
-
-/* Cope with old-style linux recursive lock initialization by adding */
-/* skipped internal declaration from pthread.h */
-#ifdef linux
-#ifndef PTHREAD_MUTEX_RECURSIVE
-extern int pthread_mutexattr_setkind_np __P ((pthread_mutexattr_t *__attr,
- int __kind));
-#define PTHREAD_MUTEX_RECURSIVE PTHREAD_MUTEX_RECURSIVE_NP
-#define pthread_mutexattr_settype(x,y) pthread_mutexattr_setkind_np(x,y)
-#endif
-#endif
-
-static int pthread_init_lock (MLOCK_T *sl) {
- pthread_mutexattr_t attr;
- if (pthread_mutexattr_init(&attr)) return 1;
- if (pthread_mutexattr_settype(&attr, PTHREAD_MUTEX_RECURSIVE)) return 1;
- if (pthread_mutex_init(sl, &attr)) return 1;
- if (pthread_mutexattr_destroy(&attr)) return 1;
- return 0;
-}
-
-#else /* WIN32 */
-/* Win32 critical sections */
-#define MLOCK_T CRITICAL_SECTION
-#define CURRENT_THREAD GetCurrentThreadId()
-#define INITIAL_LOCK(s) (!InitializeCriticalSectionAndSpinCount((s), 0x80000000|4000))
-#define ACQUIRE_LOCK(s) (EnterCriticalSection(sl), 0)
-#define RELEASE_LOCK(s) LeaveCriticalSection(sl)
-#define TRY_LOCK(s) TryEnterCriticalSection(sl)
-#define NEED_GLOBAL_LOCK_INIT
-
-static MLOCK_T malloc_global_mutex;
-static volatile long malloc_global_mutex_status;
-
-/* Use spin loop to initialize global lock */
-static void init_malloc_global_mutex() {
- for (;;) {
- long stat = malloc_global_mutex_status;
- if (stat > 0)
- return;
- /* transition to < 0 while initializing, then to > 0) */
- if (stat == 0 &&
- interlockedcompareexchange(&malloc_global_mutex_status, -1, 0) == 0) {
- InitializeCriticalSection(&malloc_global_mutex);
- interlockedexchange(&malloc_global_mutex_status,1);
- return;
- }
- SleepEx(0, FALSE);
- }
-}
-
-#endif /* WIN32 */
-#endif /* USE_SPIN_LOCKS */
-#endif /* USE_LOCKS == 1 */
-
-/* ----------------------- User-defined locks ------------------------ */
-
-#if USE_LOCKS > 1
-/* Define your own lock implementation here */
-/* #define INITIAL_LOCK(sl) ... */
-/* #define ACQUIRE_LOCK(sl) ... */
-/* #define RELEASE_LOCK(sl) ... */
-/* #define TRY_LOCK(sl) ... */
-/* static MLOCK_T malloc_global_mutex = ... */
-#endif /* USE_LOCKS > 1 */
-
-/* ----------------------- Lock-based state ------------------------ */
-
-#if USE_LOCKS
-#define USE_LOCK_BIT (2U)
-#else /* USE_LOCKS */
-#define USE_LOCK_BIT (0U)
-#define INITIAL_LOCK(l)
-#endif /* USE_LOCKS */
-
-#if USE_LOCKS
-#ifndef ACQUIRE_MALLOC_GLOBAL_LOCK
-#define ACQUIRE_MALLOC_GLOBAL_LOCK() ACQUIRE_LOCK(&malloc_global_mutex);
-#endif
-#ifndef RELEASE_MALLOC_GLOBAL_LOCK
-#define RELEASE_MALLOC_GLOBAL_LOCK() RELEASE_LOCK(&malloc_global_mutex);
-#endif
-#else /* USE_LOCKS */
-#define ACQUIRE_MALLOC_GLOBAL_LOCK()
-#define RELEASE_MALLOC_GLOBAL_LOCK()
-#endif /* USE_LOCKS */
-
-
-/* ----------------------- Chunk representations ------------------------ */
-
-/*
- (The following includes lightly edited explanations by Colin Plumb.)
-
- The malloc_chunk declaration below is misleading (but accurate and
- necessary). It declares a "view" into memory allowing access to
- necessary fields at known offsets from a given base.
-
- Chunks of memory are maintained using a `boundary tag' method as
- originally described by Knuth. (See the paper by Paul Wilson
- ftp://ftp.cs.utexas.edu/pub/garbage/allocsrv.ps for a survey of such
- techniques.) Sizes of free chunks are stored both in the front of
- each chunk and at the end. This makes consolidating fragmented
- chunks into bigger chunks fast. The head fields also hold bits
- representing whether chunks are free or in use.
-
- Here are some pictures to make it clearer. They are "exploded" to
- show that the state of a chunk can be thought of as extending from
- the high 31 bits of the head field of its header through the
- prev_foot and PINUSE_BIT bit of the following chunk header.
-
- A chunk that's in use looks like:
-
- chunk-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- | Size of previous chunk (if P = 0) |
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |P|
- | Size of this chunk 1| +-+
- mem-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- | |
- +- -+
- | |
- +- -+
- | :
- +- size - sizeof(size_t) available payload bytes -+
- : |
- chunk-> +- -+
- | |
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |1|
- | Size of next chunk (may or may not be in use) | +-+
- mem-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
-
- And if it's free, it looks like this:
-
- chunk-> +- -+
- | User payload (must be in use, or we would have merged!) |
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |P|
- | Size of this chunk 0| +-+
- mem-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- | Next pointer |
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- | Prev pointer |
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- | :
- +- size - sizeof(struct chunk) unused bytes -+
- : |
- chunk-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- | Size of this chunk |
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0|
- | Size of next chunk (must be in use, or we would have merged)| +-+
- mem-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- | :
- +- User payload -+
- : |
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- |0|
- +-+
- Note that since we always merge adjacent free chunks, the chunks
- adjacent to a free chunk must be in use.
-
- Given a pointer to a chunk (which can be derived trivially from the
- payload pointer) we can, in O(1) time, find out whether the adjacent
- chunks are free, and if so, unlink them from the lists that they
- are on and merge them with the current chunk.
-
- Chunks always begin on even word boundaries, so the mem portion
- (which is returned to the user) is also on an even word boundary, and
- thus at least double-word aligned.
-
- The P (PINUSE_BIT) bit, stored in the unused low-order bit of the
- chunk size (which is always a multiple of two words), is an in-use
- bit for the *previous* chunk. If that bit is *clear*, then the
- word before the current chunk size contains the previous chunk
- size, and can be used to find the front of the previous chunk.
- The very first chunk allocated always has this bit set, preventing
- access to non-existent (or non-owned) memory. If pinuse is set for
- any given chunk, then you CANNOT determine the size of the
- previous chunk, and might even get a memory addressing fault when
- trying to do so.
-
- The C (CINUSE_BIT) bit, stored in the unused second-lowest bit of
- the chunk size redundantly records whether the current chunk is
- inuse (unless the chunk is mmapped). This redundancy enables usage
- checks within free and realloc, and reduces indirection when freeing
- and consolidating chunks.
-
- Each freshly allocated chunk must have both cinuse and pinuse set.
- That is, each allocated chunk borders either a previously allocated
- and still in-use chunk, or the base of its memory arena. This is
- ensured by making all allocations from the the `lowest' part of any
- found chunk. Further, no free chunk physically borders another one,
- so each free chunk is known to be preceded and followed by either
- inuse chunks or the ends of memory.
-
- Note that the `foot' of the current chunk is actually represented
- as the prev_foot of the NEXT chunk. This makes it easier to
- deal with alignments etc but can be very confusing when trying
- to extend or adapt this code.
-
- The exceptions to all this are
-
- 1. The special chunk `top' is the top-most available chunk (i.e.,
- the one bordering the end of available memory). It is treated
- specially. Top is never included in any bin, is used only if
- no other chunk is available, and is released back to the
- system if it is very large (see M_TRIM_THRESHOLD). In effect,
- the top chunk is treated as larger (and thus less well
- fitting) than any other available chunk. The top chunk
- doesn't update its trailing size field since there is no next
- contiguous chunk that would have to index off it. However,
- space is still allocated for it (TOP_FOOT_SIZE) to enable
- separation or merging when space is extended.
-
- 3. Chunks allocated via mmap, have both cinuse and pinuse bits
- cleared in their head fields. Because they are allocated
- one-by-one, each must carry its own prev_foot field, which is
- also used to hold the offset this chunk has within its mmapped
- region, which is needed to preserve alignment. Each mmapped
- chunk is trailed by the first two fields of a fake next-chunk
- for sake of usage checks.
-
-*/
-
-struct malloc_chunk {
- size_t prev_foot; /* Size of previous chunk (if free). */
- size_t head; /* Size and inuse bits. */
- struct malloc_chunk* fd; /* double links -- used only if free. */
- struct malloc_chunk* bk;
-};
-
-typedef struct malloc_chunk mchunk;
-typedef struct malloc_chunk* mchunkptr;
-typedef struct malloc_chunk* sbinptr; /* The type of bins of chunks */
-typedef unsigned int bindex_t; /* Described below */
-typedef unsigned int binmap_t; /* Described below */
-typedef unsigned int flag_t; /* The type of various bit flag sets */
-
-/* ------------------- Chunks sizes and alignments ----------------------- */
-
-#define MCHUNK_SIZE (sizeof(mchunk))
-
-#if FOOTERS
-#define CHUNK_OVERHEAD (TWO_SIZE_T_SIZES)
-#else /* FOOTERS */
-#define CHUNK_OVERHEAD (SIZE_T_SIZE)
-#endif /* FOOTERS */
-
-/* MMapped chunks need a second word of overhead ... */
-#define MMAP_CHUNK_OVERHEAD (TWO_SIZE_T_SIZES)
-/* ... and additional padding for fake next-chunk at foot */
-#define MMAP_FOOT_PAD (FOUR_SIZE_T_SIZES)
-
-/* The smallest size we can malloc is an aligned minimal chunk */
-#define MIN_CHUNK_SIZE\
- ((MCHUNK_SIZE + CHUNK_ALIGN_MASK) & ~CHUNK_ALIGN_MASK)
-
-/* conversion from malloc headers to user pointers, and back */
-#define chunk2mem(p) ((void*)((char*)(p) + TWO_SIZE_T_SIZES))
-#define mem2chunk(mem) ((mchunkptr)((char*)(mem) - TWO_SIZE_T_SIZES))
-/* chunk associated with aligned address A */
-#define align_as_chunk(A) (mchunkptr)((A) + align_offset(chunk2mem(A)))
-
-/* Bounds on request (not chunk) sizes. */
-#define MAX_REQUEST ((-MIN_CHUNK_SIZE) << 2)
-#define MIN_REQUEST (MIN_CHUNK_SIZE - CHUNK_OVERHEAD - SIZE_T_ONE)
-
-/* pad request bytes into a usable size */
-#define pad_request(req) \
- (((req) + CHUNK_OVERHEAD + CHUNK_ALIGN_MASK) & ~CHUNK_ALIGN_MASK)
-
-/* pad request, checking for minimum (but not maximum) */
-#define request2size(req) \
- (((req) < MIN_REQUEST)? MIN_CHUNK_SIZE : pad_request(req))
-
-
-/* ------------------ Operations on head and foot fields ----------------- */
-
-/*
- The head field of a chunk is or'ed with PINUSE_BIT when previous
- adjacent chunk in use, and or'ed with CINUSE_BIT if this chunk is in
- use, unless mmapped, in which case both bits are cleared.
-
- FLAG4_BIT is not used by this malloc, but might be useful in extensions.
-*/
-
-#define PINUSE_BIT (SIZE_T_ONE)
-#define CINUSE_BIT (SIZE_T_TWO)
-#define FLAG4_BIT (SIZE_T_FOUR)
-#define INUSE_BITS (PINUSE_BIT|CINUSE_BIT)
-#define FLAG_BITS (PINUSE_BIT|CINUSE_BIT|FLAG4_BIT)
-
-/* Head value for fenceposts */
-#define FENCEPOST_HEAD (INUSE_BITS|SIZE_T_SIZE)
-
-/* extraction of fields from head words */
-#define cinuse(p) ((p)->head & CINUSE_BIT)
-#define pinuse(p) ((p)->head & PINUSE_BIT)
-#define is_inuse(p) (((p)->head & INUSE_BITS) != PINUSE_BIT)
-#define is_mmapped(p) (((p)->head & INUSE_BITS) == 0)
-
-#define chunksize(p) ((p)->head & ~(FLAG_BITS))
-
-#define clear_pinuse(p) ((p)->head &= ~PINUSE_BIT)
-
-/* Treat space at ptr +/- offset as a chunk */
-#define chunk_plus_offset(p, s) ((mchunkptr)(((char*)(p)) + (s)))
-#define chunk_minus_offset(p, s) ((mchunkptr)(((char*)(p)) - (s)))
-
-/* Ptr to next or previous physical malloc_chunk. */
-#define next_chunk(p) ((mchunkptr)( ((char*)(p)) + ((p)->head & ~FLAG_BITS)))
-#define prev_chunk(p) ((mchunkptr)( ((char*)(p)) - ((p)->prev_foot) ))
-
-/* extract next chunk's pinuse bit */
-#define next_pinuse(p) ((next_chunk(p)->head) & PINUSE_BIT)
-
-/* Get/set size at footer */
-#define get_foot(p, s) (((mchunkptr)((char*)(p) + (s)))->prev_foot)
-#define set_foot(p, s) (((mchunkptr)((char*)(p) + (s)))->prev_foot = (s))
-
-/* Set size, pinuse bit, and foot */
-#define set_size_and_pinuse_of_free_chunk(p, s)\
- ((p)->head = (s|PINUSE_BIT), set_foot(p, s))
-
-/* Set size, pinuse bit, foot, and clear next pinuse */
-#define set_free_with_pinuse(p, s, n)\
- (clear_pinuse(n), set_size_and_pinuse_of_free_chunk(p, s))
-
-/* Get the internal overhead associated with chunk p */
-#define overhead_for(p)\
- (is_mmapped(p)? MMAP_CHUNK_OVERHEAD : CHUNK_OVERHEAD)
-
-/* Return true if malloced space is not necessarily cleared */
-#if MMAP_CLEARS
-#define calloc_must_clear(p) (!is_mmapped(p))
-#else /* MMAP_CLEARS */
-#define calloc_must_clear(p) (1)
-#endif /* MMAP_CLEARS */
-
-/* ---------------------- Overlaid data structures ----------------------- */
-
-/*
- When chunks are not in use, they are treated as nodes of either
- lists or trees.
-
- "Small" chunks are stored in circular doubly-linked lists, and look
- like this:
-
- chunk-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- | Size of previous chunk |
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- `head:' | Size of chunk, in bytes |P|
- mem-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- | Forward pointer to next chunk in list |
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- | Back pointer to previous chunk in list |
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- | Unused space (may be 0 bytes long) .
- . .
- . |
-nextchunk-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- `foot:' | Size of chunk, in bytes |
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
-
- Larger chunks are kept in a form of bitwise digital trees (aka
- tries) keyed on chunksizes. Because malloc_tree_chunks are only for
- free chunks greater than 256 bytes, their size doesn't impose any
- constraints on user chunk sizes. Each node looks like:
-
- chunk-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- | Size of previous chunk |
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- `head:' | Size of chunk, in bytes |P|
- mem-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- | Forward pointer to next chunk of same size |
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- | Back pointer to previous chunk of same size |
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- | Pointer to left child (child[0]) |
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- | Pointer to right child (child[1]) |
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- | Pointer to parent |
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- | bin index of this chunk |
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- | Unused space .
- . |
-nextchunk-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
- `foot:' | Size of chunk, in bytes |
- +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
-
- Each tree holding treenodes is a tree of unique chunk sizes. Chunks
- of the same size are arranged in a circularly-linked list, with only
- the oldest chunk (the next to be used, in our FIFO ordering)
- actually in the tree. (Tree members are distinguished by a non-null
- parent pointer.) If a chunk with the same size an an existing node
- is inserted, it is linked off the existing node using pointers that
- work in the same way as fd/bk pointers of small chunks.
-
- Each tree contains a power of 2 sized range of chunk sizes (the
- smallest is 0x100 <= x < 0x180), which is is divided in half at each
- tree level, with the chunks in the smaller half of the range (0x100
- <= x < 0x140 for the top nose) in the left subtree and the larger
- half (0x140 <= x < 0x180) in the right subtree. This is, of course,
- done by inspecting individual bits.
-
- Using these rules, each node's left subtree contains all smaller
- sizes than its right subtree. However, the node at the root of each
- subtree has no particular ordering relationship to either. (The
- dividing line between the subtree sizes is based on trie relation.)
- If we remove the last chunk of a given size from the interior of the
- tree, we need to replace it with a leaf node. The tree ordering
- rules permit a node to be replaced by any leaf below it.
-
- The smallest chunk in a tree (a common operation in a best-fit
- allocator) can be found by walking a path to the leftmost leaf in
- the tree. Unlike a usual binary tree, where we follow left child
- pointers until we reach a null, here we follow the right child
- pointer any time the left one is null, until we reach a leaf with
- both child pointers null. The smallest chunk in the tree will be
- somewhere along that path.
-
- The worst case number of steps to add, find, or remove a node is
- bounded by the number of bits differentiating chunks within
- bins. Under current bin calculations, this ranges from 6 up to 21
- (for 32 bit sizes) or up to 53 (for 64 bit sizes). The typical case
- is of course much better.
-*/
-
-struct malloc_tree_chunk {
- /* The first four fields must be compatible with malloc_chunk */
- size_t prev_foot;
- size_t head;
- struct malloc_tree_chunk* fd;
- struct malloc_tree_chunk* bk;
-
- struct malloc_tree_chunk* child[2];
- struct malloc_tree_chunk* parent;
- bindex_t index;
-};
-
-typedef struct malloc_tree_chunk tchunk;
-typedef struct malloc_tree_chunk* tchunkptr;
-typedef struct malloc_tree_chunk* tbinptr; /* The type of bins of trees */
-
-/* A little helper macro for trees */
-#define leftmost_child(t) ((t)->child[0] != 0? (t)->child[0] : (t)->child[1])
-
-/* ----------------------------- Segments -------------------------------- */
-
-/*
- Each malloc space may include non-contiguous segments, held in a
- list headed by an embedded malloc_segment record representing the
- top-most space. Segments also include flags holding properties of
- the space. Large chunks that are directly allocated by mmap are not
- included in this list. They are instead independently created and
- destroyed without otherwise keeping track of them.
-
- Segment management mainly comes into play for spaces allocated by
- MMAP. Any call to MMAP might or might not return memory that is
- adjacent to an existing segment. MORECORE normally contiguously
- extends the current space, so this space is almost always adjacent,
- which is simpler and faster to deal with. (This is why MORECORE is
- used preferentially to MMAP when both are available -- see
- sys_alloc.) When allocating using MMAP, we don't use any of the
- hinting mechanisms (inconsistently) supported in various
- implementations of unix mmap, or distinguish reserving from
- committing memory. Instead, we just ask for space, and exploit
- contiguity when we get it. It is probably possible to do
- better than this on some systems, but no general scheme seems
- to be significantly better.
-
- Management entails a simpler variant of the consolidation scheme
- used for chunks to reduce fragmentation -- new adjacent memory is
- normally prepended or appended to an existing segment. However,
- there are limitations compared to chunk consolidation that mostly
- reflect the fact that segment processing is relatively infrequent
- (occurring only when getting memory from system) and that we
- don't expect to have huge numbers of segments:
-
- * Segments are not indexed, so traversal requires linear scans. (It
- would be possible to index these, but is not worth the extra
- overhead and complexity for most programs on most platforms.)
- * New segments are only appended to old ones when holding top-most
- memory; if they cannot be prepended to others, they are held in
- different segments.
-
- Except for the top-most segment of an mstate, each segment record
- is kept at the tail of its segment. Segments are added by pushing
- segment records onto the list headed by &mstate.seg for the
- containing mstate.
-
- Segment flags control allocation/merge/deallocation policies:
- * If EXTERN_BIT set, then we did not allocate this segment,
- and so should not try to deallocate or merge with others.
- (This currently holds only for the initial segment passed
- into create_mspace_with_base.)
- * If USE_MMAP_BIT set, the segment may be merged with
- other surrounding mmapped segments and trimmed/de-allocated
- using munmap.
- * If neither bit is set, then the segment was obtained using
- MORECORE so can be merged with surrounding MORECORE'd segments
- and deallocated/trimmed using MORECORE with negative arguments.
-*/
-
-struct malloc_segment {
- char* base; /* base address */
- size_t size; /* allocated size */
- struct malloc_segment* next; /* ptr to next segment */
- flag_t sflags; /* mmap and extern flag */
-};
-
-#define is_mmapped_segment(S) ((S)->sflags & USE_MMAP_BIT)
-#define is_extern_segment(S) ((S)->sflags & EXTERN_BIT)
-
-typedef struct malloc_segment msegment;
-typedef struct malloc_segment* msegmentptr;
-
-/* ---------------------------- malloc_state ----------------------------- */
-
-/*
- A malloc_state holds all of the bookkeeping for a space.
- The main fields are:
-
- Top
- The topmost chunk of the currently active segment. Its size is
- cached in topsize. The actual size of topmost space is
- topsize+TOP_FOOT_SIZE, which includes space reserved for adding
- fenceposts and segment records if necessary when getting more
- space from the system. The size at which to autotrim top is
- cached from mparams in trim_check, except that it is disabled if
- an autotrim fails.
-
- Designated victim (dv)
- This is the preferred chunk for servicing small requests that
- don't have exact fits. It is normally the chunk split off most
- recently to service another small request. Its size is cached in
- dvsize. The link fields of this chunk are not maintained since it
- is not kept in a bin.
-
- SmallBins
- An array of bin headers for free chunks. These bins hold chunks
- with sizes less than MIN_LARGE_SIZE bytes. Each bin contains
- chunks of all the same size, spaced 8 bytes apart. To simplify
- use in double-linked lists, each bin header acts as a malloc_chunk
- pointing to the real first node, if it exists (else pointing to
- itself). This avoids special-casing for headers. But to avoid
- waste, we allocate only the fd/bk pointers of bins, and then use
- repositioning tricks to treat these as the fields of a chunk.
-
- TreeBins
- Treebins are pointers to the roots of trees holding a range of
- sizes. There are 2 equally spaced treebins for each power of two
- from TREE_SHIFT to TREE_SHIFT+16. The last bin holds anything
- larger.
-
- Bin maps
- There is one bit map for small bins ("smallmap") and one for
- treebins ("treemap). Each bin sets its bit when non-empty, and
- clears the bit when empty. Bit operations are then used to avoid
- bin-by-bin searching -- nearly all "search" is done without ever
- looking at bins that won't be selected. The bit maps
- conservatively use 32 bits per map word, even if on 64bit system.
- For a good description of some of the bit-based techniques used
- here, see Henry S. Warren Jr's book "Hacker's Delight" (and
- supplement at http://hackersdelight.org/). Many of these are
- intended to reduce the branchiness of paths through malloc etc, as
- well as to reduce the number of memory locations read or written.
-
- Segments
- A list of segments headed by an embedded malloc_segment record
- representing the initial space.
-
- Address check support
- The least_addr field is the least address ever obtained from
- MORECORE or MMAP. Attempted frees and reallocs of any address less
- than this are trapped (unless INSECURE is defined).
-
- Magic tag
- A cross-check field that should always hold same value as mparams.magic.
-
- Flags
- Bits recording whether to use MMAP, locks, or contiguous MORECORE
-
- Statistics
- Each space keeps track of current and maximum system memory
- obtained via MORECORE or MMAP.
-
- Trim support
- Fields holding the amount of unused topmost memory that should trigger
- timming, and a counter to force periodic scanning to release unused
- non-topmost segments.
-
- Locking
- If USE_LOCKS is defined, the "mutex" lock is acquired and released
- around every public call using this mspace.
-
- Extension support
- A void* pointer and a size_t field that can be used to help implement
- extensions to this malloc.
-*/
-
-/* Bin types, widths and sizes */
-#define NSMALLBINS (32U)
-#define NTREEBINS (32U)
-#define SMALLBIN_SHIFT (3U)
-#define SMALLBIN_WIDTH (SIZE_T_ONE << SMALLBIN_SHIFT)
-#define TREEBIN_SHIFT (8U)
-#define MIN_LARGE_SIZE (SIZE_T_ONE << TREEBIN_SHIFT)
-#define MAX_SMALL_SIZE (MIN_LARGE_SIZE - SIZE_T_ONE)
-#define MAX_SMALL_REQUEST (MAX_SMALL_SIZE - CHUNK_ALIGN_MASK - CHUNK_OVERHEAD)
-
-struct malloc_state {
- binmap_t smallmap;
- binmap_t treemap;
- size_t dvsize;
- size_t topsize;
- char* least_addr;
- mchunkptr dv;
- mchunkptr top;
- size_t trim_check;
- size_t release_checks;
- size_t magic;
- mchunkptr smallbins[(NSMALLBINS+1)*2];
- tbinptr treebins[NTREEBINS];
- size_t footprint;
- size_t max_footprint;
- flag_t mflags;
-#if USE_LOCKS
- MLOCK_T mutex; /* locate lock among fields that rarely change */
-#endif /* USE_LOCKS */
- msegment seg;
- void* extp; /* Unused but available for extensions */
- size_t exts;
-};
-
-typedef struct malloc_state* mstate;
-
-/* ------------- Global malloc_state and malloc_params ------------------- */
-
-/*
- malloc_params holds global properties, including those that can be
- dynamically set using mallopt. There is a single instance, mparams,
- initialized in init_mparams. Note that the non-zeroness of "magic"
- also serves as an initialization flag.
-*/
-
-struct malloc_params {
- volatile size_t magic;
- size_t page_size;
- size_t granularity;
- size_t mmap_threshold;
- size_t trim_threshold;
- flag_t default_mflags;
-};
-
-static struct malloc_params mparams;
-
-/* Ensure mparams initialized */
-#define ensure_initialization() (void)(mparams.magic != 0 || init_mparams())
-
-#if !ONLY_MSPACES
-
-/* The global malloc_state used for all non-"mspace" calls */
-static struct malloc_state _gm_;
-#define gm (&_gm_)
-#define is_global(M) ((M) == &_gm_)
-
-#endif /* !ONLY_MSPACES */
-
-#define is_initialized(M) ((M)->top != 0)
-
-/* -------------------------- system alloc setup ------------------------- */
-
-/* Operations on mflags */
-
-#define use_lock(M) ((M)->mflags & USE_LOCK_BIT)
-#define enable_lock(M) ((M)->mflags |= USE_LOCK_BIT)
-#define disable_lock(M) ((M)->mflags &= ~USE_LOCK_BIT)
-
-#define use_mmap(M) ((M)->mflags & USE_MMAP_BIT)
-#define enable_mmap(M) ((M)->mflags |= USE_MMAP_BIT)
-#define disable_mmap(M) ((M)->mflags &= ~USE_MMAP_BIT)
-
-#define use_noncontiguous(M) ((M)->mflags & USE_NONCONTIGUOUS_BIT)
-#define disable_contiguous(M) ((M)->mflags |= USE_NONCONTIGUOUS_BIT)
-
-#define set_lock(M,L)\
- ((M)->mflags = (L)?\
- ((M)->mflags | USE_LOCK_BIT) :\
- ((M)->mflags & ~USE_LOCK_BIT))
-
-/* page-align a size */
-#define page_align(S)\
- (((S) + (mparams.page_size - SIZE_T_ONE)) & ~(mparams.page_size - SIZE_T_ONE))
-
-/* granularity-align a size */
-#define granularity_align(S)\
- (((S) + (mparams.granularity - SIZE_T_ONE))\
- & ~(mparams.granularity - SIZE_T_ONE))
-
-
-/* For mmap, use granularity alignment on windows, else page-align */
-#ifdef WIN32
-#define mmap_align(S) granularity_align(S)
-#else
-#define mmap_align(S) page_align(S)
-#endif
-
-/* For sys_alloc, enough padding to ensure can malloc request on success */
-#define SYS_ALLOC_PADDING (TOP_FOOT_SIZE + MALLOC_ALIGNMENT)
-
-#define is_page_aligned(S)\
- (((size_t)(S) & (mparams.page_size - SIZE_T_ONE)) == 0)
-#define is_granularity_aligned(S)\
- (((size_t)(S) & (mparams.granularity - SIZE_T_ONE)) == 0)
-
-/* True if segment S holds address A */
-#define segment_holds(S, A)\
- ((char*)(A) >= S->base && (char*)(A) < S->base + S->size)
-
-/* Return segment holding given address */
-static msegmentptr segment_holding(mstate m, char* addr) {
- msegmentptr sp = &m->seg;
- for (;;) {
- if (addr >= sp->base && addr < sp->base + sp->size)
- return sp;
- if ((sp = sp->next) == 0)
- return 0;
- }
-}
-
-/* Return true if segment contains a segment link */
-static int has_segment_link(mstate m, msegmentptr ss) {
- msegmentptr sp = &m->seg;
- for (;;) {
- if ((char*)sp >= ss->base && (char*)sp < ss->base + ss->size)
- return 1;
- if ((sp = sp->next) == 0)
- return 0;
- }
-}
-
-#ifndef MORECORE_CANNOT_TRIM
-#define should_trim(M,s) ((s) > (M)->trim_check)
-#else /* MORECORE_CANNOT_TRIM */
-#define should_trim(M,s) (0)
-#endif /* MORECORE_CANNOT_TRIM */
-
-/*
- TOP_FOOT_SIZE is padding at the end of a segment, including space
- that may be needed to place segment records and fenceposts when new
- noncontiguous segments are added.
-*/
-#define TOP_FOOT_SIZE\
- (align_offset(chunk2mem(0))+pad_request(sizeof(struct malloc_segment))+MIN_CHUNK_SIZE)
-
-
-/* ------------------------------- Hooks -------------------------------- */
-
-/*
- PREACTION should be defined to return 0 on success, and nonzero on
- failure. If you are not using locking, you can redefine these to do
- anything you like.
-*/
-
-#if USE_LOCKS
-
-#define PREACTION(M) ((use_lock(M))? ACQUIRE_LOCK(&(M)->mutex) : 0)
-#define POSTACTION(M) { if (use_lock(M)) RELEASE_LOCK(&(M)->mutex); }
-#else /* USE_LOCKS */
-
-#ifndef PREACTION
-#define PREACTION(M) (0)
-#endif /* PREACTION */
-
-#ifndef POSTACTION
-#define POSTACTION(M)
-#endif /* POSTACTION */
-
-#endif /* USE_LOCKS */
-
-/*
- CORRUPTION_ERROR_ACTION is triggered upon detected bad addresses.
- USAGE_ERROR_ACTION is triggered on detected bad frees and
- reallocs. The argument p is an address that might have triggered the
- fault. It is ignored by the two predefined actions, but might be
- useful in custom actions that try to help diagnose errors.
-*/
-
-#if PROCEED_ON_ERROR
-
-/* A count of the number of corruption errors causing resets */
-int malloc_corruption_error_count;
-
-/* default corruption action */
-static void reset_on_error(mstate m);
-
-#define CORRUPTION_ERROR_ACTION(m) reset_on_error(m)
-#define USAGE_ERROR_ACTION(m, p)
-
-#else /* PROCEED_ON_ERROR */
-
-#ifndef CORRUPTION_ERROR_ACTION
-#define CORRUPTION_ERROR_ACTION(m) ABORT
-#endif /* CORRUPTION_ERROR_ACTION */
-
-#ifndef USAGE_ERROR_ACTION
-#define USAGE_ERROR_ACTION(m,p) ABORT
-#endif /* USAGE_ERROR_ACTION */
-
-#endif /* PROCEED_ON_ERROR */
-
-/* -------------------------- Debugging setup ---------------------------- */
-
-#if ! DEBUG
-
-#define check_free_chunk(M,P)
-#define check_inuse_chunk(M,P)
-#define check_malloced_chunk(M,P,N)
-#define check_mmapped_chunk(M,P)
-#define check_malloc_state(M)
-#define check_top_chunk(M,P)
-
-#else /* DEBUG */
-#define check_free_chunk(M,P) do_check_free_chunk(M,P)
-#define check_inuse_chunk(M,P) do_check_inuse_chunk(M,P)
-#define check_top_chunk(M,P) do_check_top_chunk(M,P)
-#define check_malloced_chunk(M,P,N) do_check_malloced_chunk(M,P,N)
-#define check_mmapped_chunk(M,P) do_check_mmapped_chunk(M,P)
-#define check_malloc_state(M) do_check_malloc_state(M)
-
-static void do_check_any_chunk(mstate m, mchunkptr p);
-static void do_check_top_chunk(mstate m, mchunkptr p);
-static void do_check_mmapped_chunk(mstate m, mchunkptr p);
-static void do_check_inuse_chunk(mstate m, mchunkptr p);
-static void do_check_free_chunk(mstate m, mchunkptr p);
-static void do_check_malloced_chunk(mstate m, void* mem, size_t s);
-static void do_check_tree(mstate m, tchunkptr t);
-static void do_check_treebin(mstate m, bindex_t i);
-static void do_check_smallbin(mstate m, bindex_t i);
-static void do_check_malloc_state(mstate m);
-static int bin_find(mstate m, mchunkptr x);
-static size_t traverse_and_check(mstate m);
-#endif /* DEBUG */
-
-/* ---------------------------- Indexing Bins ---------------------------- */
-
-#define is_small(s) (((s) >> SMALLBIN_SHIFT) < NSMALLBINS)
-#define small_index(s) ((s) >> SMALLBIN_SHIFT)
-#define small_index2size(i) ((i) << SMALLBIN_SHIFT)
-#define MIN_SMALL_INDEX (small_index(MIN_CHUNK_SIZE))
-
-/* addressing by index. See above about smallbin repositioning */
-#define smallbin_at(M, i) ((sbinptr)((char*)&((M)->smallbins[(i)<<1])))
-#define treebin_at(M,i) (&((M)->treebins[i]))
-
-/* assign tree index for size S to variable I. Use x86 asm if possible */
-#if defined(__GNUC__) && (defined(__i386__) || defined(__x86_64__))
-#define compute_tree_index(S, I)\
-{\
- unsigned int X = S >> TREEBIN_SHIFT;\
- if (X == 0)\
- I = 0;\
- else if (X > 0xFFFF)\
- I = NTREEBINS-1;\
- else {\
- unsigned int K;\
- __asm__("bsrl\t%1, %0\n\t" : "=r" (K) : "g" (X));\
- I = (bindex_t)((K << 1) + ((S >> (K + (TREEBIN_SHIFT-1)) & 1)));\
- }\
-}
-
-#elif defined (__INTEL_COMPILER)
-#define compute_tree_index(S, I)\
-{\
- size_t X = S >> TREEBIN_SHIFT;\
- if (X == 0)\
- I = 0;\
- else if (X > 0xFFFF)\
- I = NTREEBINS-1;\
- else {\
- unsigned int K = _bit_scan_reverse (X); \
- I = (bindex_t)((K << 1) + ((S >> (K + (TREEBIN_SHIFT-1)) & 1)));\
- }\
-}
-
-#elif defined(_MSC_VER) && _MSC_VER>=1300
-#define compute_tree_index(S, I)\
-{\
- size_t X = S >> TREEBIN_SHIFT;\
- if (X == 0)\
- I = 0;\
- else if (X > 0xFFFF)\
- I = NTREEBINS-1;\
- else {\
- unsigned int K;\
- _BitScanReverse((DWORD *) &K, X);\
- I = (bindex_t)((K << 1) + ((S >> (K + (TREEBIN_SHIFT-1)) & 1)));\
- }\
-}
-
-#else /* GNUC */
-#define compute_tree_index(S, I)\
-{\
- size_t X = S >> TREEBIN_SHIFT;\
- if (X == 0)\
- I = 0;\
- else if (X > 0xFFFF)\
- I = NTREEBINS-1;\
- else {\
- unsigned int Y = (unsigned int)X;\
- unsigned int N = ((Y - 0x100) >> 16) & 8;\
- unsigned int K = (((Y <<= N) - 0x1000) >> 16) & 4;\
- N += K;\
- N += K = (((Y <<= K) - 0x4000) >> 16) & 2;\
- K = 14 - N + ((Y <<= K) >> 15);\
- I = (K << 1) + ((S >> (K + (TREEBIN_SHIFT-1)) & 1));\
- }\
-}
-#endif /* GNUC */
-
-/* Bit representing maximum resolved size in a treebin at i */
-#define bit_for_tree_index(i) \
- (i == NTREEBINS-1)? (SIZE_T_BITSIZE-1) : (((i) >> 1) + TREEBIN_SHIFT - 2)
-
-/* Shift placing maximum resolved bit in a treebin at i as sign bit */
-#define leftshift_for_tree_index(i) \
- ((i == NTREEBINS-1)? 0 : \
- ((SIZE_T_BITSIZE-SIZE_T_ONE) - (((i) >> 1) + TREEBIN_SHIFT - 2)))
-
-/* The size of the smallest chunk held in bin with index i */
-#define minsize_for_tree_index(i) \
- ((SIZE_T_ONE << (((i) >> 1) + TREEBIN_SHIFT)) | \
- (((size_t)((i) & SIZE_T_ONE)) << (((i) >> 1) + TREEBIN_SHIFT - 1)))
-
-
-/* ------------------------ Operations on bin maps ----------------------- */
-
-/* bit corresponding to given index */
-#define idx2bit(i) ((binmap_t)(1) << (i))
-
-/* Mark/Clear bits with given index */
-#define mark_smallmap(M,i) ((M)->smallmap |= idx2bit(i))
-#define clear_smallmap(M,i) ((M)->smallmap &= ~idx2bit(i))
-#define smallmap_is_marked(M,i) ((M)->smallmap & idx2bit(i))
-
-#define mark_treemap(M,i) ((M)->treemap |= idx2bit(i))
-#define clear_treemap(M,i) ((M)->treemap &= ~idx2bit(i))
-#define treemap_is_marked(M,i) ((M)->treemap & idx2bit(i))
-
-/* isolate the least set bit of a bitmap */
-#define least_bit(x) ((x) & -(x))
-
-/* mask with all bits to left of least bit of x on */
-#define left_bits(x) ((x<<1) | -(x<<1))
-
-/* mask with all bits to left of or equal to least bit of x on */
-#define same_or_left_bits(x) ((x) | -(x))
-
-/* index corresponding to given bit. Use x86 asm if possible */
-
-#if defined(__GNUC__) && (defined(__i386__) || defined(__x86_64__))
-#define compute_bit2idx(X, I)\
-{\
- unsigned int J;\
- __asm__("bsfl\t%1, %0\n\t" : "=r" (J) : "g" (X));\
- I = (bindex_t)J;\
-}
-
-#elif defined (__INTEL_COMPILER)
-#define compute_bit2idx(X, I)\
-{\
- unsigned int J;\
- J = _bit_scan_forward (X); \
- I = (bindex_t)J;\
-}
-
-#elif defined(_MSC_VER) && _MSC_VER>=1300
-#define compute_bit2idx(X, I)\
-{\
- unsigned int J;\
- _BitScanForward((DWORD *) &J, X);\
- I = (bindex_t)J;\
-}
-
-#elif USE_BUILTIN_FFS
-#define compute_bit2idx(X, I) I = ffs(X)-1
-
-#else
-#define compute_bit2idx(X, I)\
-{\
- unsigned int Y = X - 1;\
- unsigned int K = Y >> (16-4) & 16;\
- unsigned int N = K; Y >>= K;\
- N += K = Y >> (8-3) & 8; Y >>= K;\
- N += K = Y >> (4-2) & 4; Y >>= K;\
- N += K = Y >> (2-1) & 2; Y >>= K;\
- N += K = Y >> (1-0) & 1; Y >>= K;\
- I = (bindex_t)(N + Y);\
-}
-#endif /* GNUC */
-
-
-/* ----------------------- Runtime Check Support ------------------------- */
-
-/*
- For security, the main invariant is that malloc/free/etc never
- writes to a static address other than malloc_state, unless static
- malloc_state itself has been corrupted, which cannot occur via
- malloc (because of these checks). In essence this means that we
- believe all pointers, sizes, maps etc held in malloc_state, but
- check all of those linked or offsetted from other embedded data
- structures. These checks are interspersed with main code in a way
- that tends to minimize their run-time cost.
-
- When FOOTERS is defined, in addition to range checking, we also
- verify footer fields of inuse chunks, which can be used guarantee
- that the mstate controlling malloc/free is intact. This is a
- streamlined version of the approach described by William Robertson
- et al in "Run-time Detection of Heap-based Overflows" LISA'03
- http://www.usenix.org/events/lisa03/tech/robertson.html The footer
- of an inuse chunk holds the xor of its mstate and a random seed,
- that is checked upon calls to free() and realloc(). This is
- (probablistically) unguessable from outside the program, but can be
- computed by any code successfully malloc'ing any chunk, so does not
- itself provide protection against code that has already broken
- security through some other means. Unlike Robertson et al, we
- always dynamically check addresses of all offset chunks (previous,
- next, etc). This turns out to be cheaper than relying on hashes.
-*/
-
-#if !INSECURE
-/* Check if address a is at least as high as any from MORECORE or MMAP */
-#define ok_address(M, a) ((char*)(a) >= (M)->least_addr)
-/* Check if address of next chunk n is higher than base chunk p */
-#define ok_next(p, n) ((char*)(p) < (char*)(n))
-/* Check if p has inuse status */
-#define ok_inuse(p) is_inuse(p)
-/* Check if p has its pinuse bit on */
-#define ok_pinuse(p) pinuse(p)
-
-#else /* !INSECURE */
-#define ok_address(M, a) (1)
-#define ok_next(b, n) (1)
-#define ok_inuse(p) (1)
-#define ok_pinuse(p) (1)
-#endif /* !INSECURE */
-
-#if (FOOTERS && !INSECURE)
-/* Check if (alleged) mstate m has expected magic field */
-#define ok_magic(M) ((M)->magic == mparams.magic)
-#else /* (FOOTERS && !INSECURE) */
-#define ok_magic(M) (1)
-#endif /* (FOOTERS && !INSECURE) */
-
-
-/* In gcc, use __builtin_expect to minimize impact of checks */
-#if !INSECURE
-#if defined(__GNUC__) && __GNUC__ >= 3
-#define RTCHECK(e) __builtin_expect(e, 1)
-#else /* GNUC */
-#define RTCHECK(e) (e)
-#endif /* GNUC */
-#else /* !INSECURE */
-#define RTCHECK(e) (1)
-#endif /* !INSECURE */
-
-/* macros to set up inuse chunks with or without footers */
-
-#if !FOOTERS
-
-#define mark_inuse_foot(M,p,s)
-
-/* Macros for setting head/foot of non-mmapped chunks */
-
-/* Set cinuse bit and pinuse bit of next chunk */
-#define set_inuse(M,p,s)\
- ((p)->head = (((p)->head & PINUSE_BIT)|s|CINUSE_BIT),\
- ((mchunkptr)(((char*)(p)) + (s)))->head |= PINUSE_BIT)
-
-/* Set cinuse and pinuse of this chunk and pinuse of next chunk */
-#define set_inuse_and_pinuse(M,p,s)\
- ((p)->head = (s|PINUSE_BIT|CINUSE_BIT),\
- ((mchunkptr)(((char*)(p)) + (s)))->head |= PINUSE_BIT)
-
-/* Set size, cinuse and pinuse bit of this chunk */
-#define set_size_and_pinuse_of_inuse_chunk(M, p, s)\
- ((p)->head = (s|PINUSE_BIT|CINUSE_BIT))
-
-#else /* FOOTERS */
-
-/* Set foot of inuse chunk to be xor of mstate and seed */
-#define mark_inuse_foot(M,p,s)\
- (((mchunkptr)((char*)(p) + (s)))->prev_foot = ((size_t)(M) ^ mparams.magic))
-
-#define get_mstate_for(p)\
- ((mstate)(((mchunkptr)((char*)(p) +\
- (chunksize(p))))->prev_foot ^ mparams.magic))
-
-#define set_inuse(M,p,s)\
- ((p)->head = (((p)->head & PINUSE_BIT)|s|CINUSE_BIT),\
- (((mchunkptr)(((char*)(p)) + (s)))->head |= PINUSE_BIT), \
- mark_inuse_foot(M,p,s))
-
-#define set_inuse_and_pinuse(M,p,s)\
- ((p)->head = (s|PINUSE_BIT|CINUSE_BIT),\
- (((mchunkptr)(((char*)(p)) + (s)))->head |= PINUSE_BIT),\
- mark_inuse_foot(M,p,s))
-
-#define set_size_and_pinuse_of_inuse_chunk(M, p, s)\
- ((p)->head = (s|PINUSE_BIT|CINUSE_BIT),\
- mark_inuse_foot(M, p, s))
-
-#endif /* !FOOTERS */
-
-/* ---------------------------- setting mparams -------------------------- */
-
-/* Initialize mparams */
-static int init_mparams(void) {
-#ifdef NEED_GLOBAL_LOCK_INIT
- if (malloc_global_mutex_status <= 0)
- init_malloc_global_mutex();
-#endif
-
- ACQUIRE_MALLOC_GLOBAL_LOCK();
- if (mparams.magic == 0) {
- size_t magic;
- size_t psize;
- size_t gsize;
-
-#ifndef WIN32
- psize = malloc_getpagesize;
- gsize = ((DEFAULT_GRANULARITY != 0)? DEFAULT_GRANULARITY : psize);
-#else /* WIN32 */
- {
- SYSTEM_INFO system_info;
- GetSystemInfo(&system_info);
- psize = system_info.dwPageSize;
- gsize = ((DEFAULT_GRANULARITY != 0)?
