1 // This file is part of Eigen, a lightweight C++ template library 2 // for linear algebra. 3 // 4 // Copyright (C) 2010 Gael Guennebaud <gael.guennebaud (at) inria.fr> 5 // 6 // This Source Code Form is subject to the terms of the Mozilla 7 // Public License v. 2.0. If a copy of the MPL was not distributed 8 // with this file, You can obtain one at http://mozilla.org/MPL/2.0/. 9 10 #ifndef EIGEN_PARALLELIZER_H 11 #define EIGEN_PARALLELIZER_H 12 13 namespace Eigen { 14 15 namespace internal { 16 17 /** \internal */ 18 inline void manage_multi_threading(Action action, int* v) 19 { 20 static EIGEN_UNUSED int m_maxThreads = -1; 21 22 if(action==SetAction) 23 { 24 eigen_internal_assert(v!=0); 25 m_maxThreads = *v; 26 } 27 else if(action==GetAction) 28 { 29 eigen_internal_assert(v!=0); 30 #ifdef EIGEN_HAS_OPENMP 31 if(m_maxThreads>0) 32 *v = m_maxThreads; 33 else 34 *v = omp_get_max_threads(); 35 #else 36 *v = 1; 37 #endif 38 } 39 else 40 { 41 eigen_internal_assert(false); 42 } 43 } 44 45 } 46 47 /** Must be call first when calling Eigen from multiple threads */ 48 inline void initParallel() 49 { 50 int nbt; 51 internal::manage_multi_threading(GetAction, &nbt); 52 std::ptrdiff_t l1, l2; 53 internal::manage_caching_sizes(GetAction, &l1, &l2); 54 } 55 56 /** \returns the max number of threads reserved for Eigen 57 * \sa setNbThreads */ 58 inline int nbThreads() 59 { 60 int ret; 61 internal::manage_multi_threading(GetAction, &ret); 62 return ret; 63 } 64 65 /** Sets the max number of threads reserved for Eigen 66 * \sa nbThreads */ 67 inline void setNbThreads(int v) 68 { 69 internal::manage_multi_threading(SetAction, &v); 70 } 71 72 namespace internal { 73 74 template<typename Index> struct GemmParallelInfo 75 { 76 GemmParallelInfo() : sync(-1), users(0), rhs_start(0), rhs_length(0) {} 77 78 int volatile sync; 79 int volatile users; 80 81 Index rhs_start; 82 Index rhs_length; 83 }; 84 85 template<bool Condition, typename Functor, typename Index> 86 void parallelize_gemm(const Functor& func, Index rows, Index cols, bool transpose) 87 { 88 // TODO when EIGEN_USE_BLAS is defined, 89 // we should still enable OMP for other scalar types 90 #if !(defined (EIGEN_HAS_OPENMP)) || defined (EIGEN_USE_BLAS) 91 // FIXME the transpose variable is only needed to properly split 92 // the matrix product when multithreading is enabled. This is a temporary 93 // fix to support row-major destination matrices. This whole 94 // parallelizer mechanism has to be redisigned anyway. 95 EIGEN_UNUSED_VARIABLE(transpose); 96 func(0,rows, 0,cols); 97 #else 98 99 // Dynamically check whether we should enable or disable OpenMP. 100 // The conditions are: 101 // - the max number of threads we can create is greater than 1 102 // - we are not already in a parallel code 103 // - the sizes are large enough 104 105 // 1- are we already in a parallel session? 106 // FIXME omp_get_num_threads()>1 only works for openmp, what if the user does not use openmp? 107 if((!Condition) || (omp_get_num_threads()>1)) 108 return func(0,rows, 0,cols); 109 110 Index size = transpose ? cols : rows; 111 112 // 2- compute the maximal number of threads from the size of the product: 113 // FIXME this has to be fine tuned 114 Index max_threads = std::max<Index>(1,size / 32); 115 116 // 3 - compute the number of threads we are going to use 117 Index threads = std::min<Index>(nbThreads(), max_threads); 118 119 if(threads==1) 120 return func(0,rows, 0,cols); 121 122 Eigen::initParallel(); 123 func.initParallelSession(); 124 125 if(transpose) 126 std::swap(rows,cols); 127 128 Index blockCols = (cols / threads) & ~Index(0x3); 129 Index blockRows = (rows / threads) & ~Index(0x7); 130 131 GemmParallelInfo<Index>* info = new GemmParallelInfo<Index>[threads]; 132 133 #pragma omp parallel for schedule(static,1) num_threads(threads) 134 for(Index i=0; i<threads; ++i) 135 { 136 Index r0 = i*blockRows; 137 Index actualBlockRows = (i+1==threads) ? rows-r0 : blockRows; 138 139 Index c0 = i*blockCols; 140 Index actualBlockCols = (i+1==threads) ? cols-c0 : blockCols; 141 142 info[i].rhs_start = c0; 143 info[i].rhs_length = actualBlockCols; 144 145 if(transpose) 146 func(0, cols, r0, actualBlockRows, info); 147 else 148 func(r0, actualBlockRows, 0,cols, info); 149 } 150 151 delete[] info; 152 #endif 153 } 154 155 } // end namespace internal 156 157 } // end namespace Eigen 158 159 #endif // EIGEN_PARALLELIZER_H 160
