/usr/include/m4ri/strassen.h is in libm4ri-dev 20130416-6.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 | /**
* \file strassen.h
*
* \brief Matrix operations using Strassen's formulas including
* Winograd's improvements.
*
* \author Gregory Bard <bard@fordham.edu>
* \author Martin Albrecht <M.R.Albrecht@rhul.ac.uk>
*/
#ifndef M4RI_STRASSEN_H
#define M4RI_STRASSEN_H
/*******************************************************************
*
* M4RI: Linear Algebra over GF(2)
*
* Copyright (C) 2008 Martin Albrecht <M.R.Albrecht@rhul.ac.uk>
* Copyright (C) 2008 Clement Pernet <pernet@math.washington.edu>
*
* Distributed under the terms of the GNU General Public License (GPL)
* version 2 or higher.
*
* This code 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
* General Public License for more details.
*
* The full text of the GPL is available at:
*
* http://www.gnu.org/licenses/
*
********************************************************************/
#include <math.h>
#include <m4ri/mzd.h>
#include <m4ri/brilliantrussian.h>
/**
* \brief Matrix multiplication via the Strassen-Winograd matrix
* multiplication algorithm, i.e. compute C = AB.
*
* This is the wrapper function including bounds checks. See
* _mzd_mul_even for implementation details.
*
* \param C Preallocated product matrix, may be NULL for automatic creation.
* \param A Input matrix A
* \param B Input matrix B
* \param cutoff Minimal dimension for Strassen recursion.
*/
mzd_t *mzd_mul(mzd_t *C, mzd_t const *A, mzd_t const *B, int cutoff);
/**
* \brief Matrix multiplication and in-place addition via the
* Strassen-Winograd matrix multiplication algorithm, i.e. compute
* C = C+ AB.
*
* This is the wrapper function including bounds checks. See
* _mzd_addmul_even for implementation details.
*
* \param C product matrix
* \param A Input matrix A
* \param B Input matrix B
* \param cutoff Minimal dimension for Strassen recursion.
*/
mzd_t *mzd_addmul(mzd_t *C, mzd_t const *A, mzd_t const *B, int cutoff);
/**
* \brief Matrix multiplication via the Strassen-Winograd matrix
* multiplication algorithm, i.e. compute C = AB.
*
* This is the actual implementation. Any matrix where either the
* number of rows or the number of columns is smaller than cutoff is
* processed using the M4RM algorithm.
*
* \param C Preallocated product matrix, may be NULL for automatic creation.
* \param A Input matrix A
* \param B Input matrix B
* \param cutoff Minimal dimension for Strassen recursion.
*
* \note This implementation is heavily inspired by the function
* strassen_window_multiply_c in Sage 3.0; For reference see
* http://www.sagemath.org
*/
mzd_t *_mzd_mul_even(mzd_t *C, mzd_t const *A, mzd_t const *B, int cutoff);
/**
* \brief Matrix multiplication and in-place addition via the
* Strassen-Winograd matrix multiplication algorithm, i.e. compute
* C = C+ AB.
*
* This is the actual implementation. Any matrix where either the
* number of rows or the number of columns is smaller than cutoff is
* processed using the M4RM algorithm.
*
* \param C Preallocated product matrix, may be NULL for automatic creation.
* \param A Input matrix A
* \param B Input matrix B
* \param cutoff Minimal dimension for Strassen recursion.
*
* \note This implementation is heavily inspired by the function
* strassen_window_multiply_c in Sage 3.0; For reference see
* http://www.sagemath.org
*/
mzd_t *_mzd_addmul_even(mzd_t *C, mzd_t const *A, mzd_t const *B, int cutoff);
/**
* \brief Matrix multiplication and in-place addition via the
* Strassen-Winograd matrix multiplication algorithm, i.e. compute
* C = C + AB.
*
* The matrices A and B are respectively m x k and k x n, and can be not
* aligned on the m4ri_radix grid.
*
* \param C Preallocated product matrix, may be NULL for automatic creation.
* \param A Input matrix A
* \param B Input matrix B
* \param cutoff Minimal dimension for Strassen recursion.
*
*/
mzd_t *_mzd_addmul (mzd_t *C, mzd_t const *A, mzd_t const *B, int cutoff);
/**
* C = A*B + C for matrices with offsets != 0
*
* This is scratch code.
*
* \internal
*/
mzd_t *_mzd_addmul_weird_weird (mzd_t *C, mzd_t const *A, mzd_t const *B);
/**
* C = A*B + C for A with offset == 0 and B with offset != 0.
*
* This is scratch code.
*
* \internal
*/
mzd_t *_mzd_addmul_weird_even (mzd_t *C, mzd_t const *A, mzd_t const *B, int cutoff);
/**
* C = A*B + C for A with offset != 0 and B with offset == 0.
*
* This is scratch code.
*
* \internal
*/
mzd_t *_mzd_addmul_even_weird (mzd_t *C, mzd_t const *A, mzd_t const *B, int cutoff);
/**
* The default cutoff for Strassen-Winograd multiplication. It should
* hold hold that 2 * (n^2)/8 fits into the L2 cache.
*/
#ifndef __M4RI_STRASSEN_MUL_CUTOFF
#define __M4RI_STRASSEN_MUL_CUTOFF MIN(((int)sqrt((double)(4 * __M4RI_CPU_L3_CACHE))), 4096)
#endif
#endif // M4RI_STRASSEN_H
|