- DEFAULT_GRANULARITY : system_info.dwAllocationGranularity);
- }
-#endif /* WIN32 */
-
- /* Sanity-check configuration:
- size_t must be unsigned and as wide as pointer type.
- ints must be at least 4 bytes.
- alignment must be at least 8.
- Alignment, min chunk size, and page size must all be powers of 2.
- */
- if ((sizeof(size_t) != sizeof(char*)) ||
- (MAX_SIZE_T < MIN_CHUNK_SIZE) ||
- (sizeof(int) < 4) ||
- (MALLOC_ALIGNMENT < (size_t)8U) ||
- ((MALLOC_ALIGNMENT & (MALLOC_ALIGNMENT-SIZE_T_ONE)) != 0) ||
- ((MCHUNK_SIZE & (MCHUNK_SIZE-SIZE_T_ONE)) != 0) ||
- ((gsize & (gsize-SIZE_T_ONE)) != 0) ||
- ((psize & (psize-SIZE_T_ONE)) != 0))
- ABORT;
-
- mparams.granularity = gsize;
- mparams.page_size = psize;
- mparams.mmap_threshold = DEFAULT_MMAP_THRESHOLD;
- mparams.trim_threshold = DEFAULT_TRIM_THRESHOLD;
-#if MORECORE_CONTIGUOUS
- mparams.default_mflags = USE_LOCK_BIT|USE_MMAP_BIT;
-#else /* MORECORE_CONTIGUOUS */
- mparams.default_mflags = USE_LOCK_BIT|USE_MMAP_BIT|USE_NONCONTIGUOUS_BIT;
-#endif /* MORECORE_CONTIGUOUS */
-
-#if !ONLY_MSPACES
- /* Set up lock for main malloc area */
- gm->mflags = mparams.default_mflags;
- INITIAL_LOCK(&gm->mutex);
-#endif
-
- {
-#if USE_DEV_RANDOM
- int fd;
- unsigned char buf[sizeof(size_t)];
- /* Try to use /dev/urandom, else fall back on using time */
- if ((fd = open("/dev/urandom", O_RDONLY)) >= 0 &&
- read(fd, buf, sizeof(buf)) == sizeof(buf)) {
- magic = *((size_t *) buf);
- close(fd);
- }
- else
-#endif /* USE_DEV_RANDOM */
-#ifdef WIN32
- magic = (size_t)(GetTickCount() ^ (size_t)0x55555555U);
-#else
- magic = (size_t)(time(0) ^ (size_t)0x55555555U);
-#endif
- magic |= (size_t)8U; /* ensure nonzero */
- magic &= ~(size_t)7U; /* improve chances of fault for bad values */
- mparams.magic = magic;
- }
- }
-
- RELEASE_MALLOC_GLOBAL_LOCK();
- return 1;
-}
-
-/* support for mallopt */
-static int change_mparam(int param_number, int value) {
- size_t val;
- ensure_initialization();
- val = (value == -1)? MAX_SIZE_T : (size_t)value;
- switch(param_number) {
- case M_TRIM_THRESHOLD:
- mparams.trim_threshold = val;
- return 1;
- case M_GRANULARITY:
- if (val >= mparams.page_size && ((val & (val-1)) == 0)) {
- mparams.granularity = val;
- return 1;
- }
- else
- return 0;
- case M_MMAP_THRESHOLD:
- mparams.mmap_threshold = val;
- return 1;
- default:
- return 0;
- }
-}
-
-#if DEBUG
-/* ------------------------- Debugging Support --------------------------- */
-
-/* Check properties of any chunk, whether free, inuse, mmapped etc */
-static void do_check_any_chunk(mstate m, mchunkptr p) {
- assert((is_aligned(chunk2mem(p))) || (p->head == FENCEPOST_HEAD));
- assert(ok_address(m, p));
-}
-
-/* Check properties of top chunk */
-static void do_check_top_chunk(mstate m, mchunkptr p) {
- msegmentptr sp = segment_holding(m, (char*)p);
- size_t sz = p->head & ~INUSE_BITS; /* third-lowest bit can be set! */
- assert(sp != 0);
- assert((is_aligned(chunk2mem(p))) || (p->head == FENCEPOST_HEAD));
- assert(ok_address(m, p));
- assert(sz == m->topsize);
- assert(sz > 0);
- assert(sz == ((sp->base + sp->size) - (char*)p) - TOP_FOOT_SIZE);
- assert(pinuse(p));
- assert(!pinuse(chunk_plus_offset(p, sz)));
-}
-
-/* Check properties of (inuse) mmapped chunks */
-static void do_check_mmapped_chunk(mstate m, mchunkptr p) {
- size_t sz = chunksize(p);
- size_t len = (sz + (p->prev_foot) + MMAP_FOOT_PAD);
- assert(is_mmapped(p));
- assert(use_mmap(m));
- assert((is_aligned(chunk2mem(p))) || (p->head == FENCEPOST_HEAD));
- assert(ok_address(m, p));
- assert(!is_small(sz));
- assert((len & (mparams.page_size-SIZE_T_ONE)) == 0);
- assert(chunk_plus_offset(p, sz)->head == FENCEPOST_HEAD);
- assert(chunk_plus_offset(p, sz+SIZE_T_SIZE)->head == 0);
-}
-
-/* Check properties of inuse chunks */
-static void do_check_inuse_chunk(mstate m, mchunkptr p) {
- do_check_any_chunk(m, p);
- assert(is_inuse(p));
- assert(next_pinuse(p));
- /* If not pinuse and not mmapped, previous chunk has OK offset */
- assert(is_mmapped(p) || pinuse(p) || next_chunk(prev_chunk(p)) == p);
- if (is_mmapped(p))
- do_check_mmapped_chunk(m, p);
-}
-
-/* Check properties of free chunks */
-static void do_check_free_chunk(mstate m, mchunkptr p) {
- size_t sz = chunksize(p);
- mchunkptr next = chunk_plus_offset(p, sz);
- do_check_any_chunk(m, p);
- assert(!is_inuse(p));
- assert(!next_pinuse(p));
- assert (!is_mmapped(p));
- if (p != m->dv && p != m->top) {
- if (sz >= MIN_CHUNK_SIZE) {
- assert((sz & CHUNK_ALIGN_MASK) == 0);
- assert(is_aligned(chunk2mem(p)));
- assert(next->prev_foot == sz);
- assert(pinuse(p));
- assert (next == m->top || is_inuse(next));
- assert(p->fd->bk == p);
- assert(p->bk->fd == p);
- }
- else /* markers are always of size SIZE_T_SIZE */
- assert(sz == SIZE_T_SIZE);
- }
-}
-
-/* Check properties of malloced chunks at the point they are malloced */
-static void do_check_malloced_chunk(mstate m, void* mem, size_t s) {
- if (mem != 0) {
- mchunkptr p = mem2chunk(mem);
- size_t sz = p->head & ~INUSE_BITS;
- do_check_inuse_chunk(m, p);
- assert((sz & CHUNK_ALIGN_MASK) == 0);
- assert(sz >= MIN_CHUNK_SIZE);
- assert(sz >= s);
- /* unless mmapped, size is less than MIN_CHUNK_SIZE more than request */
- assert(is_mmapped(p) || sz < (s + MIN_CHUNK_SIZE));
- }
-}
-
-/* Check a tree and its subtrees. */
-static void do_check_tree(mstate m, tchunkptr t) {
- tchunkptr head = 0;
- tchunkptr u = t;
- bindex_t tindex = t->index;
- size_t tsize = chunksize(t);
- bindex_t idx;
- compute_tree_index(tsize, idx);
- assert(tindex == idx);
- assert(tsize >= MIN_LARGE_SIZE);
- assert(tsize >= minsize_for_tree_index(idx));
- assert((idx == NTREEBINS-1) || (tsize < minsize_for_tree_index((idx+1))));
-
- do { /* traverse through chain of same-sized nodes */
- do_check_any_chunk(m, ((mchunkptr)u));
- assert(u->index == tindex);
- assert(chunksize(u) == tsize);
- assert(!is_inuse(u));
- assert(!next_pinuse(u));
- assert(u->fd->bk == u);
- assert(u->bk->fd == u);
- if (u->parent == 0) {
- assert(u->child[0] == 0);
- assert(u->child[1] == 0);
- }
- else {
- assert(head == 0); /* only one node on chain has parent */
- head = u;
- assert(u->parent != u);
- assert (u->parent->child[0] == u ||
- u->parent->child[1] == u ||
- *((tbinptr*)(u->parent)) == u);
- if (u->child[0] != 0) {
- assert(u->child[0]->parent == u);
- assert(u->child[0] != u);
- do_check_tree(m, u->child[0]);
- }
- if (u->child[1] != 0) {
- assert(u->child[1]->parent == u);
- assert(u->child[1] != u);
- do_check_tree(m, u->child[1]);
- }
- if (u->child[0] != 0 && u->child[1] != 0) {
- assert(chunksize(u->child[0]) < chunksize(u->child[1]));
- }
- }
- u = u->fd;
- } while (u != t);
- assert(head != 0);
-}
-
-/* Check all the chunks in a treebin. */
-static void do_check_treebin(mstate m, bindex_t i) {
- tbinptr* tb = treebin_at(m, i);
- tchunkptr t = *tb;
- int empty = (m->treemap & (1U << i)) == 0;
- if (t == 0)
- assert(empty);
- if (!empty)
- do_check_tree(m, t);
-}
-
-/* Check all the chunks in a smallbin. */
-static void do_check_smallbin(mstate m, bindex_t i) {
- sbinptr b = smallbin_at(m, i);
- mchunkptr p = b->bk;
- unsigned int empty = (m->smallmap & (1U << i)) == 0;
- if (p == b)
- assert(empty);
- if (!empty) {
- for (; p != b; p = p->bk) {
- size_t size = chunksize(p);
- mchunkptr q;
- /* each chunk claims to be free */
- do_check_free_chunk(m, p);
- /* chunk belongs in bin */
- assert(small_index(size) == i);
- assert(p->bk == b || chunksize(p->bk) == chunksize(p));
- /* chunk is followed by an inuse chunk */
- q = next_chunk(p);
- if (q->head != FENCEPOST_HEAD)
- do_check_inuse_chunk(m, q);
- }
- }
-}
-
-/* Find x in a bin. Used in other check functions. */
-static int bin_find(mstate m, mchunkptr x) {
- size_t size = chunksize(x);
- if (is_small(size)) {
- bindex_t sidx = small_index(size);
- sbinptr b = smallbin_at(m, sidx);
- if (smallmap_is_marked(m, sidx)) {
- mchunkptr p = b;
- do {
- if (p == x)
- return 1;
- } while ((p = p->fd) != b);
- }
- }
- else {
- bindex_t tidx;
- compute_tree_index(size, tidx);
- if (treemap_is_marked(m, tidx)) {
- tchunkptr t = *treebin_at(m, tidx);
- size_t sizebits = size << leftshift_for_tree_index(tidx);
- while (t != 0 && chunksize(t) != size) {
- t = t->child[(sizebits >> (SIZE_T_BITSIZE-SIZE_T_ONE)) & 1];
- sizebits <<= 1;
- }
- if (t != 0) {
- tchunkptr u = t;
- do {
- if (u == (tchunkptr)x)
- return 1;
- } while ((u = u->fd) != t);
- }
- }
- }
- return 0;
-}
-
-/* Traverse each chunk and check it; return total */
-static size_t traverse_and_check(mstate m) {
- size_t sum = 0;
- if (is_initialized(m)) {
- msegmentptr s = &m->seg;
- sum += m->topsize + TOP_FOOT_SIZE;
- while (s != 0) {
- mchunkptr q = align_as_chunk(s->base);
- mchunkptr lastq = 0;
- assert(pinuse(q));
- while (segment_holds(s, q) &&
- q != m->top && q->head != FENCEPOST_HEAD) {
- sum += chunksize(q);
- if (is_inuse(q)) {
- assert(!bin_find(m, q));
- do_check_inuse_chunk(m, q);
- }
- else {
- assert(q == m->dv || bin_find(m, q));
- assert(lastq == 0 || is_inuse(lastq)); /* Not 2 consecutive free */
- do_check_free_chunk(m, q);
- }
- lastq = q;
- q = next_chunk(q);
- }
- s = s->next;
- }
- }
- return sum;
-}
-
-/* Check all properties of malloc_state. */
-static void do_check_malloc_state(mstate m) {
- bindex_t i;
- size_t total;
- /* check bins */
- for (i = 0; i < NSMALLBINS; ++i)
- do_check_smallbin(m, i);
- for (i = 0; i < NTREEBINS; ++i)
- do_check_treebin(m, i);
-
- if (m->dvsize != 0) { /* check dv chunk */
- do_check_any_chunk(m, m->dv);
- assert(m->dvsize == chunksize(m->dv));
- assert(m->dvsize >= MIN_CHUNK_SIZE);
- assert(bin_find(m, m->dv) == 0);
- }
-
- if (m->top != 0) { /* check top chunk */
- do_check_top_chunk(m, m->top);
- /*assert(m->topsize == chunksize(m->top)); redundant */
- assert(m->topsize > 0);
- assert(bin_find(m, m->top) == 0);
- }
-
- total = traverse_and_check(m);
- assert(total <= m->footprint);
- assert(m->footprint <= m->max_footprint);
-}
-#endif /* DEBUG */
-
-/* ----------------------------- statistics ------------------------------ */
-
-#if !NO_MALLINFO
-static struct mallinfo internal_mallinfo(mstate m) {
- struct mallinfo nm = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
- ensure_initialization();
- if (!PREACTION(m)) {
- check_malloc_state(m);
- if (is_initialized(m)) {
- size_t nfree = SIZE_T_ONE; /* top always free */
- size_t mfree = m->topsize + TOP_FOOT_SIZE;
- size_t sum = mfree;
- msegmentptr s = &m->seg;
- while (s != 0) {
- mchunkptr q = align_as_chunk(s->base);
- while (segment_holds(s, q) &&
- q != m->top && q->head != FENCEPOST_HEAD) {
- size_t sz = chunksize(q);
- sum += sz;
- if (!is_inuse(q)) {
- mfree += sz;
- ++nfree;
- }
- q = next_chunk(q);
- }
- s = s->next;
- }
-
- nm.arena = sum;
- nm.ordblks = nfree;
- nm.hblkhd = m->footprint - sum;
- nm.usmblks = m->max_footprint;
- nm.uordblks = m->footprint - mfree;
- nm.fordblks = mfree;
- nm.keepcost = m->topsize;
- }
-
- POSTACTION(m);
- }
- return nm;
-}
-#endif /* !NO_MALLINFO */
-
-static void internal_malloc_stats(mstate m) {
- ensure_initialization();
- if (!PREACTION(m)) {
- size_t maxfp = 0;
- size_t fp = 0;
- size_t used = 0;
- check_malloc_state(m);
- if (is_initialized(m)) {
- msegmentptr s = &m->seg;
- maxfp = m->max_footprint;
- fp = m->footprint;
- used = fp - (m->topsize + TOP_FOOT_SIZE);
-
- while (s != 0) {
- mchunkptr q = align_as_chunk(s->base);
- while (segment_holds(s, q) &&
- q != m->top && q->head != FENCEPOST_HEAD) {
- if (!is_inuse(q))
- used -= chunksize(q);
- q = next_chunk(q);
- }
- s = s->next;
- }
- }
-
- fprintf(stderr, "max system bytes = %10lu\n", (unsigned long)(maxfp));
- fprintf(stderr, "system bytes = %10lu\n", (unsigned long)(fp));
- fprintf(stderr, "in use bytes = %10lu\n", (unsigned long)(used));
-
- POSTACTION(m);
- }
-}
-
-/* ----------------------- Operations on smallbins ----------------------- */
-
-/*
- Various forms of linking and unlinking are defined as macros. Even
- the ones for trees, which are very long but have very short typical
- paths. This is ugly but reduces reliance on inlining support of
- compilers.
-*/
-
-/* Link a free chunk into a smallbin */
-#define insert_small_chunk(M, P, S) {\
- bindex_t I = small_index(S);\
- mchunkptr B = smallbin_at(M, I);\
- mchunkptr F = B;\
- assert(S >= MIN_CHUNK_SIZE);\
- if (!smallmap_is_marked(M, I))\
- mark_smallmap(M, I);\
- else if (RTCHECK(ok_address(M, B->fd)))\
- F = B->fd;\
- else {\
- CORRUPTION_ERROR_ACTION(M);\
- }\
- B->fd = P;\
- F->bk = P;\
- P->fd = F;\
- P->bk = B;\
-}
-
-/* Unlink a chunk from a smallbin */
-#define unlink_small_chunk(M, P, S) {\
- mchunkptr F = P->fd;\
- mchunkptr B = P->bk;\
- bindex_t I = small_index(S);\
- assert(P != B);\
- assert(P != F);\
- assert(chunksize(P) == small_index2size(I));\
- if (F == B)\
- clear_smallmap(M, I);\
- else if (RTCHECK((F == smallbin_at(M,I) || ok_address(M, F)) &&\
- (B == smallbin_at(M,I) || ok_address(M, B)))) {\
- F->bk = B;\
- B->fd = F;\
- }\
- else {\
- CORRUPTION_ERROR_ACTION(M);\
- }\
-}
-
-/* Unlink the first chunk from a smallbin */
-#define unlink_first_small_chunk(M, B, P, I) {\
- mchunkptr F = P->fd;\
- assert(P != B);\
- assert(P != F);\
- assert(chunksize(P) == small_index2size(I));\
- if (B == F)\
- clear_smallmap(M, I);\
- else if (RTCHECK(ok_address(M, F))) {\
- B->fd = F;\
- F->bk = B;\
- }\
- else {\
- CORRUPTION_ERROR_ACTION(M);\
- }\
-}
-
-
-
-/* Replace dv node, binning the old one */
-/* Used only when dvsize known to be small */
-#define replace_dv(M, P, S) {\
- size_t DVS = M->dvsize;\
- if (DVS != 0) {\
- mchunkptr DV = M->dv;\
- assert(is_small(DVS));\
- insert_small_chunk(M, DV, DVS);\
- }\
- M->dvsize = S;\
- M->dv = P;\
-}
-
-/* ------------------------- Operations on trees ------------------------- */
-
-/* Insert chunk into tree */
-#define insert_large_chunk(M, X, S) {\
- tbinptr* H;\
- bindex_t I;\
- compute_tree_index(S, I);\
- H = treebin_at(M, I);\
- X->index = I;\
- X->child[0] = X->child[1] = 0;\
- if (!treemap_is_marked(M, I)) {\
- mark_treemap(M, I);\
- *H = X;\
- X->parent = (tchunkptr)H;\
- X->fd = X->bk = X;\
- }\
- else {\
- tchunkptr T = *H;\
- size_t K = S << leftshift_for_tree_index(I);\
- for (;;) {\
- if (chunksize(T) != S) {\
- tchunkptr* C = &(T->child[(K >> (SIZE_T_BITSIZE-SIZE_T_ONE)) & 1]);\
- K <<= 1;\
- if (*C != 0)\
- T = *C;\
- else if (RTCHECK(ok_address(M, C))) {\
- *C = X;\
- X->parent = T;\
- X->fd = X->bk = X;\
- break;\
- }\
- else {\
- CORRUPTION_ERROR_ACTION(M);\
- break;\
- }\
- }\
- else {\
- tchunkptr F = T->fd;\
- if (RTCHECK(ok_address(M, T) && ok_address(M, F))) {\
- T->fd = F->bk = X;\
- X->fd = F;\
- X->bk = T;\
- X->parent = 0;\
- break;\
- }\
- else {\
- CORRUPTION_ERROR_ACTION(M);\
- break;\
- }\
- }\
- }\
- }\
-}
-
-/*
- Unlink steps:
-
- 1. If x is a chained node, unlink it from its same-sized fd/bk links
- and choose its bk node as its replacement.
- 2. If x was the last node of its size, but not a leaf node, it must
- be replaced with a leaf node (not merely one with an open left or
- right), to make sure that lefts and rights of descendents
- correspond properly to bit masks. We use the rightmost descendent
- of x. We could use any other leaf, but this is easy to locate and
- tends to counteract removal of leftmosts elsewhere, and so keeps
- paths shorter than minimally guaranteed. This doesn't loop much
- because on average a node in a tree is near the bottom.
- 3. If x is the base of a chain (i.e., has parent links) relink
- x's parent and children to x's replacement (or null if none).
-*/
-
-#define unlink_large_chunk(M, X) {\
- tchunkptr XP = X->parent;\
- tchunkptr R;\
- if (X->bk != X) {\
- tchunkptr F = X->fd;\
- R = X->bk;\
- if (RTCHECK(ok_address(M, F))) {\
- F->bk = R;\
- R->fd = F;\
- }\
- else {\
- CORRUPTION_ERROR_ACTION(M);\
- }\
- }\
- else {\
- tchunkptr* RP;\
- if (((R = *(RP = &(X->child[1]))) != 0) ||\
- ((R = *(RP = &(X->child[0]))) != 0)) {\
- tchunkptr* CP;\
- while ((*(CP = &(R->child[1])) != 0) ||\
- (*(CP = &(R->child[0])) != 0)) {\
- R = *(RP = CP);\
- }\
- if (RTCHECK(ok_address(M, RP)))\
- *RP = 0;\
- else {\
- CORRUPTION_ERROR_ACTION(M);\
- }\
- }\
- }\
- if (XP != 0) {\
- tbinptr* H = treebin_at(M, X->index);\
- if (X == *H) {\
- if ((*H = R) == 0) \
- clear_treemap(M, X->index);\
- }\
- else if (RTCHECK(ok_address(M, XP))) {\
- if (XP->child[0] == X) \
- XP->child[0] = R;\
- else \
- XP->child[1] = R;\
- }\
- else\
- CORRUPTION_ERROR_ACTION(M);\
- if (R != 0) {\
- if (RTCHECK(ok_address(M, R))) {\
- tchunkptr C0, C1;\
- R->parent = XP;\
- if ((C0 = X->child[0]) != 0) {\
- if (RTCHECK(ok_address(M, C0))) {\
- R->child[0] = C0;\
- C0->parent = R;\
- }\
- else\
- CORRUPTION_ERROR_ACTION(M);\
- }\
- if ((C1 = X->child[1]) != 0) {\
- if (RTCHECK(ok_address(M, C1))) {\
- R->child[1] = C1;\
- C1->parent = R;\
- }\
- else\
- CORRUPTION_ERROR_ACTION(M);\
- }\
- }\
- else\
- CORRUPTION_ERROR_ACTION(M);\
- }\
- }\
-}
-
-/* Relays to large vs small bin operations */
-
-#define insert_chunk(M, P, S)\
- if (is_small(S)) insert_small_chunk(M, P, S)\
- else { tchunkptr TP = (tchunkptr)(P); insert_large_chunk(M, TP, S); }
-
-#define unlink_chunk(M, P, S)\
- if (is_small(S)) unlink_small_chunk(M, P, S)\
- else { tchunkptr TP = (tchunkptr)(P); unlink_large_chunk(M, TP); }
-
-
-/* Relays to internal calls to malloc/free from realloc, memalign etc */
-
-#if ONLY_MSPACES
-#define internal_malloc(m, b) mspace_malloc(m, b)
-#define internal_free(m, mem) mspace_free(m,mem);
-#else /* ONLY_MSPACES */
-#if MSPACES
-#define internal_malloc(m, b)\
- (m == gm)? dlmalloc(b) : mspace_malloc(m, b)
-#define internal_free(m, mem)\
- if (m == gm) dlfree(mem); else mspace_free(m,mem);
-#else /* MSPACES */
-#define internal_malloc(m, b) dlmalloc(b)
-#define internal_free(m, mem) dlfree(mem)
-#endif /* MSPACES */
-#endif /* ONLY_MSPACES */
-
-/* ----------------------- Direct-mmapping chunks ----------------------- */
-
-/*
- Directly mmapped chunks are set up with an offset to the start of
- the mmapped region stored in the prev_foot field of the chunk. This
- allows reconstruction of the required argument to MUNMAP when freed,
- and also allows adjustment of the returned chunk to meet alignment
- requirements (especially in memalign).
-*/
-
-/* Malloc using mmap */
-static void* mmap_alloc(mstate m, size_t nb) {
- size_t mmsize = mmap_align(nb + SIX_SIZE_T_SIZES + CHUNK_ALIGN_MASK);
- if (mmsize > nb) { /* Check for wrap around 0 */
- char* mm = (char*)(CALL_DIRECT_MMAP(mmsize));
- if (mm != CMFAIL) {
- size_t offset = align_offset(chunk2mem(mm));
- size_t psize = mmsize - offset - MMAP_FOOT_PAD;
- mchunkptr p = (mchunkptr)(mm + offset);
- p->prev_foot = offset;
- p->head = psize;
- mark_inuse_foot(m, p, psize);
- chunk_plus_offset(p, psize)->head = FENCEPOST_HEAD;
- chunk_plus_offset(p, psize+SIZE_T_SIZE)->head = 0;
-
- if (m->least_addr == 0 || mm < m->least_addr)
- m->least_addr = mm;
- if ((m->footprint += mmsize) > m->max_footprint)
- m->max_footprint = m->footprint;
- assert(is_aligned(chunk2mem(p)));
- check_mmapped_chunk(m, p);
- return chunk2mem(p);
- }
- }
- return 0;
-}
-
-/* Realloc using mmap */
-static mchunkptr mmap_resize(mstate m, mchunkptr oldp, size_t nb) {
- size_t oldsize = chunksize(oldp);
- if (is_small(nb)) /* Can't shrink mmap regions below small size */
- return 0;
- /* Keep old chunk if big enough but not too big */
- if (oldsize >= nb + SIZE_T_SIZE &&
- (oldsize - nb) <= (mparams.granularity << 1))
- return oldp;
- else {
- size_t offset = oldp->prev_foot;
- size_t oldmmsize = oldsize + offset + MMAP_FOOT_PAD;
- size_t newmmsize = mmap_align(nb + SIX_SIZE_T_SIZES + CHUNK_ALIGN_MASK);
- char* cp = (char*)CALL_MREMAP((char*)oldp - offset,
- oldmmsize, newmmsize, 1);
- if (cp != CMFAIL) {
- mchunkptr newp = (mchunkptr)(cp + offset);
- size_t psize = newmmsize - offset - MMAP_FOOT_PAD;
- newp->head = psize;
- mark_inuse_foot(m, newp, psize);
- chunk_plus_offset(newp, psize)->head = FENCEPOST_HEAD;
- chunk_plus_offset(newp, psize+SIZE_T_SIZE)->head = 0;
-
- if (cp < m->least_addr)
- m->least_addr = cp;
- if ((m->footprint += newmmsize - oldmmsize) > m->max_footprint)
- m->max_footprint = m->footprint;
- check_mmapped_chunk(m, newp);
- return newp;
- }
- }
- return 0;
-}
-
-/* -------------------------- mspace management -------------------------- */
-
-/* Initialize top chunk and its size */
-static void init_top(mstate m, mchunkptr p, size_t psize) {
- /* Ensure alignment */
- size_t offset = align_offset(chunk2mem(p));
- p = (mchunkptr)((char*)p + offset);
- psize -= offset;
-
- m->top = p;
- m->topsize = psize;
- p->head = psize | PINUSE_BIT;
- /* set size of fake trailing chunk holding overhead space only once */
- chunk_plus_offset(p, psize)->head = TOP_FOOT_SIZE;
- m->trim_check = mparams.trim_threshold; /* reset on each update */
-}
-
-/* Initialize bins for a new mstate that is otherwise zeroed out */
-static void init_bins(mstate m) {
- /* Establish circular links for smallbins */
- bindex_t i;
- for (i = 0; i < NSMALLBINS; ++i) {
- sbinptr bin = smallbin_at(m,i);
- bin->fd = bin->bk = bin;
- }
-}
-
-#if PROCEED_ON_ERROR
-
-/* default corruption action */
-static void reset_on_error(mstate m) {
- int i;
- ++malloc_corruption_error_count;
- /* Reinitialize fields to forget about all memory */
- m->smallbins = m->treebins = 0;
- m->dvsize = m->topsize = 0;
- m->seg.base = 0;
- m->seg.size = 0;
- m->seg.next = 0;
- m->top = m->dv = 0;
- for (i = 0; i < NTREEBINS; ++i)
- *treebin_at(m, i) = 0;
- init_bins(m);
-}
-#endif /* PROCEED_ON_ERROR */
-
-/* Allocate chunk and prepend remainder with chunk in successor base. */
-static void* prepend_alloc(mstate m, char* newbase, char* oldbase,
- size_t nb) {
- mchunkptr p = align_as_chunk(newbase);
- mchunkptr oldfirst = align_as_chunk(oldbase);
- size_t psize = (char*)oldfirst - (char*)p;
- mchunkptr q = chunk_plus_offset(p, nb);
- size_t qsize = psize - nb;
- set_size_and_pinuse_of_inuse_chunk(m, p, nb);
-
- assert((char*)oldfirst > (char*)q);
- assert(pinuse(oldfirst));
- assert(qsize >= MIN_CHUNK_SIZE);
-
- /* consolidate remainder with first chunk of old base */
- if (oldfirst == m->top) {
- size_t tsize = m->topsize += qsize;
- m->top = q;
- q->head = tsize | PINUSE_BIT;
- check_top_chunk(m, q);
- }
- else if (oldfirst == m->dv) {
- size_t dsize = m->dvsize += qsize;
- m->dv = q;
- set_size_and_pinuse_of_free_chunk(q, dsize);
- }
- else {
- if (!is_inuse(oldfirst)) {
- size_t nsize = chunksize(oldfirst);
- unlink_chunk(m, oldfirst, nsize);
- oldfirst = chunk_plus_offset(oldfirst, nsize);
- qsize += nsize;
- }
- set_free_with_pinuse(q, qsize, oldfirst);
- insert_chunk(m, q, qsize);
- check_free_chunk(m, q);
- }
-
- check_malloced_chunk(m, chunk2mem(p), nb);
- return chunk2mem(p);
-}
-
-/* Add a segment to hold a new noncontiguous region */
-static void add_segment(mstate m, char* tbase, size_t tsize, flag_t mmapped) {
- /* Determine locations and sizes of segment, fenceposts, old top */
- char* old_top = (char*)m->top;
- msegmentptr oldsp = segment_holding(m, old_top);
- char* old_end = oldsp->base + oldsp->size;
- size_t ssize = pad_request(sizeof(struct malloc_segment));
- char* rawsp = old_end - (ssize + FOUR_SIZE_T_SIZES + CHUNK_ALIGN_MASK);
- size_t offset = align_offset(chunk2mem(rawsp));
- char* asp = rawsp + offset;
- char* csp = (asp < (old_top + MIN_CHUNK_SIZE))? old_top : asp;
- mchunkptr sp = (mchunkptr)csp;
- msegmentptr ss = (msegmentptr)(chunk2mem(sp));
- mchunkptr tnext = chunk_plus_offset(sp, ssize);
- mchunkptr p = tnext;
- int nfences = 0;
-
- /* reset top to new space */
- init_top(m, (mchunkptr)tbase, tsize - TOP_FOOT_SIZE);
-
- /* Set up segment record */
- assert(is_aligned(ss));
- set_size_and_pinuse_of_inuse_chunk(m, sp, ssize);
- *ss = m->seg; /* Push current record */
- m->seg.base = tbase;
- m->seg.size = tsize;
- m->seg.sflags = mmapped;
- m->seg.next = ss;
-
- /* Insert trailing fenceposts */
- for (;;) {
- mchunkptr nextp = chunk_plus_offset(p, SIZE_T_SIZE);
- p->head = FENCEPOST_HEAD;
- ++nfences;
- if ((char*)(&(nextp->head)) < old_end)
- p = nextp;
- else
- break;
- }
- assert(nfences >= 2);
-
- /* Insert the rest of old top into a bin as an ordinary free chunk */
- if (csp != old_top) {
- mchunkptr q = (mchunkptr)old_top;
- size_t psize = csp - old_top;
- mchunkptr tn = chunk_plus_offset(q, psize);
- set_free_with_pinuse(q, psize, tn);
- insert_chunk(m, q, psize);
- }
-
- check_top_chunk(m, m->top);
-}
-
-/* -------------------------- System allocation -------------------------- */
-
-/* Get memory from system using MORECORE or MMAP */
-static void* sys_alloc(mstate m, size_t nb) {
- char* tbase = CMFAIL;
- size_t tsize = 0;
- flag_t mmap_flag = 0;
-
- ensure_initialization();
-
- /* Directly map large chunks, but only if already initialized */
- if (use_mmap(m) && nb >= mparams.mmap_threshold && m->topsize != 0) {
- void* mem = mmap_alloc(m, nb);
- if (mem != 0)
- return mem;
- }
-
- /*
- Try getting memory in any of three ways (in most-preferred to
- least-preferred order):
- 1. A call to MORECORE that can normally contiguously extend memory.
- (disabled if not MORECORE_CONTIGUOUS or not HAVE_MORECORE or
- or main space is mmapped or a previous contiguous call failed)
- 2. A call to MMAP new space (disabled if not HAVE_MMAP).
- Note that under the default settings, if MORECORE is unable to
- fulfill a request, and HAVE_MMAP is true, then mmap is
- used as a noncontiguous system allocator. This is a useful backup
- strategy for systems with holes in address spaces -- in this case
- sbrk cannot contiguously expand the heap, but mmap may be able to
- find space.
- 3. A call to MORECORE that cannot usually contiguously extend memory.
- (disabled if not HAVE_MORECORE)
-
- In all cases, we need to request enough bytes from system to ensure
- we can malloc nb bytes upon success, so pad with enough space for
- top_foot, plus alignment-pad to make sure we don't lose bytes if
- not on boundary, and round this up to a granularity unit.
- */
-
- if (MORECORE_CONTIGUOUS && !use_noncontiguous(m)) {
- char* br = CMFAIL;
- msegmentptr ss = (m->top == 0)? 0 : segment_holding(m, (char*)m->top);
- size_t asize = 0;
- ACQUIRE_MALLOC_GLOBAL_LOCK();
-
- if (ss == 0) { /* First time through or recovery */
- char* base = (char*)CALL_MORECORE(0);
- if (base != CMFAIL) {
- asize = granularity_align(nb + SYS_ALLOC_PADDING);
- /* Adjust to end on a page boundary */
- if (!is_page_aligned(base))
- asize += (page_align((size_t)base) - (size_t)base);
- /* Can't call MORECORE if size is negative when treated as signed */
- if (asize < HALF_MAX_SIZE_T &&
- (br = (char*)(CALL_MORECORE(asize))) == base) {
- tbase = base;
- tsize = asize;
- }
- }
- }
- else {
- /* Subtract out existing available top space from MORECORE request. */
- asize = granularity_align(nb - m->topsize + SYS_ALLOC_PADDING);
- /* Use mem here only if it did continuously extend old space */
- if (asize < HALF_MAX_SIZE_T &&
- (br = (char*)(CALL_MORECORE(asize))) == ss->base+ss->size) {
- tbase = br;
- tsize = asize;
- }
- }
-
- if (tbase == CMFAIL) { /* Cope with partial failure */
- if (br != CMFAIL) { /* Try to use/extend the space we did get */
- if (asize < HALF_MAX_SIZE_T &&
- asize < nb + SYS_ALLOC_PADDING) {
- size_t esize = granularity_align(nb + SYS_ALLOC_PADDING - asize);
- if (esize < HALF_MAX_SIZE_T) {
- char* end = (char*)CALL_MORECORE(esize);
- if (end != CMFAIL)
- asize += esize;
- else { /* Can't use; try to release */
- (void) CALL_MORECORE(-asize);
- br = CMFAIL;
- }
- }
- }
- }
- if (br != CMFAIL) { /* Use the space we did get */
- tbase = br;
- tsize = asize;
- }
- else
- disable_contiguous(m); /* Don't try contiguous path in the future */
- }
-
- RELEASE_MALLOC_GLOBAL_LOCK();
- }
-
- if (HAVE_MMAP && tbase == CMFAIL) { /* Try MMAP */
- size_t rsize = granularity_align(nb + SYS_ALLOC_PADDING);
- if (rsize > nb) { /* Fail if wraps around zero */
- char* mp = (char*)(CALL_MMAP(rsize));
- if (mp != CMFAIL) {
- tbase = mp;
- tsize = rsize;
- mmap_flag = USE_MMAP_BIT;
- }
- }
- }
-
- if (HAVE_MORECORE && tbase == CMFAIL) { /* Try noncontiguous MORECORE */
- size_t asize = granularity_align(nb + SYS_ALLOC_PADDING);
- if (asize < HALF_MAX_SIZE_T) {
- char* br = CMFAIL;
- char* end = CMFAIL;
- ACQUIRE_MALLOC_GLOBAL_LOCK();
- br = (char*)(CALL_MORECORE(asize));
- end = (char*)(CALL_MORECORE(0));
- RELEASE_MALLOC_GLOBAL_LOCK();
- if (br != CMFAIL && end != CMFAIL && br < end) {
- size_t ssize = end - br;
- if (ssize > nb + TOP_FOOT_SIZE) {
- tbase = br;
- tsize = ssize;
- }
- }
- }
- }
-
- if (tbase != CMFAIL) {
-
- if ((m->footprint += tsize) > m->max_footprint)
- m->max_footprint = m->footprint;
-
- if (!is_initialized(m)) { /* first-time initialization */
- if (m->least_addr == 0 || tbase < m->least_addr)
- m->least_addr = tbase;
- m->seg.base = tbase;
- m->seg.size = tsize;
- m->seg.sflags = mmap_flag;
- m->magic = mparams.magic;
- m->release_checks = MAX_RELEASE_CHECK_RATE;
- init_bins(m);
-#if !ONLY_MSPACES
- if (is_global(m))
- init_top(m, (mchunkptr)tbase, tsize - TOP_FOOT_SIZE);
- else
-#endif
- {
- /* Offset top by embedded malloc_state */
- mchunkptr mn = next_chunk(mem2chunk(m));
- init_top(m, mn, (size_t)((tbase + tsize) - (char*)mn) -TOP_FOOT_SIZE);
- }
- }
-
- else {
- /* Try to merge with an existing segment */
- msegmentptr sp = &m->seg;
- /* Only consider most recent segment if traversal suppressed */
- while (sp != 0 && tbase != sp->base + sp->size)
- sp = (NO_SEGMENT_TRAVERSAL) ? 0 : sp->next;
- if (sp != 0 &&
- !is_extern_segment(sp) &&
- (sp->sflags & USE_MMAP_BIT) == mmap_flag &&
- segment_holds(sp, m->top)) { /* append */
- sp->size += tsize;
- init_top(m, m->top, m->topsize + tsize);
- }
- else {
- if (tbase < m->least_addr)
- m->least_addr = tbase;
- sp = &m->seg;
- while (sp != 0 && sp->base != tbase + tsize)
- sp = (NO_SEGMENT_TRAVERSAL) ? 0 : sp->next;
- if (sp != 0 &&
- !is_extern_segment(sp) &&
- (sp->sflags & USE_MMAP_BIT) == mmap_flag) {
- char* oldbase = sp->base;
- sp->base = tbase;
- sp->size += tsize;
- return prepend_alloc(m, tbase, oldbase, nb);
- }
- else
- add_segment(m, tbase, tsize, mmap_flag);
- }
- }
-
- if (nb < m->topsize) { /* Allocate from new or extended top space */
- size_t rsize = m->topsize -= nb;
- mchunkptr p = m->top;
- mchunkptr r = m->top = chunk_plus_offset(p, nb);
- r->head = rsize | PINUSE_BIT;
- set_size_and_pinuse_of_inuse_chunk(m, p, nb);
- check_top_chunk(m, m->top);
- check_malloced_chunk(m, chunk2mem(p), nb);
- return chunk2mem(p);
- }
- }
-
- MALLOC_FAILURE_ACTION;
- return 0;
-}
-
-/* ----------------------- system deallocation -------------------------- */
-
-/* Unmap and unlink any mmapped segments that don't contain used chunks */
-static size_t release_unused_segments(mstate m) {
- size_t released = 0;
- int nsegs = 0;
- msegmentptr pred = &m->seg;
- msegmentptr sp = pred->next;
- while (sp != 0) {
- char* base = sp->base;
- size_t size = sp->size;
- msegmentptr next = sp->next;
- ++nsegs;
- if (is_mmapped_segment(sp) && !is_extern_segment(sp)) {
- mchunkptr p = align_as_chunk(base);
- size_t psize = chunksize(p);
- /* Can unmap if first chunk holds entire segment and not pinned */
- if (!is_inuse(p) && (char*)p + psize >= base + size - TOP_FOOT_SIZE) {
- tchunkptr tp = (tchunkptr)p;
- assert(segment_holds(sp, (char*)sp));
- if (p == m->dv) {
- m->dv = 0;
- m->dvsize = 0;
- }
- else {
- unlink_large_chunk(m, tp);
- }
- if (CALL_MUNMAP(base, size) == 0) {
- released += size;
- m->footprint -= size;
- /* unlink obsoleted record */
- sp = pred;
- sp->next = next;
- }
- else { /* back out if cannot unmap */
- insert_large_chunk(m, tp, psize);
- }
- }
- }
- if (NO_SEGMENT_TRAVERSAL) /* scan only first segment */
- break;
- pred = sp;
- sp = next;
- }
- /* Reset check counter */
- m->release_checks = ((nsegs > MAX_RELEASE_CHECK_RATE)?
- nsegs : MAX_RELEASE_CHECK_RATE);
- return released;
-}
-
-static int sys_trim(mstate m, size_t pad) {
- size_t released = 0;
- ensure_initialization();
- if (pad < MAX_REQUEST && is_initialized(m)) {
- pad += TOP_FOOT_SIZE; /* ensure enough room for segment overhead */
-
- if (m->topsize > pad) {
- /* Shrink top space in granularity-size units, keeping at least one */
- size_t unit = mparams.granularity;
- size_t extra = ((m->topsize - pad + (unit - SIZE_T_ONE)) / unit -
- SIZE_T_ONE) * unit;
- msegmentptr sp = segment_holding(m, (char*)m->top);
-
- if (!is_extern_segment(sp)) {
- if (is_mmapped_segment(sp)) {
- if (HAVE_MMAP &&
- sp->size >= extra &&
- !has_segment_link(m, sp)) { /* can't shrink if pinned */
- size_t newsize = sp->size - extra;
- /* Prefer mremap, fall back to munmap */
- if ((CALL_MREMAP(sp->base, sp->size, newsize, 0) != MFAIL) ||
- (CALL_MUNMAP(sp->base + newsize, extra) == 0)) {
- released = extra;
- }
- }
- }
- else if (HAVE_MORECORE) {
- if (extra >= HALF_MAX_SIZE_T) /* Avoid wrapping negative */
- extra = (HALF_MAX_SIZE_T) + SIZE_T_ONE - unit;
- ACQUIRE_MALLOC_GLOBAL_LOCK();
- {
- /* Make sure end of memory is where we last set it. */
- char* old_br = (char*)(CALL_MORECORE(0));
- if (old_br == sp->base + sp->size) {
- char* rel_br = (char*)(CALL_MORECORE(-extra));
- char* new_br = (char*)(CALL_MORECORE(0));
- if (rel_br != CMFAIL && new_br < old_br)
- released = old_br - new_br;
- }
- }
- RELEASE_MALLOC_GLOBAL_LOCK();
- }
- }
-
- if (released != 0) {
- sp->size -= released;
- m->footprint -= released;
- init_top(m, m->top, m->topsize - released);
- check_top_chunk(m, m->top);
- }
- }
-
- /* Unmap any unused mmapped segments */
- if (HAVE_MMAP)
- released += release_unused_segments(m);
-
- /* On failure, disable autotrim to avoid repeated failed future calls */
- if (released == 0 && m->topsize > m->trim_check)
- m->trim_check = MAX_SIZE_T;
- }
-
- return (released != 0)? 1 : 0;
-}
-
-
-/* ---------------------------- malloc support --------------------------- */
-
-/* allocate a large request from the best fitting chunk in a treebin */
-static void* tmalloc_large(mstate m, size_t nb) {
- tchunkptr v = 0;
- size_t rsize = -nb; /* Unsigned negation */
- tchunkptr t;
- bindex_t idx;
- compute_tree_index(nb, idx);
- if ((t = *treebin_at(m, idx)) != 0) {
- /* Traverse tree for this bin looking for node with size == nb */
- size_t sizebits = nb << leftshift_for_tree_index(idx);
- tchunkptr rst = 0; /* The deepest untaken right subtree */
- for (;;) {
- tchunkptr rt;
- size_t trem = chunksize(t) - nb;
- if (trem < rsize) {
- v = t;
- if ((rsize = trem) == 0)
- break;
- }
- rt = t->child[1];
- t = t->child[(sizebits >> (SIZE_T_BITSIZE-SIZE_T_ONE)) & 1];
- if (rt != 0 && rt != t)
- rst = rt;
- if (t == 0) {
- t = rst; /* set t to least subtree holding sizes > nb */
- break;
- }
- sizebits <<= 1;
- }
- }
- if (t == 0 && v == 0) { /* set t to root of next non-empty treebin */
- binmap_t leftbits = left_bits(idx2bit(idx)) & m->treemap;
- if (leftbits != 0) {
- bindex_t i;
- binmap_t leastbit = least_bit(leftbits);
- compute_bit2idx(leastbit, i);
- t = *treebin_at(m, i);
- }
- }
-
- while (t != 0) { /* find smallest of tree or subtree */
- size_t trem = chunksize(t) - nb;
- if (trem < rsize) {
- rsize = trem;
- v = t;
- }
- t = leftmost_child(t);
- }
-
- /* If dv is a better fit, return 0 so malloc will use it */
- if (v != 0 && rsize < (size_t)(m->dvsize - nb)) {
- if (RTCHECK(ok_address(m, v))) { /* split */
- mchunkptr r = chunk_plus_offset(v, nb);
- assert(chunksize(v) == rsize + nb);
- if (RTCHECK(ok_next(v, r))) {
- unlink_large_chunk(m, v);
- if (rsize < MIN_CHUNK_SIZE)
- set_inuse_and_pinuse(m, v, (rsize + nb));
- else {
- set_size_and_pinuse_of_inuse_chunk(m, v, nb);
- set_size_and_pinuse_of_free_chunk(r, rsize);
- insert_chunk(m, r, rsize);
- }
- return chunk2mem(v);
- }
- }
- CORRUPTION_ERROR_ACTION(m);
- }
- return 0;
-}
-
-/* allocate a small request from the best fitting chunk in a treebin */
-static void* tmalloc_small(mstate m, size_t nb) {
- tchunkptr t, v;
- size_t rsize;
- bindex_t i;
- binmap_t leastbit = least_bit(m->treemap);
- compute_bit2idx(leastbit, i);
- v = t = *treebin_at(m, i);
- rsize = chunksize(t) - nb;
-
- while ((t = leftmost_child(t)) != 0) {
- size_t trem = chunksize(t) - nb;
- if (trem < rsize) {
- rsize = trem;
- v = t;
- }
- }
-
- if (RTCHECK(ok_address(m, v))) {
- mchunkptr r = chunk_plus_offset(v, nb);
- assert(chunksize(v) == rsize + nb);
- if (RTCHECK(ok_next(v, r))) {
- unlink_large_chunk(m, v);
- if (rsize < MIN_CHUNK_SIZE)
- set_inuse_and_pinuse(m, v, (rsize + nb));
- else {
- set_size_and_pinuse_of_inuse_chunk(m, v, nb);
- set_size_and_pinuse_of_free_chunk(r, rsize);
- replace_dv(m, r, rsize);
- }
- return chunk2mem(v);
- }
- }
-
- CORRUPTION_ERROR_ACTION(m);
- return 0;
-}
-
-/* --------------------------- realloc support --------------------------- */
-
-static void* internal_realloc(mstate m, void* oldmem, size_t bytes) {
- if (bytes >= MAX_REQUEST) {
- MALLOC_FAILURE_ACTION;
- return 0;
- }
- if (!PREACTION(m)) {
- mchunkptr oldp = mem2chunk(oldmem);
- size_t oldsize = chunksize(oldp);
- mchunkptr next = chunk_plus_offset(oldp, oldsize);
- mchunkptr newp = 0;
- void* extra = 0;
-
- /* Try to either shrink or extend into top. Else malloc-copy-free */
-
- if (RTCHECK(ok_address(m, oldp) && ok_inuse(oldp) &&
- ok_next(oldp, next) && ok_pinuse(next))) {
- size_t nb = request2size(bytes);
- if (is_mmapped(oldp))
- newp = mmap_resize(m, oldp, nb);
- else if (oldsize >= nb) { /* already big enough */
- size_t rsize = oldsize - nb;
- newp = oldp;
- if (rsize >= MIN_CHUNK_SIZE) {
- mchunkptr remainder = chunk_plus_offset(newp, nb);
- set_inuse(m, newp, nb);
- set_inuse_and_pinuse(m, remainder, rsize);
- extra = chunk2mem(remainder);
- }
- }
- else if (next == m->top && oldsize + m->topsize > nb) {
- /* Expand into top */
- size_t newsize = oldsize + m->topsize;
- size_t newtopsize = newsize - nb;
- mchunkptr newtop = chunk_plus_offset(oldp, nb);
- set_inuse(m, oldp, nb);
- newtop->head = newtopsize |PINUSE_BIT;
- m->top = newtop;
- m->topsize = newtopsize;
- newp = oldp;
- }
- }
- else {
- USAGE_ERROR_ACTION(m, oldmem);
- POSTACTION(m);
- return 0;
- }
-#if DEBUG
- if (newp != 0) {
- check_inuse_chunk(m, newp); /* Check requires lock */
- }
-#endif
-
- POSTACTION(m);
-
- if (newp != 0) {
- if (extra != 0) {
- internal_free(m, extra);
- }
- return chunk2mem(newp);
- }
- else {
- void* newmem = internal_malloc(m, bytes);
- if (newmem != 0) {
- size_t oc = oldsize - overhead_for(oldp);
- memcpy(newmem, oldmem, (oc < bytes)? oc : bytes);
- internal_free(m, oldmem);
- }
- return newmem;
- }
- }
- return 0;
-}
-
-/* --------------------------- memalign support -------------------------- */
-
-static void* internal_memalign(mstate m, size_t alignment, size_t bytes) {
- if (alignment <= MALLOC_ALIGNMENT) /* Can just use malloc */
- return internal_malloc(m, bytes);
- if (alignment < MIN_CHUNK_SIZE) /* must be at least a minimum chunk size */
- alignment = MIN_CHUNK_SIZE;
- if ((alignment & (alignment-SIZE_T_ONE)) != 0) {/* Ensure a power of 2 */
- size_t a = MALLOC_ALIGNMENT << 1;
- while (a < alignment) a <<= 1;
- alignment = a;
- }
-
- if (bytes >= MAX_REQUEST - alignment) {
- if (m != 0) { /* Test isn't needed but avoids compiler warning */
- MALLOC_FAILURE_ACTION;
- }
- }
- else {
- size_t nb = request2size(bytes);
- size_t req = nb + alignment + MIN_CHUNK_SIZE - CHUNK_OVERHEAD;
- char* mem = (char*)internal_malloc(m, req);
- if (mem != 0) {
- void* leader = 0;
- void* trailer = 0;
- mchunkptr p = mem2chunk(mem);
-
- if (PREACTION(m)) return 0;
- if ((((size_t)(mem)) % alignment) != 0) { /* misaligned */
- /*
- Find an aligned spot inside chunk. Since we need to give
- back leading space in a chunk of at least MIN_CHUNK_SIZE, if
- the first calculation places us at a spot with less than
- MIN_CHUNK_SIZE leader, we can move to the next aligned spot.
- We've allocated enough total room so that this is always
- possible.
- */
- char* br = (char*)mem2chunk((size_t)(((size_t)(mem +
- alignment -
- SIZE_T_ONE)) &
- -alignment));
- char* pos = ((size_t)(br - (char*)(p)) >= MIN_CHUNK_SIZE)?
- br : br+alignment;
- mchunkptr newp = (mchunkptr)pos;
- size_t leadsize = pos - (char*)(p);
- size_t newsize = chunksize(p) - leadsize;
-
- if (is_mmapped(p)) { /* For mmapped chunks, just adjust offset */
- newp->prev_foot = p->prev_foot + leadsize;
- newp->head = newsize;
- }
- else { /* Otherwise, give back leader, use the rest */
- set_inuse(m, newp, newsize);
- set_inuse(m, p, leadsize);
- leader = chunk2mem(p);
- }
- p = newp;
- }
-
- /* Give back spare room at the end */
- if (!is_mmapped(p)) {
- size_t size = chunksize(p);
- if (size > nb + MIN_CHUNK_SIZE) {
- size_t remainder_size = size - nb;
- mchunkptr remainder = chunk_plus_offset(p, nb);
- set_inuse(m, p, nb);
- set_inuse(m, remainder, remainder_size);
- trailer = chunk2mem(remainder);
- }
- }
-
- assert (chunksize(p) >= nb);
- assert((((size_t)(chunk2mem(p))) % alignment) == 0);
- check_inuse_chunk(m, p);
- POSTACTION(m);
- if (leader != 0) {
- internal_free(m, leader);
- }
- if (trailer != 0) {
- internal_free(m, trailer);
- }
- return chunk2mem(p);
- }
- }
- return 0;
-}
-
-/* ------------------------ comalloc/coalloc support --------------------- */
-
-static void** ialloc(mstate m,
- size_t n_elements,
- size_t* sizes,
- int opts,
- void* chunks[]) {
- /*
- This provides common support for independent_X routines, handling
- all of the combinations that can result.
-
- The opts arg has:
- bit 0 set if all elements are same size (using sizes[0])
- bit 1 set if elements should be zeroed
- */
-
- size_t element_size; /* chunksize of each element, if all same */
- size_t contents_size; /* total size of elements */
- size_t array_size; /* request size of pointer array */
- void* mem; /* malloced aggregate space */
- mchunkptr p; /* corresponding chunk */
- size_t remainder_size; /* remaining bytes while splitting */
- void** marray; /* either "chunks" or malloced ptr array */
- mchunkptr array_chunk; /* chunk for malloced ptr array */
- flag_t was_enabled; /* to disable mmap */
- size_t size;
- size_t i;
-
- ensure_initialization();
- /* compute array length, if needed */
- if (chunks != 0) {
- if (n_elements == 0)
- return chunks; /* nothing to do */
- marray = chunks;
- array_size = 0;
- }
- else {
- /* if empty req, must still return chunk representing empty array */
- if (n_elements == 0)
- return (void**)internal_malloc(m, 0);
- marray = 0;
- array_size = request2size(n_elements * (sizeof(void*)));
- }
-
- /* compute total element size */
- if (opts & 0x1) { /* all-same-size */
- element_size = request2size(*sizes);
- contents_size = n_elements * element_size;
- }
- else { /* add up all the sizes */
- element_size = 0;
- contents_size = 0;
- for (i = 0; i != n_elements; ++i)
- contents_size += request2size(sizes[i]);
- }
-
- size = contents_size + array_size;
-
- /*
- Allocate the aggregate chunk. First disable direct-mmapping so
- malloc won't use it, since we would not be able to later
- free/realloc space internal to a segregated mmap region.
- */
- was_enabled = use_mmap(m);
- disable_mmap(m);
- mem = internal_malloc(m, size - CHUNK_OVERHEAD);
- if (was_enabled)
- enable_mmap(m);
- if (mem == 0)
- return 0;
-
- if (PREACTION(m)) return 0;
- p = mem2chunk(mem);
- remainder_size = chunksize(p);
-
- assert(!is_mmapped(p));
-
- if (opts & 0x2) { /* optionally clear the elements */
- memset((size_t*)mem, 0, remainder_size - SIZE_T_SIZE - array_size);
- }
-
- /* If not provided, allocate the pointer array as final part of chunk */
- if (marray == 0) {
- size_t array_chunk_size;
- array_chunk = chunk_plus_offset(p, contents_size);
- array_chunk_size = remainder_size - contents_size;
- marray = (void**) (chunk2mem(array_chunk));
- set_size_and_pinuse_of_inuse_chunk(m, array_chunk, array_chunk_size);
- remainder_size = contents_size;
- }
-
- /* split out elements */
- for (i = 0; ; ++i) {
- marray[i] = chunk2mem(p);
- if (i != n_elements-1) {
- if (element_size != 0)
- size = element_size;
- else
- size = request2size(sizes[i]);
- remainder_size -= size;
- set_size_and_pinuse_of_inuse_chunk(m, p, size);
- p = chunk_plus_offset(p, size);
- }
- else { /* the final element absorbs any overallocation slop */
- set_size_and_pinuse_of_inuse_chunk(m, p, remainder_size);
- break;
- }
- }
-
-#if DEBUG
- if (marray != chunks) {
- /* final element must have exactly exhausted chunk */
- if (element_size != 0) {
- assert(remainder_size == element_size);
- }
- else {
- assert(remainder_size == request2size(sizes[i]));
- }
- check_inuse_chunk(m, mem2chunk(marray));
- }
- for (i = 0; i != n_elements; ++i)
- check_inuse_chunk(m, mem2chunk(marray[i]));
-
-#endif /* DEBUG */
-
- POSTACTION(m);
- return marray;
-}
-
-
-/* -------------------------- public routines ---------------------------- */
-
-#if !ONLY_MSPACES
-
-void* dlmalloc(size_t bytes) {
- /*
- Basic algorithm:
- If a small request (< 256 bytes minus per-chunk overhead):
- 1. If one exists, use a remainderless chunk in associated smallbin.
- (Remainderless means that there are too few excess bytes to
- represent as a chunk.)
- 2. If it is big enough, use the dv chunk, which is normally the
- chunk adjacent to the one used for the most recent small request.
- 3. If one exists, split the smallest available chunk in a bin,
- saving remainder in dv.
- 4. If it is big enough, use the top chunk.
- 5. If available, get memory from system and use it
- Otherwise, for a large request:
- 1. Find the smallest available binned chunk that fits, and use it
- if it is better fitting than dv chunk, splitting if necessary.
- 2. If better fitting than any binned chunk, use the dv chunk.
- 3. If it is big enough, use the top chunk.
- 4. If request size >= mmap threshold, try to directly mmap this chunk.
- 5. If available, get memory from system and use it
-
- The ugly goto's here ensure that postaction occurs along all paths.
- */
-
-#if USE_LOCKS
- ensure_initialization(); /* initialize in sys_alloc if not using locks */
-#endif
-
- if (!PREACTION(gm)) {
- void* mem;
- size_t nb;
- if (bytes <= MAX_SMALL_REQUEST) {
- bindex_t idx;
- binmap_t smallbits;
- nb = (bytes < MIN_REQUEST)? MIN_CHUNK_SIZE : pad_request(bytes);
- idx = small_index(nb);
- smallbits = gm->smallmap >> idx;
-
- if ((smallbits & 0x3U) != 0) { /* Remainderless fit to a smallbin. */
- mchunkptr b, p;
- idx += ~smallbits & 1; /* Uses next bin if idx empty */
- b = smallbin_at(gm, idx);
- p = b->fd;
- assert(chunksize(p) == small_index2size(idx));
- unlink_first_small_chunk(gm, b, p, idx);
- set_inuse_and_pinuse(gm, p, small_index2size(idx));
- mem = chunk2mem(p);
- check_malloced_chunk(gm, mem, nb);
- goto postaction;
- }
-
- else if (nb > gm->dvsize) {
- if (smallbits != 0) { /* Use chunk in next nonempty smallbin */
- mchunkptr b, p, r;
- size_t rsize;
- bindex_t i;
- binmap_t leftbits = (smallbits << idx) & left_bits(idx2bit(idx));
- binmap_t leastbit = least_bit(leftbits);
- compute_bit2idx(leastbit, i);
- b = smallbin_at(gm, i);
- p = b->fd;
- assert(chunksize(p) == small_index2size(i));
- unlink_first_small_chunk(gm, b, p, i);
- rsize = small_index2size(i) - nb;
- /* Fit here cannot be remainderless if 4byte sizes */
- if (SIZE_T_SIZE != 4 && rsize < MIN_CHUNK_SIZE)
- set_inuse_and_pinuse(gm, p, small_index2size(i));
- else {
- set_size_and_pinuse_of_inuse_chunk(gm, p, nb);
- r = chunk_plus_offset(p, nb);
- set_size_and_pinuse_of_free_chunk(r, rsize);
- replace_dv(gm, r, rsize);
- }
- mem = chunk2mem(p);
- check_malloced_chunk(gm, mem, nb);
- goto postaction;
- }
-
- else if (gm->treemap != 0 && (mem = tmalloc_small(gm, nb)) != 0) {
- check_malloced_chunk(gm, mem, nb);
- goto postaction;
- }
- }
- }
- else if (bytes >= MAX_REQUEST)
- nb = MAX_SIZE_T; /* Too big to allocate. Force failure (in sys alloc) */
- else {
- nb = pad_request(bytes);
- if (gm->treemap != 0 && (mem = tmalloc_large(gm, nb)) != 0) {
- check_malloced_chunk(gm, mem, nb);
- goto postaction;
- }
- }
-
- if (nb <= gm->dvsize) {
- size_t rsize = gm->dvsize - nb;
- mchunkptr p = gm->dv;
- if (rsize >= MIN_CHUNK_SIZE) { /* split dv */
- mchunkptr r = gm->dv = chunk_plus_offset(p, nb);
- gm->dvsize = rsize;
- set_size_and_pinuse_of_free_chunk(r, rsize);
- set_size_and_pinuse_of_inuse_chunk(gm, p, nb);
- }
- else { /* exhaust dv */
- size_t dvs = gm->dvsize;
- gm->dvsize = 0;
- gm->dv = 0;
- set_inuse_and_pinuse(gm, p, dvs);
- }
- mem = chunk2mem(p);
- check_malloced_chunk(gm, mem, nb);
- goto postaction;
- }
-
- else if (nb < gm->topsize) { /* Split top */
- size_t rsize = gm->topsize -= nb;
- mchunkptr p = gm->top;
- mchunkptr r = gm->top = chunk_plus_offset(p, nb);
- r->head = rsize | PINUSE_BIT;
- set_size_and_pinuse_of_inuse_chunk(gm, p, nb);
- mem = chunk2mem(p);
- check_top_chunk(gm, gm->top);
- check_malloced_chunk(gm, mem, nb);
- goto postaction;
- }
-
- mem = sys_alloc(gm, nb);
-
- postaction:
- POSTACTION(gm);
- return mem;
- }
-
- return 0;
-}
-
-void dlfree(void* mem) {
- /*
- Consolidate freed chunks with preceeding or succeeding bordering
- free chunks, if they exist, and then place in a bin. Intermixed
- with special cases for top, dv, mmapped chunks, and usage errors.
- */
-
- if (mem != 0) {
- mchunkptr p = mem2chunk(mem);
-#if FOOTERS
- mstate fm = get_mstate_for(p);
- if (!ok_magic(fm)) {
- USAGE_ERROR_ACTION(fm, p);
- return;
- }
-#else /* FOOTERS */
-#define fm gm
-#endif /* FOOTERS */
- if (!PREACTION(fm)) {
- check_inuse_chunk(fm, p);
- if (RTCHECK(ok_address(fm, p) && ok_inuse(p))) {
- size_t psize = chunksize(p);
- mchunkptr next = chunk_plus_offset(p, psize);
- if (!pinuse(p)) {
- size_t prevsize = p->prev_foot;
- if (is_mmapped(p)) {
- psize += prevsize + MMAP_FOOT_PAD;
- if (CALL_MUNMAP((char*)p - prevsize, psize) == 0)
- fm->footprint -= psize;
- goto postaction;
- }
- else {
- mchunkptr prev = chunk_minus_offset(p, prevsize);
- psize += prevsize;
- p = prev;
- if (RTCHECK(ok_address(fm, prev))) { /* consolidate backward */
- if (p != fm->dv) {
- unlink_chunk(fm, p, prevsize);
- }
- else if ((next->head & INUSE_BITS) == INUSE_BITS) {
- fm->dvsize = psize;
- set_free_with_pinuse(p, psize, next);
- goto postaction;
- }
- }
- else
- goto erroraction;
- }
- }
-
- if (RTCHECK(ok_next(p, next) && ok_pinuse(next))) {
- if (!cinuse(next)) { /* consolidate forward */
- if (next == fm->top) {
- size_t tsize = fm->topsize += psize;
- fm->top = p;
- p->head = tsize | PINUSE_BIT;
- if (p == fm->dv) {
- fm->dv = 0;
- fm->dvsize = 0;
- }
- if (should_trim(fm, tsize))
- sys_trim(fm, 0);
- goto postaction;
- }
- else if (next == fm->dv) {
- size_t dsize = fm->dvsize += psize;
- fm->dv = p;
- set_size_and_pinuse_of_free_chunk(p, dsize);
- goto postaction;
- }
- else {
- size_t nsize = chunksize(next);
- psize += nsize;
- unlink_chunk(fm, next, nsize);
- set_size_and_pinuse_of_free_chunk(p, psize);
- if (p == fm->dv) {
- fm->dvsize = psize;
- goto postaction;
- }
- }
- }
- else
- set_free_with_pinuse(p, psize, next);
-
- if (is_small(psize)) {
- insert_small_chunk(fm, p, psize);
- check_free_chunk(fm, p);
- }
- else {
- tchunkptr tp = (tchunkptr)p;
- insert_large_chunk(fm, tp, psize);
- check_free_chunk(fm, p);
- if (--fm->release_checks == 0)
- release_unused_segments(fm);
- }
- goto postaction;
- }
- }
- erroraction:
- USAGE_ERROR_ACTION(fm, p);
- postaction:
- POSTACTION(fm);
- }
- }
-#if !FOOTERS
-#undef fm
-#endif /* FOOTERS */
-}
-
-void* dlcalloc(size_t n_elements, size_t elem_size) {
- void* mem;
- size_t req = 0;
- if (n_elements != 0) {
- req = n_elements * elem_size;
- if (((n_elements | elem_size) & ~(size_t)0xffff) &&
- (req / n_elements != elem_size))
- req = MAX_SIZE_T; /* force downstream failure on overflow */
- }
- mem = dlmalloc(req);
- if (mem != 0 && calloc_must_clear(mem2chunk(mem)))
- memset(mem, 0, req);
- return mem;
-}
-
-void* dlrealloc(void* oldmem, size_t bytes) {
- if (oldmem == 0)
- return dlmalloc(bytes);
-#ifdef REALLOC_ZERO_BYTES_FREES
- if (bytes == 0) {
- dlfree(oldmem);
- return 0;
- }
-#endif /* REALLOC_ZERO_BYTES_FREES */
- else {
-#if ! FOOTERS
- mstate m = gm;
-#else /* FOOTERS */
- mstate m = get_mstate_for(mem2chunk(oldmem));
- if (!ok_magic(m)) {
- USAGE_ERROR_ACTION(m, oldmem);
- return 0;
- }
-#endif /* FOOTERS */
- return internal_realloc(m, oldmem, bytes);
- }
-}
-
-void* dlmemalign(size_t alignment, size_t bytes) {
- return internal_memalign(gm, alignment, bytes);
-}
-
-void** dlindependent_calloc(size_t n_elements, size_t elem_size,
- void* chunks[]) {
- size_t sz = elem_size; /* serves as 1-element array */
- return ialloc(gm, n_elements, &sz, 3, chunks);
-}
-
-void** dlindependent_comalloc(size_t n_elements, size_t sizes[],
- void* chunks[]) {
- return ialloc(gm, n_elements, sizes, 0, chunks);
-}
-
-void* dlvalloc(size_t bytes) {
- size_t pagesz;
- ensure_initialization();
- pagesz = mparams.page_size;
- return dlmemalign(pagesz, bytes);
-}
-
-void* dlpvalloc(size_t bytes) {
- size_t pagesz;
- ensure_initialization();
- pagesz = mparams.page_size;
- return dlmemalign(pagesz, (bytes + pagesz - SIZE_T_ONE) & ~(pagesz - SIZE_T_ONE));
-}
-
-int dlmalloc_trim(size_t pad) {
- int result = 0;
- ensure_initialization();
- if (!PREACTION(gm)) {
- result = sys_trim(gm, pad);
- POSTACTION(gm);
- }
- return result;
-}
-
-size_t dlmalloc_footprint(void) {
- return gm->footprint;
-}
-
-size_t dlmalloc_max_footprint(void) {
- return gm->max_footprint;
-}
-
-#if !NO_MALLINFO
-struct mallinfo dlmallinfo(void) {
- return internal_mallinfo(gm);
-}
-#endif /* NO_MALLINFO */
-
-void dlmalloc_stats() {
- internal_malloc_stats(gm);
-}
-
-int dlmallopt(int param_number, int value) {
- return change_mparam(param_number, value);
-}
-
-#endif /* !ONLY_MSPACES */
-
-size_t dlmalloc_usable_size(void* mem) {
- if (mem != 0) {
- mchunkptr p = mem2chunk(mem);
- if (is_inuse(p))
- return chunksize(p) - overhead_for(p);
- }
- return 0;
-}
-
-/* ----------------------------- user mspaces ---------------------------- */
-
-#if MSPACES
-
-static mstate init_user_mstate(char* tbase, size_t tsize) {
- size_t msize = pad_request(sizeof(struct malloc_state));
- mchunkptr mn;
- mchunkptr msp = align_as_chunk(tbase);
- mstate m = (mstate)(chunk2mem(msp));
- memset(m, 0, msize);
- INITIAL_LOCK(&m->mutex);
- msp->head = (msize|INUSE_BITS);
- m->seg.base = m->least_addr = tbase;
- m->seg.size = m->footprint = m->max_footprint = tsize;
- m->magic = mparams.magic;
- m->release_checks = MAX_RELEASE_CHECK_RATE;
- m->mflags = mparams.default_mflags;
- m->extp = 0;
- m->exts = 0;
- disable_contiguous(m);
- init_bins(m);
- mn = next_chunk(mem2chunk(m));
- init_top(m, mn, (size_t)((tbase + tsize) - (char*)mn) - TOP_FOOT_SIZE);
- check_top_chunk(m, m->top);
- return m;
-}
-
-mspace create_mspace(size_t capacity, int locked) {
- mstate m = 0;
- size_t msize;
- ensure_initialization();
- msize = pad_request(sizeof(struct malloc_state));
- if (capacity < (size_t) -(msize + TOP_FOOT_SIZE + mparams.page_size)) {
- size_t rs = ((capacity == 0)? mparams.granularity :
- (capacity + TOP_FOOT_SIZE + msize));
- size_t tsize = granularity_align(rs);
- char* tbase = (char*)(CALL_MMAP(tsize));
- if (tbase != CMFAIL) {
- m = init_user_mstate(tbase, tsize);
- m->seg.sflags = USE_MMAP_BIT;
- set_lock(m, locked);
- }
- }
- return (mspace)m;
-}
-
-mspace create_mspace_with_base(void* base, size_t capacity, int locked) {
- mstate m = 0;
- size_t msize;
- ensure_initialization();
- msize = pad_request(sizeof(struct malloc_state));
- if (capacity > msize + TOP_FOOT_SIZE &&
- capacity < (size_t) -(msize + TOP_FOOT_SIZE + mparams.page_size)) {
- m = init_user_mstate((char*)base, capacity);
- m->seg.sflags = EXTERN_BIT;
- set_lock(m, locked);
- }
- return (mspace)m;
-}
-
-int mspace_track_large_chunks(mspace msp, int enable) {
- int ret = 0;
- mstate ms = (mstate)msp;
- if (!PREACTION(ms)) {
- if (!use_mmap(ms))
- ret = 1;
- if (!enable)
- enable_mmap(ms);
- else
- disable_mmap(ms);
- POSTACTION(ms);
- }
- return ret;
-}
-
-size_t destroy_mspace(mspace msp) {
- size_t freed = 0;
- mstate ms = (mstate)msp;
- if (ok_magic(ms)) {
- msegmentptr sp = &ms->seg;
- while (sp != 0) {
- char* base = sp->base;
- size_t size = sp->size;
- flag_t flag = sp->sflags;
- sp = sp->next;
- if ((flag & USE_MMAP_BIT) && !(flag & EXTERN_BIT) &&
- CALL_MUNMAP(base, size) == 0)
- freed += size;
- }
- }
- else {
- USAGE_ERROR_ACTION(ms,ms);
- }
- return freed;
-}
-
-/*
- mspace versions of routines are near-clones of the global
- versions. This is not so nice but better than the alternatives.
-*/
-
-
-void* mspace_malloc(mspace msp, size_t bytes) {
- mstate ms = (mstate)msp;
- if (!ok_magic(ms)) {
- USAGE_ERROR_ACTION(ms,ms);
- return 0;
- }
- if (!PREACTION(ms)) {
- void* mem;
- size_t nb;
- if (bytes <= MAX_SMALL_REQUEST) {
- bindex_t idx;
- binmap_t smallbits;
- nb = (bytes < MIN_REQUEST)? MIN_CHUNK_SIZE : pad_request(bytes);
- idx = small_index(nb);
- smallbits = ms->smallmap >> idx;
-
- if ((smallbits & 0x3U) != 0) { /* Remainderless fit to a smallbin. */
- mchunkptr b, p;
- idx += ~smallbits & 1; /* Uses next bin if idx empty */
- b = smallbin_at(ms, idx);
- p = b->fd;
- assert(chunksize(p) == small_index2size(idx));
- unlink_first_small_chunk(ms, b, p, idx);
- set_inuse_and_pinuse(ms, p, small_index2size(idx));
- mem = chunk2mem(p);
- check_malloced_chunk(ms, mem, nb);
- goto postaction;
- }
-
- else if (nb > ms->dvsize) {
- if (smallbits != 0) { /* Use chunk in next nonempty smallbin */
- mchunkptr b, p, r;
- size_t rsize;
- bindex_t i;
- binmap_t leftbits = (smallbits << idx) & left_bits(idx2bit(idx));
- binmap_t leastbit = least_bit(leftbits);
- compute_bit2idx(leastbit, i);
- b = smallbin_at(ms, i);
- p = b->fd;
- assert(chunksize(p) == small_index2size(i));
- unlink_first_small_chunk(ms, b, p, i);
- rsize = small_index2size(i) - nb;
- /* Fit here cannot be remainderless if 4byte sizes */
- if (SIZE_T_SIZE != 4 && rsize < MIN_CHUNK_SIZE)
- set_inuse_and_pinuse(ms, p, small_index2size(i));
- else {
- set_size_and_pinuse_of_inuse_chunk(ms, p, nb);
- r = chunk_plus_offset(p, nb);
- set_size_and_pinuse_of_free_chunk(r, rsize);
- replace_dv(ms, r, rsize);
- }
- mem = chunk2mem(p);
- check_malloced_chunk(ms, mem, nb);
- goto postaction;
- }
-
- else if (ms->treemap != 0 && (mem = tmalloc_small(ms, nb)) != 0) {
- check_malloced_chunk(ms, mem, nb);
- goto postaction;
- }
- }
- }
- else if (bytes >= MAX_REQUEST)
- nb = MAX_SIZE_T; /* Too big to allocate. Force failure (in sys alloc) */
- else {
- nb = pad_request(bytes);
- if (ms->treemap != 0 && (mem = tmalloc_large(ms, nb)) != 0) {
- check_malloced_chunk(ms, mem, nb);
- goto postaction;
- }
- }
-
- if (nb <= ms->dvsize) {
- size_t rsize = ms->dvsize - nb;
- mchunkptr p = ms->dv;
- if (rsize >= MIN_CHUNK_SIZE) { /* split dv */
- mchunkptr r = ms->dv = chunk_plus_offset(p, nb);
- ms->dvsize = rsize;
- set_size_and_pinuse_of_free_chunk(r, rsize);
- set_size_and_pinuse_of_inuse_chunk(ms, p, nb);
- }
- else { /* exhaust dv */
- size_t dvs = ms->dvsize;
- ms->dvsize = 0;
- ms->dv = 0;
- set_inuse_and_pinuse(ms, p, dvs);
- }
- mem = chunk2mem(p);
- check_malloced_chunk(ms, mem, nb);
- goto postaction;
- }
-
- else if (nb < ms->topsize) { /* Split top */
- size_t rsize = ms->topsize -= nb;
- mchunkptr p = ms->top;
- mchunkptr r = ms->top = chunk_plus_offset(p, nb);
- r->head = rsize | PINUSE_BIT;
- set_size_and_pinuse_of_inuse_chunk(ms, p, nb);
- mem = chunk2mem(p);
- check_top_chunk(ms, ms->top);
- check_malloced_chunk(ms, mem, nb);
- goto postaction;
- }
-
- mem = sys_alloc(ms, nb);
-
- postaction:
- POSTACTION(ms);
- return mem;
- }
-
- return 0;
-}
-
-void mspace_free(mspace msp, void* mem) {
- if (mem != 0) {
- mchunkptr p = mem2chunk(mem);
-#if FOOTERS
- mstate fm = get_mstate_for(p);
- msp = msp; /* placate people compiling -Wunused */
-#else /* FOOTERS */
- mstate fm = (mstate)msp;
-#endif /* FOOTERS */
- if (!ok_magic(fm)) {
- USAGE_ERROR_ACTION(fm, p);
- return;
- }
- if (!PREACTION(fm)) {
- check_inuse_chunk(fm, p);
- if (RTCHECK(ok_address(fm, p) && ok_inuse(p))) {
- size_t psize = chunksize(p);
- mchunkptr next = chunk_plus_offset(p, psize);
- if (!pinuse(p)) {
- size_t prevsize = p->prev_foot;
- if (is_mmapped(p)) {
- psize += prevsize + MMAP_FOOT_PAD;
- if (CALL_MUNMAP((char*)p - prevsize, psize) == 0)
- fm->footprint -= psize;
- goto postaction;
- }
- else {
- mchunkptr prev = chunk_minus_offset(p, prevsize);
- psize += prevsize;
- p = prev;
- if (RTCHECK(ok_address(fm, prev))) { /* consolidate backward */
- if (p != fm->dv) {
- unlink_chunk(fm, p, prevsize);
- }
- else if ((next->head & INUSE_BITS) == INUSE_BITS) {
- fm->dvsize = psize;
- set_free_with_pinuse(p, psize, next);
- goto postaction;
- }
- }
- else
- goto erroraction;
- }
- }
-
- if (RTCHECK(ok_next(p, next) && ok_pinuse(next))) {
- if (!cinuse(next)) { /* consolidate forward */
- if (next == fm->top) {
- size_t tsize = fm->topsize += psize;
- fm->top = p;
- p->head = tsize | PINUSE_BIT;
- if (p == fm->dv) {
- fm->dv = 0;
- fm->dvsize = 0;
- }
- if (should_trim(fm, tsize))
- sys_trim(fm, 0);
- goto postaction;
- }
- else if (next == fm->dv) {
- size_t dsize = fm->dvsize += psize;
- fm->dv = p;
- set_size_and_pinuse_of_free_chunk(p, dsize);
- goto postaction;
- }
- else {
- size_t nsize = chunksize(next);
- psize += nsize;
- unlink_chunk(fm, next, nsize);
- set_size_and_pinuse_of_free_chunk(p, psize);
- if (p == fm->dv) {
- fm->dvsize = psize;
- goto postaction;
- }
- }
- }
- else
- set_free_with_pinuse(p, psize, next);
-
- if (is_small(psize)) {
- insert_small_chunk(fm, p, psize);
- check_free_chunk(fm, p);
- }
- else {
- tchunkptr tp = (tchunkptr)p;
- insert_large_chunk(fm, tp, psize);
- check_free_chunk(fm, p);
- if (--fm->release_checks == 0)
- release_unused_segments(fm);
- }
- goto postaction;
- }
- }
- erroraction:
- USAGE_ERROR_ACTION(fm, p);
- postaction:
- POSTACTION(fm);
- }
- }
-}
-
-void* mspace_calloc(mspace msp, size_t n_elements, size_t elem_size) {
- void* mem;
- size_t req = 0;
- mstate ms = (mstate)msp;
- if (!ok_magic(ms)) {
- USAGE_ERROR_ACTION(ms,ms);
- return 0;
- }
- if (n_elements != 0) {
- req = n_elements * elem_size;
- if (((n_elements | elem_size) & ~(size_t)0xffff) &&
- (req / n_elements != elem_size))
- req = MAX_SIZE_T; /* force downstream failure on overflow */
- }
- mem = internal_malloc(ms, req);
- if (mem != 0 && calloc_must_clear(mem2chunk(mem)))
- memset(mem, 0, req);
- return mem;
-}
-
-void* mspace_realloc(mspace msp, void* oldmem, size_t bytes) {
- if (oldmem == 0)
- return mspace_malloc(msp, bytes);
-#ifdef REALLOC_ZERO_BYTES_FREES
- if (bytes == 0) {
- mspace_free(msp, oldmem);
- return 0;
- }
-#endif /* REALLOC_ZERO_BYTES_FREES */
- else {
-#if FOOTERS
- mchunkptr p = mem2chunk(oldmem);
- mstate ms = get_mstate_for(p);
-#else /* FOOTERS */
- mstate ms = (mstate)msp;
-#endif /* FOOTERS */
- if (!ok_magic(ms)) {
- USAGE_ERROR_ACTION(ms,ms);
- return 0;
- }
- return internal_realloc(ms, oldmem, bytes);
- }
-}
-
-void* mspace_memalign(mspace msp, size_t alignment, size_t bytes) {
- mstate ms = (mstate)msp;
- if (!ok_magic(ms)) {
- USAGE_ERROR_ACTION(ms,ms);
- return 0;
- }
- return internal_memalign(ms, alignment, bytes);
-}
-
-void** mspace_independent_calloc(mspace msp, size_t n_elements,
- size_t elem_size, void* chunks[]) {
- size_t sz = elem_size; /* serves as 1-element array */
- mstate ms = (mstate)msp;
- if (!ok_magic(ms)) {
- USAGE_ERROR_ACTION(ms,ms);
- return 0;
- }
- return ialloc(ms, n_elements, &sz, 3, chunks);
-}
-
-void** mspace_independent_comalloc(mspace msp, size_t n_elements,
- size_t sizes[], void* chunks[]) {
- mstate ms = (mstate)msp;
- if (!ok_magic(ms)) {
- USAGE_ERROR_ACTION(ms,ms);
- return 0;
- }
- return ialloc(ms, n_elements, sizes, 0, chunks);
-}
-
-int mspace_trim(mspace msp, size_t pad) {
- int result = 0;
- mstate ms = (mstate)msp;
- if (ok_magic(ms)) {
- if (!PREACTION(ms)) {
- result = sys_trim(ms, pad);
- POSTACTION(ms);
- }
- }
- else {
- USAGE_ERROR_ACTION(ms,ms);
- }
- return result;
-}
-
-void mspace_malloc_stats(mspace msp) {
- mstate ms = (mstate)msp;
- if (ok_magic(ms)) {
- internal_malloc_stats(ms);
- }
- else {
- USAGE_ERROR_ACTION(ms,ms);
- }
-}
-
-size_t mspace_footprint(mspace msp) {
- size_t result = 0;
- mstate ms = (mstate)msp;
- if (ok_magic(ms)) {
- result = ms->footprint;
- }
- else {
- USAGE_ERROR_ACTION(ms,ms);
- }
- return result;
-}
-
-
-size_t mspace_max_footprint(mspace msp) {
- size_t result = 0;
- mstate ms = (mstate)msp;
- if (ok_magic(ms)) {
- result = ms->max_footprint;
- }
- else {
- USAGE_ERROR_ACTION(ms,ms);
- }
- return result;
-}
-
-
-#if !NO_MALLINFO
-struct mallinfo mspace_mallinfo(mspace msp) {
- mstate ms = (mstate)msp;
- if (!ok_magic(ms)) {
- USAGE_ERROR_ACTION(ms,ms);
- }
- return internal_mallinfo(ms);
-}
-#endif /* NO_MALLINFO */
-
-size_t mspace_usable_size(void* mem) {
- if (mem != 0) {
- mchunkptr p = mem2chunk(mem);
- if (is_inuse(p))
- return chunksize(p) - overhead_for(p);
- }
- return 0;
-}
-
-int mspace_mallopt(int param_number, int value) {
- return change_mparam(param_number, value);
-}
-
-#endif /* MSPACES */
-
-
-/* -------------------- Alternative MORECORE functions ------------------- */
-
-/*
- Guidelines for creating a custom version of MORECORE:
-
- * For best performance, MORECORE should allocate in multiples of pagesize.
- * MORECORE may allocate more memory than requested. (Or even less,
- but this will usually result in a malloc failure.)
- * MORECORE must not allocate memory when given argument zero, but
- instead return one past the end address of memory from previous
- nonzero call.
- * For best performance, consecutive calls to MORECORE with positive
- arguments should return increasing addresses, indicating that
- space has been contiguously extended.
- * Even though consecutive calls to MORECORE need not return contiguous
- addresses, it must be OK for malloc'ed chunks to span multiple
- regions in those cases where they do happen to be contiguous.
- * MORECORE need not handle negative arguments -- it may instead
- just return MFAIL when given negative arguments.
- Negative arguments are always multiples of pagesize. MORECORE
- must not misinterpret negative args as large positive unsigned
- args. You can suppress all such calls from even occurring by defining
- MORECORE_CANNOT_TRIM,
-
- As an example alternative MORECORE, here is a custom allocator
- kindly contributed for pre-OSX macOS. It uses virtually but not
- necessarily physically contiguous non-paged memory (locked in,
- present and won't get swapped out). You can use it by uncommenting
- this section, adding some #includes, and setting up the appropriate
- defines above:
-
- #define MORECORE osMoreCore
-
- There is also a shutdown routine that should somehow be called for
- cleanup upon program exit.
-
- #define MAX_POOL_ENTRIES 100
- #define MINIMUM_MORECORE_SIZE (64 * 1024U)
- static int next_os_pool;
- void *our_os_pools[MAX_POOL_ENTRIES];
-
- void *osMoreCore(int size)
- {
- void *ptr = 0;
- static void *sbrk_top = 0;
-
- if (size > 0)
- {
- if (size < MINIMUM_MORECORE_SIZE)
- size = MINIMUM_MORECORE_SIZE;
- if (CurrentExecutionLevel() == kTaskLevel)
- ptr = PoolAllocateResident(size + RM_PAGE_SIZE, 0);
- if (ptr == 0)
- {
- return (void *) MFAIL;
- }
- // save ptrs so they can be freed during cleanup
- our_os_pools[next_os_pool] = ptr;
- next_os_pool++;
- ptr = (void *) ((((size_t) ptr) + RM_PAGE_MASK) & ~RM_PAGE_MASK);
- sbrk_top = (char *) ptr + size;
- return ptr;
- }
- else if (size < 0)
- {
- // we don't currently support shrink behavior
- return (void *) MFAIL;
- }
- else
- {
- return sbrk_top;
- }
- }
-
- // cleanup any allocated memory pools
- // called as last thing before shutting down driver
-
- void osCleanupMem(void)
- {
- void **ptr;
-
- for (ptr = our_os_pools; ptr < &our_os_pools[MAX_POOL_ENTRIES]; ptr++)
- if (*ptr)
- {
- PoolDeallocate(*ptr);
- *ptr = 0;
- }
- }
-
-*/
-
-
-/* -----------------------------------------------------------------------
-History:
- V2.8.4 Wed May 27 09:56:23 2009 Doug Lea (dl at gee)
- * Use zeros instead of prev foot for is_mmapped
- * Add mspace_track_large_chunks; thanks to Jean Brouwers
- * Fix set_inuse in internal_realloc; thanks to Jean Brouwers
- * Fix insufficient sys_alloc padding when using 16byte alignment
- * Fix bad error check in mspace_footprint
- * Adaptations for ptmalloc; thanks to Wolfram Gloger.
- * Reentrant spin locks; thanks to Earl Chew and others
- * Win32 improvements; thanks to Niall Douglas and Earl Chew
- * Add NO_SEGMENT_TRAVERSAL and MAX_RELEASE_CHECK_RATE options
- * Extension hook in malloc_state
- * Various small adjustments to reduce warnings on some compilers
- * Various configuration extensions/changes for more platforms. Thanks
- to all who contributed these.
-
- V2.8.3 Thu Sep 22 11:16:32 2005 Doug Lea (dl at gee)
- * Add max_footprint functions
- * Ensure all appropriate literals are size_t
- * Fix conditional compilation problem for some #define settings
- * Avoid concatenating segments with the one provided
- in create_mspace_with_base
- * Rename some variables to avoid compiler shadowing warnings
- * Use explicit lock initialization.
- * Better handling of sbrk interference.
- * Simplify and fix segment insertion, trimming and mspace_destroy
- * Reinstate REALLOC_ZERO_BYTES_FREES option from 2.7.x
- * Thanks especially to Dennis Flanagan for help on these.
-
- V2.8.2 Sun Jun 12 16:01:10 2005 Doug Lea (dl at gee)
- * Fix memalign brace error.
-
- V2.8.1 Wed Jun 8 16:11:46 2005 Doug Lea (dl at gee)
- * Fix improper #endif nesting in C++
- * Add explicit casts needed for C++
-
- V2.8.0 Mon May 30 14:09:02 2005 Doug Lea (dl at gee)
- * Use trees for large bins
- * Support mspaces
- * Use segments to unify sbrk-based and mmap-based system allocation,
- removing need for emulation on most platforms without sbrk.
- * Default safety checks
- * Optional footer checks. Thanks to William Robertson for the idea.
- * Internal code refactoring
- * Incorporate suggestions and platform-specific changes.
- Thanks to Dennis Flanagan, Colin Plumb, Niall Douglas,
- Aaron Bachmann, Emery Berger, and others.
- * Speed up non-fastbin processing enough to remove fastbins.
- * Remove useless cfree() to avoid conflicts with other apps.
- * Remove internal memcpy, memset. Compilers handle builtins better.
- * Remove some options that no one ever used and rename others.
-
- V2.7.2 Sat Aug 17 09:07:30 2002 Doug Lea (dl at gee)
- * Fix malloc_state bitmap array misdeclaration
-
- V2.7.1 Thu Jul 25 10:58:03 2002 Doug Lea (dl at gee)
- * Allow tuning of FIRST_SORTED_BIN_SIZE
- * Use PTR_UINT as type for all ptr->int casts. Thanks to John Belmonte.
- * Better detection and support for non-contiguousness of MORECORE.
- Thanks to Andreas Mueller, Conal Walsh, and Wolfram Gloger
- * Bypass most of malloc if no frees. Thanks To Emery Berger.
- * Fix freeing of old top non-contiguous chunk im sysmalloc.
- * Raised default trim and map thresholds to 256K.
- * Fix mmap-related #defines. Thanks to Lubos Lunak.
- * Fix copy macros; added LACKS_FCNTL_H. Thanks to Neal Walfield.
- * Branch-free bin calculation
- * Default trim and mmap thresholds now 256K.
-
- V2.7.0 Sun Mar 11 14:14:06 2001 Doug Lea (dl at gee)
- * Introduce independent_comalloc and independent_calloc.
- Thanks to Michael Pachos for motivation and help.
- * Make optional .h file available
- * Allow > 2GB requests on 32bit systems.
- * new WIN32 sbrk, mmap, munmap, lock code from <Walter@GeNeSys-e.de>.
- Thanks also to Andreas Mueller <a.mueller at paradatec.de>,
- and Anonymous.
- * Allow override of MALLOC_ALIGNMENT (Thanks to Ruud Waij for
- helping test this.)
- * memalign: check alignment arg
- * realloc: don't try to shift chunks backwards, since this
- leads to more fragmentation in some programs and doesn't
- seem to help in any others.
- * Collect all cases in malloc requiring system memory into sysmalloc
- * Use mmap as backup to sbrk
- * Place all internal state in malloc_state
- * Introduce fastbins (although similar to 2.5.1)
- * Many minor tunings and cosmetic improvements
- * Introduce USE_PUBLIC_MALLOC_WRAPPERS, USE_MALLOC_LOCK
- * Introduce MALLOC_FAILURE_ACTION, MORECORE_CONTIGUOUS
- Thanks to Tony E. Bennett <tbennett@nvidia.com> and others.
- * Include errno.h to support default failure action.
-
- V2.6.6 Sun Dec 5 07:42:19 1999 Doug Lea (dl at gee)
- * return null for negative arguments
- * Added Several WIN32 cleanups from Martin C. Fong <mcfong at yahoo.com>
- * Add 'LACKS_SYS_PARAM_H' for those systems without 'sys/param.h'
- (e.g. WIN32 platforms)
- * Cleanup header file inclusion for WIN32 platforms
- * Cleanup code to avoid Microsoft Visual C++ compiler complaints
- * Add 'USE_DL_PREFIX' to quickly allow co-existence with existing
- memory allocation routines
- * Set 'malloc_getpagesize' for WIN32 platforms (needs more work)
- * Use 'assert' rather than 'ASSERT' in WIN32 code to conform to
- usage of 'assert' in non-WIN32 code
- * Improve WIN32 'sbrk()' emulation's 'findRegion()' routine to
- avoid infinite loop
- * Always call 'fREe()' rather than 'free()'
-
- V2.6.5 Wed Jun 17 15:57:31 1998 Doug Lea (dl at gee)
- * Fixed ordering problem with boundary-stamping
-
- V2.6.3 Sun May 19 08:17:58 1996 Doug Lea (dl at gee)
- * Added pvalloc, as recommended by H.J. Liu
- * Added 64bit pointer support mainly from Wolfram Gloger
- * Added anonymously donated WIN32 sbrk emulation
- * Malloc, calloc, getpagesize: add optimizations from Raymond Nijssen
- * malloc_extend_top: fix mask error that caused wastage after
- foreign sbrks
- * Add linux mremap support code from HJ Liu
-
- V2.6.2 Tue Dec 5 06:52:55 1995 Doug Lea (dl at gee)
- * Integrated most documentation with the code.
- * Add support for mmap, with help from
- Wolfram Gloger (Gloger@lrz.uni-muenchen.de).
- * Use last_remainder in more cases.
- * Pack bins using idea from colin@nyx10.cs.du.edu
- * Use ordered bins instead of best-fit threshhold
- * Eliminate block-local decls to simplify tracing and debugging.
- * Support another case of realloc via move into top
- * Fix error occuring when initial sbrk_base not word-aligned.
- * Rely on page size for units instead of SBRK_UNIT to
- avoid surprises about sbrk alignment conventions.
- * Add mallinfo, mallopt. Thanks to Raymond Nijssen
- (raymond@es.ele.tue.nl) for the suggestion.
- * Add `pad' argument to malloc_trim and top_pad mallopt parameter.
- * More precautions for cases where other routines call sbrk,
- courtesy of Wolfram Gloger (Gloger@lrz.uni-muenchen.de).
- * Added macros etc., allowing use in linux libc from
- H.J. Lu (hjl@gnu.ai.mit.edu)
- * Inverted this history list
-
- V2.6.1 Sat Dec 2 14:10:57 1995 Doug Lea (dl at gee)
- * Re-tuned and fixed to behave more nicely with V2.6.0 changes.
- * Removed all preallocation code since under current scheme
- the work required to undo bad preallocations exceeds
- the work saved in good cases for most test programs.
- * No longer use return list or unconsolidated bins since
- no scheme using them consistently outperforms those that don't
- given above changes.
- * Use best fit for very large chunks to prevent some worst-cases.
- * Added some support for debugging
-
- V2.6.0 Sat Nov 4 07:05:23 1995 Doug Lea (dl at gee)
- * Removed footers when chunks are in use. Thanks to
- Paul Wilson (wilson@cs.texas.edu) for the suggestion.
-
- V2.5.4 Wed Nov 1 07:54:51 1995 Doug Lea (dl at gee)
- * Added malloc_trim, with help from Wolfram Gloger
- (wmglo@Dent.MED.Uni-Muenchen.DE).
-
- V2.5.3 Tue Apr 26 10:16:01 1994 Doug Lea (dl at g)
-
- V2.5.2 Tue Apr 5 16:20:40 1994 Doug Lea (dl at g)
- * realloc: try to expand in both directions
- * malloc: swap order of clean-bin strategy;
- * realloc: only conditionally expand backwards
- * Try not to scavenge used bins
- * Use bin counts as a guide to preallocation
- * Occasionally bin return list chunks in first scan
- * Add a few optimizations from colin@nyx10.cs.du.edu
-
- V2.5.1 Sat Aug 14 15:40:43 1993 Doug Lea (dl at g)
- * faster bin computation & slightly different binning
- * merged all consolidations to one part of malloc proper
- (eliminating old malloc_find_space & malloc_clean_bin)
- * Scan 2 returns chunks (not just 1)
- * Propagate failure in realloc if malloc returns 0
- * Add stuff to allow compilation on non-ANSI compilers
- from kpv@research.att.com
-
- V2.5 Sat Aug 7 07:41:59 1993 Doug Lea (dl at g.oswego.edu)
- * removed potential for odd address access in prev_chunk
- * removed dependency on getpagesize.h
- * misc cosmetics and a bit more internal documentation
- * anticosmetics: mangled names in macros to evade debugger strangeness
- * tested on sparc, hp-700, dec-mips, rs6000
- with gcc & native cc (hp, dec only) allowing
- Detlefs & Zorn comparison study (in SIGPLAN Notices.)
-
- Trial version Fri Aug 28 13:14:29 1992 Doug Lea (dl at g.oswego.edu)
- * Based loosely on libg++-1.2X malloc. (It retains some of the overall
- structure of old version, but most details differ.)
-
-*/
-
diff --git a/unsupported/test/mpreal/dlmalloc.h b/unsupported/test/mpreal/dlmalloc.h deleted file mode 100755 index a90dcb6f5..000000000 --- a/unsupported/test/mpreal/dlmalloc.h +++ /dev/null @@ -1,562 +0,0 @@ -/*
- Default header file for malloc-2.8.x, written by Doug Lea
- and released to the public domain, as explained at
- http://creativecommons.org/licenses/publicdomain.
-
- last update: Wed May 27 14:25:17 2009 Doug Lea (dl at gee)
-
- This header is for ANSI C/C++ only. You can set any of
- the following #defines before including:
-
- * If USE_DL_PREFIX is defined, it is assumed that malloc.c
- was also compiled with this option, so all routines
- have names starting with "dl".
-
- * If HAVE_USR_INCLUDE_MALLOC_H is defined, it is assumed that this
- file will be #included AFTER <malloc.h>. This is needed only if
- your system defines a struct mallinfo that is incompatible with the
- standard one declared here. Otherwise, you can include this file
- INSTEAD of your system system <malloc.h>. At least on ANSI, all
- declarations should be compatible with system versions
-
- * If MSPACES is defined, declarations for mspace versions are included.
-*/
-
-#ifndef MALLOC_280_H
-#define MALLOC_280_H
-
-#define USE_DL_PREFIX
-
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-#include <stddef.h> /* for size_t */
-
-#ifndef ONLY_MSPACES
-#define ONLY_MSPACES 0 /* define to a value */
-#endif /* ONLY_MSPACES */
-#ifndef NO_MALLINFO
-#define NO_MALLINFO 0
-#endif /* NO_MALLINFO */
-
-
-#if !ONLY_MSPACES
-
-#ifndef USE_DL_PREFIX
-#define dlcalloc calloc
-#define dlfree free
-#define dlmalloc malloc
-#define dlmemalign memalign
-#define dlrealloc realloc
-#define dlvalloc valloc
-#define dlpvalloc pvalloc
-#define dlmallinfo mallinfo
-#define dlmallopt mallopt
-#define dlmalloc_trim malloc_trim
-#define dlmalloc_stats malloc_stats
-#define dlmalloc_usable_size malloc_usable_size
-#define dlmalloc_footprint malloc_footprint
-#define dlindependent_calloc independent_calloc
-#define dlindependent_comalloc independent_comalloc
-#endif /* USE_DL_PREFIX */
-#if !NO_MALLINFO
-#ifndef HAVE_USR_INCLUDE_MALLOC_H
-#ifndef _MALLOC_H
-#ifndef MALLINFO_FIELD_TYPE
-#define MALLINFO_FIELD_TYPE size_t
-#endif /* MALLINFO_FIELD_TYPE */
-#ifndef STRUCT_MALLINFO_DECLARED
-#define STRUCT_MALLINFO_DECLARED 1
-struct mallinfo {
- MALLINFO_FIELD_TYPE arena; /* non-mmapped space allocated from system */
- MALLINFO_FIELD_TYPE ordblks; /* number of free chunks */
- MALLINFO_FIELD_TYPE smblks; /* always 0 */
- MALLINFO_FIELD_TYPE hblks; /* always 0 */
- MALLINFO_FIELD_TYPE hblkhd; /* space in mmapped regions */
- MALLINFO_FIELD_TYPE usmblks; /* maximum total allocated space */
- MALLINFO_FIELD_TYPE fsmblks; /* always 0 */
- MALLINFO_FIELD_TYPE uordblks; /* total allocated space */
- MALLINFO_FIELD_TYPE fordblks; /* total free space */
- MALLINFO_FIELD_TYPE keepcost; /* releasable (via malloc_trim) space */
-};
-#endif /* STRUCT_MALLINFO_DECLARED */
-#endif /* _MALLOC_H */
-#endif /* HAVE_USR_INCLUDE_MALLOC_H */
-#endif /* !NO_MALLINFO */
-
-/*
- malloc(size_t n)
- Returns a pointer to a newly allocated chunk of at least n bytes, or
- null if no space is available, in which case errno is set to ENOMEM
- on ANSI C systems.
-
- If n is zero, malloc returns a minimum-sized chunk. (The minimum
- size is 16 bytes on most 32bit systems, and 32 bytes on 64bit
- systems.) Note that size_t is an unsigned type, so calls with
- arguments that would be negative if signed are interpreted as
- requests for huge amounts of space, which will often fail. The
- maximum supported value of n differs across systems, but is in all
- cases less than the maximum representable value of a size_t.
-*/
-void* dlmalloc(size_t);
-
-/*
- free(void* p)
- Releases the chunk of memory pointed to by p, that had been previously
- allocated using malloc or a related routine such as realloc.
- It has no effect if p is null. If p was not malloced or already
- freed, free(p) will by default cuase the current program to abort.
-*/
-void dlfree(void*);
-
-/*
- calloc(size_t n_elements, size_t element_size);
- Returns a pointer to n_elements * element_size bytes, with all locations
- set to zero.
-*/
-void* dlcalloc(size_t, size_t);
-
-/*
- realloc(void* p, size_t n)
- Returns a pointer to a chunk of size n that contains the same data
- as does chunk p up to the minimum of (n, p's size) bytes, or null
- if no space is available.
-
- The returned pointer may or may not be the same as p. The algorithm
- prefers extending p in most cases when possible, otherwise it
- employs the equivalent of a malloc-copy-free sequence.
-
- If p is null, realloc is equivalent to malloc.
-
- If space is not available, realloc returns null, errno is set (if on
- ANSI) and p is NOT freed.
-
- if n is for fewer bytes than already held by p, the newly unused
- space is lopped off and freed if possible. realloc with a size
- argument of zero (re)allocates a minimum-sized chunk.
-
- The old unix realloc convention of allowing the last-free'd chunk
- to be used as an argument to realloc is not supported.
-*/
-
-void* dlrealloc(void*, size_t);
-
-/*
- memalign(size_t alignment, size_t n);
- Returns a pointer to a newly allocated chunk of n bytes, aligned
- in accord with the alignment argument.
-
- The alignment argument should be a power of two. If the argument is
- not a power of two, the nearest greater power is used.
- 8-byte alignment is guaranteed by normal malloc calls, so don't
- bother calling memalign with an argument of 8 or less.
-
- Overreliance on memalign is a sure way to fragment space.
-*/
-void* dlmemalign(size_t, size_t);
-
-/*
- valloc(size_t n);
- Equivalent to memalign(pagesize, n), where pagesize is the page
- size of the system. If the pagesize is unknown, 4096 is used.
-*/
-void* dlvalloc(size_t);
-
-/*
- mallopt(int parameter_number, int parameter_value)
- Sets tunable parameters The format is to provide a
- (parameter-number, parameter-value) pair. mallopt then sets the
- corresponding parameter to the argument value if it can (i.e., so
- long as the value is meaningful), and returns 1 if successful else
- 0. SVID/XPG/ANSI defines four standard param numbers for mallopt,
- normally defined in malloc.h. None of these are use in this malloc,
- so setting them has no effect. But this malloc also supports other
- options in mallopt:
-
- Symbol param # default allowed param values
- M_TRIM_THRESHOLD -1 2*1024*1024 any (-1U disables trimming)
- M_GRANULARITY -2 page size any power of 2 >= page size
- M_MMAP_THRESHOLD -3 256*1024 any (or 0 if no MMAP support)
-*/
-int dlmallopt(int, int);
-
-#define M_TRIM_THRESHOLD (-1)
-#define M_GRANULARITY (-2)
-#define M_MMAP_THRESHOLD (-3)
-
-
-/*
- malloc_footprint();
- Returns the number of bytes obtained from the system. The total
- number of bytes allocated by malloc, realloc etc., is less than this
- value. Unlike mallinfo, this function returns only a precomputed
- result, so can be called frequently to monitor memory consumption.
- Even if locks are otherwise defined, this function does not use them,
- so results might not be up to date.
-*/
-size_t dlmalloc_footprint();
-
-#if !NO_MALLINFO
-/*
- mallinfo()
- Returns (by copy) a struct containing various summary statistics:
-
- arena: current total non-mmapped bytes allocated from system
- ordblks: the number of free chunks
- smblks: always zero.
- hblks: current number of mmapped regions
- hblkhd: total bytes held in mmapped regions
- usmblks: the maximum total allocated space. This will be greater
- than current total if trimming has occurred.
- fsmblks: always zero
- uordblks: current total allocated space (normal or mmapped)
- fordblks: total free space
- keepcost: the maximum number of bytes that could ideally be released
- back to system via malloc_trim. ("ideally" means that
- it ignores page restrictions etc.)
-
- Because these fields are ints, but internal bookkeeping may
- be kept as longs, the reported values may wrap around zero and
- thus be inaccurate.
-*/
-
-struct mallinfo dlmallinfo(void);
-#endif /* NO_MALLINFO */
-
-/*
- independent_calloc(size_t n_elements, size_t element_size, void* chunks[]);
-
- independent_calloc is similar to calloc, but instead of returning a
- single cleared space, it returns an array of pointers to n_elements
- independent elements that can hold contents of size elem_size, each
- of which starts out cleared, and can be independently freed,
- realloc'ed etc. The elements are guaranteed to be adjacently
- allocated (this is not guaranteed to occur with multiple callocs or
- mallocs), which may also improve cache locality in some
- applications.
-
- The "chunks" argument is optional (i.e., may be null, which is
- probably the most typical usage). If it is null, the returned array
- is itself dynamically allocated and should also be freed when it is
- no longer needed. Otherwise, the chunks array must be of at least
- n_elements in length. It is filled in with the pointers to the
- chunks.
-
- In either case, independent_calloc returns this pointer array, or
- null if the allocation failed. If n_elements is zero and "chunks"
- is null, it returns a chunk representing an array with zero elements
- (which should be freed if not wanted).
-
- Each element must be individually freed when it is no longer
- needed. If you'd like to instead be able to free all at once, you
- should instead use regular calloc and assign pointers into this
- space to represent elements. (In this case though, you cannot
- independently free elements.)
-
- independent_calloc simplifies and speeds up implementations of many
- kinds of pools. It may also be useful when constructing large data
- structures that initially have a fixed number of fixed-sized nodes,
- but the number is not known at compile time, and some of the nodes
- may later need to be freed. For example:
-
- struct Node { int item; struct Node* next; };
-
- struct Node* build_list() {
- struct Node** pool;
- int n = read_number_of_nodes_needed();
- if (n <= 0) return 0;
- pool = (struct Node**)(independent_calloc(n, sizeof(struct Node), 0);
- if (pool == 0) die();
- // organize into a linked list...
- struct Node* first = pool[0];
- for (i = 0; i < n-1; ++i)
- pool[i]->next = pool[i+1];
- free(pool); // Can now free the array (or not, if it is needed later)
- return first;
- }
-*/
-void** dlindependent_calloc(size_t, size_t, void**);
-
-/*
- independent_comalloc(size_t n_elements, size_t sizes[], void* chunks[]);
-
- independent_comalloc allocates, all at once, a set of n_elements
- chunks with sizes indicated in the "sizes" array. It returns
- an array of pointers to these elements, each of which can be
- independently freed, realloc'ed etc. The elements are guaranteed to
- be adjacently allocated (this is not guaranteed to occur with
- multiple callocs or mallocs), which may also improve cache locality
- in some applications.
-
- The "chunks" argument is optional (i.e., may be null). If it is null
- the returned array is itself dynamically allocated and should also
- be freed when it is no longer needed. Otherwise, the chunks array
- must be of at least n_elements in length. It is filled in with the
- pointers to the chunks.
-
- In either case, independent_comalloc returns this pointer array, or
- null if the allocation failed. If n_elements is zero and chunks is
- null, it returns a chunk representing an array with zero elements
- (which should be freed if not wanted).
-
- Each element must be individually freed when it is no longer
- needed. If you'd like to instead be able to free all at once, you
- should instead use a single regular malloc, and assign pointers at
- particular offsets in the aggregate space. (In this case though, you
- cannot independently free elements.)
-
- independent_comallac differs from independent_calloc in that each
- element may have a different size, and also that it does not
- automatically clear elements.
-
- independent_comalloc can be used to speed up allocation in cases
- where several structs or objects must always be allocated at the
- same time. For example:
-
- struct Head { ... }
- struct Foot { ... }
-
- void send_message(char* msg) {
- int msglen = strlen(msg);
- size_t sizes[3] = { sizeof(struct Head), msglen, sizeof(struct Foot) };
- void* chunks[3];
- if (independent_comalloc(3, sizes, chunks) == 0)
- die();
- struct Head* head = (struct Head*)(chunks[0]);
- char* body = (char*)(chunks[1]);
- struct Foot* foot = (struct Foot*)(chunks[2]);
- // ...
- }
-
- In general though, independent_comalloc is worth using only for
- larger values of n_elements. For small values, you probably won't
- detect enough difference from series of malloc calls to bother.
-
- Overuse of independent_comalloc can increase overall memory usage,
- since it cannot reuse existing noncontiguous small chunks that
- might be available for some of the elements.
-*/
-void** dlindependent_comalloc(size_t, size_t*, void**);
-
-
-/*
- pvalloc(size_t n);
- Equivalent to valloc(minimum-page-that-holds(n)), that is,
- round up n to nearest pagesize.
- */
-void* dlpvalloc(size_t);
-
-/*
- malloc_trim(size_t pad);
-
- If possible, gives memory back to the system (via negative arguments
- to sbrk) if there is unused memory at the `high' end of the malloc
- pool or in unused MMAP segments. You can call this after freeing
- large blocks of memory to potentially reduce the system-level memory
- requirements of a program. However, it cannot guarantee to reduce
- memory. Under some allocation patterns, some large free blocks of
- memory will be locked between two used chunks, so they cannot be
- given back to the system.
-
- The `pad' argument to malloc_trim represents the amount of free
- trailing space to leave untrimmed. If this argument is zero, only
- the minimum amount of memory to maintain internal data structures
- will be left. Non-zero arguments can be supplied to maintain enough
- trailing space to service future expected allocations without having
- to re-obtain memory from the system.
-
- Malloc_trim returns 1 if it actually released any memory, else 0.
-*/
-int dlmalloc_trim(size_t);
-
-/*
- malloc_stats();
- Prints on stderr the amount of space obtained from the system (both
- via sbrk and mmap), the maximum amount (which may be more than
- current if malloc_trim and/or munmap got called), and the current
- number of bytes allocated via malloc (or realloc, etc) but not yet
- freed. Note that this is the number of bytes allocated, not the
- number requested. It will be larger than the number requested
- because of alignment and bookkeeping overhead. Because it includes
- alignment wastage as being in use, this figure may be greater than
- zero even when no user-level chunks are allocated.
-
- The reported current and maximum system memory can be inaccurate if
- a program makes other calls to system memory allocation functions
- (normally sbrk) outside of malloc.
-
- malloc_stats prints only the most commonly interesting statistics.
- More information can be obtained by calling mallinfo.
-*/
-void dlmalloc_stats();
-
-#endif /* !ONLY_MSPACES */
-
-/*
- malloc_usable_size(void* p);
-
- Returns the number of bytes you can actually use in
- an allocated chunk, which may be more than you requested (although
- often not) due to alignment and minimum size constraints.
- You can use this many bytes without worrying about
- overwriting other allocated objects. This is not a particularly great
- programming practice. malloc_usable_size can be more useful in
- debugging and assertions, for example:
-
- p = malloc(n);
- assert(malloc_usable_size(p) >= 256);
-*/
-size_t dlmalloc_usable_size(void*);
-
-
-#if MSPACES
-
-/*
- mspace is an opaque type representing an independent
- region of space that supports mspace_malloc, etc.
-*/
-typedef void* mspace;
-
-/*
- create_mspace creates and returns a new independent space with the
- given initial capacity, or, if 0, the default granularity size. It
- returns null if there is no system memory available to create the
- space. If argument locked is non-zero, the space uses a separate
- lock to control access. The capacity of the space will grow
- dynamically as needed to service mspace_malloc requests. You can
- control the sizes of incremental increases of this space by
- compiling with a different DEFAULT_GRANULARITY or dynamically
- setting with mallopt(M_GRANULARITY, value).
-*/
-mspace create_mspace(size_t capacity, int locked);
-
-/*
- destroy_mspace destroys the given space, and attempts to return all
- of its memory back to the system, returning the total number of
- bytes freed. After destruction, the results of access to all memory
- used by the space become undefined.
-*/
-size_t destroy_mspace(mspace msp);
-
-/*
- create_mspace_with_base uses the memory supplied as the initial base
- of a new mspace. Part (less than 128*sizeof(size_t) bytes) of this
- space is used for bookkeeping, so the capacity must be at least this
- large. (Otherwise 0 is returned.) When this initial space is
- exhausted, additional memory will be obtained from the system.
- Destroying this space will deallocate all additionally allocated
- space (if possible) but not the initial base.
-*/
-mspace create_mspace_with_base(void* base, size_t capacity, int locked);
-
-/*
- mspace_track_large_chunks controls whether requests for large chunks
- are allocated in their own untracked mmapped regions, separate from
- others in this mspace. By default large chunks are not tracked,
- which reduces fragmentation. However, such chunks are not
- necessarily released to the system upon destroy_mspace. Enabling
- tracking by setting to true may increase fragmentation, but avoids
- leakage when relying on destroy_mspace to release all memory
- allocated using this space. The function returns the previous
- setting.
-*/
-int mspace_track_large_chunks(mspace msp, int enable);
-
-/*
- mspace_malloc behaves as malloc, but operates within
- the given space.
-*/
-void* mspace_malloc(mspace msp, size_t bytes);
-
-/*
- mspace_free behaves as free, but operates within
- the given space.
-
- If compiled with FOOTERS==1, mspace_free is not actually needed.
- free may be called instead of mspace_free because freed chunks from
- any space are handled by their originating spaces.
-*/
-void mspace_free(mspace msp, void* mem);
-
-/*
- mspace_realloc behaves as realloc, but operates within
- the given space.
-
- If compiled with FOOTERS==1, mspace_realloc is not actually
- needed. realloc may be called instead of mspace_realloc because
- realloced chunks from any space are handled by their originating
- spaces.
-*/
-void* mspace_realloc(mspace msp, void* mem, size_t newsize);
-
-/*
- mspace_calloc behaves as calloc, but operates within
- the given space.
-*/
-void* mspace_calloc(mspace msp, size_t n_elements, size_t elem_size);
-
-/*
- mspace_memalign behaves as memalign, but operates within
- the given space.
-*/
-void* mspace_memalign(mspace msp, size_t alignment, size_t bytes);
-
-/*
- mspace_independent_calloc behaves as independent_calloc, but
- operates within the given space.
-*/
-void** mspace_independent_calloc(mspace msp, size_t n_elements,
- size_t elem_size, void* chunks[]);
-
-/*
- mspace_independent_comalloc behaves as independent_comalloc, but
- operates within the given space.
-*/
-void** mspace_independent_comalloc(mspace msp, size_t n_elements,
- size_t sizes[], void* chunks[]);
-
-/*
- mspace_footprint() returns the number of bytes obtained from the
- system for this space.
-*/
-size_t mspace_footprint(mspace msp);
-
-
-#if !NO_MALLINFO
-/*
- mspace_mallinfo behaves as mallinfo, but reports properties of
- the given space.
-*/
-struct mallinfo mspace_mallinfo(mspace msp);
-#endif /* NO_MALLINFO */
-
-/*
- malloc_usable_size(void* p) behaves the same as malloc_usable_size;
-*/
- size_t mspace_usable_size(void* mem);
-
-/*
- mspace_malloc_stats behaves as malloc_stats, but reports
- properties of the given space.
-*/
-void mspace_malloc_stats(mspace msp);
-
-/*
- mspace_trim behaves as malloc_trim, but
- operates within the given space.
-*/
-int mspace_trim(mspace msp, size_t pad);
-
-/*
- An alias for mallopt.
-*/
-int mspace_mallopt(int, int);
-
-#endif /* MSPACES */
-
-#ifdef __cplusplus
-}; /* end of extern "C" */
-#endif
-
-#endif /* MALLOC_280_H */
diff --git a/unsupported/test/mpreal/mpreal.cpp b/unsupported/test/mpreal/mpreal.cpp deleted file mode 100644 index 5c23544ef..000000000 --- a/unsupported/test/mpreal/mpreal.cpp +++ /dev/null @@ -1,597 +0,0 @@ -/*
- Multi-precision real number class. C++ interface fo MPFR library.
- Project homepage: http://www.holoborodko.com/pavel/
- Contact e-mail: pavel@holoborodko.com
-
- Copyright (c) 2008-2011 Pavel Holoborodko
-
- Core Developers:
- Pavel Holoborodko, Dmitriy Gubanov, Konstantin Holoborodko.
-
- Contributors:
- Brian Gladman, Helmut Jarausch, Fokko Beekhof, Ulrich Mutze,
- Heinz van Saanen, Pere Constans, Peter van Hoof, Gael Guennebaud,
- Tsai Chia Cheng, Alexei Zubanov.
-
- ****************************************************************************
- This library is free software; you can redistribute it and/or
- modify it under the terms of the GNU Lesser General Public
- License as published by the Free Software Foundation; either
- version 2.1 of the License, or (at your option) any later version.
-
- This library is distributed in the hope that it will be useful,
- but WITHOUT ANY WARRANTY; without even the implied warranty of
- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- Lesser General Public License for more details.
-
- You should have received a copy of the GNU Lesser General Public
- License along with this library; if not, write to the Free Software
- Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
-
- ****************************************************************************
- ****************************************************************************
- Redistribution and use in source and binary forms, with or without
- modification, are permitted provided that the following conditions
- are met:
-
- 1. Redistributions of source code must retain the above copyright
- notice, this list of conditions and the following disclaimer.
-
- 2. Redistributions in binary form must reproduce the above copyright
- notice, this list of conditions and the following disclaimer in the
- documentation and/or other materials provided with the distribution.
-
- 3. The name of the author may be used to endorse or promote products
- derived from this software without specific prior written permission.
-
- THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
- ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- SUCH DAMAGE.
-*/
-#include <cstring>
-#include "mpreal.h"
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
-#include "dlmalloc.h"
-#endif
-
-using std::ws;
-using std::cerr;
-using std::endl;
-using std::string;
-using std::ostream;
-using std::istream;
-
-namespace mpfr{
-
-mp_rnd_t mpreal::default_rnd = MPFR_RNDN; //(mpfr_get_default_rounding_mode)();
-mp_prec_t mpreal::default_prec = 64; //(mpfr_get_default_prec)();
-int mpreal::default_base = 10;
-int mpreal::double_bits = -1;
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
-bool mpreal::is_custom_malloc = false;
-#endif
-
-// Default constructor: creates mp number and initializes it to 0.
-mpreal::mpreal()
-{
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
- set_custom_malloc();
-#endif
-
- mpfr_init2(mp,default_prec);
- mpfr_set_ui(mp,0,default_rnd);
-
- MPREAL_MSVC_DEBUGVIEW_CODE;
-}
-
-mpreal::mpreal(const mpreal& u)
-{
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
- set_custom_malloc();
-#endif
-
- mpfr_init2(mp,mpfr_get_prec(u.mp));
- mpfr_set(mp,u.mp,default_rnd);
-
- MPREAL_MSVC_DEBUGVIEW_CODE;
-}
-
-mpreal::mpreal(const mpfr_t u)
-{
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
- set_custom_malloc();
-#endif
-
- mpfr_init2(mp,mpfr_get_prec(u));
- mpfr_set(mp,u,default_rnd);
-
- MPREAL_MSVC_DEBUGVIEW_CODE;
-}
-
-mpreal::mpreal(const mpf_t u)
-{
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
- set_custom_malloc();
-#endif
-
- mpfr_init2(mp,(mp_prec_t) mpf_get_prec(u)); // (gmp: mp_bitcnt_t) unsigned long -> long (mpfr: mp_prec_t)
- mpfr_set_f(mp,u,default_rnd);
-
- MPREAL_MSVC_DEBUGVIEW_CODE;
-}
-
-mpreal::mpreal(const mpz_t u, mp_prec_t prec, mp_rnd_t mode)
-{
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
- set_custom_malloc();
-#endif
-
- mpfr_init2(mp,prec);
- mpfr_set_z(mp,u,mode);
-
- MPREAL_MSVC_DEBUGVIEW_CODE;
-}
-
-mpreal::mpreal(const mpq_t u, mp_prec_t prec, mp_rnd_t mode)
-{
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
- set_custom_malloc();
-#endif
-
- mpfr_init2(mp,prec);
- mpfr_set_q(mp,u,mode);
-
- MPREAL_MSVC_DEBUGVIEW_CODE;
-}
-
-mpreal::mpreal(const double u, mp_prec_t prec, mp_rnd_t mode)
-{
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
- set_custom_malloc();
-#endif
-
- if(double_bits == -1 || fits_in_bits(u, double_bits))
- {
- mpfr_init2(mp,prec);
- mpfr_set_d(mp,u,mode);
-
- MPREAL_MSVC_DEBUGVIEW_CODE;
- }
- else
- throw conversion_overflow();
-}
-
-mpreal::mpreal(const long double u, mp_prec_t prec, mp_rnd_t mode)
-{
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
- set_custom_malloc();
-#endif
-
- mpfr_init2(mp,prec);
- mpfr_set_ld(mp,u,mode);
-
- MPREAL_MSVC_DEBUGVIEW_CODE;
-}
-
-mpreal::mpreal(const unsigned long int u, mp_prec_t prec, mp_rnd_t mode)
-{
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
- set_custom_malloc();
-#endif
-
- mpfr_init2(mp,prec);
- mpfr_set_ui(mp,u,mode);
-
- MPREAL_MSVC_DEBUGVIEW_CODE;
-}
-
-mpreal::mpreal(const unsigned int u, mp_prec_t prec, mp_rnd_t mode)
-{
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
- set_custom_malloc();
-#endif
-
- mpfr_init2(mp,prec);
- mpfr_set_ui(mp,u,mode);
-
- MPREAL_MSVC_DEBUGVIEW_CODE;
-}
-
-mpreal::mpreal(const long int u, mp_prec_t prec, mp_rnd_t mode)
-{
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
- set_custom_malloc();
-#endif
-
- mpfr_init2(mp,prec);
- mpfr_set_si(mp,u,mode);
-
- MPREAL_MSVC_DEBUGVIEW_CODE;
-}
-
-mpreal::mpreal(const int u, mp_prec_t prec, mp_rnd_t mode)
-{
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
- set_custom_malloc();
-#endif
-
- mpfr_init2(mp,prec);
- mpfr_set_si(mp,u,mode);
-
- MPREAL_MSVC_DEBUGVIEW_CODE;
-}
-
-#if defined (MPREAL_HAVE_INT64_SUPPORT)
-mpreal::mpreal(const uint64_t u, mp_prec_t prec, mp_rnd_t mode)
-{
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
- set_custom_malloc();
-#endif
-
- mpfr_init2(mp,prec);
- mpfr_set_uj(mp, u, mode);
-
- MPREAL_MSVC_DEBUGVIEW_CODE;
-}
-
-mpreal::mpreal(const int64_t u, mp_prec_t prec, mp_rnd_t mode)
-{
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
- set_custom_malloc();
-#endif
-
- mpfr_init2(mp,prec);
- mpfr_set_sj(mp, u, mode);
-
- MPREAL_MSVC_DEBUGVIEW_CODE;
-}
-#endif
-
-mpreal::mpreal(const char* s, mp_prec_t prec, int base, mp_rnd_t mode)
-{
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
- set_custom_malloc();
-#endif
-
- mpfr_init2(mp,prec);
- mpfr_set_str(mp, s, base, mode);
-
- MPREAL_MSVC_DEBUGVIEW_CODE;
-}
-
-mpreal::mpreal(const std::string& s, mp_prec_t prec, int base, mp_rnd_t mode)
-{
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
- set_custom_malloc();
-#endif
-
- mpfr_init2(mp,prec);
- mpfr_set_str(mp, s.c_str(), base, mode);
-
- MPREAL_MSVC_DEBUGVIEW_CODE;
-}
-
-mpreal::~mpreal()
-{
- mpfr_clear(mp);
-}
-
-// Operators - Assignment
-mpreal& mpreal::operator=(const char* s)
-{
- mpfr_t t;
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
- set_custom_malloc();
-#endif
-
- if(0==mpfr_init_set_str(t,s,default_base,default_rnd))
- {
- // We will rewrite mp anyway, so flash it and resize
- mpfr_set_prec(mp,mpfr_get_prec(t));
- mpfr_set(mp,t,mpreal::default_rnd);
- mpfr_clear(t);
-
- MPREAL_MSVC_DEBUGVIEW_CODE;
-
- }else{
- mpfr_clear(t);
- }
-
- return *this;
-}
-
-const mpreal fma (const mpreal& v1, const mpreal& v2, const mpreal& v3, mp_rnd_t rnd_mode)
-{
- mpreal a;
- mp_prec_t p1, p2, p3;
-
- p1 = v1.get_prec();
- p2 = v2.get_prec();
- p3 = v3.get_prec();
-
- a.set_prec(p3>p2?(p3>p1?p3:p1):(p2>p1?p2:p1));
-
- mpfr_fma(a.mp,v1.mp,v2.mp,v3.mp,rnd_mode);
- return a;
-}
-
-const mpreal fms (const mpreal& v1, const mpreal& v2, const mpreal& v3, mp_rnd_t rnd_mode)
-{
- mpreal a;
- mp_prec_t p1, p2, p3;
-
- p1 = v1.get_prec();
- p2 = v2.get_prec();
- p3 = v3.get_prec();
-
- a.set_prec(p3>p2?(p3>p1?p3:p1):(p2>p1?p2:p1));
-
- mpfr_fms(a.mp,v1.mp,v2.mp,v3.mp,rnd_mode);
- return a;
-}
-
-const mpreal agm (const mpreal& v1, const mpreal& v2, mp_rnd_t rnd_mode)
-{
- mpreal a;
- mp_prec_t p1, p2;
-
- p1 = v1.get_prec();
- p2 = v2.get_prec();
-
- a.set_prec(p1>p2?p1:p2);
-
- mpfr_agm(a.mp, v1.mp, v2.mp, rnd_mode);
-
- return a;
-}
-
-const mpreal sum (const mpreal tab[], unsigned long int n, mp_rnd_t rnd_mode)
-{
- mpreal x;
- mpfr_ptr* t;
- unsigned long int i;
-
- t = new mpfr_ptr[n];
- for (i=0;i<n;i++) t[i] = (mpfr_ptr)tab[i].mp;
- mpfr_sum(x.mp,t,n,rnd_mode);
- delete[] t;
- return x;
-}
-
-const mpreal remquo (long* q, const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode)
-{
- mpreal a;
- mp_prec_t yp, xp;
-
- yp = y.get_prec();
- xp = x.get_prec();
-
- a.set_prec(yp>xp?yp:xp);
-
- mpfr_remquo(a.mp,q, x.mp, y.mp, rnd_mode);
-
- return a;
-}
-
-template <class T>
-std::string toString(T t, std::ios_base & (*f)(std::ios_base&))
-{
- std::ostringstream oss;
- oss << f << t;
- return oss.str();
-}
-
-#if (MPFR_VERSION >= MPFR_VERSION_NUM(2,4,0))
-
-std::string mpreal::toString(const std::string& format) const
-{
- char *s = NULL;
- string out;
-
- if( !format.empty() )
- {
- if(!(mpfr_asprintf(&s,format.c_str(),mp) < 0))
- {
- out = std::string(s);
-
- mpfr_free_str(s);
- }
- }
-
- return out;
-}
-
-#endif
-
-std::string mpreal::toString(int n, int b, mp_rnd_t mode) const
-{
- (void)b;
- (void)mode;
-#if (MPFR_VERSION >= MPFR_VERSION_NUM(2,4,0))
-
- // Use MPFR native function for output
- char format[128];
- int digits;
-
- digits = n > 0 ? n : bits2digits(mpfr_get_prec(mp));
-
- sprintf(format,"%%.%dRNg",digits); // Default format
-
- return toString(std::string(format));
-
-#else
-
- char *s, *ns = NULL;
- size_t slen, nslen;
- mp_exp_t exp;
- string out;
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
- set_custom_malloc();
-#endif
-
- if(mpfr_inf_p(mp))
- {
- if(mpfr_sgn(mp)>0) return "+Inf";
- else return "-Inf";
- }
-
- if(mpfr_zero_p(mp)) return "0";
- if(mpfr_nan_p(mp)) return "NaN";
-
- s = mpfr_get_str(NULL,&exp,b,0,mp,mode);
- ns = mpfr_get_str(NULL,&exp,b,n,mp,mode);
-
- if(s!=NULL && ns!=NULL)
- {
- slen = strlen(s);
- nslen = strlen(ns);
- if(nslen<=slen)
- {
- mpfr_free_str(s);
- s = ns;
- slen = nslen;
- }
- else {
- mpfr_free_str(ns);
- }
-
- // Make human eye-friendly formatting if possible
- if (exp>0 && static_cast<size_t>(exp)<slen)
- {
- if(s[0]=='-')
- {
- // Remove zeros starting from right end
- char* ptr = s+slen-1;
- while (*ptr=='0' && ptr>s+exp) ptr--;
-
- if(ptr==s+exp) out = string(s,exp+1);
- else out = string(s,exp+1)+'.'+string(s+exp+1,ptr-(s+exp+1)+1);
-
- //out = string(s,exp+1)+'.'+string(s+exp+1);
- }
- else
- {
- // Remove zeros starting from right end
- char* ptr = s+slen-1;
- while (*ptr=='0' && ptr>s+exp-1) ptr--;
-
- if(ptr==s+exp-1) out = string(s,exp);
- else out = string(s,exp)+'.'+string(s+exp,ptr-(s+exp)+1);
-
- //out = string(s,exp)+'.'+string(s+exp);
- }
-
- }else{ // exp<0 || exp>slen
- if(s[0]=='-')
- {
- // Remove zeros starting from right end
- char* ptr = s+slen-1;
- while (*ptr=='0' && ptr>s+1) ptr--;
-
- if(ptr==s+1) out = string(s,2);
- else out = string(s,2)+'.'+string(s+2,ptr-(s+2)+1);
-
- //out = string(s,2)+'.'+string(s+2);
- }
- else
- {
- // Remove zeros starting from right end
- char* ptr = s+slen-1;
- while (*ptr=='0' && ptr>s) ptr--;
-
- if(ptr==s) out = string(s,1);
- else out = string(s,1)+'.'+string(s+1,ptr-(s+1)+1);
-
- //out = string(s,1)+'.'+string(s+1);
- }
-
- // Make final string
- if(--exp)
- {
- if(exp>0) out += "e+"+mpfr::toString<mp_exp_t>(exp,std::dec);
- else out += "e"+mpfr::toString<mp_exp_t>(exp,std::dec);
- }
- }
-
- mpfr_free_str(s);
- return out;
- }else{
- return "conversion error!";
- }
-#endif
-}
-
-
-//////////////////////////////////////////////////////////////////////////
-// I/O
-ostream& operator<<(ostream& os, const mpreal& v)
-{
- return os<<v.toString(static_cast<int>(os.precision()));
-}
-
-istream& operator>>(istream &is, mpreal& v)
-{
- string tmp;
- is >> tmp;
- mpfr_set_str(v.mp, tmp.c_str(),mpreal::default_base,mpreal::default_rnd);
- return is;
-}
-
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
- // Optimized dynamic memory allocation/(re-)deallocation.
- void * mpreal::mpreal_allocate(size_t alloc_size)
- {
- return(dlmalloc(alloc_size));
- }
-
- void * mpreal::mpreal_reallocate(void *ptr, size_t old_size, size_t new_size)
- {
- return(dlrealloc(ptr,new_size));
- }
-
- void mpreal::mpreal_free(void *ptr, size_t size)
- {
- dlfree(ptr);
- }
-
- inline void mpreal::set_custom_malloc(void)
- {
- if(!is_custom_malloc)
- {
- mp_set_memory_functions(mpreal_allocate,mpreal_reallocate,mpreal_free);
- is_custom_malloc = true;
- }
- }
-#endif
-
-}
-
diff --git a/unsupported/test/mpreal/mpreal.h b/unsupported/test/mpreal/mpreal.h index c640af947..ef0a6a9f0 100644 --- a/unsupported/test/mpreal/mpreal.h +++ b/unsupported/test/mpreal/mpreal.h @@ -1,59 +1,59 @@ /*
- Multi-precision real number class. C++ interface for MPFR library.
- Project homepage: http://www.holoborodko.com/pavel/
- Contact e-mail: pavel@holoborodko.com
-
- Copyright (c) 2008-2012 Pavel Holoborodko
-
- Core Developers:
- Pavel Holoborodko, Dmitriy Gubanov, Konstantin Holoborodko.
-
- Contributors:
- Brian Gladman, Helmut Jarausch, Fokko Beekhof, Ulrich Mutze,
- Heinz van Saanen, Pere Constans, Peter van Hoof, Gael Guennebaud,
- Tsai Chia Cheng, Alexei Zubanov.
-
- ****************************************************************************
- This library is free software; you can redistribute it and/or
- modify it under the terms of the GNU Lesser General Public
- License as published by the Free Software Foundation; either
- version 2.1 of the License, or (at your option) any later version.
-
- This library is distributed in the hope that it will be useful,
- but WITHOUT ANY WARRANTY; without even the implied warranty of
- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- Lesser General Public License for more details.
-
- You should have received a copy of the GNU Lesser General Public
- License along with this library; if not, write to the Free Software
- Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
-
- ****************************************************************************
- Redistribution and use in source and binary forms, with or without
- modification, are permitted provided that the following conditions
- are met:
-
- 1. Redistributions of source code must retain the above copyright
- notice, this list of conditions and the following disclaimer.
-
- 2. Redistributions in binary form must reproduce the above copyright
- notice, this list of conditions and the following disclaimer in the
- documentation and/or other materials provided with the distribution.
-
- 3. The name of the author may be used to endorse or promote products
- derived from this software without specific prior written permission.
-
- THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
- ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- SUCH DAMAGE.
+ Multi-precision floating point number class for C++.
+ Based on MPFR library: http://mpfr.org
+
+ Project homepage: http://www.holoborodko.com/pavel/mpfr
+ Contact e-mail: pavel@holoborodko.com
+
+ Copyright (c) 2008-2012 Pavel Holoborodko
+
+ Contributors:
+ Dmitriy Gubanov, Konstantin Holoborodko, Brian Gladman,
+ Helmut Jarausch, Fokko Beekhof, Ulrich Mutze, Heinz van Saanen,
+ Pere Constans, Peter van Hoof, Gael Guennebaud, Tsai Chia Cheng,
+ Alexei Zubanov, Jauhien Piatlicki, Victor Berger, John Westwood.
+
+ ****************************************************************************
+ This library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ This library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with this library; if not, write to the Free Software
+ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+
+ ****************************************************************************
+ Redistribution and use in source and binary forms, with or without
+ modification, are permitted provided that the following conditions
+ are met:
+
+ 1. Redistributions of source code must retain the above copyright
+ notice, this list of conditions and the following disclaimer.
+
+ 2. Redistributions in binary form must reproduce the above copyright
+ notice, this list of conditions and the following disclaimer in the
+ documentation and/or other materials provided with the distribution.
+
+ 3. The name of the author may not be used to endorse or promote products
+ derived from this software without specific prior written permission.
+
+ THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ SUCH DAMAGE.
*/
#ifndef __MPREAL_H__
@@ -65,627 +65,785 @@ #include <stdexcept>
#include <cfloat>
#include <cmath>
+#include <limits>
// Options
-#define MPREAL_HAVE_INT64_SUPPORT // int64_t support: available only for MSVC 2010 & GCC
-#define MPREAL_HAVE_MSVC_DEBUGVIEW // Enable Debugger Visualizer (valid only for MSVC in "Debug" builds)
+#define MPREAL_HAVE_INT64_SUPPORT // Enable int64_t support if possible. Available only for MSVC 2010 & GCC.
+#define MPREAL_HAVE_MSVC_DEBUGVIEW // Enable Debugger Visualizer for "Debug" builds in MSVC.
// Detect compiler using signatures from http://predef.sourceforge.net/
#if defined(__GNUC__) && defined(__INTEL_COMPILER)
- #define IsInf(x) isinf(x) // Intel ICC compiler on Linux
-
-#elif defined(_MSC_VER) // Microsoft Visual C++
- #define IsInf(x) (!_finite(x))
+ #define IsInf(x) isinf(x) // Intel ICC compiler on Linux
-#elif defined(__GNUC__)
- #define IsInf(x) std::isinf(x) // GNU C/C++
+#elif defined(_MSC_VER) // Microsoft Visual C++
+ #define IsInf(x) (!_finite(x))
#else
- #define IsInf(x) std::isinf(x) // Unknown compiler, just hope for C99 conformance
+ #define IsInf(x) std::isinf(x) // GNU C/C++ (and/or other compilers), just hope for C99 conformance
#endif
#if defined(MPREAL_HAVE_INT64_SUPPORT)
-
- #define MPFR_USE_INTMAX_T // should be defined before mpfr.h
-
- #if defined(_MSC_VER) // <stdint.h> is available only in msvc2010!
- #if (_MSC_VER >= 1600)
- #include <stdint.h>
- #else // MPFR relies on intmax_t which is available only in msvc2010
- #undef MPREAL_HAVE_INT64_SUPPORT // Besides, MPFR - MPIR have to be compiled with msvc2010
- #undef MPFR_USE_INTMAX_T // Since we cannot detect this, disable x64 by default
- // Someone should change this manually if needed.
- #endif
- #endif
-
- #if defined (__MINGW32__) || defined(__MINGW64__)
- #include <stdint.h> // equivalent to msvc2010
- #elif defined (__GNUC__)
- #if defined(__amd64__) || defined(__amd64) || defined(__x86_64__) || defined(__x86_64)
- #undef MPREAL_HAVE_INT64_SUPPORT // remove all shaman dances for x64 builds since
- #undef MPFR_USE_INTMAX_T // GCC already support x64 as of "long int" is 64-bit integer, nothing left to do
- #else
- #include <stdint.h> // use int64_t, uint64_t otherwise.
- #endif
- #endif
+
+ #define MPFR_USE_INTMAX_T // Should be defined before mpfr.h
+
+ #if defined(_MSC_VER) // MSVC + Windows
+ #if (_MSC_VER >= 1600)
+ #include <stdint.h> // <stdint.h> is available only in msvc2010!
+
+ #else // MPFR relies on intmax_t which is available only in msvc2010
+ #undef MPREAL_HAVE_INT64_SUPPORT // Besides, MPFR & MPIR have to be compiled with msvc2010
+ #undef MPFR_USE_INTMAX_T // Since we cannot detect this, disable x64 by default
+ // Someone should change this manually if needed.
+ #endif
+
+ #elif defined (__GNUC__) && defined(__linux__)
+ #if defined(__amd64__) || defined(__amd64) || defined(__x86_64__) || defined(__x86_64) || defined(__ia64) || defined(__itanium__) || defined(_M_IA64)
+ #undef MPREAL_HAVE_INT64_SUPPORT // Remove all shaman dances for x64 builds since
+ #undef MPFR_USE_INTMAX_T // GCC already supports x64 as of "long int" is 64-bit integer, nothing left to do
+ #else
+ #include <stdint.h> // use int64_t, uint64_t otherwise
+ #endif
+
+ #else
+ #include <stdint.h> // rely on int64_t, uint64_t in all other cases, Mac OSX, etc.
+ #endif
#endif
#if defined(MPREAL_HAVE_MSVC_DEBUGVIEW) && defined(_MSC_VER) && defined(_DEBUG)
-#define MPREAL_MSVC_DEBUGVIEW_CODE DebugView = toString()
- #define MPREAL_MSVC_DEBUGVIEW_DATA std::string DebugView
+#define MPREAL_MSVC_DEBUGVIEW_CODE DebugView = toString();
+ #define MPREAL_MSVC_DEBUGVIEW_DATA std::string DebugView;
#else
- #define MPREAL_MSVC_DEBUGVIEW_CODE
- #define MPREAL_MSVC_DEBUGVIEW_DATA
+ #define MPREAL_MSVC_DEBUGVIEW_CODE
+ #define MPREAL_MSVC_DEBUGVIEW_DATA
#endif
#include <mpfr.h>
#if (MPFR_VERSION < MPFR_VERSION_NUM(3,0,0))
- #include <cstdlib> // needed for random()
+ #include <cstdlib> // Needed for random()
+#endif
+
+// Less important options
+#define MPREAL_DOUBLE_BITS_OVERFLOW -1 // Triggers overflow exception during conversion to double if mpreal
+ // cannot fit in MPREAL_DOUBLE_BITS_OVERFLOW bits
+ // = -1 disables overflow checks (default)
+#if defined(__GNUC__)
+ #define MPREAL_PERMISSIVE_EXPR __extension__
+#else
+ #define MPREAL_PERMISSIVE_EXPR
#endif
namespace mpfr {
class mpreal {
private:
- mpfr_t mp;
-
-public:
- static mp_rnd_t default_rnd;
- static mp_prec_t default_prec;
- static int default_base;
- static int double_bits;
+ mpfr_t mp;
public:
- // Constructors && type conversion
- mpreal();
- mpreal(const mpreal& u);
- mpreal(const mpfr_t u);
- mpreal(const mpf_t u);
- mpreal(const mpz_t u, mp_prec_t prec = default_prec, mp_rnd_t mode = default_rnd);
- mpreal(const mpq_t u, mp_prec_t prec = default_prec, mp_rnd_t mode = default_rnd);
- mpreal(const double u, mp_prec_t prec = default_prec, mp_rnd_t mode = default_rnd);
- mpreal(const long double u, mp_prec_t prec = default_prec, mp_rnd_t mode = default_rnd);
- mpreal(const unsigned long int u, mp_prec_t prec = default_prec, mp_rnd_t mode = default_rnd);
- mpreal(const unsigned int u, mp_prec_t prec = default_prec, mp_rnd_t mode = default_rnd);
- mpreal(const long int u, mp_prec_t prec = default_prec, mp_rnd_t mode = default_rnd);
- mpreal(const int u, mp_prec_t prec = default_prec, mp_rnd_t mode = default_rnd);
+
+ // Get default rounding mode & precision
+ inline static mp_rnd_t get_default_rnd() { return (mp_rnd_t)(mpfr_get_default_rounding_mode()); }
+ inline static mp_prec_t get_default_prec() { return mpfr_get_default_prec(); }
+
+ // Constructors && type conversions
+ mpreal();
+ mpreal(const mpreal& u);
+ mpreal(const mpfr_t u);
+ mpreal(const mpf_t u);
+ mpreal(const mpz_t u, mp_prec_t prec = mpreal::get_default_prec(), mp_rnd_t mode = mpreal::get_default_rnd());
+ mpreal(const mpq_t u, mp_prec_t prec = mpreal::get_default_prec(), mp_rnd_t mode = mpreal::get_default_rnd());
+ mpreal(const double u, mp_prec_t prec = mpreal::get_default_prec(), mp_rnd_t mode = mpreal::get_default_rnd());
+ mpreal(const long double u, mp_prec_t prec = mpreal::get_default_prec(), mp_rnd_t mode = mpreal::get_default_rnd());
+ mpreal(const unsigned long int u, mp_prec_t prec = mpreal::get_default_prec(), mp_rnd_t mode = mpreal::get_default_rnd());
+ mpreal(const unsigned int u, mp_prec_t prec = mpreal::get_default_prec(), mp_rnd_t mode = mpreal::get_default_rnd());
+ mpreal(const long int u, mp_prec_t prec = mpreal::get_default_prec(), mp_rnd_t mode = mpreal::get_default_rnd());
+ mpreal(const int u, mp_prec_t prec = mpreal::get_default_prec(), mp_rnd_t mode = mpreal::get_default_rnd());
#if defined (MPREAL_HAVE_INT64_SUPPORT)
- mpreal(const uint64_t u, mp_prec_t prec = default_prec, mp_rnd_t mode = default_rnd);
- mpreal(const int64_t u, mp_prec_t prec = default_prec, mp_rnd_t mode = default_rnd);
+ mpreal(const uint64_t u, mp_prec_t prec = mpreal::get_default_prec(), mp_rnd_t mode = mpreal::get_default_rnd());
+ mpreal(const int64_t u, mp_prec_t prec = mpreal::get_default_prec(), mp_rnd_t mode = mpreal::get_default_rnd());
#endif
- mpreal(const char* s, mp_prec_t prec = default_prec, int base = default_base, mp_rnd_t mode = default_rnd);
- mpreal(const std::string& s, mp_prec_t prec = default_prec, int base = default_base, mp_rnd_t mode = default_rnd);
-
- ~mpreal();
-
- // Operations
- // =
- // +, -, *, /, ++, --, <<, >>
- // *=, +=, -=, /=,
- // <, >, ==, <=, >=
-
- // =
- mpreal& operator=(const mpreal& v);
- mpreal& operator=(const mpf_t v);
- mpreal& operator=(const mpz_t v);
- mpreal& operator=(const mpq_t v);
- mpreal& operator=(const long double v);
- mpreal& operator=(const double v);
- mpreal& operator=(const unsigned long int v);
- mpreal& operator=(const unsigned int v);
- mpreal& operator=(const long int v);
- mpreal& operator=(const int v);
- mpreal& operator=(const char* s);
-
- // +
- mpreal& operator+=(const mpreal& v);
- mpreal& operator+=(const mpf_t v);
- mpreal& operator+=(const mpz_t v);
- mpreal& operator+=(const mpq_t v);
- mpreal& operator+=(const long double u);
- mpreal& operator+=(const double u);
- mpreal& operator+=(const unsigned long int u);
- mpreal& operator+=(const unsigned int u);
- mpreal& operator+=(const long int u);
- mpreal& operator+=(const int u);
+ mpreal(const char* s, mp_prec_t prec = mpreal::get_default_prec(), int base = 10, mp_rnd_t mode = mpreal::get_default_rnd());
+ mpreal(const std::string& s, mp_prec_t prec = mpreal::get_default_prec(), int base = 10, mp_rnd_t mode = mpreal::get_default_rnd());
+
+ ~mpreal();
+
+ // Operations
+ // =
+ // +, -, *, /, ++, --, <<, >>
+ // *=, +=, -=, /=,
+ // <, >, ==, <=, >=
+
+ // =
+ mpreal& operator=(const mpreal& v);
+ mpreal& operator=(const mpf_t v);
+ mpreal& operator=(const mpz_t v);
+ mpreal& operator=(const mpq_t v);
+ mpreal& operator=(const long double v);
+ mpreal& operator=(const double v);
+ mpreal& operator=(const unsigned long int v);
+ mpreal& operator=(const unsigned int v);
+ mpreal& operator=(const long int v);
+ mpreal& operator=(const int v);
+ mpreal& operator=(const char* s);
+ mpreal& operator=(const std::string& s);
+
+ // +
+ mpreal& operator+=(const mpreal& v);
+ mpreal& operator+=(const mpf_t v);
+ mpreal& operator+=(const mpz_t v);
+ mpreal& operator+=(const mpq_t v);
+ mpreal& operator+=(const long double u);
+ mpreal& operator+=(const double u);
+ mpreal& operator+=(const unsigned long int u);
+ mpreal& operator+=(const unsigned int u);
+ mpreal& operator+=(const long int u);
+ mpreal& operator+=(const int u);
#if defined (MPREAL_HAVE_INT64_SUPPORT)
- mpreal& operator+=(const int64_t u);
- mpreal& operator+=(const uint64_t u);
- mpreal& operator-=(const int64_t u);
- mpreal& operator-=(const uint64_t u);
- mpreal& operator*=(const int64_t u);
- mpreal& operator*=(const uint64_t u);
- mpreal& operator/=(const int64_t u);
- mpreal& operator/=(const uint64_t u);
+ mpreal& operator+=(const int64_t u);
+ mpreal& operator+=(const uint64_t u);
+ mpreal& operator-=(const int64_t u);
+ mpreal& operator-=(const uint64_t u);
+ mpreal& operator*=(const int64_t u);
+ mpreal& operator*=(const uint64_t u);
+ mpreal& operator/=(const int64_t u);
+ mpreal& operator/=(const uint64_t u);
#endif
- const mpreal operator+() const;
- mpreal& operator++ ();
- const mpreal operator++ (int);
-
- // -
- mpreal& operator-=(const mpreal& v);
- mpreal& operator-=(const mpz_t v);
- mpreal& operator-=(const mpq_t v);
- mpreal& operator-=(const long double u);
- mpreal& operator-=(const double u);
- mpreal& operator-=(const unsigned long int u);
- mpreal& operator-=(const unsigned int u);
- mpreal& operator-=(const long int u);
- mpreal& operator-=(const int u);
- const mpreal operator-() const;
- friend const mpreal operator-(const unsigned long int b, const mpreal& a);
- friend const mpreal operator-(const unsigned int b, const mpreal& a);
- friend const mpreal operator-(const long int b, const mpreal& a);
- friend const mpreal operator-(const int b, const mpreal& a);
- friend const mpreal operator-(const double b, const mpreal& a);
- mpreal& operator-- ();
- const mpreal operator-- (int);
-
- // *
- mpreal& operator*=(const mpreal& v);
- mpreal& operator*=(const mpz_t v);
- mpreal& operator*=(const mpq_t v);
- mpreal& operator*=(const long double v);
- mpreal& operator*=(const double v);
- mpreal& operator*=(const unsigned long int v);
- mpreal& operator*=(const unsigned int v);
- mpreal& operator*=(const long int v);
- mpreal& operator*=(const int v);
-
- // /
- mpreal& operator/=(const mpreal& v);
- mpreal& operator/=(const mpz_t v);
- mpreal& operator/=(const mpq_t v);
- mpreal& operator/=(const long double v);
- mpreal& operator/=(const double v);
- mpreal& operator/=(const unsigned long int v);
- mpreal& operator/=(const unsigned int v);
- mpreal& operator/=(const long int v);
- mpreal& operator/=(const int v);
- friend const mpreal operator/(const unsigned long int b, const mpreal& a);
- friend const mpreal operator/(const unsigned int b, const mpreal& a);
- friend const mpreal operator/(const long int b, const mpreal& a);
- friend const mpreal operator/(const int b, const mpreal& a);
- friend const mpreal operator/(const double b, const mpreal& a);
-
- //<<= Fast Multiplication by 2^u
- mpreal& operator<<=(const unsigned long int u);
- mpreal& operator<<=(const unsigned int u);
- mpreal& operator<<=(const long int u);
- mpreal& operator<<=(const int u);
-
- //>>= Fast Division by 2^u
- mpreal& operator>>=(const unsigned long int u);
- mpreal& operator>>=(const unsigned int u);
- mpreal& operator>>=(const long int u);
- mpreal& operator>>=(const int u);
-
- // Boolean Operators
- friend bool operator > (const mpreal& a, const mpreal& b);
- friend bool operator >= (const mpreal& a, const mpreal& b);
- friend bool operator < (const mpreal& a, const mpreal& b);
- friend bool operator <= (const mpreal& a, const mpreal& b);
- friend bool operator == (const mpreal& a, const mpreal& b);
- friend bool operator != (const mpreal& a, const mpreal& b);
-
- // Optimized specializations for boolean operators
- friend bool operator == (const mpreal& a, const unsigned long int b);
- friend bool operator == (const mpreal& a, const unsigned int b);
- friend bool operator == (const mpreal& a, const long int b);
- friend bool operator == (const mpreal& a, const int b);
- friend bool operator == (const mpreal& a, const long double b);
- friend bool operator == (const mpreal& a, const double b);
-
- // Type Conversion operators
- long toLong() const;
- unsigned long toULong() const;
- double toDouble() const;
- long double toLDouble() const;
+ const mpreal operator+() const;
+ mpreal& operator++ ();
+ const mpreal operator++ (int);
+
+ // -
+ mpreal& operator-=(const mpreal& v);
+ mpreal& operator-=(const mpz_t v);
+ mpreal& operator-=(const mpq_t v);
+ mpreal& operator-=(const long double u);
+ mpreal& operator-=(const double u);
+ mpreal& operator-=(const unsigned long int u);
+ mpreal& operator-=(const unsigned int u);
+ mpreal& operator-=(const long int u);
+ mpreal& operator-=(const int u);
+ const mpreal operator-() const;
+ friend const mpreal operator-(const unsigned long int b, const mpreal& a);
+ friend const mpreal operator-(const unsigned int b, const mpreal& a);
+ friend const mpreal operator-(const long int b, const mpreal& a);
+ friend const mpreal operator-(const int b, const mpreal& a);
+ friend const mpreal operator-(const double b, const mpreal& a);
+ mpreal& operator-- ();
+ const mpreal operator-- (int);
+
+ // *
+ mpreal& operator*=(const mpreal& v);
+ mpreal& operator*=(const mpz_t v);
+ mpreal& operator*=(const mpq_t v);
+ mpreal& operator*=(const long double v);
+ mpreal& operator*=(const double v);
+ mpreal& operator*=(const unsigned long int v);
+ mpreal& operator*=(const unsigned int v);
+ mpreal& operator*=(const long int v);
+ mpreal& operator*=(const int v);
+
+ // /
+ mpreal& operator/=(const mpreal& v);
+ mpreal& operator/=(const mpz_t v);
+ mpreal& operator/=(const mpq_t v);
+ mpreal& operator/=(const long double v);
+ mpreal& operator/=(const double v);
+ mpreal& operator/=(const unsigned long int v);
+ mpreal& operator/=(const unsigned int v);
+ mpreal& operator/=(const long int v);
+ mpreal& operator/=(const int v);
+ friend const mpreal operator/(const unsigned long int b, const mpreal& a);
+ friend const mpreal operator/(const unsigned int b, const mpreal& a);
+ friend const mpreal operator/(const long int b, const mpreal& a);
+ friend const mpreal operator/(const int b, const mpreal& a);
+ friend const mpreal operator/(const double b, const mpreal& a);
+
+ //<<= Fast Multiplication by 2^u
+ mpreal& operator<<=(const unsigned long int u);
+ mpreal& operator<<=(const unsigned int u);
+ mpreal& operator<<=(const long int u);
+ mpreal& operator<<=(const int u);
+
+ //>>= Fast Division by 2^u
+ mpreal& operator>>=(const unsigned long int u);
+ mpreal& operator>>=(const unsigned int u);
+ mpreal& operator>>=(const long int u);
+ mpreal& operator>>=(const int u);
+
+ // Boolean Operators
+ friend bool operator > (const mpreal& a, const mpreal& b);
+ friend bool operator >= (const mpreal& a, const mpreal& b);
+ friend bool operator < (const mpreal& a, const mpreal& b);
+ friend bool operator <= (const mpreal& a, const mpreal& b);
+ friend bool operator == (const mpreal& a, const mpreal& b);
+ friend bool operator != (const mpreal& a, const mpreal& b);
+
+ // Optimized specializations for boolean operators
+ friend bool operator == (const mpreal& a, const unsigned long int b);
+ friend bool operator == (const mpreal& a, const unsigned int b);
+ friend bool operator == (const mpreal& a, const long int b);
+ friend bool operator == (const mpreal& a, const int b);
+ friend bool operator == (const mpreal& a, const long double b);
+ friend bool operator == (const mpreal& a, const double b);
+
+ // Type Conversion operators
+ long toLong (mp_rnd_t mode = GMP_RNDZ) const;
+ unsigned long toULong (mp_rnd_t mode = GMP_RNDZ) const;
+ double toDouble (mp_rnd_t mode = GMP_RNDN) const;
+ long double toLDouble (mp_rnd_t mode = GMP_RNDN) const;
#if defined (MPREAL_HAVE_INT64_SUPPORT)
- int64_t toInt64() const;
- uint64_t toUInt64() const;
+ int64_t toInt64 (mp_rnd_t mode = GMP_RNDZ) const;
+ uint64_t toUInt64 (mp_rnd_t mode = GMP_RNDZ) const;
#endif
- // Get raw pointers
- ::mpfr_ptr mpfr_ptr();
- ::mpfr_srcptr mpfr_srcptr() const;
+ // Get raw pointers so that mpreal can be directly used in raw mpfr_* functions
+ ::mpfr_ptr mpfr_ptr();
+ ::mpfr_srcptr mpfr_ptr() const;
+ ::mpfr_srcptr mpfr_srcptr() const;
- // Convert mpreal to string with n significant digits in base b
- // n = 0 -> convert with the maximum available digits
- std::string toString(int n = 0, int b = default_base, mp_rnd_t mode = default_rnd) const;
+ // Convert mpreal to string with n significant digits in base b
+ // n = 0 -> convert with the maximum available digits
+ std::string toString(int n = 0, int b = 10, mp_rnd_t mode = mpreal::get_default_rnd()) const;
#if (MPFR_VERSION >= MPFR_VERSION_NUM(2,4,0))
- std::string toString(const std::string& format) const;
+ std::string toString(const std::string& format) const;
#endif
- // Math Functions
- friend const mpreal sqr (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal sqrt(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal sqrt(const unsigned long int v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal cbrt(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal root(const mpreal& v, unsigned long int k, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal pow (const mpreal& a, const mpreal& b, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal pow (const mpreal& a, const mpz_t b, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal pow (const mpreal& a, const unsigned long int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal pow (const mpreal& a, const long int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal pow (const unsigned long int a, const mpreal& b, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal pow (const unsigned long int a, const unsigned long int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal fabs(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
-
- friend const mpreal abs(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal dim(const mpreal& a, const mpreal& b, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend inline const mpreal mul_2ui(const mpreal& v, unsigned long int k, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend inline const mpreal mul_2si(const mpreal& v, long int k, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend inline const mpreal div_2ui(const mpreal& v, unsigned long int k, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend inline const mpreal div_2si(const mpreal& v, long int k, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend int cmpabs(const mpreal& a,const mpreal& b);
-
- friend const mpreal log (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal log2 (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal log10(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal exp (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal exp2 (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal exp10(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal log1p(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal expm1(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
-
- friend const mpreal cos(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal sin(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal tan(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal sec(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal csc(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal cot(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend int sin_cos(mpreal& s, mpreal& c, const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
-
- friend const mpreal acos (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal asin (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal atan (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal atan2 (const mpreal& y, const mpreal& x, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal acot (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal asec (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal acsc (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
-
- friend const mpreal cosh (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal sinh (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal tanh (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal sech (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal csch (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal coth (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal acosh (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal asinh (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal atanh (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal acoth (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal asech (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal acsch (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
-
- friend const mpreal hypot (const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode = mpreal::default_rnd);
-
- friend const mpreal fac_ui (unsigned long int v, mp_prec_t prec = mpreal::default_prec, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal eint (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
-
- friend const mpreal gamma (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal lngamma (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal lgamma (const mpreal& v, int *signp = 0, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal zeta (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal erf (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal erfc (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal besselj0 (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal besselj1 (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal besseljn (long n, const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal bessely0 (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal bessely1 (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal besselyn (long n, const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal fma (const mpreal& v1, const mpreal& v2, const mpreal& v3, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal fms (const mpreal& v1, const mpreal& v2, const mpreal& v3, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal agm (const mpreal& v1, const mpreal& v2, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal sum (const mpreal tab[], unsigned long int n, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend int sgn(const mpreal& v); // -1 or +1
+ // Math Functions
+ friend const mpreal sqr (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal sqrt(const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal sqrt(const unsigned long int v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal cbrt(const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal root(const mpreal& v, unsigned long int k, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal pow (const mpreal& a, const mpreal& b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal pow (const mpreal& a, const mpz_t b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal pow (const mpreal& a, const unsigned long int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal pow (const mpreal& a, const long int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal pow (const unsigned long int a, const mpreal& b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal pow (const unsigned long int a, const unsigned long int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal fabs(const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+
+ friend const mpreal abs(const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal dim(const mpreal& a, const mpreal& b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend inline const mpreal mul_2ui(const mpreal& v, unsigned long int k, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend inline const mpreal mul_2si(const mpreal& v, long int k, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend inline const mpreal div_2ui(const mpreal& v, unsigned long int k, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend inline const mpreal div_2si(const mpreal& v, long int k, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend int cmpabs(const mpreal& a,const mpreal& b);
+
+ friend const mpreal log (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal log2 (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal log10(const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal exp (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal exp2 (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal exp10(const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal log1p(const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal expm1(const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+
+ friend const mpreal cos(const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal sin(const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal tan(const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal sec(const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal csc(const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal cot(const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend int sin_cos(mpreal& s, mpreal& c, const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+
+ friend const mpreal acos (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal asin (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal atan (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal atan2 (const mpreal& y, const mpreal& x, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal acot (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal asec (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal acsc (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+
+ friend const mpreal cosh (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal sinh (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal tanh (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal sech (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal csch (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal coth (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal acosh (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal asinh (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal atanh (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal acoth (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal asech (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal acsch (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+
+ friend const mpreal hypot (const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+
+ friend const mpreal fac_ui (unsigned long int v, mp_prec_t prec = mpreal::get_default_prec(), mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal eint (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+
+ friend const mpreal gamma (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal lngamma (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal lgamma (const mpreal& v, int *signp = 0, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal zeta (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal erf (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal erfc (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal besselj0 (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal besselj1 (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal besseljn (long n, const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal bessely0 (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal bessely1 (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal besselyn (long n, const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal fma (const mpreal& v1, const mpreal& v2, const mpreal& v3, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal fms (const mpreal& v1, const mpreal& v2, const mpreal& v3, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal agm (const mpreal& v1, const mpreal& v2, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal sum (const mpreal tab[], unsigned long int n, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend int sgn(const mpreal& v); // returns -1 or +1
// MPFR 2.4.0 Specifics
#if (MPFR_VERSION >= MPFR_VERSION_NUM(2,4,0))
- friend int sinh_cosh(mpreal& s, mpreal& c, const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal li2(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal fmod (const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal rec_sqrt(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
+ friend int sinh_cosh (mpreal& s, mpreal& c, const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal li2 (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal fmod (const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal rec_sqrt (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+
+ // MATLAB's semantic equivalents
+ friend const mpreal rem (const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode = mpreal::get_default_rnd()); // Remainder after division
+ friend const mpreal mod (const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode = mpreal::get_default_rnd()); // Modulus after division
#endif
// MPFR 3.0.0 Specifics
#if (MPFR_VERSION >= MPFR_VERSION_NUM(3,0,0))
- friend const mpreal digamma(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal ai(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal urandom (gmp_randstate_t& state,mp_rnd_t rnd_mode = mpreal::default_rnd); // use gmp_randinit_default() to init state, gmp_randclear() to clear
- friend bool isregular(const mpreal& v);
+ friend const mpreal digamma (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal ai (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal urandom (gmp_randstate_t& state, mp_rnd_t rnd_mode = mpreal::get_default_rnd()); // use gmp_randinit_default() to init state, gmp_randclear() to clear
#endif
-
- // Uniformly distributed random number generation in [0,1] using
- // Mersenne-Twister algorithm by default.
- // Use parameter to setup seed, e.g.: random((unsigned)time(NULL))
- // Check urandom() for more precise control.
- friend const mpreal random(unsigned int seed = 0);
-
- // Exponent and mantissa manipulation
- friend const mpreal frexp(const mpreal& v, mp_exp_t* exp);
- friend const mpreal ldexp(const mpreal& v, mp_exp_t exp);
-
- // Splits mpreal value into fractional and integer parts.
- // Returns fractional part and stores integer part in n.
- friend const mpreal modf(const mpreal& v, mpreal& n);
-
- // Constants
- // don't forget to call mpfr_free_cache() for every thread where you are using const-functions
- friend const mpreal const_log2 (mp_prec_t prec = mpreal::default_prec, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal const_pi (mp_prec_t prec = mpreal::default_prec, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal const_euler (mp_prec_t prec = mpreal::default_prec, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal const_catalan (mp_prec_t prec = mpreal::default_prec, mp_rnd_t rnd_mode = mpreal::default_rnd);
- // returns +inf iff sign>=0 otherwise -inf
- friend const mpreal const_infinity(int sign = 1, mp_prec_t prec = mpreal::default_prec, mp_rnd_t rnd_mode = mpreal::default_rnd);
-
- // Output/ Input
- friend std::ostream& operator<<(std::ostream& os, const mpreal& v);
+
+ // Uniformly distributed random number generation in [0,1] using
+ // Mersenne-Twister algorithm by default.
+ // Use parameter to setup seed, e.g.: random((unsigned)time(NULL))
+ // Check urandom() for more precise control.
+ friend const mpreal random(unsigned int seed = 0);
+
+ // Exponent and mantissa manipulation
+ friend const mpreal frexp(const mpreal& v, mp_exp_t* exp);
+ friend const mpreal ldexp(const mpreal& v, mp_exp_t exp);
+
+ // Splits mpreal value into fractional and integer parts.
+ // Returns fractional part and stores integer part in n.
+ friend const mpreal modf(const mpreal& v, mpreal& n);
+
+ // Constants
+ // don't forget to call mpfr_free_cache() for every thread where you are using const-functions
+ friend const mpreal const_log2 (mp_prec_t prec = mpreal::get_default_prec(), mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal const_pi (mp_prec_t prec = mpreal::get_default_prec(), mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal const_euler (mp_prec_t prec = mpreal::get_default_prec(), mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal const_catalan (mp_prec_t prec = mpreal::get_default_prec(), mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+
+ // returns +inf iff sign>=0 otherwise -inf
+ friend const mpreal const_infinity(int sign = 1, mp_prec_t prec = mpreal::get_default_prec(), mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+
+ // Output/ Input
+ friend std::ostream& operator<<(std::ostream& os, const mpreal& v);
friend std::istream& operator>>(std::istream& is, mpreal& v);
- // Integer Related Functions
- friend const mpreal rint (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal ceil (const mpreal& v);
- friend const mpreal floor(const mpreal& v);
- friend const mpreal round(const mpreal& v);
- friend const mpreal trunc(const mpreal& v);
- friend const mpreal rint_ceil (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal rint_floor(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal rint_round(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal rint_trunc(const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal frac (const mpreal& v, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal remainder (const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode = mpreal::default_rnd);
- friend const mpreal remquo (long* q, const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode = mpreal::default_rnd);
-
- // Miscellaneous Functions
- friend const mpreal nexttoward (const mpreal& x, const mpreal& y);
- friend const mpreal nextabove (const mpreal& x);
- friend const mpreal nextbelow (const mpreal& x);
-
- // use gmp_randinit_default() to init state, gmp_randclear() to clear
- friend const mpreal urandomb (gmp_randstate_t& state);
+ // Integer Related Functions
+ friend const mpreal rint (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal ceil (const mpreal& v);
+ friend const mpreal floor(const mpreal& v);
+ friend const mpreal round(const mpreal& v);
+ friend const mpreal trunc(const mpreal& v);
+ friend const mpreal rint_ceil (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal rint_floor (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal rint_round (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal rint_trunc (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal frac (const mpreal& v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal remainder ( const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal remquo (long* q, const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+
+ // Miscellaneous Functions
+ friend const mpreal nexttoward (const mpreal& x, const mpreal& y);
+ friend const mpreal nextabove (const mpreal& x);
+ friend const mpreal nextbelow (const mpreal& x);
+
+ // use gmp_randinit_default() to init state, gmp_randclear() to clear
+ friend const mpreal urandomb (gmp_randstate_t& state);
// MPFR < 2.4.2 Specifics
#if (MPFR_VERSION <= MPFR_VERSION_NUM(2,4,2))
- friend const mpreal random2 (mp_size_t size, mp_exp_t exp);
+ friend const mpreal random2 (mp_size_t size, mp_exp_t exp);
#endif
- // Instance Checkers
- friend bool isnan (const mpreal& v);
- friend bool isinf (const mpreal& v);
- friend bool isfinite(const mpreal& v);
-
- friend bool isnum(const mpreal& v);
- friend bool iszero(const mpreal& v);
- friend bool isint(const mpreal& v);
-
- // Set/Get instance properties
- inline mp_prec_t get_prec() const;
- inline void set_prec(mp_prec_t prec, mp_rnd_t rnd_mode = default_rnd); // Change precision with rounding mode
-
- // Aliases for get_prec(), set_prec() - needed for compatibility with std::complex<mpreal> interface
- inline mpreal& setPrecision(int Precision, mp_rnd_t RoundingMode = (mpfr_get_default_rounding_mode)());
- inline int getPrecision() const;
-
- // Set mpreal to +/- inf, NaN, +/-0
- mpreal& setInf (int Sign = +1);
- mpreal& setNan ();
- mpreal& setZero (int Sign = +1);
- mpreal& setSign (int Sign, mp_rnd_t RoundingMode = (mpfr_get_default_rounding_mode)());
-
- //Exponent
- mp_exp_t get_exp();
- int set_exp(mp_exp_t e);
- int check_range (int t, mp_rnd_t rnd_mode = default_rnd);
- int subnormalize (int t,mp_rnd_t rnd_mode = default_rnd);
-
- // Inexact conversion from float
- inline bool fits_in_bits(double x, int n);
-
- // Set/Get global properties
- static void set_default_prec(mp_prec_t prec);
- static mp_prec_t get_default_prec();
- static void set_default_base(int base);
- static int get_default_base();
- static void set_double_bits(int dbits);
- static int get_double_bits();
- static void set_default_rnd(mp_rnd_t rnd_mode);
- static mp_rnd_t get_default_rnd();
- static mp_exp_t get_emin (void);
- static mp_exp_t get_emax (void);
- static mp_exp_t get_emin_min (void);
- static mp_exp_t get_emin_max (void);
- static mp_exp_t get_emax_min (void);
- static mp_exp_t get_emax_max (void);
- static int set_emin (mp_exp_t exp);
- static int set_emax (mp_exp_t exp);
-
- // Efficient swapping of two mpreal values
- friend void swap(mpreal& x, mpreal& y);
-
- //Min Max - macros is evil. Needed for systems which defines max and min globally as macros (e.g. Windows)
- //Hope that globally defined macros use > < operations only
- friend const mpreal fmax(const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode = default_rnd);
- friend const mpreal fmin(const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode = default_rnd);
-
-#if defined (MPREAL_HAVE_CUSTOM_MPFR_MALLOC)
+ // Instance Checkers
+ friend bool isnan (const mpreal& v);
+ friend bool isinf (const mpreal& v);
+ friend bool isfinite (const mpreal& v);
-private:
- // Optimized dynamic memory allocation/(re-)deallocation.
- static bool is_custom_malloc;
- static void *mpreal_allocate (size_t alloc_size);
- static void *mpreal_reallocate (void *ptr, size_t old_size, size_t new_size);
- static void mpreal_free (void *ptr, size_t size);
- inline static void set_custom_malloc (void);
+ friend bool isnum (const mpreal& v);
+ friend bool iszero (const mpreal& v);
+ friend bool isint (const mpreal& v);
+#if (MPFR_VERSION >= MPFR_VERSION_NUM(3,0,0))
+ friend bool isregular(const mpreal& v);
#endif
+ // Set/Get instance properties
+ inline mp_prec_t get_prec() const;
+ inline void set_prec(mp_prec_t prec, mp_rnd_t rnd_mode = get_default_rnd()); // Change precision with rounding mode
+
+ // Aliases for get_prec(), set_prec() - needed for compatibility with std::complex<mpreal> interface
+ inline mpreal& setPrecision(int Precision, mp_rnd_t RoundingMode = get_default_rnd());
+ inline int getPrecision() const;
+
+ // Set mpreal to +/- inf, NaN, +/-0
+ mpreal& setInf (int Sign = +1);
+ mpreal& setNan ();
+ mpreal& setZero (int Sign = +1);
+ mpreal& setSign (int Sign, mp_rnd_t RoundingMode = get_default_rnd());
+
+ //Exponent
+ mp_exp_t get_exp();
+ int set_exp(mp_exp_t e);
+ int check_range (int t, mp_rnd_t rnd_mode = get_default_rnd());
+ int subnormalize (int t,mp_rnd_t rnd_mode = get_default_rnd());
+
+ // Inexact conversion from float
+ inline bool fits_in_bits(double x, int n);
+
+ // Set/Get global properties
+ static void set_default_prec(mp_prec_t prec);
+ static void set_default_rnd(mp_rnd_t rnd_mode);
+
+ static mp_exp_t get_emin (void);
+ static mp_exp_t get_emax (void);
+ static mp_exp_t get_emin_min (void);
+ static mp_exp_t get_emin_max (void);
+ static mp_exp_t get_emax_min (void);
+ static mp_exp_t get_emax_max (void);
+ static int set_emin (mp_exp_t exp);
+ static int set_emax (mp_exp_t exp);
+
+ // Efficient swapping of two mpreal values - needed for std algorithms
+ friend void swap(mpreal& x, mpreal& y);
+
+ friend const mpreal fmax(const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+ friend const mpreal fmin(const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
private:
- // Human friendly Debug Preview in Visual Studio.
- // Put one of these lines:
- //
- // mpfr::mpreal=<DebugView> ; Show value only
- // mpfr::mpreal=<DebugView>, <mp[0]._mpfr_prec,u>bits ; Show value & precision
- //
- // at the beginning of
- // [Visual Studio Installation Folder]\Common7\Packages\Debugger\autoexp.dat
- MPREAL_MSVC_DEBUGVIEW_DATA
+ // Human friendly Debug Preview in Visual Studio.
+ // Put one of these lines:
+ //
+ // mpfr::mpreal=<DebugView> ; Show value only
+ // mpfr::mpreal=<DebugView>, <mp[0]._mpfr_prec,u>bits ; Show value & precision
+ //
+ // at the beginning of
+ // [Visual Studio Installation Folder]\Common7\Packages\Debugger\autoexp.dat
+ MPREAL_MSVC_DEBUGVIEW_DATA
};
//////////////////////////////////////////////////////////////////////////
// Exceptions
class conversion_overflow : public std::exception {
public:
- std::string why() { return "inexact conversion from floating point"; }
+ std::string why() { return "inexact conversion from floating point"; }
};
+//////////////////////////////////////////////////////////////////////////
+// Constructors & converters
+// Default constructor: creates mp number and initializes it to 0.
+inline mpreal::mpreal()
+{
+ mpfr_init2(mp,mpreal::get_default_prec());
+ mpfr_set_ui(mp,0,mpreal::get_default_rnd());
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+}
+
+inline mpreal::mpreal(const mpreal& u)
+{
+ mpfr_init2(mp,mpfr_get_prec(u.mp));
+ mpfr_set(mp,u.mp,mpreal::get_default_rnd());
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+}
+
+inline mpreal::mpreal(const mpfr_t u)
+{
+ mpfr_init2(mp,mpfr_get_prec(u));
+ mpfr_set(mp,u,mpreal::get_default_rnd());
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+}
+
+inline mpreal::mpreal(const mpf_t u)
+{
+ mpfr_init2(mp,(mp_prec_t) mpf_get_prec(u)); // (gmp: mp_bitcnt_t) unsigned long -> long (mpfr: mp_prec_t)
+ mpfr_set_f(mp,u,mpreal::get_default_rnd());
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+}
+
+inline mpreal::mpreal(const mpz_t u, mp_prec_t prec, mp_rnd_t mode)
+{
+ mpfr_init2(mp,prec);
+ mpfr_set_z(mp,u,mode);
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+}
+
+inline mpreal::mpreal(const mpq_t u, mp_prec_t prec, mp_rnd_t mode)
+{
+ mpfr_init2(mp,prec);
+ mpfr_set_q(mp,u,mode);
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+}
+
+inline mpreal::mpreal(const double u, mp_prec_t prec, mp_rnd_t mode)
+{
+ mpfr_init2(mp, prec);
+
+#if (MPREAL_DOUBLE_BITS_OVERFLOW > -1)
+ if(fits_in_bits(u, MPREAL_DOUBLE_BITS_OVERFLOW))
+ {
+ mpfr_set_d(mp, u, mode);
+ }else
+ throw conversion_overflow();
+#else
+ mpfr_set_d(mp, u, mode);
+#endif
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+}
+
+inline mpreal::mpreal(const long double u, mp_prec_t prec, mp_rnd_t mode)
+{
+ mpfr_init2(mp,prec);
+ mpfr_set_ld(mp,u,mode);
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+}
+
+inline mpreal::mpreal(const unsigned long int u, mp_prec_t prec, mp_rnd_t mode)
+{
+ mpfr_init2(mp,prec);
+ mpfr_set_ui(mp,u,mode);
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+}
+
+inline mpreal::mpreal(const unsigned int u, mp_prec_t prec, mp_rnd_t mode)
+{
+ mpfr_init2(mp,prec);
+ mpfr_set_ui(mp,u,mode);
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+}
+
+inline mpreal::mpreal(const long int u, mp_prec_t prec, mp_rnd_t mode)
+{
+ mpfr_init2(mp,prec);
+ mpfr_set_si(mp,u,mode);
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+}
+
+inline mpreal::mpreal(const int u, mp_prec_t prec, mp_rnd_t mode)
+{
+ mpfr_init2(mp,prec);
+ mpfr_set_si(mp,u,mode);
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+}
+
+#if defined (MPREAL_HAVE_INT64_SUPPORT)
+inline mpreal::mpreal(const uint64_t u, mp_prec_t prec, mp_rnd_t mode)
+{
+ mpfr_init2(mp,prec);
+ mpfr_set_uj(mp, u, mode);
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+}
+
+inline mpreal::mpreal(const int64_t u, mp_prec_t prec, mp_rnd_t mode)
+{
+ mpfr_init2(mp,prec);
+ mpfr_set_sj(mp, u, mode);
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+}
+#endif
+
+inline mpreal::mpreal(const char* s, mp_prec_t prec, int base, mp_rnd_t mode)
+{
+ mpfr_init2(mp, prec);
+ mpfr_set_str(mp, s, base, mode);
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+}
+
+inline mpreal::mpreal(const std::string& s, mp_prec_t prec, int base, mp_rnd_t mode)
+{
+ mpfr_init2(mp, prec);
+ mpfr_set_str(mp, s.c_str(), base, mode);
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+}
+
+inline mpreal::~mpreal()
+{
+ mpfr_clear(mp);
+}
+
+// internal namespace needed for template magic
namespace internal{
- // Use SFINAE to restrict arithmetic operations instantiation only for numeric types
- // This is needed for smooth integration with libraries based on expression templates
- template <typename ArgumentType> struct result_type {};
-
- template <> struct result_type<mpreal> {typedef mpreal type;};
- template <> struct result_type<mpz_t> {typedef mpreal type;};
- template <> struct result_type<mpq_t> {typedef mpreal type;};
- template <> struct result_type<long double> {typedef mpreal type;};
- template <> struct result_type<double> {typedef mpreal type;};
- template <> struct result_type<unsigned long int> {typedef mpreal type;};
- template <> struct result_type<unsigned int> {typedef mpreal type;};
- template <> struct result_type<long int> {typedef mpreal type;};
- template <> struct result_type<int> {typedef mpreal type;};
+ // Use SFINAE to restrict arithmetic operations instantiation only for numeric types
+ // This is needed for smooth integration with libraries based on expression templates, like Eigen.
+ // TODO: Do the same for boolean operators.
+ template <typename ArgumentType> struct result_type {};
+
+ template <> struct result_type<mpreal> {typedef mpreal type;};
+ template <> struct result_type<mpz_t> {typedef mpreal type;};
+ template <> struct result_type<mpq_t> {typedef mpreal type;};
+ template <> struct result_type<long double> {typedef mpreal type;};
+ template <> struct result_type<double> {typedef mpreal type;};
+ template <> struct result_type<unsigned long int> {typedef mpreal type;};
+ template <> struct result_type<unsigned int> {typedef mpreal type;};
+ template <> struct result_type<long int> {typedef mpreal type;};
+ template <> struct result_type<int> {typedef mpreal type;};
#if defined (MPREAL_HAVE_INT64_SUPPORT)
- template <> struct result_type<int64_t > {typedef mpreal type;};
- template <> struct result_type<uint64_t > {typedef mpreal type;};
+ template <> struct result_type<int64_t > {typedef mpreal type;};
+ template <> struct result_type<uint64_t > {typedef mpreal type;};
#endif
}
// + Addition
template <typename Rhs>
inline const typename internal::result_type<Rhs>::type
- operator+(const mpreal& lhs, const Rhs& rhs){ return mpreal(lhs) += rhs; }
+ operator+(const mpreal& lhs, const Rhs& rhs){ return mpreal(lhs) += rhs; }
template <typename Lhs>
inline const typename internal::result_type<Lhs>::type
- operator+(const Lhs& lhs, const mpreal& rhs){ return mpreal(rhs) += lhs; }
+ operator+(const Lhs& lhs, const mpreal& rhs){ return mpreal(rhs) += lhs; }
// - Subtraction
template <typename Rhs>
inline const typename internal::result_type<Rhs>::type
- operator-(const mpreal& lhs, const Rhs& rhs){ return mpreal(lhs) -= rhs; }
+ operator-(const mpreal& lhs, const Rhs& rhs){ return mpreal(lhs) -= rhs; }
template <typename Lhs>
inline const typename internal::result_type<Lhs>::type
- operator-(const Lhs& lhs, const mpreal& rhs){ return mpreal(lhs) -= rhs; }
+ operator-(const Lhs& lhs, const mpreal& rhs){ return mpreal(lhs) -= rhs; }
// * Multiplication
template <typename Rhs>
inline const typename internal::result_type<Rhs>::type
- operator*(const mpreal& lhs, const Rhs& rhs){ return mpreal(lhs) *= rhs; }
+ operator*(const mpreal& lhs, const Rhs& rhs){ return mpreal(lhs) *= rhs; }
template <typename Lhs>
inline const typename internal::result_type<Lhs>::type
- operator*(const Lhs& lhs, const mpreal& rhs){ return mpreal(rhs) *= lhs; }
+ operator*(const Lhs& lhs, const mpreal& rhs){ return mpreal(rhs) *= lhs; }
// / Division
template <typename Rhs>
inline const typename internal::result_type<Rhs>::type
- operator/(const mpreal& lhs, const Rhs& rhs){ return mpreal(lhs) /= rhs; }
+ operator/(const mpreal& lhs, const Rhs& rhs){ return mpreal(lhs) /= rhs; }
template <typename Lhs>
inline const typename internal::result_type<Lhs>::type
- operator/(const Lhs& lhs, const mpreal& rhs){ return mpreal(lhs) /= rhs; }
+ operator/(const Lhs& lhs, const mpreal& rhs){ return mpreal(lhs) /= rhs; }
//////////////////////////////////////////////////////////////////////////
// sqrt
-const mpreal sqrt(const unsigned int v, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal sqrt(const long int v, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal sqrt(const int v, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal sqrt(const long double v, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal sqrt(const double v, mp_rnd_t rnd_mode = mpreal::default_rnd);
+const mpreal sqrt(const unsigned int v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal sqrt(const long int v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal sqrt(const int v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal sqrt(const long double v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal sqrt(const double v, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
//////////////////////////////////////////////////////////////////////////
// pow
-const mpreal pow(const mpreal& a, const unsigned int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const mpreal& a, const int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const mpreal& a, const long double b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const mpreal& a, const double b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-
-const mpreal pow(const unsigned int a, const mpreal& b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const long int a, const mpreal& b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const int a, const mpreal& b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const long double a, const mpreal& b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const double a, const mpreal& b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-
-const mpreal pow(const unsigned long int a, const unsigned int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const unsigned long int a, const long int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const unsigned long int a, const int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const unsigned long int a, const long double b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const unsigned long int a, const double b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-
-const mpreal pow(const unsigned int a, const unsigned long int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const unsigned int a, const unsigned int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const unsigned int a, const long int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const unsigned int a, const int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const unsigned int a, const long double b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const unsigned int a, const double b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-
-const mpreal pow(const long int a, const unsigned long int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const long int a, const unsigned int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const long int a, const long int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const long int a, const int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const long int a, const long double b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const long int a, const double b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-
-const mpreal pow(const int a, const unsigned long int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const int a, const unsigned int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const int a, const long int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const int a, const int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const int a, const long double b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const int a, const double b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-
-const mpreal pow(const long double a, const long double b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const long double a, const unsigned long int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const long double a, const unsigned int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const long double a, const long int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const long double a, const int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-
-const mpreal pow(const double a, const double b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const double a, const unsigned long int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const double a, const unsigned int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const double a, const long int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
-const mpreal pow(const double a, const int b, mp_rnd_t rnd_mode = mpreal::default_rnd);
+const mpreal pow(const mpreal& a, const unsigned int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const mpreal& a, const int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const mpreal& a, const long double b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const mpreal& a, const double b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+
+const mpreal pow(const unsigned int a, const mpreal& b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const long int a, const mpreal& b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const int a, const mpreal& b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const long double a, const mpreal& b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const double a, const mpreal& b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+
+const mpreal pow(const unsigned long int a, const unsigned int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const unsigned long int a, const long int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const unsigned long int a, const int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const unsigned long int a, const long double b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const unsigned long int a, const double b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+
+const mpreal pow(const unsigned int a, const unsigned long int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const unsigned int a, const unsigned int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const unsigned int a, const long int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const unsigned int a, const int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const unsigned int a, const long double b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const unsigned int a, const double b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+
+const mpreal pow(const long int a, const unsigned long int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const long int a, const unsigned int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const long int a, const long int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const long int a, const int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const long int a, const long double b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const long int a, const double b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+
+const mpreal pow(const int a, const unsigned long int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const int a, const unsigned int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const int a, const long int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const int a, const int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const int a, const long double b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const int a, const double b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+
+const mpreal pow(const long double a, const long double b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const long double a, const unsigned long int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const long double a, const unsigned int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const long double a, const long int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const long double a, const int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+
+const mpreal pow(const double a, const double b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const double a, const unsigned long int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const double a, const unsigned int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const double a, const long int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
+const mpreal pow(const double a, const int b, mp_rnd_t rnd_mode = mpreal::get_default_rnd());
//////////////////////////////////////////////////////////////////////////
// Estimate machine epsilon for the given precision
// Returns smallest eps such that 1.0 + eps != 1.0
-inline const mpreal machine_epsilon(mp_prec_t prec = mpreal::get_default_prec());
+inline mpreal machine_epsilon(mp_prec_t prec = mpreal::get_default_prec());
-// Returns the positive distance from abs(x) to the next larger in magnitude floating point number of the same precision as x
-inline const mpreal machine_epsilon(const mpreal& x);
+// Returns smallest eps such that x + eps != x (relative machine epsilon)
+inline mpreal machine_epsilon(const mpreal& x);
-inline const mpreal mpreal_min(mp_prec_t prec = mpreal::get_default_prec());
-inline const mpreal mpreal_max(mp_prec_t prec = mpreal::get_default_prec());
+// Gives max & min values for the required precision,
+// minval is 'safe' meaning 1 / minval does not overflow
+// maxval is 'safe' meaning 1 / maxval does not underflow
+inline mpreal minval(mp_prec_t prec = mpreal::get_default_prec());
+inline mpreal maxval(mp_prec_t prec = mpreal::get_default_prec());
+
+// 'Dirty' equality check 1: |a-b| < min{|a|,|b|} * eps
inline bool isEqualFuzzy(const mpreal& a, const mpreal& b, const mpreal& eps);
+
+// 'Dirty' equality check 2: |a-b| < min{|a|,|b|} * eps( min{|a|,|b|} )
+inline bool isEqualFuzzy(const mpreal& a, const mpreal& b);
+
+// 'Bitwise' equality check
+// maxUlps - a and b can be apart by maxUlps binary numbers.
inline bool isEqualUlps(const mpreal& a, const mpreal& b, int maxUlps);
//////////////////////////////////////////////////////////////////////////
-// Bits - decimal digits relation
-// bits = ceil(digits*log[2](10))
-// digits = floor(bits*log[10](2))
+// Convert precision in 'bits' to decimal digits and vice versa.
+// bits = ceil(digits*log[2](10))
+// digits = floor(bits*log[10](2))
inline mp_prec_t digits2bits(int d);
-inline int bits2digits(mp_prec_t b);
+inline int bits2digits(mp_prec_t b);
//////////////////////////////////////////////////////////////////////////
// min, max
@@ -700,541 +858,586 @@ const mpreal (min)(const mpreal& x, const mpreal& y); // Operators - Assignment
inline mpreal& mpreal::operator=(const mpreal& v)
{
- if (this != &v)
- {
- mpfr_clear(mp);
- mpfr_init2(mp,mpfr_get_prec(v.mp));
- mpfr_set(mp,v.mp,default_rnd);
+ if (this != &v)
+ {
+ mp_prec_t tp = mpfr_get_prec(mp);
+ mp_prec_t vp = mpfr_get_prec(v.mp);
+
+ if(tp != vp){
+ mpfr_clear(mp);
+ mpfr_init2(mp, vp);
+ }
+
+ mpfr_set(mp, v.mp, mpreal::get_default_rnd());
- MPREAL_MSVC_DEBUGVIEW_CODE;
- }
- return *this;
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ }
+ return *this;
}
inline mpreal& mpreal::operator=(const mpf_t v)
{
- mpfr_set_f(mp,v,default_rnd);
-
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_set_f(mp, v, mpreal::get_default_rnd());
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator=(const mpz_t v)
{
- mpfr_set_z(mp,v,default_rnd);
-
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_set_z(mp, v, mpreal::get_default_rnd());
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator=(const mpq_t v)
{
- mpfr_set_q(mp,v,default_rnd);
+ mpfr_set_q(mp, v, mpreal::get_default_rnd());
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
-inline mpreal& mpreal::operator=(const long double v)
-{
- mpfr_set_ld(mp,v,default_rnd);
+inline mpreal& mpreal::operator=(const long double v)
+{
+ mpfr_set_ld(mp, v, mpreal::get_default_rnd());
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
+}
+
+inline mpreal& mpreal::operator=(const double v)
+{
+#if (MPREAL_DOUBLE_BITS_OVERFLOW > -1)
+ if(fits_in_bits(v, MPREAL_DOUBLE_BITS_OVERFLOW))
+ {
+ mpfr_set_d(mp,v,mpreal::get_default_rnd());
+ }else
+ throw conversion_overflow();
+#else
+ mpfr_set_d(mp,v,mpreal::get_default_rnd());
+#endif
MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ return *this;
}
-inline mpreal& mpreal::operator=(const double v)
-{
- if(double_bits == -1 || fits_in_bits(v, double_bits))
- {
- mpfr_set_d(mp,v,default_rnd);
+inline mpreal& mpreal::operator=(const unsigned long int v)
+{
+ mpfr_set_ui(mp, v, mpreal::get_default_rnd());
- MPREAL_MSVC_DEBUGVIEW_CODE;
- }
- else
- throw conversion_overflow();
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
+}
- return *this;
+inline mpreal& mpreal::operator=(const unsigned int v)
+{
+ mpfr_set_ui(mp, v, mpreal::get_default_rnd());
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
-inline mpreal& mpreal::operator=(const unsigned long int v)
-{
- mpfr_set_ui(mp,v,default_rnd);
+inline mpreal& mpreal::operator=(const long int v)
+{
+ mpfr_set_si(mp, v, mpreal::get_default_rnd());
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
-inline mpreal& mpreal::operator=(const unsigned int v)
-{
- mpfr_set_ui(mp,v,default_rnd);
+inline mpreal& mpreal::operator=(const int v)
+{
+ mpfr_set_si(mp, v, mpreal::get_default_rnd());
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
-inline mpreal& mpreal::operator=(const long int v)
-{
- mpfr_set_si(mp,v,default_rnd);
+inline mpreal& mpreal::operator=(const char* s)
+{
+ // Use other converters for more precise control on base & precision & rounding:
+ //
+ // mpreal(const char* s, mp_prec_t prec, int base, mp_rnd_t mode)
+ // mpreal(const std::string& s,mp_prec_t prec, int base, mp_rnd_t mode)
+ //
+ // Here we assume base = 10 and we use precision of target variable.
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_t t;
+
+ mpfr_init2(t, mpfr_get_prec(mp));
+
+ if(0 == mpfr_set_str(t, s, 10, mpreal::get_default_rnd()))
+ {
+ mpfr_set(mp, t, mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ }
+
+ mpfr_clear(t);
+ return *this;
}
-inline mpreal& mpreal::operator=(const int v)
-{
- mpfr_set_si(mp,v,default_rnd);
+inline mpreal& mpreal::operator=(const std::string& s)
+{
+ // Use other converters for more precise control on base & precision & rounding:
+ //
+ // mpreal(const char* s, mp_prec_t prec, int base, mp_rnd_t mode)
+ // mpreal(const std::string& s,mp_prec_t prec, int base, mp_rnd_t mode)
+ //
+ // Here we assume base = 10 and we use precision of target variable.
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_t t;
+
+ mpfr_init2(t, mpfr_get_prec(mp));
+
+ if(0 == mpfr_set_str(t, s.c_str(), 10, mpreal::get_default_rnd()))
+ {
+ mpfr_set(mp, t, mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ }
+
+ mpfr_clear(t);
+ return *this;
}
+
//////////////////////////////////////////////////////////////////////////
// + Addition
inline mpreal& mpreal::operator+=(const mpreal& v)
{
- mpfr_add(mp,mp,v.mp,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_add(mp,mp,v.mp,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator+=(const mpf_t u)
{
- *this += mpreal(u);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ *this += mpreal(u);
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator+=(const mpz_t u)
{
- mpfr_add_z(mp,mp,u,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_add_z(mp,mp,u,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator+=(const mpq_t u)
{
- mpfr_add_q(mp,mp,u,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_add_q(mp,mp,u,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator+= (const long double u)
{
- *this += mpreal(u);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ *this += mpreal(u);
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator+= (const double u)
{
#if (MPFR_VERSION >= MPFR_VERSION_NUM(2,4,0))
- mpfr_add_d(mp,mp,u,default_rnd);
+ mpfr_add_d(mp,mp,u,mpreal::get_default_rnd());
#else
- *this += mpreal(u);
+ *this += mpreal(u);
#endif
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator+=(const unsigned long int u)
{
- mpfr_add_ui(mp,mp,u,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_add_ui(mp,mp,u,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator+=(const unsigned int u)
{
- mpfr_add_ui(mp,mp,u,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_add_ui(mp,mp,u,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator+=(const long int u)
{
- mpfr_add_si(mp,mp,u,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_add_si(mp,mp,u,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator+=(const int u)
{
- mpfr_add_si(mp,mp,u,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_add_si(mp,mp,u,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
#if defined (MPREAL_HAVE_INT64_SUPPORT)
-inline mpreal& mpreal::operator+=(const int64_t u){ *this += mpreal(u); MPREAL_MSVC_DEBUGVIEW_CODE; return *this; }
-inline mpreal& mpreal::operator+=(const uint64_t u){ *this += mpreal(u); MPREAL_MSVC_DEBUGVIEW_CODE; return *this; }
-inline mpreal& mpreal::operator-=(const int64_t u){ *this -= mpreal(u); MPREAL_MSVC_DEBUGVIEW_CODE; return *this; }
-inline mpreal& mpreal::operator-=(const uint64_t u){ *this -= mpreal(u); MPREAL_MSVC_DEBUGVIEW_CODE; return *this; }
-inline mpreal& mpreal::operator*=(const int64_t u){ *this *= mpreal(u); MPREAL_MSVC_DEBUGVIEW_CODE; return *this; }
-inline mpreal& mpreal::operator*=(const uint64_t u){ *this *= mpreal(u); MPREAL_MSVC_DEBUGVIEW_CODE; return *this; }
-inline mpreal& mpreal::operator/=(const int64_t u){ *this /= mpreal(u); MPREAL_MSVC_DEBUGVIEW_CODE; return *this; }
-inline mpreal& mpreal::operator/=(const uint64_t u){ *this /= mpreal(u); MPREAL_MSVC_DEBUGVIEW_CODE; return *this; }
+inline mpreal& mpreal::operator+=(const int64_t u){ *this += mpreal(u); MPREAL_MSVC_DEBUGVIEW_CODE; return *this; }
+inline mpreal& mpreal::operator+=(const uint64_t u){ *this += mpreal(u); MPREAL_MSVC_DEBUGVIEW_CODE; return *this; }
+inline mpreal& mpreal::operator-=(const int64_t u){ *this -= mpreal(u); MPREAL_MSVC_DEBUGVIEW_CODE; return *this; }
+inline mpreal& mpreal::operator-=(const uint64_t u){ *this -= mpreal(u); MPREAL_MSVC_DEBUGVIEW_CODE; return *this; }
+inline mpreal& mpreal::operator*=(const int64_t u){ *this *= mpreal(u); MPREAL_MSVC_DEBUGVIEW_CODE; return *this; }
+inline mpreal& mpreal::operator*=(const uint64_t u){ *this *= mpreal(u); MPREAL_MSVC_DEBUGVIEW_CODE; return *this; }
+inline mpreal& mpreal::operator/=(const int64_t u){ *this /= mpreal(u); MPREAL_MSVC_DEBUGVIEW_CODE; return *this; }
+inline mpreal& mpreal::operator/=(const uint64_t u){ *this /= mpreal(u); MPREAL_MSVC_DEBUGVIEW_CODE; return *this; }
#endif
-inline const mpreal mpreal::operator+()const { return mpreal(*this); }
+inline const mpreal mpreal::operator+()const { return mpreal(*this); }
inline const mpreal operator+(const mpreal& a, const mpreal& b)
{
- // prec(a+b) = max(prec(a),prec(b))
- if(a.get_prec()>b.get_prec()) return mpreal(a) += b;
- else return mpreal(b) += a;
+ mpreal c(0, (std::max)(mpfr_get_prec(a.mpfr_ptr()), mpfr_get_prec(b.mpfr_ptr())));
+ mpfr_add(c.mpfr_ptr(), a.mpfr_srcptr(), b.mpfr_srcptr(), mpreal::get_default_rnd());
+ return c;
}
inline mpreal& mpreal::operator++()
{
- return *this += 1;
+ return *this += 1;
}
inline const mpreal mpreal::operator++ (int)
{
- mpreal x(*this);
- *this += 1;
- return x;
+ mpreal x(*this);
+ *this += 1;
+ return x;
}
inline mpreal& mpreal::operator--()
{
- return *this -= 1;
+ return *this -= 1;
}
inline const mpreal mpreal::operator-- (int)
{
- mpreal x(*this);
- *this -= 1;
- return x;
+ mpreal x(*this);
+ *this -= 1;
+ return x;
}
//////////////////////////////////////////////////////////////////////////
// - Subtraction
-inline mpreal& mpreal::operator-= (const mpreal& v)
+inline mpreal& mpreal::operator-=(const mpreal& v)
{
- mpfr_sub(mp,mp,v.mp,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_sub(mp,mp,v.mp,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator-=(const mpz_t v)
{
- mpfr_sub_z(mp,mp,v,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_sub_z(mp,mp,v,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator-=(const mpq_t v)
{
- mpfr_sub_q(mp,mp,v,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_sub_q(mp,mp,v,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator-=(const long double v)
{
- *this -= mpreal(v);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ *this -= mpreal(v);
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator-=(const double v)
{
#if (MPFR_VERSION >= MPFR_VERSION_NUM(2,4,0))
- mpfr_sub_d(mp,mp,v,default_rnd);
+ mpfr_sub_d(mp,mp,v,mpreal::get_default_rnd());
#else
- *this -= mpreal(v);
+ *this -= mpreal(v);
#endif
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator-=(const unsigned long int v)
{
- mpfr_sub_ui(mp,mp,v,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_sub_ui(mp,mp,v,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator-=(const unsigned int v)
{
- mpfr_sub_ui(mp,mp,v,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_sub_ui(mp,mp,v,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator-=(const long int v)
{
- mpfr_sub_si(mp,mp,v,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_sub_si(mp,mp,v,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator-=(const int v)
{
- mpfr_sub_si(mp,mp,v,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_sub_si(mp,mp,v,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline const mpreal mpreal::operator-()const
{
- mpreal u(*this);
- mpfr_neg(u.mp,u.mp,default_rnd);
- return u;
+ mpreal u(*this);
+ mpfr_neg(u.mp,u.mp,mpreal::get_default_rnd());
+ return u;
}
inline const mpreal operator-(const mpreal& a, const mpreal& b)
{
- // prec(a-b) = max(prec(a),prec(b))
- if(a.getPrecision() >= b.getPrecision())
- {
- return mpreal(a) -= b;
- }else{
- mpreal x(a);
- x.setPrecision(b.getPrecision());
- return x -= b;
- }
+ mpreal c(0, (std::max)(mpfr_get_prec(a.mpfr_ptr()), mpfr_get_prec(b.mpfr_ptr())));
+ mpfr_sub(c.mpfr_ptr(), a.mpfr_srcptr(), b.mpfr_srcptr(), mpreal::get_default_rnd());
+ return c;
}
inline const mpreal operator-(const double b, const mpreal& a)
{
#if (MPFR_VERSION >= MPFR_VERSION_NUM(2,4,0))
- mpreal x(a);
- mpfr_d_sub(x.mp,b,a.mp,mpreal::default_rnd);
- return x;
+ mpreal x(0, mpfr_get_prec(a.mpfr_ptr()));
+ mpfr_d_sub(x.mpfr_ptr(), b, a.mpfr_srcptr(), mpreal::get_default_rnd());
+ return x;
#else
- return mpreal(b) -= a;
+ mpreal x(b, mpfr_get_prec(a.mpfr_ptr()));
+ x -= a;
+ return x;
#endif
}
inline const mpreal operator-(const unsigned long int b, const mpreal& a)
{
- mpreal x(a);
- mpfr_ui_sub(x.mp,b,a.mp,mpreal::default_rnd);
- return x;
+ mpreal x(0, mpfr_get_prec(a.mpfr_ptr()));
+ mpfr_ui_sub(x.mpfr_ptr(), b, a.mpfr_srcptr(), mpreal::get_default_rnd());
+ return x;
}
inline const mpreal operator-(const unsigned int b, const mpreal& a)
{
- mpreal x(a);
- mpfr_ui_sub(x.mp,b,a.mp,mpreal::default_rnd);
- return x;
+ mpreal x(0, mpfr_get_prec(a.mpfr_ptr()));
+ mpfr_ui_sub(x.mpfr_ptr(), b, a.mpfr_srcptr(), mpreal::get_default_rnd());
+ return x;
}
inline const mpreal operator-(const long int b, const mpreal& a)
{
- mpreal x(a);
- mpfr_si_sub(x.mp,b,a.mp,mpreal::default_rnd);
- return x;
+ mpreal x(0, mpfr_get_prec(a.mpfr_ptr()));
+ mpfr_si_sub(x.mpfr_ptr(), b, a.mpfr_srcptr(), mpreal::get_default_rnd());
+ return x;
}
inline const mpreal operator-(const int b, const mpreal& a)
{
- mpreal x(a);
- mpfr_si_sub(x.mp,b,a.mp,mpreal::default_rnd);
- return x;
+ mpreal x(0, mpfr_get_prec(a.mpfr_ptr()));
+ mpfr_si_sub(x.mpfr_ptr(), b, a.mpfr_srcptr(), mpreal::get_default_rnd());
+ return x;
}
//////////////////////////////////////////////////////////////////////////
// * Multiplication
inline mpreal& mpreal::operator*= (const mpreal& v)
{
- mpfr_mul(mp,mp,v.mp,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_mul(mp,mp,v.mp,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator*=(const mpz_t v)
{
- mpfr_mul_z(mp,mp,v,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_mul_z(mp,mp,v,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator*=(const mpq_t v)
{
- mpfr_mul_q(mp,mp,v,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_mul_q(mp,mp,v,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator*=(const long double v)
{
- *this *= mpreal(v);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ *this *= mpreal(v);
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator*=(const double v)
{
#if (MPFR_VERSION >= MPFR_VERSION_NUM(2,4,0))
- mpfr_mul_d(mp,mp,v,default_rnd);
+ mpfr_mul_d(mp,mp,v,mpreal::get_default_rnd());
#else
- *this *= mpreal(v);
+ *this *= mpreal(v);
#endif
-
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator*=(const unsigned long int v)
{
- mpfr_mul_ui(mp,mp,v,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_mul_ui(mp,mp,v,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator*=(const unsigned int v)
{
- mpfr_mul_ui(mp,mp,v,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_mul_ui(mp,mp,v,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator*=(const long int v)
{
- mpfr_mul_si(mp,mp,v,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_mul_si(mp,mp,v,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator*=(const int v)
{
- mpfr_mul_si(mp,mp,v,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_mul_si(mp,mp,v,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline const mpreal operator*(const mpreal& a, const mpreal& b)
{
- // prec(a*b) = max(prec(a),prec(b))
- if(a.getPrecision() >= b.getPrecision()) return mpreal(a) *= b;
- else return mpreal(b) *= a;
+ mpreal c(0, (std::max)(mpfr_get_prec(a.mpfr_ptr()), mpfr_get_prec(b.mpfr_ptr())));
+ mpfr_mul(c.mpfr_ptr(), a.mpfr_srcptr(), b.mpfr_srcptr(), mpreal::get_default_rnd());
+ return c;
}
//////////////////////////////////////////////////////////////////////////
// / Division
inline mpreal& mpreal::operator/=(const mpreal& v)
{
- mpfr_div(mp,mp,v.mp,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_div(mp,mp,v.mp,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator/=(const mpz_t v)
{
- mpfr_div_z(mp,mp,v,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_div_z(mp,mp,v,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator/=(const mpq_t v)
{
- mpfr_div_q(mp,mp,v,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_div_q(mp,mp,v,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator/=(const long double v)
{
- *this /= mpreal(v);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ *this /= mpreal(v);
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator/=(const double v)
{
#if (MPFR_VERSION >= MPFR_VERSION_NUM(2,4,0))
- mpfr_div_d(mp,mp,v,default_rnd);
+ mpfr_div_d(mp,mp,v,mpreal::get_default_rnd());
#else
- *this /= mpreal(v);
+ *this /= mpreal(v);
#endif
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator/=(const unsigned long int v)
{
- mpfr_div_ui(mp,mp,v,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_div_ui(mp,mp,v,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator/=(const unsigned int v)
{
- mpfr_div_ui(mp,mp,v,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_div_ui(mp,mp,v,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator/=(const long int v)
{
- mpfr_div_si(mp,mp,v,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_div_si(mp,mp,v,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator/=(const int v)
{
- mpfr_div_si(mp,mp,v,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_div_si(mp,mp,v,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline const mpreal operator/(const mpreal& a, const mpreal& b)
{
- // prec(a/b) = max(prec(a),prec(b))
- if(a.getPrecision() >= b.getPrecision())
- {
- return mpreal(a) /= b;
- }else{
-
- mpreal x(a);
- x.setPrecision(b.getPrecision());
- return x /= b;
- }
+ mpreal c(0, (std::max)(mpfr_get_prec(a.mpfr_ptr()), mpfr_get_prec(b.mpfr_ptr())));
+ mpfr_div(c.mpfr_ptr(), a.mpfr_srcptr(), b.mpfr_srcptr(), mpreal::get_default_rnd());
+ return c;
}
inline const mpreal operator/(const unsigned long int b, const mpreal& a)
{
- mpreal x(a);
- mpfr_ui_div(x.mp,b,a.mp,mpreal::default_rnd);
- return x;
+ mpreal x(0, mpfr_get_prec(a.mpfr_ptr()));
+ mpfr_ui_div(x.mpfr_ptr(), b, a.mpfr_srcptr(), mpreal::get_default_rnd());
+ return x;
}
inline const mpreal operator/(const unsigned int b, const mpreal& a)
{
- mpreal x(a);
- mpfr_ui_div(x.mp,b,a.mp,mpreal::default_rnd);
- return x;
+ mpreal x(0, mpfr_get_prec(a.mpfr_ptr()));
+ mpfr_ui_div(x.mpfr_ptr(), b, a.mpfr_srcptr(), mpreal::get_default_rnd());
+ return x;
}
inline const mpreal operator/(const long int b, const mpreal& a)
{
- mpreal x(a);
- mpfr_si_div(x.mp,b,a.mp,mpreal::default_rnd);
- return x;
+ mpreal x(0, mpfr_get_prec(a.mpfr_ptr()));
+ mpfr_si_div(x.mpfr_ptr(), b, a.mpfr_srcptr(),mpreal::get_default_rnd());
+ return x;
}
inline const mpreal operator/(const int b, const mpreal& a)
{
- mpreal x(a);
- mpfr_si_div(x.mp,b,a.mp,mpreal::default_rnd);
- return x;
+ mpreal x(0, mpfr_get_prec(a.mpfr_ptr()));
+ mpfr_si_div(x.mpfr_ptr(), b, a.mpfr_srcptr(),mpreal::get_default_rnd());
+ return x;
}
inline const mpreal operator/(const double b, const mpreal& a)
{
#if (MPFR_VERSION >= MPFR_VERSION_NUM(2,4,0))
- mpreal x(a);
- mpfr_d_div(x.mp,b,a.mp,mpreal::default_rnd);
- return x;
+ mpreal x(0, mpfr_get_prec(a.mpfr_ptr()));
+ mpfr_d_div(x.mpfr_ptr(), b, a.mpfr_srcptr(),mpreal::get_default_rnd());
+ return x;
#else
- return mpreal(b) /= a;
+ mpreal x(0, mpfr_get_prec(a.mpfr_ptr()));
+ x /= a;
+ return x;
#endif
}
@@ -1242,1106 +1445,1034 @@ inline const mpreal operator/(const double b, const mpreal& a) // Shifts operators - Multiplication/Division by power of 2
inline mpreal& mpreal::operator<<=(const unsigned long int u)
{
- mpfr_mul_2ui(mp,mp,u,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_mul_2ui(mp,mp,u,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator<<=(const unsigned int u)
{
- mpfr_mul_2ui(mp,mp,static_cast<unsigned long int>(u),default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_mul_2ui(mp,mp,static_cast<unsigned long int>(u),mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator<<=(const long int u)
{
- mpfr_mul_2si(mp,mp,u,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_mul_2si(mp,mp,u,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator<<=(const int u)
{
- mpfr_mul_2si(mp,mp,static_cast<long int>(u),default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_mul_2si(mp,mp,static_cast<long int>(u),mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator>>=(const unsigned long int u)
{
- mpfr_div_2ui(mp,mp,u,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_div_2ui(mp,mp,u,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator>>=(const unsigned int u)
{
- mpfr_div_2ui(mp,mp,static_cast<unsigned long int>(u),default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_div_2ui(mp,mp,static_cast<unsigned long int>(u),mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator>>=(const long int u)
{
- mpfr_div_2si(mp,mp,u,default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_div_2si(mp,mp,u,mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::operator>>=(const int u)
{
- mpfr_div_2si(mp,mp,static_cast<long int>(u),default_rnd);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_div_2si(mp,mp,static_cast<long int>(u),mpreal::get_default_rnd());
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline const mpreal operator<<(const mpreal& v, const unsigned long int k)
{
- return mul_2ui(v,k);
+ return mul_2ui(v,k);
}
inline const mpreal operator<<(const mpreal& v, const unsigned int k)
{
- return mul_2ui(v,static_cast<unsigned long int>(k));
+ return mul_2ui(v,static_cast<unsigned long int>(k));
}
inline const mpreal operator<<(const mpreal& v, const long int k)
{
- return mul_2si(v,k);
+ return mul_2si(v,k);
}
inline const mpreal operator<<(const mpreal& v, const int k)
{
- return mul_2si(v,static_cast<long int>(k));
+ return mul_2si(v,static_cast<long int>(k));
}
inline const mpreal operator>>(const mpreal& v, const unsigned long int k)
{
- return div_2ui(v,k);
+ return div_2ui(v,k);
}
inline const mpreal operator>>(const mpreal& v, const long int k)
{
- return div_2si(v,k);
+ return div_2si(v,k);
}
inline const mpreal operator>>(const mpreal& v, const unsigned int k)
{
- return div_2ui(v,static_cast<unsigned long int>(k));
+ return div_2ui(v,static_cast<unsigned long int>(k));
}
inline const mpreal operator>>(const mpreal& v, const int k)
{
- return div_2si(v,static_cast<long int>(k));
+ return div_2si(v,static_cast<long int>(k));
}
// mul_2ui
inline const mpreal mul_2ui(const mpreal& v, unsigned long int k, mp_rnd_t rnd_mode)
{
- mpreal x(v);
- mpfr_mul_2ui(x.mp,v.mp,k,rnd_mode);
- return x;
+ mpreal x(v);
+ mpfr_mul_2ui(x.mp,v.mp,k,rnd_mode);
+ return x;
}
// mul_2si
inline const mpreal mul_2si(const mpreal& v, long int k, mp_rnd_t rnd_mode)
{
- mpreal x(v);
- mpfr_mul_2si(x.mp,v.mp,k,rnd_mode);
- return x;
+ mpreal x(v);
+ mpfr_mul_2si(x.mp,v.mp,k,rnd_mode);
+ return x;
}
inline const mpreal div_2ui(const mpreal& v, unsigned long int k, mp_rnd_t rnd_mode)
{
- mpreal x(v);
- mpfr_div_2ui(x.mp,v.mp,k,rnd_mode);
- return x;
+ mpreal x(v);
+ mpfr_div_2ui(x.mp,v.mp,k,rnd_mode);
+ return x;
}
inline const mpreal div_2si(const mpreal& v, long int k, mp_rnd_t rnd_mode)
{
- mpreal x(v);
- mpfr_div_2si(x.mp,v.mp,k,rnd_mode);
- return x;
+ mpreal x(v);
+ mpfr_div_2si(x.mp,v.mp,k,rnd_mode);
+ return x;
}
//////////////////////////////////////////////////////////////////////////
//Boolean operators
-inline bool operator > (const mpreal& a, const mpreal& b){ return (mpfr_greater_p(a.mp,b.mp) !=0); }
-inline bool operator >= (const mpreal& a, const mpreal& b){ return (mpfr_greaterequal_p(a.mp,b.mp) !=0); }
-inline bool operator < (const mpreal& a, const mpreal& b){ return (mpfr_less_p(a.mp,b.mp) !=0); }
-inline bool operator <= (const mpreal& a, const mpreal& b){ return (mpfr_lessequal_p(a.mp,b.mp) !=0); }
-inline bool operator == (const mpreal& a, const mpreal& b){ return (mpfr_equal_p(a.mp,b.mp) !=0); }
-inline bool operator != (const mpreal& a, const mpreal& b){ return (mpfr_lessgreater_p(a.mp,b.mp) !=0); }
-
-inline bool operator == (const mpreal& a, const unsigned long int b ){ return (mpfr_cmp_ui(a.mp,b) == 0); }
-inline bool operator == (const mpreal& a, const unsigned int b ){ return (mpfr_cmp_ui(a.mp,b) == 0); }
-inline bool operator == (const mpreal& a, const long int b ){ return (mpfr_cmp_si(a.mp,b) == 0); }
-inline bool operator == (const mpreal& a, const int b ){ return (mpfr_cmp_si(a.mp,b) == 0); }
-inline bool operator == (const mpreal& a, const long double b ){ return (mpfr_cmp_ld(a.mp,b) == 0); }
-inline bool operator == (const mpreal& a, const double b ){ return (mpfr_cmp_d(a.mp,b) == 0); }
-
-
-inline bool isnan (const mpreal& v){ return (mpfr_nan_p(v.mp) != 0); }
-inline bool isinf (const mpreal& v){ return (mpfr_inf_p(v.mp) != 0); }
-inline bool isfinite(const mpreal& v){ return (mpfr_number_p(v.mp) != 0); }
-inline bool iszero (const mpreal& v){ return (mpfr_zero_p(v.mp) != 0); }
-inline bool isint (const mpreal& v){ return (mpfr_integer_p(v.mp) != 0); }
+inline bool operator > (const mpreal& a, const mpreal& b){ return (mpfr_greater_p(a.mp,b.mp) !=0); }
+inline bool operator >= (const mpreal& a, const mpreal& b){ return (mpfr_greaterequal_p(a.mp,b.mp) !=0); }
+inline bool operator < (const mpreal& a, const mpreal& b){ return (mpfr_less_p(a.mp,b.mp) !=0); }
+inline bool operator <= (const mpreal& a, const mpreal& b){ return (mpfr_lessequal_p(a.mp,b.mp) !=0); }
+inline bool operator == (const mpreal& a, const mpreal& b){ return (mpfr_equal_p(a.mp,b.mp) !=0); }
+inline bool operator != (const mpreal& a, const mpreal& b){ return (mpfr_lessgreater_p(a.mp,b.mp) !=0); }
+
+inline bool operator == (const mpreal& a, const unsigned long int b ){ return (mpfr_cmp_ui(a.mp,b) == 0); }
+inline bool operator == (const mpreal& a, const unsigned int b ){ return (mpfr_cmp_ui(a.mp,b) == 0); }
+inline bool operator == (const mpreal& a, const long int b ){ return (mpfr_cmp_si(a.mp,b) == 0); }
+inline bool operator == (const mpreal& a, const int b ){ return (mpfr_cmp_si(a.mp,b) == 0); }
+inline bool operator == (const mpreal& a, const long double b ){ return (mpfr_cmp_ld(a.mp,b) == 0); }
+inline bool operator == (const mpreal& a, const double b ){ return (mpfr_cmp_d(a.mp,b) == 0); }
+
+
+inline bool isnan (const mpreal& v){ return (mpfr_nan_p(v.mp) != 0); }
+inline bool isinf (const mpreal& v){ return (mpfr_inf_p(v.mp) != 0); }
+inline bool isfinite (const mpreal& v){ return (mpfr_number_p(v.mp) != 0); }
+inline bool iszero (const mpreal& v){ return (mpfr_zero_p(v.mp) != 0); }
+inline bool isint (const mpreal& v){ return (mpfr_integer_p(v.mp) != 0); }
#if (MPFR_VERSION >= MPFR_VERSION_NUM(3,0,0))
-inline bool isregular(const mpreal& v){ return (mpfr_regular_p(v.mp));}
+inline bool isregular(const mpreal& v){ return (mpfr_regular_p(v.mp));}
#endif
//////////////////////////////////////////////////////////////////////////
// Type Converters
-inline long mpreal::toLong() const { return mpfr_get_si(mp,GMP_RNDZ); }
-inline unsigned long mpreal::toULong() const { return mpfr_get_ui(mp,GMP_RNDZ); }
-inline double mpreal::toDouble() const { return mpfr_get_d(mp,default_rnd); }
-inline long double mpreal::toLDouble() const { return mpfr_get_ld(mp,default_rnd); }
+inline long mpreal::toLong (mp_rnd_t mode) const { return mpfr_get_si(mp, mode); }
+inline unsigned long mpreal::toULong (mp_rnd_t mode) const { return mpfr_get_ui(mp, mode); }
+inline double mpreal::toDouble (mp_rnd_t mode) const { return mpfr_get_d (mp, mode); }
+inline long double mpreal::toLDouble(mp_rnd_t mode) const { return mpfr_get_ld(mp, mode); }
#if defined (MPREAL_HAVE_INT64_SUPPORT)
-inline int64_t mpreal::toInt64() const{ return mpfr_get_sj(mp,GMP_RNDZ); }
-inline uint64_t mpreal::toUInt64() const{ return mpfr_get_uj(mp,GMP_RNDZ); }
+inline int64_t mpreal::toInt64 (mp_rnd_t mode) const{ return mpfr_get_sj(mp, mode); }
+inline uint64_t mpreal::toUInt64(mp_rnd_t mode) const{ return mpfr_get_uj(mp, mode); }
+#endif
+
+inline ::mpfr_ptr mpreal::mpfr_ptr() { return mp; }
+inline ::mpfr_srcptr mpreal::mpfr_ptr() const { return mp; }
+inline ::mpfr_srcptr mpreal::mpfr_srcptr() const { return mp; }
+
+template <class T>
+inline std::string toString(T t, std::ios_base & (*f)(std::ios_base&))
+{
+ std::ostringstream oss;
+ oss << f << t;
+ return oss.str();
+}
+
+#if (MPFR_VERSION >= MPFR_VERSION_NUM(2,4,0))
+
+inline std::string mpreal::toString(const std::string& format) const
+{
+ char *s = NULL;
+ std::string out;
+
+ if( !format.empty() )
+ {
+ if(!(mpfr_asprintf(&s,format.c_str(),mp) < 0))
+ {
+ out = std::string(s);
+
+ mpfr_free_str(s);
+ }
+ }
+
+ return out;
+}
+
+#endif
+
+inline std::string mpreal::toString(int n, int b, mp_rnd_t mode) const
+{
+ (void)b;
+ (void)mode;
+
+#if (MPFR_VERSION >= MPFR_VERSION_NUM(2,4,0))
+
+ // Use MPFR native function for output
+ char format[128];
+ int digits;
+
+ digits = n > 0 ? n : bits2digits(mpfr_get_prec(mp));
+
+ sprintf(format,"%%.%dRNg",digits); // Default format
+
+ return toString(std::string(format));
+
+#else
+
+ char *s, *ns = NULL;
+ size_t slen, nslen;
+ mp_exp_t exp;
+ std::string out;
+
+ if(mpfr_inf_p(mp))
+ {
+ if(mpfr_sgn(mp)>0) return "+Inf";
+ else return "-Inf";
+ }
+
+ if(mpfr_zero_p(mp)) return "0";
+ if(mpfr_nan_p(mp)) return "NaN";
+
+ s = mpfr_get_str(NULL,&exp,b,0,mp,mode);
+ ns = mpfr_get_str(NULL,&exp,b,n,mp,mode);
+
+ if(s!=NULL && ns!=NULL)
+ {
+ slen = strlen(s);
+ nslen = strlen(ns);
+ if(nslen<=slen)
+ {
+ mpfr_free_str(s);
+ s = ns;
+ slen = nslen;
+ }
+ else {
+ mpfr_free_str(ns);
+ }
+
+ // Make human eye-friendly formatting if possible
+ if (exp>0 && static_cast<size_t>(exp)<slen)
+ {
+ if(s[0]=='-')
+ {
+ // Remove zeros starting from right end
+ char* ptr = s+slen-1;
+ while (*ptr=='0' && ptr>s+exp) ptr--;
+
+ if(ptr==s+exp) out = std::string(s,exp+1);
+ else out = std::string(s,exp+1)+'.'+std::string(s+exp+1,ptr-(s+exp+1)+1);
+
+ //out = string(s,exp+1)+'.'+string(s+exp+1);
+ }
+ else
+ {
+ // Remove zeros starting from right end
+ char* ptr = s+slen-1;
+ while (*ptr=='0' && ptr>s+exp-1) ptr--;
+
+ if(ptr==s+exp-1) out = std::string(s,exp);
+ else out = std::string(s,exp)+'.'+std::string(s+exp,ptr-(s+exp)+1);
+
+ //out = string(s,exp)+'.'+string(s+exp);
+ }
+
+ }else{ // exp<0 || exp>slen
+ if(s[0]=='-')
+ {
+ // Remove zeros starting from right end
+ char* ptr = s+slen-1;
+ while (*ptr=='0' && ptr>s+1) ptr--;
+
+ if(ptr==s+1) out = std::string(s,2);
+ else out = std::string(s,2)+'.'+std::string(s+2,ptr-(s+2)+1);
+
+ //out = string(s,2)+'.'+string(s+2);
+ }
+ else
+ {
+ // Remove zeros starting from right end
+ char* ptr = s+slen-1;
+ while (*ptr=='0' && ptr>s) ptr--;
+
+ if(ptr==s) out = std::string(s,1);
+ else out = std::string(s,1)+'.'+std::string(s+1,ptr-(s+1)+1);
+
+ //out = string(s,1)+'.'+string(s+1);
+ }
+
+ // Make final string
+ if(--exp)
+ {
+ if(exp>0) out += "e+"+mpfr::toString<mp_exp_t>(exp,std::dec);
+ else out += "e"+mpfr::toString<mp_exp_t>(exp,std::dec);
+ }
+ }
+
+ mpfr_free_str(s);
+ return out;
+ }else{
+ return "conversion error!";
+ }
#endif
+}
+
+
+//////////////////////////////////////////////////////////////////////////
+// I/O
+inline std::ostream& operator<<(std::ostream& os, const mpreal& v)
+{
+ return os<<v.toString(static_cast<int>(os.precision()));
+}
-inline ::mpfr_ptr mpreal::mpfr_ptr() { return mp; }
-inline ::mpfr_srcptr mpreal::mpfr_srcptr() const { return const_cast< ::mpfr_srcptr >(mp); }
+inline std::istream& operator>>(std::istream &is, mpreal& v)
+{
+ // ToDo, use cout::hexfloat and other flags to setup base
+ std::string tmp;
+ is >> tmp;
+ mpfr_set_str(v.mp, tmp.c_str(), 10, mpreal::get_default_rnd());
+ return is;
+}
//////////////////////////////////////////////////////////////////////////
-// Bits - decimal digits relation
-// bits = ceil(digits*log[2](10))
-// digits = floor(bits*log[10](2))
+// Bits - decimal digits relation
+// bits = ceil(digits*log[2](10))
+// digits = floor(bits*log[10](2))
inline mp_prec_t digits2bits(int d)
{
- const double LOG2_10 = 3.3219280948873624;
-
- d = 10>d?10:d;
+ const double LOG2_10 = 3.3219280948873624;
- return (mp_prec_t)std::ceil((d)*LOG2_10);
+ return (mp_prec_t) std::ceil( d * LOG2_10 );
}
inline int bits2digits(mp_prec_t b)
{
- const double LOG10_2 = 0.30102999566398119;
+ const double LOG10_2 = 0.30102999566398119;
- b = 34>b?34:b;
-
- return (int)std::floor((b)*LOG10_2);
+ return (int) std::floor( b * LOG10_2 );
}
//////////////////////////////////////////////////////////////////////////
// Set/Get number properties
inline int sgn(const mpreal& v)
{
- int r = mpfr_signbit(v.mp);
- return (r>0?-1:1);
+ int r = mpfr_signbit(v.mp);
+ return (r>0?-1:1);
}
inline mpreal& mpreal::setSign(int sign, mp_rnd_t RoundingMode)
{
- mpfr_setsign(mp,mp,(sign<0?1:0),RoundingMode);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_setsign(mp,mp,(sign < 0 ? 1 : 0),RoundingMode);
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline int mpreal::getPrecision() const
{
- return mpfr_get_prec(mp);
+ return mpfr_get_prec(mp);
}
inline mpreal& mpreal::setPrecision(int Precision, mp_rnd_t RoundingMode)
{
- mpfr_prec_round(mp,Precision, RoundingMode);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_prec_round(mp, Precision, RoundingMode);
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mpreal& mpreal::setInf(int sign)
{
- mpfr_set_inf(mp,sign);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
-}
+ mpfr_set_inf(mp,sign);
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
+}
inline mpreal& mpreal::setNan()
{
- mpfr_set_nan(mp);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+ mpfr_set_nan(mp);
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
-inline mpreal& mpreal::setZero(int sign)
+inline mpreal& mpreal::setZero(int sign)
{
- mpfr_set_zero(mp,sign);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return *this;
+
+#if (MPFR_VERSION >= MPFR_VERSION_NUM(3,0,0))
+ mpfr_set_zero(mp, sign);
+#else
+ mpfr_set_si(mp, 0, (mpfr_get_default_rounding_mode)());
+ setSign(sign);
+#endif
+
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return *this;
}
inline mp_prec_t mpreal::get_prec() const
{
- return mpfr_get_prec(mp);
+ return mpfr_get_prec(mp);
}
inline void mpreal::set_prec(mp_prec_t prec, mp_rnd_t rnd_mode)
{
- mpfr_prec_round(mp,prec,rnd_mode);
- MPREAL_MSVC_DEBUGVIEW_CODE;
+ mpfr_prec_round(mp,prec,rnd_mode);
+ MPREAL_MSVC_DEBUGVIEW_CODE;
}
inline mp_exp_t mpreal::get_exp ()
{
- return mpfr_get_exp(mp);
+ return mpfr_get_exp(mp);
}
inline int mpreal::set_exp (mp_exp_t e)
{
- int x = mpfr_set_exp(mp, e);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return x;
+ int x = mpfr_set_exp(mp, e);
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return x;
}
inline const mpreal frexp(const mpreal& v, mp_exp_t* exp)
{
- mpreal x(v);
- *exp = x.get_exp();
- x.set_exp(0);
- return x;
+ mpreal x(v);
+ *exp = x.get_exp();
+ x.set_exp(0);
+ return x;
}
inline const mpreal ldexp(const mpreal& v, mp_exp_t exp)
{
- mpreal x(v);
+ mpreal x(v);
- // rounding is not important since we just increasing the exponent
- mpfr_mul_2si(x.mp,x.mp,exp,mpreal::default_rnd);
- return x;
+ // rounding is not important since we just increasing the exponent
+ mpfr_mul_2si(x.mp,x.mp,exp,mpreal::get_default_rnd());
+ return x;
}
-inline const mpreal machine_epsilon(mp_prec_t prec)
+inline mpreal machine_epsilon(mp_prec_t prec)
{
- // the smallest eps such that 1.0+eps != 1.0
- // depends (of cause) on the precision
- return machine_epsilon(mpreal(1,prec));
+ /* the smallest eps such that 1 + eps != 1 */
+ return machine_epsilon(mpreal(1, prec));
}
-inline const mpreal machine_epsilon(const mpreal& x)
-{
- if( x < 0)
- {
- return nextabove(-x)+x;
- }else{
- return nextabove(x)-x;
- }
+inline mpreal machine_epsilon(const mpreal& x)
+{
+ /* the smallest eps such that x + eps != x */
+ if( x < 0)
+ {
+ return nextabove(-x)+x;
+ }else{
+ return nextabove(x)-x;
+ }
}
-inline const mpreal mpreal_min(mp_prec_t prec)
+// minval is 'safe' meaning 1 / minval does not overflow
+inline mpreal minval(mp_prec_t prec)
{
- // min = 1/2*2^emin = 2^(emin-1)
-
- return mpreal(1,prec) << mpreal::get_emin()-1;
+ /* min = 1/2 * 2^emin = 2^(emin - 1) */
+ return mpreal(1, prec) << mpreal::get_emin()-1;
}
-inline const mpreal mpreal_max(mp_prec_t prec)
+// maxval is 'safe' meaning 1 / maxval does not underflow
+inline mpreal maxval(mp_prec_t prec)
{
- // max = (1-eps)*2^emax, assume eps = 0?,
- // and use emax-1 to prevent value to be +inf
- // max = 2^(emax-1)
-
- return mpreal(1,prec) << mpreal::get_emax()-1;
+ /* max = (1 - eps) * 2^emax, eps is machine epsilon */
+ return (mpreal(1, prec) - machine_epsilon(prec)) << mpreal::get_emax();
}
inline bool isEqualUlps(const mpreal& a, const mpreal& b, int maxUlps)
{
- /*
- maxUlps - a and b can be apart by maxUlps binary numbers.
- */
return abs(a - b) <= machine_epsilon((max)(abs(a), abs(b))) * maxUlps;
}
inline bool isEqualFuzzy(const mpreal& a, const mpreal& b, const mpreal& eps)
{
- return abs(a - b) <= (min)(abs(a), abs(b)) * eps;
+ return abs(a - b) <= (min)(abs(a), abs(b)) * eps;
}
inline bool isEqualFuzzy(const mpreal& a, const mpreal& b)
{
- return isEqualFuzzy(a,b,machine_epsilon((std::min)(abs(a), abs(b))));
+ return isEqualFuzzy(a, b, machine_epsilon((min)(abs(a), abs(b))));
}
inline const mpreal modf(const mpreal& v, mpreal& n)
{
- mpreal frac(v);
+ mpreal frac(v);
- // rounding is not important since we are using the same number
- mpfr_frac(frac.mp,frac.mp,mpreal::default_rnd);
- mpfr_trunc(n.mp,v.mp);
- return frac;
+ // rounding is not important since we are using the same number
+ mpfr_frac(frac.mp,frac.mp,mpreal::get_default_rnd());
+ mpfr_trunc(n.mp,v.mp);
+ return frac;
}
inline int mpreal::check_range (int t, mp_rnd_t rnd_mode)
{
- return mpfr_check_range(mp,t,rnd_mode);
+ return mpfr_check_range(mp,t,rnd_mode);
}
inline int mpreal::subnormalize (int t,mp_rnd_t rnd_mode)
{
- int r = mpfr_subnormalize(mp,t,rnd_mode);
- MPREAL_MSVC_DEBUGVIEW_CODE;
- return r;
+ int r = mpfr_subnormalize(mp,t,rnd_mode);
+ MPREAL_MSVC_DEBUGVIEW_CODE;
+ return r;
}
inline mp_exp_t mpreal::get_emin (void)
{
- return mpfr_get_emin();
+ return mpfr_get_emin();
}
inline int mpreal::set_emin (mp_exp_t exp)
{
- return mpfr_set_emin(exp);
+ return mpfr_set_emin(exp);
}
inline mp_exp_t mpreal::get_emax (void)
{
- return mpfr_get_emax();
+ return mpfr_get_emax();
}
inline int mpreal::set_emax (mp_exp_t exp)
{
- return mpfr_set_emax(exp);
+ return mpfr_set_emax(exp);
}
inline mp_exp_t mpreal::get_emin_min (void)
{
- return mpfr_get_emin_min();
+ return mpfr_get_emin_min();
}
inline mp_exp_t mpreal::get_emin_max (void)
{
- return mpfr_get_emin_max();
+ return mpfr_get_emin_max();
}
inline mp_exp_t mpreal::get_emax_min (void)
{
- return mpfr_get_emax_min();
+ return mpfr_get_emax_min();
}
inline mp_exp_t mpreal::get_emax_max (void)
{
- return mpfr_get_emax_max();
+ return mpfr_get_emax_max();
}
//////////////////////////////////////////////////////////////////////////
// Mathematical Functions
//////////////////////////////////////////////////////////////////////////
-inline const mpreal sqr(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_sqr(x.mp,x.mp,rnd_mode);
- return x;
-}
+#define MPREAL_UNARY_MATH_FUNCTION_BODY(f) \
+ mpreal y(0, mpfr_get_prec(x.mpfr_srcptr())); \
+ mpfr_##f(y.mpfr_ptr(), x.mpfr_srcptr(), r); \
+ return y;
-inline const mpreal sqrt(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_sqrt(x.mp,x.mp,rnd_mode);
- return x;
-}
+inline const mpreal sqr (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(sqr ); }
+inline const mpreal sqrt (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(sqrt); }
-inline const mpreal sqrt(const unsigned long int v, mp_rnd_t rnd_mode)
+inline const mpreal sqrt(const unsigned long int x, mp_rnd_t r)
{
- mpreal x;
- mpfr_sqrt_ui(x.mp,v,rnd_mode);
- return x;
+ mpreal y;
+ mpfr_sqrt_ui(y.mpfr_ptr(), x, r);
+ return y;
}
inline const mpreal sqrt(const unsigned int v, mp_rnd_t rnd_mode)
{
- return sqrt(static_cast<unsigned long int>(v),rnd_mode);
+ return sqrt(static_cast<unsigned long int>(v),rnd_mode);
}
inline const mpreal sqrt(const long int v, mp_rnd_t rnd_mode)
{
- if (v>=0) return sqrt(static_cast<unsigned long int>(v),rnd_mode);
- else return mpreal().setNan(); // NaN
+ if (v>=0) return sqrt(static_cast<unsigned long int>(v),rnd_mode);
+ else return mpreal().setNan(); // NaN
}
inline const mpreal sqrt(const int v, mp_rnd_t rnd_mode)
{
- if (v>=0) return sqrt(static_cast<unsigned long int>(v),rnd_mode);
- else return mpreal().setNan(); // NaN
+ if (v>=0) return sqrt(static_cast<unsigned long int>(v),rnd_mode);
+ else return mpreal().setNan(); // NaN
}
-inline const mpreal sqrt(const long double v, mp_rnd_t rnd_mode)
+inline const mpreal root(const mpreal& x, unsigned long int k, mp_rnd_t r)
{
- return sqrt(mpreal(v),rnd_mode);
+ mpreal y(0, mpfr_get_prec(x.mpfr_srcptr()));
+ mpfr_root(y.mpfr_ptr(), x.mpfr_srcptr(), k, r);
+ return y;
}
-inline const mpreal sqrt(const double v, mp_rnd_t rnd_mode)
+inline const mpreal dim(const mpreal& a, const mpreal& b, mp_rnd_t r)
{
- return sqrt(mpreal(v),rnd_mode);
-}
-
-inline const mpreal cbrt(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_cbrt(x.mp,x.mp,rnd_mode);
- return x;
-}
-
-inline const mpreal root(const mpreal& v, unsigned long int k, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_root(x.mp,x.mp,k,rnd_mode);
- return x;
-}
-
-inline const mpreal fabs(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_abs(x.mp,x.mp,rnd_mode);
- return x;
-}
-
-inline const mpreal abs(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_abs(x.mp,x.mp,rnd_mode);
- return x;
-}
-
-inline const mpreal dim(const mpreal& a, const mpreal& b, mp_rnd_t rnd_mode)
-{
- mpreal x(a);
- mpfr_dim(x.mp,a.mp,b.mp,rnd_mode);
- return x;
+ mpreal y(0, mpfr_get_prec(a.mpfr_srcptr()));
+ mpfr_dim(y.mpfr_ptr(), a.mpfr_srcptr(), b.mpfr_srcptr(), r);
+ return y;
}
inline int cmpabs(const mpreal& a,const mpreal& b)
{
- return mpfr_cmpabs(a.mp,b.mp);
-}
-
-inline const mpreal log (const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_log(x.mp,v.mp,rnd_mode);
- return x;
-}
-
-inline const mpreal log2(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_log2(x.mp,v.mp,rnd_mode);
- return x;
+ return mpfr_cmpabs(a.mp,b.mp);
}
-inline const mpreal log10(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_log10(x.mp,v.mp,rnd_mode);
- return x;
-}
-
-inline const mpreal exp(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_exp(x.mp,v.mp,rnd_mode);
- return x;
-}
-
-inline const mpreal exp2(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_exp2(x.mp,v.mp,rnd_mode);
- return x;
-}
-
-inline const mpreal exp10(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_exp10(x.mp,v.mp,rnd_mode);
- return x;
-}
-
-inline const mpreal cos(const mpreal& v, mp_rnd_t rnd_mode)
+inline int sin_cos(mpreal& s, mpreal& c, const mpreal& v, mp_rnd_t rnd_mode)
{
- mpreal x(v);
- mpfr_cos(x.mp,v.mp,rnd_mode);
- return x;
-}
+ return mpfr_sin_cos(s.mp,c.mp,v.mp,rnd_mode);
+}
+
+inline const mpreal sqrt (const long double v, mp_rnd_t rnd_mode) { return sqrt(mpreal(v),rnd_mode); }
+inline const mpreal sqrt (const double v, mp_rnd_t rnd_mode) { return sqrt(mpreal(v),rnd_mode); }
+
+inline const mpreal cbrt (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(cbrt ); }
+inline const mpreal fabs (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(abs ); }
+inline const mpreal abs (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(abs ); }
+inline const mpreal log (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(log ); }
+inline const mpreal log2 (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(log2 ); }
+inline const mpreal log10 (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(log10); }
+inline const mpreal exp (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(exp ); }
+inline const mpreal exp2 (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(exp2 ); }
+inline const mpreal exp10 (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(exp10); }
+inline const mpreal cos (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(cos ); }
+inline const mpreal sin (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(sin ); }
+inline const mpreal tan (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(tan ); }
+inline const mpreal sec (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(sec ); }
+inline const mpreal csc (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(csc ); }
+inline const mpreal cot (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(cot ); }
+inline const mpreal acos (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(acos); }
+inline const mpreal asin (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(asin); }
+inline const mpreal atan (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(atan); }
+
+inline const mpreal acot (const mpreal& v, mp_rnd_t r) { return atan (1/v, r); }
+inline const mpreal asec (const mpreal& v, mp_rnd_t r) { return acos (1/v, r); }
+inline const mpreal acsc (const mpreal& v, mp_rnd_t r) { return asin (1/v, r); }
+inline const mpreal acoth (const mpreal& v, mp_rnd_t r) { return atanh(1/v, r); }
+inline const mpreal asech (const mpreal& v, mp_rnd_t r) { return acosh(1/v, r); }
+inline const mpreal acsch (const mpreal& v, mp_rnd_t r) { return asinh(1/v, r); }
+
+inline const mpreal cosh (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(cosh ); }
+inline const mpreal sinh (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(sinh ); }
+inline const mpreal tanh (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(tanh ); }
+inline const mpreal sech (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(sech ); }
+inline const mpreal csch (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(csch ); }
+inline const mpreal coth (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(coth ); }
+inline const mpreal acosh (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(acosh); }
+inline const mpreal asinh (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(asinh); }
+inline const mpreal atanh (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(atanh); }
+
+inline const mpreal log1p (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(log1p ); }
+inline const mpreal expm1 (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(expm1 ); }
+inline const mpreal eint (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(eint ); }
+inline const mpreal gamma (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(gamma ); }
+inline const mpreal lngamma (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(lngamma); }
+inline const mpreal zeta (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(zeta ); }
+inline const mpreal erf (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(erf ); }
+inline const mpreal erfc (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(erfc ); }
+inline const mpreal besselj0(const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(j0 ); }
+inline const mpreal besselj1(const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(j1 ); }
+inline const mpreal bessely0(const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(y0 ); }
+inline const mpreal bessely1(const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(y1 ); }
-inline const mpreal sin(const mpreal& v, mp_rnd_t rnd_mode)
+inline const mpreal atan2 (const mpreal& y, const mpreal& x, mp_rnd_t rnd_mode)
{
- mpreal x(v);
- mpfr_sin(x.mp,v.mp,rnd_mode);
- return x;
-}
+ mpreal a;
+ mp_prec_t yp, xp;
-inline const mpreal tan(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_tan(x.mp,v.mp,rnd_mode);
- return x;
-}
+ yp = y.get_prec();
+ xp = x.get_prec();
-inline const mpreal sec(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_sec(x.mp,v.mp,rnd_mode);
- return x;
-}
+ a.set_prec(yp>xp?yp:xp);
-inline const mpreal csc(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_csc(x.mp,v.mp,rnd_mode);
- return x;
-}
+ mpfr_atan2(a.mp, y.mp, x.mp, rnd_mode);
-inline const mpreal cot(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_cot(x.mp,v.mp,rnd_mode);
- return x;
+ return a;
}
-inline int sin_cos(mpreal& s, mpreal& c, const mpreal& v, mp_rnd_t rnd_mode)
+inline const mpreal hypot (const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode)
{
- return mpfr_sin_cos(s.mp,c.mp,v.mp,rnd_mode);
-}
+ mpreal a;
+ mp_prec_t yp, xp;
-inline const mpreal acos (const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_acos(x.mp,v.mp,rnd_mode);
- return x;
-}
+ yp = y.get_prec();
+ xp = x.get_prec();
-inline const mpreal asin (const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_asin(x.mp,v.mp,rnd_mode);
- return x;
-}
+ a.set_prec(yp>xp?yp:xp);
-inline const mpreal atan (const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_atan(x.mp,v.mp,rnd_mode);
- return x;
-}
+ mpfr_hypot(a.mp, x.mp, y.mp, rnd_mode);
-inline const mpreal acot (const mpreal& v, mp_rnd_t rnd_mode)
-{
- return atan(1/v, rnd_mode);
+ return a;
}
-inline const mpreal asec (const mpreal& v, mp_rnd_t rnd_mode)
-{
- return acos(1/v, rnd_mode);
-}
+inline const mpreal remainder (const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode)
+{
+ mpreal a;
+ mp_prec_t yp, xp;
-inline const mpreal acsc (const mpreal& v, mp_rnd_t rnd_mode)
-{
- return asin(1/v, rnd_mode);
-}
+ yp = y.get_prec();
+ xp = x.get_prec();
-inline const mpreal acoth (const mpreal& v, mp_rnd_t rnd_mode)
-{
- return atanh(1/v, rnd_mode);
-}
+ a.set_prec(yp>xp?yp:xp);
-inline const mpreal asech (const mpreal& v, mp_rnd_t rnd_mode)
-{
- return acosh(1/v, rnd_mode);
-}
+ mpfr_remainder(a.mp, x.mp, y.mp, rnd_mode);
-inline const mpreal acsch (const mpreal& v, mp_rnd_t rnd_mode)
-{
- return asinh(1/v, rnd_mode);
+ return a;
}
-inline const mpreal atan2 (const mpreal& y, const mpreal& x, mp_rnd_t rnd_mode)
+inline const mpreal remquo (long* q, const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode)
{
- mpreal a;
- mp_prec_t yp, xp;
+ mpreal a;
+ mp_prec_t yp, xp;
- yp = y.get_prec();
- xp = x.get_prec();
+ yp = y.get_prec();
+ xp = x.get_prec();
- a.set_prec(yp>xp?yp:xp);
+ a.set_prec(yp>xp?yp:xp);
- mpfr_atan2(a.mp, y.mp, x.mp, rnd_mode);
+ mpfr_remquo(a.mp,q, x.mp, y.mp, rnd_mode);
- return a;
+ return a;
}
-inline const mpreal cosh (const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_cosh(x.mp,v.mp,rnd_mode);
- return x;
-}
-
-inline const mpreal sinh (const mpreal& v, mp_rnd_t rnd_mode)
+inline const mpreal fac_ui (unsigned long int v, mp_prec_t prec, mp_rnd_t rnd_mode)
{
- mpreal x(v);
- mpfr_sinh(x.mp,v.mp,rnd_mode);
- return x;
+ mpreal x(0, prec);
+ mpfr_fac_ui(x.mp,v,rnd_mode);
+ return x;
}
-inline const mpreal tanh (const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_tanh(x.mp,v.mp,rnd_mode);
- return x;
-}
-inline const mpreal sech (const mpreal& v, mp_rnd_t rnd_mode)
+inline const mpreal lgamma (const mpreal& v, int *signp, mp_rnd_t rnd_mode)
{
- mpreal x(v);
- mpfr_sech(x.mp,v.mp,rnd_mode);
- return x;
-}
+ mpreal x(v);
+ int tsignp;
-inline const mpreal csch (const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_csch(x.mp,v.mp,rnd_mode);
- return x;
-}
+ if(signp) mpfr_lgamma(x.mp,signp,v.mp,rnd_mode);
+ else mpfr_lgamma(x.mp,&tsignp,v.mp,rnd_mode);
-inline const mpreal coth (const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_coth(x.mp,v.mp,rnd_mode);
- return x;
+ return x;
}
-inline const mpreal acosh (const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_acosh(x.mp,v.mp,rnd_mode);
- return x;
-}
-inline const mpreal asinh (const mpreal& v, mp_rnd_t rnd_mode)
+inline const mpreal besseljn (long n, const mpreal& x, mp_rnd_t r)
{
- mpreal x(v);
- mpfr_asinh(x.mp,v.mp,rnd_mode);
- return x;
+ mpreal y(0, mpfr_get_prec(x.mpfr_srcptr()));
+ mpfr_jn(y.mpfr_ptr(), n, x.mpfr_srcptr(), r);
+ return y;
}
-inline const mpreal atanh (const mpreal& v, mp_rnd_t rnd_mode)
+inline const mpreal besselyn (long n, const mpreal& x, mp_rnd_t r)
{
- mpreal x(v);
- mpfr_atanh(x.mp,v.mp,rnd_mode);
- return x;
+ mpreal y(0, mpfr_get_prec(x.mpfr_srcptr()));
+ mpfr_yn(y.mpfr_ptr(), n, x.mpfr_srcptr(), r);
+ return y;
}
-inline const mpreal hypot (const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode)
+inline const mpreal fma (const mpreal& v1, const mpreal& v2, const mpreal& v3, mp_rnd_t rnd_mode)
{
- mpreal a;
- mp_prec_t yp, xp;
-
- yp = y.get_prec();
- xp = x.get_prec();
-
- a.set_prec(yp>xp?yp:xp);
-
- mpfr_hypot(a.mp, x.mp, y.mp, rnd_mode);
-
- return a;
-}
+ mpreal a;
+ mp_prec_t p1, p2, p3;
-inline const mpreal remainder (const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode)
-{
- mpreal a;
- mp_prec_t yp, xp;
-
- yp = y.get_prec();
- xp = x.get_prec();
+ p1 = v1.get_prec();
+ p2 = v2.get_prec();
+ p3 = v3.get_prec();
- a.set_prec(yp>xp?yp:xp);
+ a.set_prec(p3>p2?(p3>p1?p3:p1):(p2>p1?p2:p1));
- mpfr_remainder(a.mp, x.mp, y.mp, rnd_mode);
-
- return a;
+ mpfr_fma(a.mp,v1.mp,v2.mp,v3.mp,rnd_mode);
+ return a;
}
-inline const mpreal fac_ui (unsigned long int v, mp_prec_t prec, mp_rnd_t rnd_mode)
+inline const mpreal fms (const mpreal& v1, const mpreal& v2, const mpreal& v3, mp_rnd_t rnd_mode)
{
- mpreal x(0,prec);
- mpfr_fac_ui(x.mp,v,rnd_mode);
- return x;
-}
+ mpreal a;
+ mp_prec_t p1, p2, p3;
-inline const mpreal log1p (const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_log1p(x.mp,v.mp,rnd_mode);
- return x;
-}
+ p1 = v1.get_prec();
+ p2 = v2.get_prec();
+ p3 = v3.get_prec();
-inline const mpreal expm1 (const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_expm1(x.mp,v.mp,rnd_mode);
- return x;
-}
+ a.set_prec(p3>p2?(p3>p1?p3:p1):(p2>p1?p2:p1));
-inline const mpreal eint (const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_eint(x.mp,v.mp,rnd_mode);
- return x;
+ mpfr_fms(a.mp,v1.mp,v2.mp,v3.mp,rnd_mode);
+ return a;
}
-inline const mpreal gamma (const mpreal& x, mp_rnd_t rnd_mode)
+inline const mpreal agm (const mpreal& v1, const mpreal& v2, mp_rnd_t rnd_mode)
{
- mpreal FunctionValue(x);
+ mpreal a;
+ mp_prec_t p1, p2;
- // x < 0: gamma(-x) = -pi/(x * gamma(x) * sin(pi*x))
+ p1 = v1.get_prec();
+ p2 = v2.get_prec();
- mpfr_gamma(FunctionValue.mp, x.mp, rnd_mode);
+ a.set_prec(p1>p2?p1:p2);
- return FunctionValue;
-}
+ mpfr_agm(a.mp, v1.mp, v2.mp, rnd_mode);
-inline const mpreal lngamma (const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_lngamma(x.mp,v.mp,rnd_mode);
- return x;
+ return a;
}
-inline const mpreal lgamma (const mpreal& v, int *signp, mp_rnd_t rnd_mode)
+inline const mpreal sum (const mpreal tab[], unsigned long int n, mp_rnd_t rnd_mode)
{
- mpreal x(v);
- int tsignp;
-
- if(signp)
- mpfr_lgamma(x.mp,signp,v.mp,rnd_mode);
- else
- mpfr_lgamma(x.mp,&tsignp,v.mp,rnd_mode);
+ mpreal x;
+ mpfr_ptr* t;
+ unsigned long int i;
- return x;
+ t = new mpfr_ptr[n];
+ for (i=0;i<n;i++) t[i] = (mpfr_ptr)tab[i].mp;
+ mpfr_sum(x.mp,t,n,rnd_mode);
+ delete[] t;
+ return x;
}
-inline const mpreal zeta (const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_zeta(x.mp,v.mp,rnd_mode);
- return x;
-}
-
-inline const mpreal erf (const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_erf(x.mp,v.mp,rnd_mode);
- return x;
-}
-
-inline const mpreal erfc (const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_erfc(x.mp,v.mp,rnd_mode);
- return x;
-}
+//////////////////////////////////////////////////////////////////////////
+// MPFR 2.4.0 Specifics
+#if (MPFR_VERSION >= MPFR_VERSION_NUM(2,4,0))
-inline const mpreal besselj0 (const mpreal& v, mp_rnd_t rnd_mode)
+inline int sinh_cosh(mpreal& s, mpreal& c, const mpreal& v, mp_rnd_t rnd_mode)
{
- mpreal x(v);
- mpfr_j0(x.mp,v.mp,rnd_mode);
- return x;
+ return mpfr_sinh_cosh(s.mp,c.mp,v.mp,rnd_mode);
}
-inline const mpreal besselj1 (const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_j1(x.mp,v.mp,rnd_mode);
- return x;
+inline const mpreal li2 (const mpreal& x, mp_rnd_t r)
+{
+ MPREAL_UNARY_MATH_FUNCTION_BODY(li2);
}
-inline const mpreal besseljn (long n, const mpreal& v, mp_rnd_t rnd_mode)
+inline const mpreal rem (const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode)
{
- mpreal x(v);
- mpfr_jn(x.mp,n,v.mp,rnd_mode);
- return x;
+ /* R = rem(X,Y) if Y != 0, returns X - n * Y where n = trunc(X/Y). */
+ return fmod(x, y, rnd_mode);
}
-inline const mpreal bessely0 (const mpreal& v, mp_rnd_t rnd_mode)
+inline const mpreal mod (const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode)
{
- mpreal x(v);
- mpfr_y0(x.mp,v.mp,rnd_mode);
- return x;
-}
+ (void)rnd_mode;
+
+ /*
-inline const mpreal bessely1 (const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_y1(x.mp,v.mp,rnd_mode);
- return x;
-}
+ m = mod(x,y) if y != 0, returns x - n*y where n = floor(x/y)
-inline const mpreal besselyn (long n, const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_yn(x.mp,n,v.mp,rnd_mode);
- return x;
-}
+ The following are true by convention:
+ - mod(x,0) is x
+ - mod(x,x) is 0
+ - mod(x,y) for x != y and y != 0 has the same sign as y.
+
+ */
-//////////////////////////////////////////////////////////////////////////
-// MPFR 2.4.0 Specifics
-#if (MPFR_VERSION >= MPFR_VERSION_NUM(2,4,0))
+ if(iszero(y)) return x;
+ if(x == y) return 0;
-inline int sinh_cosh(mpreal& s, mpreal& c, const mpreal& v, mp_rnd_t rnd_mode)
-{
- return mpfr_sinh_cosh(s.mp,c.mp,v.mp,rnd_mode);
-}
+ mpreal m = x - floor(x / y) * y;
+
+ m.setSign(sgn(y)); // make sure result has the same sign as Y
-inline const mpreal li2(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_li2(x.mp,v.mp,rnd_mode);
- return x;
+ return m;
}
inline const mpreal fmod (const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode)
{
- mpreal a;
- mp_prec_t yp, xp;
+ mpreal a;
+ mp_prec_t yp, xp;
- yp = y.get_prec();
- xp = x.get_prec();
+ yp = y.get_prec();
+ xp = x.get_prec();
- a.set_prec(yp>xp?yp:xp);
+ a.set_prec(yp>xp?yp:xp);
- mpfr_fmod(a.mp, x.mp, y.mp, rnd_mode);
+ mpfr_fmod(a.mp, x.mp, y.mp, rnd_mode);
- return a;
+ return a;
}
inline const mpreal rec_sqrt(const mpreal& v, mp_rnd_t rnd_mode)
{
- mpreal x(v);
- mpfr_rec_sqrt(x.mp,v.mp,rnd_mode);
- return x;
+ mpreal x(v);
+ mpfr_rec_sqrt(x.mp,v.mp,rnd_mode);
+ return x;
}
#endif // MPFR 2.4.0 Specifics
//////////////////////////////////////////////////////////////////////////
// MPFR 3.0.0 Specifics
#if (MPFR_VERSION >= MPFR_VERSION_NUM(3,0,0))
-
-inline const mpreal digamma(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_digamma(x.mp,v.mp,rnd_mode);
- return x;
-}
-
-inline const mpreal ai(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_ai(x.mp,v.mp,rnd_mode);
- return x;
-}
-
+inline const mpreal digamma (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(digamma); }
+inline const mpreal ai (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(ai); }
#endif // MPFR 3.0.0 Specifics
//////////////////////////////////////////////////////////////////////////
// Constants
-inline const mpreal const_log2 (mp_prec_t prec, mp_rnd_t rnd_mode)
+inline const mpreal const_log2 (mp_prec_t p, mp_rnd_t r)
{
- mpreal x;
- x.set_prec(prec);
- mpfr_const_log2(x.mp,rnd_mode);
- return x;
+ mpreal x(0, p);
+ mpfr_const_log2(x.mpfr_ptr(), r);
+ return x;
}
-inline const mpreal const_pi (mp_prec_t prec, mp_rnd_t rnd_mode)
+inline const mpreal const_pi (mp_prec_t p, mp_rnd_t r)
{
- mpreal x;
- x.set_prec(prec);
- mpfr_const_pi(x.mp,rnd_mode);
- return x;
+ mpreal x(0, p);
+ mpfr_const_pi(x.mpfr_ptr(), r);
+ return x;
}
-inline const mpreal const_euler (mp_prec_t prec, mp_rnd_t rnd_mode)
+inline const mpreal const_euler (mp_prec_t p, mp_rnd_t r)
{
- mpreal x;
- x.set_prec(prec);
- mpfr_const_euler(x.mp,rnd_mode);
- return x;
+ mpreal x(0, p);
+ mpfr_const_euler(x.mpfr_ptr(), r);
+ return x;
}
-inline const mpreal const_catalan (mp_prec_t prec, mp_rnd_t rnd_mode)
+inline const mpreal const_catalan (mp_prec_t p, mp_rnd_t r)
{
- mpreal x;
- x.set_prec(prec);
- mpfr_const_catalan(x.mp,rnd_mode);
- return x;
+ mpreal x(0, p);
+ mpfr_const_catalan(x.mpfr_ptr(), r);
+ return x;
}
-inline const mpreal const_infinity (int sign, mp_prec_t prec, mp_rnd_t rnd_mode)
+inline const mpreal const_infinity (int sign, mp_prec_t p, mp_rnd_t /*r*/)
{
- mpreal x;
- x.set_prec(prec,rnd_mode);
- mpfr_set_inf(x.mp, sign);
- return x;
+ mpreal x(0, p);
+ mpfr_set_inf(x.mpfr_ptr(), sign);
+ return x;
}
//////////////////////////////////////////////////////////////////////////
// Integer Related Functions
-inline const mpreal rint(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_rint(x.mp,v.mp,rnd_mode);
- return x;
-}
-
inline const mpreal ceil(const mpreal& v)
{
- mpreal x(v);
- mpfr_ceil(x.mp,v.mp);
- return x;
-
+ mpreal x(v);
+ mpfr_ceil(x.mp,v.mp);
+ return x;
}
inline const mpreal floor(const mpreal& v)
{
- mpreal x(v);
- mpfr_floor(x.mp,v.mp);
- return x;
+ mpreal x(v);
+ mpfr_floor(x.mp,v.mp);
+ return x;
}
inline const mpreal round(const mpreal& v)
{
- mpreal x(v);
- mpfr_round(x.mp,v.mp);
- return x;
+ mpreal x(v);
+ mpfr_round(x.mp,v.mp);
+ return x;
}
inline const mpreal trunc(const mpreal& v)
{
- mpreal x(v);
- mpfr_trunc(x.mp,v.mp);
- return x;
+ mpreal x(v);
+ mpfr_trunc(x.mp,v.mp);
+ return x;
}
-inline const mpreal rint_ceil (const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_rint_ceil(x.mp,v.mp,rnd_mode);
- return x;
-}
-
-inline const mpreal rint_floor(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_rint_floor(x.mp,v.mp,rnd_mode);
- return x;
-}
-
-inline const mpreal rint_round(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_rint_round(x.mp,v.mp,rnd_mode);
- return x;
-}
-
-inline const mpreal rint_trunc(const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_rint_trunc(x.mp,v.mp,rnd_mode);
- return x;
-}
-
-inline const mpreal frac (const mpreal& v, mp_rnd_t rnd_mode)
-{
- mpreal x(v);
- mpfr_frac(x.mp,v.mp,rnd_mode);
- return x;
-}
+inline const mpreal rint (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(rint ); }
+inline const mpreal rint_ceil (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(rint_ceil ); }
+inline const mpreal rint_floor (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(rint_floor); }
+inline const mpreal rint_round (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(rint_round); }
+inline const mpreal rint_trunc (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(rint_trunc); }
+inline const mpreal frac (const mpreal& x, mp_rnd_t r) { MPREAL_UNARY_MATH_FUNCTION_BODY(frac ); }
//////////////////////////////////////////////////////////////////////////
// Miscellaneous Functions
-inline void swap(mpreal& a, mpreal& b)
-{
- mpfr_swap(a.mp,b.mp);
-}
-
-inline const mpreal (max)(const mpreal& x, const mpreal& y)
-{
- return (x>y?x:y);
-}
-
-inline const mpreal (min)(const mpreal& x, const mpreal& y)
-{
- return (x<y?x:y);
-}
+inline void swap (mpreal& a, mpreal& b) { mpfr_swap(a.mp,b.mp); }
+inline const mpreal (max)(const mpreal& x, const mpreal& y){ return (x>y?x:y); }
+inline const mpreal (min)(const mpreal& x, const mpreal& y){ return (x<y?x:y); }
inline const mpreal fmax(const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode)
{
- mpreal a;
- mpfr_max(a.mp,x.mp,y.mp,rnd_mode);
- return a;
+ mpreal a;
+ mpfr_max(a.mp,x.mp,y.mp,rnd_mode);
+ return a;
}
inline const mpreal fmin(const mpreal& x, const mpreal& y, mp_rnd_t rnd_mode)
{
- mpreal a;
- mpfr_min(a.mp,x.mp,y.mp,rnd_mode);
- return a;
+ mpreal a;
+ mpfr_min(a.mp,x.mp,y.mp,rnd_mode);
+ return a;
}
inline const mpreal nexttoward (const mpreal& x, const mpreal& y)
{
- mpreal a(x);
- mpfr_nexttoward(a.mp,y.mp);
- return a;
+ mpreal a(x);
+ mpfr_nexttoward(a.mp,y.mp);
+ return a;
}
inline const mpreal nextabove (const mpreal& x)
{
- mpreal a(x);
- mpfr_nextabove(a.mp);
- return a;
+ mpreal a(x);
+ mpfr_nextabove(a.mp);
+ return a;
}
inline const mpreal nextbelow (const mpreal& x)
{
- mpreal a(x);
- mpfr_nextbelow(a.mp);
- return a;
+ mpreal a(x);
+ mpfr_nextbelow(a.mp);
+ return a;
}
inline const mpreal urandomb (gmp_randstate_t& state)
{
- mpreal x;
- mpfr_urandomb(x.mp,state);
- return x;
+ mpreal x;
+ mpfr_urandomb(x.mp,state);
+ return x;
}
#if (MPFR_VERSION >= MPFR_VERSION_NUM(3,0,0))
// use gmp_randinit_default() to init state, gmp_randclear() to clear
inline const mpreal urandom (gmp_randstate_t& state, mp_rnd_t rnd_mode)
{
- mpreal x;
- mpfr_urandom(x.mp,state,rnd_mode);
- return x;
+ mpreal x;
+ mpfr_urandom(x.mp,state,rnd_mode);
+ return x;
}
#endif
#if (MPFR_VERSION <= MPFR_VERSION_NUM(2,4,2))
inline const mpreal random2 (mp_size_t size, mp_exp_t exp)
{
- mpreal x;
- mpfr_random2(x.mp,size,exp);
- return x;
+ mpreal x;
+ mpfr_random2(x.mp,size,exp);
+ return x;
}
#endif
@@ -2353,22 +2484,22 @@ inline const mpreal random(unsigned int seed) {
#if (MPFR_VERSION >= MPFR_VERSION_NUM(3,0,0))
- static gmp_randstate_t state;
- static bool isFirstTime = true;
+ static gmp_randstate_t state;
+ static bool isFirstTime = true;
- if(isFirstTime)
- {
- gmp_randinit_default(state);
- gmp_randseed_ui(state,0);
- isFirstTime = false;
- }
+ if(isFirstTime)
+ {
+ gmp_randinit_default(state);
+ gmp_randseed_ui(state,0);
+ isFirstTime = false;
+ }
- if(seed != 0) gmp_randseed_ui(state,seed);
+ if(seed != 0) gmp_randseed_ui(state,seed);
- return mpfr::urandom(state);
+ return mpfr::urandom(state);
#else
- if(seed != 0) std::srand(seed);
- return mpfr::mpreal(std::rand()/(double)RAND_MAX);
+ if(seed != 0) std::srand(seed);
+ return mpfr::mpreal(std::rand()/(double)RAND_MAX);
#endif
}
@@ -2377,346 +2508,314 @@ inline const mpreal random(unsigned int seed) // Set/Get global properties
inline void mpreal::set_default_prec(mp_prec_t prec)
{
- default_prec = prec;
- mpfr_set_default_prec(prec);
-}
-
-inline mp_prec_t mpreal::get_default_prec()
-{
- return (mpfr_get_default_prec)();
-}
-
-inline void mpreal::set_default_base(int base)
-{
- default_base = base;
-}
-
-inline int mpreal::get_default_base()
-{
- return default_base;
+ mpfr_set_default_prec(prec);
}
inline void mpreal::set_default_rnd(mp_rnd_t rnd_mode)
{
- default_rnd = rnd_mode;
- mpfr_set_default_rounding_mode(rnd_mode);
-}
-
-inline mp_rnd_t mpreal::get_default_rnd()
-{
- return static_cast<mp_rnd_t>((mpfr_get_default_rounding_mode)());
-}
-
-inline void mpreal::set_double_bits(int dbits)
-{
- double_bits = dbits;
-}
-
-inline int mpreal::get_double_bits()
-{
- return double_bits;
+ mpfr_set_default_rounding_mode(rnd_mode);
}
inline bool mpreal::fits_in_bits(double x, int n)
{
- int i;
- double t;
- return IsInf(x) || (std::modf ( std::ldexp ( std::frexp ( x, &i ), n ), &t ) == 0.0);
+ int i;
+ double t;
+ return IsInf(x) || (std::modf ( std::ldexp ( std::frexp ( x, &i ), n ), &t ) == 0.0);
}
inline const mpreal pow(const mpreal& a, const mpreal& b, mp_rnd_t rnd_mode)
{
- mpreal x(a);
- mpfr_pow(x.mp,x.mp,b.mp,rnd_mode);
- return x;
+ mpreal x(a);
+ mpfr_pow(x.mp,x.mp,b.mp,rnd_mode);
+ return x;
}
inline const mpreal pow(const mpreal& a, const mpz_t b, mp_rnd_t rnd_mode)
{
- mpreal x(a);
- mpfr_pow_z(x.mp,x.mp,b,rnd_mode);
- return x;
+ mpreal x(a);
+ mpfr_pow_z(x.mp,x.mp,b,rnd_mode);
+ return x;
}
inline const mpreal pow(const mpreal& a, const unsigned long int b, mp_rnd_t rnd_mode)
{
- mpreal x(a);
- mpfr_pow_ui(x.mp,x.mp,b,rnd_mode);
- return x;
+ mpreal x(a);
+ mpfr_pow_ui(x.mp,x.mp,b,rnd_mode);
+ return x;
}
inline const mpreal pow(const mpreal& a, const unsigned int b, mp_rnd_t rnd_mode)
{
- return pow(a,static_cast<unsigned long int>(b),rnd_mode);
+ return pow(a,static_cast<unsigned long int>(b),rnd_mode);
}
inline const mpreal pow(const mpreal& a, const long int b, mp_rnd_t rnd_mode)
{
- mpreal x(a);
- mpfr_pow_si(x.mp,x.mp,b,rnd_mode);
- return x;
+ mpreal x(a);
+ mpfr_pow_si(x.mp,x.mp,b,rnd_mode);
+ return x;
}
inline const mpreal pow(const mpreal& a, const int b, mp_rnd_t rnd_mode)
{
- return pow(a,static_cast<long int>(b),rnd_mode);
+ return pow(a,static_cast<long int>(b),rnd_mode);
}
inline const mpreal pow(const mpreal& a, const long double b, mp_rnd_t rnd_mode)
{
- return pow(a,mpreal(b),rnd_mode);
+ return pow(a,mpreal(b),rnd_mode);
}
inline const mpreal pow(const mpreal& a, const double b, mp_rnd_t rnd_mode)
{
- return pow(a,mpreal(b),rnd_mode);
+ return pow(a,mpreal(b),rnd_mode);
}
inline const mpreal pow(const unsigned long int a, const mpreal& b, mp_rnd_t rnd_mode)
{
- mpreal x(a);
- mpfr_ui_pow(x.mp,a,b.mp,rnd_mode);
- return x;
+ mpreal x(a);
+ mpfr_ui_pow(x.mp,a,b.mp,rnd_mode);
+ return x;
}
inline const mpreal pow(const unsigned int a, const mpreal& b, mp_rnd_t rnd_mode)
{
- return pow(static_cast<unsigned long int>(a),b,rnd_mode);
+ return pow(static_cast<unsigned long int>(a),b,rnd_mode);
}
inline const mpreal pow(const long int a, const mpreal& b, mp_rnd_t rnd_mode)
{
- if (a>=0) return pow(static_cast<unsigned long int>(a),b,rnd_mode);
- else return pow(mpreal(a),b,rnd_mode);
+ if (a>=0) return pow(static_cast<unsigned long int>(a),b,rnd_mode);
+ else return pow(mpreal(a),b,rnd_mode);
}
inline const mpreal pow(const int a, const mpreal& b, mp_rnd_t rnd_mode)
{
- if (a>=0) return pow(static_cast<unsigned long int>(a),b,rnd_mode);
- else return pow(mpreal(a),b,rnd_mode);
+ if (a>=0) return pow(static_cast<unsigned long int>(a),b,rnd_mode);
+ else return pow(mpreal(a),b,rnd_mode);
}
inline const mpreal pow(const long double a, const mpreal& b, mp_rnd_t rnd_mode)
{
- return pow(mpreal(a),b,rnd_mode);
+ return pow(mpreal(a),b,rnd_mode);
}
inline const mpreal pow(const double a, const mpreal& b, mp_rnd_t rnd_mode)
{
- return pow(mpreal(a),b,rnd_mode);
+ return pow(mpreal(a),b,rnd_mode);
}
// pow unsigned long int
inline const mpreal pow(const unsigned long int a, const unsigned long int b, mp_rnd_t rnd_mode)
{
- mpreal x(a);
- mpfr_ui_pow_ui(x.mp,a,b,rnd_mode);
- return x;
+ mpreal x(a);
+ mpfr_ui_pow_ui(x.mp,a,b,rnd_mode);
+ return x;
}
inline const mpreal pow(const unsigned long int a, const unsigned int b, mp_rnd_t rnd_mode)
{
- return pow(a,static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
+ return pow(a,static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
}
inline const mpreal pow(const unsigned long int a, const long int b, mp_rnd_t rnd_mode)
{
- if(b>0) return pow(a,static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
- else return pow(a,mpreal(b),rnd_mode); //mpfr_ui_pow
+ if(b>0) return pow(a,static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
+ else return pow(a,mpreal(b),rnd_mode); //mpfr_ui_pow
}
inline const mpreal pow(const unsigned long int a, const int b, mp_rnd_t rnd_mode)
{
- if(b>0) return pow(a,static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
- else return pow(a,mpreal(b),rnd_mode); //mpfr_ui_pow
+ if(b>0) return pow(a,static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
+ else return pow(a,mpreal(b),rnd_mode); //mpfr_ui_pow
}
inline const mpreal pow(const unsigned long int a, const long double b, mp_rnd_t rnd_mode)
{
- return pow(a,mpreal(b),rnd_mode); //mpfr_ui_pow
+ return pow(a,mpreal(b),rnd_mode); //mpfr_ui_pow
}
inline const mpreal pow(const unsigned long int a, const double b, mp_rnd_t rnd_mode)
{
- return pow(a,mpreal(b),rnd_mode); //mpfr_ui_pow
+ return pow(a,mpreal(b),rnd_mode); //mpfr_ui_pow
}
// pow unsigned int
inline const mpreal pow(const unsigned int a, const unsigned long int b, mp_rnd_t rnd_mode)
{
- return pow(static_cast<unsigned long int>(a),b,rnd_mode); //mpfr_ui_pow_ui
+ return pow(static_cast<unsigned long int>(a),b,rnd_mode); //mpfr_ui_pow_ui
}
inline const mpreal pow(const unsigned int a, const unsigned int b, mp_rnd_t rnd_mode)
{
- return pow(static_cast<unsigned long int>(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
+ return pow(static_cast<unsigned long int>(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
}
inline const mpreal pow(const unsigned int a, const long int b, mp_rnd_t rnd_mode)
{
- if(b>0) return pow(static_cast<unsigned long int>(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
- else return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
+ if(b>0) return pow(static_cast<unsigned long int>(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
+ else return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
}
inline const mpreal pow(const unsigned int a, const int b, mp_rnd_t rnd_mode)
{
- if(b>0) return pow(static_cast<unsigned long int>(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
- else return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
+ if(b>0) return pow(static_cast<unsigned long int>(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
+ else return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
}
inline const mpreal pow(const unsigned int a, const long double b, mp_rnd_t rnd_mode)
{
- return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
+ return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
}
inline const mpreal pow(const unsigned int a, const double b, mp_rnd_t rnd_mode)
{
- return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
+ return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
}
// pow long int
inline const mpreal pow(const long int a, const unsigned long int b, mp_rnd_t rnd_mode)
{
- if (a>0) return pow(static_cast<unsigned long int>(a),b,rnd_mode); //mpfr_ui_pow_ui
- else return pow(mpreal(a),b,rnd_mode); //mpfr_pow_ui
+ if (a>0) return pow(static_cast<unsigned long int>(a),b,rnd_mode); //mpfr_ui_pow_ui
+ else return pow(mpreal(a),b,rnd_mode); //mpfr_pow_ui
}
inline const mpreal pow(const long int a, const unsigned int b, mp_rnd_t rnd_mode)
{
- if (a>0) return pow(static_cast<unsigned long int>(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
- else return pow(mpreal(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_pow_ui
+ if (a>0) return pow(static_cast<unsigned long int>(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
+ else return pow(mpreal(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_pow_ui
}
inline const mpreal pow(const long int a, const long int b, mp_rnd_t rnd_mode)
{
- if (a>0)
- {
- if(b>0) return pow(static_cast<unsigned long int>(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
- else return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
- }else{
- return pow(mpreal(a),b,rnd_mode); // mpfr_pow_si
- }
+ if (a>0)
+ {
+ if(b>0) return pow(static_cast<unsigned long int>(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
+ else return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
+ }else{
+ return pow(mpreal(a),b,rnd_mode); // mpfr_pow_si
+ }
}
inline const mpreal pow(const long int a, const int b, mp_rnd_t rnd_mode)
{
- if (a>0)
- {
- if(b>0) return pow(static_cast<unsigned long int>(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
- else return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
- }else{
- return pow(mpreal(a),static_cast<long int>(b),rnd_mode); // mpfr_pow_si
- }
+ if (a>0)
+ {
+ if(b>0) return pow(static_cast<unsigned long int>(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
+ else return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
+ }else{
+ return pow(mpreal(a),static_cast<long int>(b),rnd_mode); // mpfr_pow_si
+ }
}
inline const mpreal pow(const long int a, const long double b, mp_rnd_t rnd_mode)
{
- if (a>=0) return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
- else return pow(mpreal(a),mpreal(b),rnd_mode); //mpfr_pow
+ if (a>=0) return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
+ else return pow(mpreal(a),mpreal(b),rnd_mode); //mpfr_pow
}
inline const mpreal pow(const long int a, const double b, mp_rnd_t rnd_mode)
{
- if (a>=0) return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
- else return pow(mpreal(a),mpreal(b),rnd_mode); //mpfr_pow
+ if (a>=0) return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
+ else return pow(mpreal(a),mpreal(b),rnd_mode); //mpfr_pow
}
// pow int
inline const mpreal pow(const int a, const unsigned long int b, mp_rnd_t rnd_mode)
{
- if (a>0) return pow(static_cast<unsigned long int>(a),b,rnd_mode); //mpfr_ui_pow_ui
- else return pow(mpreal(a),b,rnd_mode); //mpfr_pow_ui
+ if (a>0) return pow(static_cast<unsigned long int>(a),b,rnd_mode); //mpfr_ui_pow_ui
+ else return pow(mpreal(a),b,rnd_mode); //mpfr_pow_ui
}
inline const mpreal pow(const int a, const unsigned int b, mp_rnd_t rnd_mode)
{
- if (a>0) return pow(static_cast<unsigned long int>(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
- else return pow(mpreal(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_pow_ui
+ if (a>0) return pow(static_cast<unsigned long int>(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
+ else return pow(mpreal(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_pow_ui
}
inline const mpreal pow(const int a, const long int b, mp_rnd_t rnd_mode)
{
- if (a>0)
- {
- if(b>0) return pow(static_cast<unsigned long int>(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
- else return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
- }else{
- return pow(mpreal(a),b,rnd_mode); // mpfr_pow_si
- }
+ if (a>0)
+ {
+ if(b>0) return pow(static_cast<unsigned long int>(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
+ else return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
+ }else{
+ return pow(mpreal(a),b,rnd_mode); // mpfr_pow_si
+ }
}
inline const mpreal pow(const int a, const int b, mp_rnd_t rnd_mode)
{
- if (a>0)
- {
- if(b>0) return pow(static_cast<unsigned long int>(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
- else return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
- }else{
- return pow(mpreal(a),static_cast<long int>(b),rnd_mode); // mpfr_pow_si
- }
+ if (a>0)
+ {
+ if(b>0) return pow(static_cast<unsigned long int>(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_ui_pow_ui
+ else return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
+ }else{
+ return pow(mpreal(a),static_cast<long int>(b),rnd_mode); // mpfr_pow_si
+ }
}
inline const mpreal pow(const int a, const long double b, mp_rnd_t rnd_mode)
{
- if (a>=0) return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
- else return pow(mpreal(a),mpreal(b),rnd_mode); //mpfr_pow
+ if (a>=0) return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
+ else return pow(mpreal(a),mpreal(b),rnd_mode); //mpfr_pow
}
inline const mpreal pow(const int a, const double b, mp_rnd_t rnd_mode)
{
- if (a>=0) return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
- else return pow(mpreal(a),mpreal(b),rnd_mode); //mpfr_pow
+ if (a>=0) return pow(static_cast<unsigned long int>(a),mpreal(b),rnd_mode); //mpfr_ui_pow
+ else return pow(mpreal(a),mpreal(b),rnd_mode); //mpfr_pow
}
// pow long double
inline const mpreal pow(const long double a, const long double b, mp_rnd_t rnd_mode)
{
- return pow(mpreal(a),mpreal(b),rnd_mode);
+ return pow(mpreal(a),mpreal(b),rnd_mode);
}
inline const mpreal pow(const long double a, const unsigned long int b, mp_rnd_t rnd_mode)
{
- return pow(mpreal(a),b,rnd_mode); //mpfr_pow_ui
+ return pow(mpreal(a),b,rnd_mode); //mpfr_pow_ui
}
inline const mpreal pow(const long double a, const unsigned int b, mp_rnd_t rnd_mode)
{
- return pow(mpreal(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_pow_ui
+ return pow(mpreal(a),static_cast<unsigned long int>(b),rnd_mode); //mpfr_pow_ui
}
inline const mpreal pow(const long double a, const long int b, mp_rnd_t rnd_mode)
{
- return pow(mpreal(a),b,rnd_mode); // mpfr_pow_si
+ return pow(mpreal(a),b,rnd_mode); // mpfr_pow_si
}
inline const mpreal pow(const long double a, const int b, mp_rnd_t rnd_mode)
{
- return pow(mpreal(a),static_cast<long int>(b),rnd_mode); // mpfr_pow_si
+ return pow(mpreal(a),static_cast<long int>(b),rnd_mode); // mpfr_pow_si
}
inline const mpreal pow(const double a, const double b, mp_rnd_t rnd_mode)
{
- return pow(mpreal(a),mpreal(b),rnd_mode);
+ return pow(mpreal(a),mpreal(b),rnd_mode);
}
inline const mpreal pow(const double a, const unsigned long int b, mp_rnd_t rnd_mode)
{
- return pow(mpreal(a),b,rnd_mode); // mpfr_pow_ui
+ return pow(mpreal(a),b,rnd_mode); // mpfr_pow_ui
}
inline const mpreal pow(const double a, const unsigned int b, mp_rnd_t rnd_mode)
{
- return pow(mpreal(a),static_cast<unsigned long int>(b),rnd_mode); // mpfr_pow_ui
+ return pow(mpreal(a),static_cast<unsigned long int>(b),rnd_mode); // mpfr_pow_ui
}
inline const mpreal pow(const double a, const long int b, mp_rnd_t rnd_mode)
{
- return pow(mpreal(a),b,rnd_mode); // mpfr_pow_si
+ return pow(mpreal(a),b,rnd_mode); // mpfr_pow_si
}
inline const mpreal pow(const double a, const int b, mp_rnd_t rnd_mode)
{
- return pow(mpreal(a),static_cast<long int>(b),rnd_mode); // mpfr_pow_si
+ return pow(mpreal(a),static_cast<long int>(b),rnd_mode); // mpfr_pow_si
}
} // End of mpfr namespace
@@ -2725,11 +2824,99 @@ inline const mpreal pow(const double a, const int b, mp_rnd_t rnd_mode) // Non-throwing swap C++ idiom: http://en.wikibooks.org/wiki/More_C%2B%2B_Idioms/Non-throwing_swap
namespace std
{
- template <>
- inline void swap(mpfr::mpreal& x, mpfr::mpreal& y)
- {
- return mpfr::swap(x, y);
- }
+ // only allowed to extend namespace std with specializations
+ template <>
+ inline void swap(mpfr::mpreal& x, mpfr::mpreal& y)
+ {
+ return mpfr::swap(x, y);
+ }
+
+ template<>
+ class numeric_limits<mpfr::mpreal>
+ {
+ public:
+ static const bool is_specialized = true;
+ static const bool is_signed = true;
+ static const bool is_integer = false;
+ static const bool is_exact = false;
+ static const int radix = 2;
+
+ static const bool has_infinity = true;
+ static const bool has_quiet_NaN = true;
+ static const bool has_signaling_NaN = true;
+
+ static const bool is_iec559 = true; // = IEEE 754
+ static const bool is_bounded = true;
+ static const bool is_modulo = false;
+ static const bool traps = true;
+ static const bool tinyness_before = true;
+
+ static const float_denorm_style has_denorm = denorm_absent;
+
+ inline static float_round_style round_style()
+ {
+ mp_rnd_t r = mpfr::mpreal::get_default_rnd();
+
+ switch (r)
+ {
+ case MPFR_RNDN: return round_to_nearest;
+ case MPFR_RNDZ: return round_toward_zero;
+ case MPFR_RNDU: return round_toward_infinity;
+ case MPFR_RNDD: return round_toward_neg_infinity;
+ default: return round_indeterminate;
+ }
+ }
+
+ inline static mpfr::mpreal (min) (mp_prec_t precision = mpfr::mpreal::get_default_prec()) { return mpfr::minval(precision); }
+ inline static mpfr::mpreal (max) (mp_prec_t precision = mpfr::mpreal::get_default_prec()) { return mpfr::maxval(precision); }
+ inline static mpfr::mpreal lowest (mp_prec_t precision = mpfr::mpreal::get_default_prec()) { return -mpfr::maxval(precision); }
+
+ // Returns smallest eps such that 1 + eps != 1 (classic machine epsilon)
+ inline static mpfr::mpreal epsilon(mp_prec_t precision = mpfr::mpreal::get_default_prec()) { return mpfr::machine_epsilon(precision); }
+
+ // Returns smallest eps such that x + eps != x (relative machine epsilon)
+ inline static mpfr::mpreal epsilon(const mpfr::mpreal& x) { return mpfr::machine_epsilon(x); }
+
+ inline static mpfr::mpreal round_error(mp_prec_t precision = mpfr::mpreal::get_default_prec())
+ {
+ mp_rnd_t r = mpfr::mpreal::get_default_rnd();
+
+ if(r == MPFR_RNDN) return mpfr::mpreal(0.5, precision);
+ else return mpfr::mpreal(1.0, precision);
+ }
+
+ inline static const mpfr::mpreal infinity() { return mpfr::const_infinity(); }
+ inline static const mpfr::mpreal quiet_NaN() { return mpfr::mpreal().setNan(); }
+ inline static const mpfr::mpreal signaling_NaN() { return mpfr::mpreal().setNan(); }
+ inline static const mpfr::mpreal denorm_min() { return (min)(); }
+
+ // Please note, exponent range is not fixed in MPFR
+ static const int min_exponent = MPFR_EMIN_DEFAULT;
+ static const int max_exponent = MPFR_EMAX_DEFAULT;
+ MPREAL_PERMISSIVE_EXPR static const int min_exponent10 = (int) (MPFR_EMIN_DEFAULT * 0.3010299956639811);
+ MPREAL_PERMISSIVE_EXPR static const int max_exponent10 = (int) (MPFR_EMAX_DEFAULT * 0.3010299956639811);
+
+ // Should be constant according to standard, but 'digits' depends on precision in MPFR
+
+ inline static int digits() { return mpfr::mpreal::get_default_prec(); }
+ inline static int digits(const mpfr::mpreal& x) { return x.getPrecision(); }
+
+ inline static int digits10(mp_prec_t precision = mpfr::mpreal::get_default_prec())
+ {
+ return mpfr::bits2digits(precision);
+ }
+
+ inline static int digits10(const mpfr::mpreal& x)
+ {
+ return mpfr::bits2digits(x.getPrecision());
+ }
+
+ inline static int max_digits10(mp_prec_t precision = mpfr::mpreal::get_default_prec())
+ {
+ return digits10(precision);
+ }
+ };
+
}
#endif /* __MPREAL_H__ */
diff --git a/unsupported/test/mpreal_support.cpp b/unsupported/test/mpreal_support.cpp index 551af9db8..bc00382be 100644 --- a/unsupported/test/mpreal_support.cpp +++ b/unsupported/test/mpreal_support.cpp @@ -5,7 +5,6 @@ #include <sstream> using namespace mpfr; -using namespace std; using namespace Eigen; void test_mpreal_support() @@ -46,8 +45,7 @@ void test_mpreal_support() // symmetric eigenvalues SelfAdjointEigenSolver<MatrixXmp> eig(S); VERIFY_IS_EQUAL(eig.info(), Success); - VERIFY_IS_APPROX((S.selfadjointView<Lower>() * eig.eigenvectors()), - eig.eigenvectors() * eig.eigenvalues().asDiagonal()); + VERIFY( (S.selfadjointView<Lower>() * eig.eigenvectors()).isApprox(eig.eigenvectors() * eig.eigenvalues().asDiagonal(), NumTraits<mpreal>::dummy_precision()*1e3) ); } { @@ -57,8 +55,3 @@ void test_mpreal_support() stream << A; } } - -extern "C" { -#include "mpreal/dlmalloc.c" -} -#include "mpreal/mpreal.cpp" diff --git a/unsupported/test/polynomialsolver.cpp b/unsupported/test/polynomialsolver.cpp index fefeaff01..13f92169e 100644 --- a/unsupported/test/polynomialsolver.cpp +++ b/unsupported/test/polynomialsolver.cpp @@ -92,6 +92,7 @@ void evalSolver( const POLYNOMIAL& pols ) template< int Deg, typename POLYNOMIAL, typename ROOTS, typename REAL_ROOTS > void evalSolverSugarFunction( const POLYNOMIAL& pols, const ROOTS& roots, const REAL_ROOTS& real_roots ) { + using std::sqrt; typedef typename POLYNOMIAL::Scalar Scalar; typedef PolynomialSolver<Scalar, Deg > PolynomialSolverType; @@ -105,17 +106,14 @@ void evalSolverSugarFunction( const POLYNOMIAL& pols, const ROOTS& roots, const typedef typename POLYNOMIAL::Scalar Scalar; typedef typename REAL_ROOTS::Scalar Real; - typedef PolynomialSolver<Scalar, Deg > PolynomialSolverType; - typedef typename PolynomialSolverType::RootsType RootsType; - typedef Matrix<Scalar,Deg,1> EvalRootsType; //Test realRoots std::vector< Real > calc_realRoots; psolve.realRoots( calc_realRoots ); VERIFY( calc_realRoots.size() == (size_t)real_roots.size() ); - const Scalar psPrec = internal::sqrt( test_precision<Scalar>() ); + const Scalar psPrec = sqrt( test_precision<Scalar>() ); for( size_t i=0; i<calc_realRoots.size(); ++i ) { @@ -130,24 +128,24 @@ void evalSolverSugarFunction( const POLYNOMIAL& pols, const ROOTS& roots, const //Test greatestRoot VERIFY( internal::isApprox( roots.array().abs().maxCoeff(), - internal::abs( psolve.greatestRoot() ), psPrec ) ); + abs( psolve.greatestRoot() ), psPrec ) ); //Test smallestRoot VERIFY( internal::isApprox( roots.array().abs().minCoeff(), - internal::abs( psolve.smallestRoot() ), psPrec ) ); + abs( psolve.smallestRoot() ), psPrec ) ); bool hasRealRoot; //Test absGreatestRealRoot Real r = psolve.absGreatestRealRoot( hasRealRoot ); VERIFY( hasRealRoot == (real_roots.size() > 0 ) ); if( hasRealRoot ){ - VERIFY( internal::isApprox( real_roots.array().abs().maxCoeff(), internal::abs(r), psPrec ) ); } + VERIFY( internal::isApprox( real_roots.array().abs().maxCoeff(), abs(r), psPrec ) ); } //Test absSmallestRealRoot r = psolve.absSmallestRealRoot( hasRealRoot ); VERIFY( hasRealRoot == (real_roots.size() > 0 ) ); if( hasRealRoot ){ - VERIFY( internal::isApprox( real_roots.array().abs().minCoeff(), internal::abs( r ), psPrec ) ); } + VERIFY( internal::isApprox( real_roots.array().abs().minCoeff(), abs( r ), psPrec ) ); } //Test greatestRealRoot r = psolve.greatestRealRoot( hasRealRoot ); diff --git a/unsupported/test/sparse_extra.cpp b/unsupported/test/sparse_extra.cpp index 5dc333424..1ee791b0f 100644 --- a/unsupported/test/sparse_extra.cpp +++ b/unsupported/test/sparse_extra.cpp @@ -17,7 +17,6 @@ template<typename SetterType,typename DenseType, typename Scalar, int Options> bool test_random_setter(SparseMatrix<Scalar,Options>& sm, const DenseType& ref, const std::vector<Vector2i>& nonzeroCoords) { - typedef SparseMatrix<Scalar,Options> SparseType; { sm.setZero(); SetterType w(sm); diff --git a/unsupported/test/splines.cpp b/unsupported/test/splines.cpp index 1043453dc..a7eb3e0c4 100644 --- a/unsupported/test/splines.cpp +++ b/unsupported/test/splines.cpp @@ -11,6 +11,8 @@ #include <unsupported/Eigen/Splines> +namespace Eigen { + // lets do some explicit instantiations and thus // force the compilation of all spline functions... template class Spline<double, 2, Dynamic>; @@ -29,6 +31,8 @@ template class Spline<float, 3, 3>; template class Spline<float, 3, 4>; template class Spline<float, 3, 5>; +} + Spline<double, 2, Dynamic> closed_spline2d() { RowVectorXd knots(12); diff --git a/unsupported/test/svd_common.h b/unsupported/test/svd_common.h new file mode 100644 index 000000000..b40c23a2b --- /dev/null +++ b/unsupported/test/svd_common.h @@ -0,0 +1,261 @@ +// This file is part of Eigen, a lightweight C++ template library +// for linear algebra. +// +// Copyright (C) 2008 Gael Guennebaud <gael.guennebaud@inria.fr> +// Copyright (C) 2009 Benoit Jacob <jacob.benoit.1@gmail.com> +// +// Copyright (C) 2013 Gauthier Brun <brun.gauthier@gmail.com> +// Copyright (C) 2013 Nicolas Carre <nicolas.carre@ensimag.fr> +// Copyright (C) 2013 Jean Ceccato <jean.ceccato@ensimag.fr> +// Copyright (C) 2013 Pierre Zoppitelli <pierre.zoppitelli@ensimag.fr> +// +// This Source Code Form is subject to the terms of the Mozilla +// Public License v. 2.0. If a copy of the MPL was not distributed +// with this file, You can obtain one at http://mozilla.org/MPL/2.0/. + +// discard stack allocation as that too bypasses malloc +#define EIGEN_STACK_ALLOCATION_LIMIT 0 +#define EIGEN_RUNTIME_NO_MALLOC + +#include "main.h" +#include <unsupported/Eigen/SVD> +#include <Eigen/LU> + + +// check if "svd" is the good image of "m" +template<typename MatrixType, typename SVD> +void svd_check_full(const MatrixType& m, const SVD& svd) +{ + typedef typename MatrixType::Index Index; + Index rows = m.rows(); + Index cols = m.cols(); + enum { + RowsAtCompileTime = MatrixType::RowsAtCompileTime, + ColsAtCompileTime = MatrixType::ColsAtCompileTime + }; + + typedef typename MatrixType::Scalar Scalar; + typedef Matrix<Scalar, RowsAtCompileTime, RowsAtCompileTime> MatrixUType; + typedef Matrix<Scalar, ColsAtCompileTime, ColsAtCompileTime> MatrixVType; + + + MatrixType sigma = MatrixType::Zero(rows, cols); + sigma.diagonal() = svd.singularValues().template cast<Scalar>(); + MatrixUType u = svd.matrixU(); + MatrixVType v = svd.matrixV(); + VERIFY_IS_APPROX(m, u * sigma * v.adjoint()); + VERIFY_IS_UNITARY(u); + VERIFY_IS_UNITARY(v); +} // end svd_check_full + + + +// Compare to a reference value +template<typename MatrixType, typename SVD> +void svd_compare_to_full(const MatrixType& m, + unsigned int computationOptions, + const SVD& referenceSvd) +{ + typedef typename MatrixType::Index Index; + Index rows = m.rows(); + Index cols = m.cols(); + Index diagSize = (std::min)(rows, cols); + + SVD svd(m, computationOptions); + + VERIFY_IS_APPROX(svd.singularValues(), referenceSvd.singularValues()); + if(computationOptions & ComputeFullU) + VERIFY_IS_APPROX(svd.matrixU(), referenceSvd.matrixU()); + if(computationOptions & ComputeThinU) + VERIFY_IS_APPROX(svd.matrixU(), referenceSvd.matrixU().leftCols(diagSize)); + if(computationOptions & ComputeFullV) + VERIFY_IS_APPROX(svd.matrixV(), referenceSvd.matrixV()); + if(computationOptions & ComputeThinV) + VERIFY_IS_APPROX(svd.matrixV(), referenceSvd.matrixV().leftCols(diagSize)); +} // end svd_compare_to_full + + + +template<typename MatrixType, typename SVD> +void svd_solve(const MatrixType& m, unsigned int computationOptions) +{ + typedef typename MatrixType::Scalar Scalar; + typedef typename MatrixType::Index Index; + Index rows = m.rows(); + Index cols = m.cols(); + + enum { + RowsAtCompileTime = MatrixType::RowsAtCompileTime, + ColsAtCompileTime = MatrixType::ColsAtCompileTime + }; + + typedef Matrix<Scalar, RowsAtCompileTime, Dynamic> RhsType; + typedef Matrix<Scalar, ColsAtCompileTime, Dynamic> SolutionType; + + RhsType rhs = RhsType::Random(rows, internal::random<Index>(1, cols)); + SVD svd(m, computationOptions); + SolutionType x = svd.solve(rhs); + // evaluate normal equation which works also for least-squares solutions + VERIFY_IS_APPROX(m.adjoint()*m*x,m.adjoint()*rhs); +} // end svd_solve + + +// test computations options +// 2 functions because Jacobisvd can return before the second function +template<typename MatrixType, typename SVD> +void svd_test_computation_options_1(const MatrixType& m, const SVD& fullSvd) +{ + svd_check_full< MatrixType, SVD >(m, fullSvd); + svd_solve< MatrixType, SVD >(m, ComputeFullU | ComputeFullV); +} + + +template<typename MatrixType, typename SVD> +void svd_test_computation_options_2(const MatrixType& m, const SVD& fullSvd) +{ + svd_compare_to_full< MatrixType, SVD >(m, ComputeFullU, fullSvd); + svd_compare_to_full< MatrixType, SVD >(m, ComputeFullV, fullSvd); + svd_compare_to_full< MatrixType, SVD >(m, 0, fullSvd); + + if (MatrixType::ColsAtCompileTime == Dynamic) { + // thin U/V are only available with dynamic number of columns + + svd_compare_to_full< MatrixType, SVD >(m, ComputeFullU|ComputeThinV, fullSvd); + svd_compare_to_full< MatrixType, SVD >(m, ComputeThinV, fullSvd); + svd_compare_to_full< MatrixType, SVD >(m, ComputeThinU|ComputeFullV, fullSvd); + svd_compare_to_full< MatrixType, SVD >(m, ComputeThinU , fullSvd); + svd_compare_to_full< MatrixType, SVD >(m, ComputeThinU|ComputeThinV, fullSvd); + svd_solve<MatrixType, SVD>(m, ComputeFullU | ComputeThinV); + svd_solve<MatrixType, SVD>(m, ComputeThinU | ComputeFullV); + svd_solve<MatrixType, SVD>(m, ComputeThinU | ComputeThinV); + + typedef typename MatrixType::Index Index; + Index diagSize = (std::min)(m.rows(), m.cols()); + SVD svd(m, ComputeThinU | ComputeThinV); + VERIFY_IS_APPROX(m, svd.matrixU().leftCols(diagSize) * svd.singularValues().asDiagonal() * svd.matrixV().leftCols(diagSize).adjoint()); + } +} + +template<typename MatrixType, typename SVD> +void svd_verify_assert(const MatrixType& m) +{ + typedef typename MatrixType::Scalar Scalar; + typedef typename MatrixType::Index Index; + Index rows = m.rows(); + Index cols = m.cols(); + + enum { + RowsAtCompileTime = MatrixType::RowsAtCompileTime, + ColsAtCompileTime = MatrixType::ColsAtCompileTime + }; + + typedef Matrix<Scalar, RowsAtCompileTime, 1> RhsType; + RhsType rhs(rows); + SVD svd; + VERIFY_RAISES_ASSERT(svd.matrixU()) + VERIFY_RAISES_ASSERT(svd.singularValues()) + VERIFY_RAISES_ASSERT(svd.matrixV()) + VERIFY_RAISES_ASSERT(svd.solve(rhs)) + MatrixType a = MatrixType::Zero(rows, cols); + a.setZero(); + svd.compute(a, 0); + VERIFY_RAISES_ASSERT(svd.matrixU()) + VERIFY_RAISES_ASSERT(svd.matrixV()) + svd.singularValues(); + VERIFY_RAISES_ASSERT(svd.solve(rhs)) + + if (ColsAtCompileTime == Dynamic) + { + svd.compute(a, ComputeThinU); + svd.matrixU(); + VERIFY_RAISES_ASSERT(svd.matrixV()) + VERIFY_RAISES_ASSERT(svd.solve(rhs)) + svd.compute(a, ComputeThinV); + svd.matrixV(); + VERIFY_RAISES_ASSERT(svd.matrixU()) + VERIFY_RAISES_ASSERT(svd.solve(rhs)) + } + else + { + VERIFY_RAISES_ASSERT(svd.compute(a, ComputeThinU)) + VERIFY_RAISES_ASSERT(svd.compute(a, ComputeThinV)) + } +} + +// work around stupid msvc error when constructing at compile time an expression that involves +// a division by zero, even if the numeric type has floating point +template<typename Scalar> +EIGEN_DONT_INLINE Scalar zero() { return Scalar(0); } + +// workaround aggressive optimization in ICC +template<typename T> EIGEN_DONT_INLINE T sub(T a, T b) { return a - b; } + + +template<typename MatrixType, typename SVD> +void svd_inf_nan() +{ + // all this function does is verify we don't iterate infinitely on nan/inf values + + SVD svd; + typedef typename MatrixType::Scalar Scalar; + Scalar some_inf = Scalar(1) / zero<Scalar>(); + VERIFY(sub(some_inf, some_inf) != sub(some_inf, some_inf)); + svd.compute(MatrixType::Constant(10,10,some_inf), ComputeFullU | ComputeFullV); + + Scalar some_nan = zero<Scalar> () / zero<Scalar> (); + VERIFY(some_nan != some_nan); + svd.compute(MatrixType::Constant(10,10,some_nan), ComputeFullU | ComputeFullV); + + MatrixType m = MatrixType::Zero(10,10); + m(internal::random<int>(0,9), internal::random<int>(0,9)) = some_inf; + svd.compute(m, ComputeFullU | ComputeFullV); + + m = MatrixType::Zero(10,10); + m(internal::random<int>(0,9), internal::random<int>(0,9)) = some_nan; + svd.compute(m, ComputeFullU | ComputeFullV); +} + + +template<typename SVD> +void svd_preallocate() +{ + Vector3f v(3.f, 2.f, 1.f); + MatrixXf m = v.asDiagonal(); + + internal::set_is_malloc_allowed(false); + VERIFY_RAISES_ASSERT(VectorXf v(10);) + SVD svd; + internal::set_is_malloc_allowed(true); + svd.compute(m); + VERIFY_IS_APPROX(svd.singularValues(), v); + + SVD svd2(3,3); + internal::set_is_malloc_allowed(false); + svd2.compute(m); + internal::set_is_malloc_allowed(true); + VERIFY_IS_APPROX(svd2.singularValues(), v); + VERIFY_RAISES_ASSERT(svd2.matrixU()); + VERIFY_RAISES_ASSERT(svd2.matrixV()); + svd2.compute(m, ComputeFullU | ComputeFullV); + VERIFY_IS_APPROX(svd2.matrixU(), Matrix3f::Identity()); + VERIFY_IS_APPROX(svd2.matrixV(), Matrix3f::Identity()); + internal::set_is_malloc_allowed(false); + svd2.compute(m); + internal::set_is_malloc_allowed(true); + + SVD svd3(3,3,ComputeFullU|ComputeFullV); + internal::set_is_malloc_allowed(false); + svd2.compute(m); + internal::set_is_malloc_allowed(true); + VERIFY_IS_APPROX(svd2.singularValues(), v); + VERIFY_IS_APPROX(svd2.matrixU(), Matrix3f::Identity()); + VERIFY_IS_APPROX(svd2.matrixV(), Matrix3f::Identity()); + internal::set_is_malloc_allowed(false); + svd2.compute(m, ComputeFullU|ComputeFullV); + internal::set_is_malloc_allowed(true); +} + + + + + |