1 // -*- C++ -*- 2 3 // Copyright (C) 2007, 2008, 2009 Free Software Foundation, Inc. 4 // 5 // This file is part of the GNU ISO C++ Library. This library is free 6 // software; you can redistribute it and/or modify it under the terms 7 // of the GNU General Public License as published by the Free Software 8 // Foundation; either version 3, or (at your option) any later 9 // version. 10 11 // This library is distributed in the hope that it will be useful, but 12 // WITHOUT ANY WARRANTY; without even the implied warranty of 13 // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU 14 // General Public License for more details. 15 16 // Under Section 7 of GPL version 3, you are granted additional 17 // permissions described in the GCC Runtime Library Exception, version 18 // 3.1, as published by the Free Software Foundation. 19 20 // You should have received a copy of the GNU General Public License and 21 // a copy of the GCC Runtime Library Exception along with this program; 22 // see the files COPYING3 and COPYING.RUNTIME respectively. If not, see 23 // <http://www.gnu.org/licenses/>. 24 25 /** @file parallel/partial_sum.h 26 * @brief Parallel implementation of std::partial_sum(), i.e. prefix 27 * sums. 28 * This file is a GNU parallel extension to the Standard C++ Library. 29 */ 30 31 // Written by Johannes Singler. 32 33 #ifndef _GLIBCXX_PARALLEL_PARTIAL_SUM_H 34 #define _GLIBCXX_PARALLEL_PARTIAL_SUM_H 1 35 36 #include <omp.h> 37 #include <new> 38 #include <bits/stl_algobase.h> 39 #include <parallel/parallel.h> 40 #include <parallel/numericfwd.h> 41 42 namespace __gnu_parallel 43 { 44 // Problem: there is no 0-element given. 45 46 /** @brief Base case prefix sum routine. 47 * @param __begin Begin iterator of input sequence. 48 * @param __end End iterator of input sequence. 49 * @param __result Begin iterator of output sequence. 50 * @param __bin_op Associative binary function. 51 * @param __value Start value. Must be passed since the neutral 52 * element is unknown in general. 53 * @return End iterator of output sequence. */ 54 template<typename _IIter, 55 typename _OutputIterator, 56 typename _BinaryOperation> 57 _OutputIterator 58 __parallel_partial_sum_basecase(_IIter __begin, _IIter __end, 59 _OutputIterator __result, 60 _BinaryOperation __bin_op, 61 typename std::iterator_traits <_IIter>::value_type __value) 62 { 63 if (__begin == __end) 64 return __result; 65 66 while (__begin != __end) 67 { 68 __value = __bin_op(__value, *__begin); 69 *__result = __value; 70 ++__result; 71 ++__begin; 72 } 73 return __result; 74 } 75 76 /** @brief Parallel partial sum implementation, two-phase approach, 77 no recursion. 78 * @param __begin Begin iterator of input sequence. 79 * @param __end End iterator of input sequence. 80 * @param __result Begin iterator of output sequence. 81 * @param __bin_op Associative binary function. 82 * @param __n Length of sequence. 83 * @param __num_threads Number of threads to use. 84 * @return End iterator of output sequence. 85 */ 86 template<typename _IIter, 87 typename _OutputIterator, 88 typename _BinaryOperation> 89 _OutputIterator 90 __parallel_partial_sum_linear(_IIter __begin, _IIter __end, 91 _OutputIterator __result, 92 _BinaryOperation __bin_op, 93 typename std::iterator_traits<_IIter>::difference_type __n) 94 { 95 typedef std::iterator_traits<_IIter> _TraitsType; 96 typedef typename _TraitsType::value_type _ValueType; 97 typedef typename _TraitsType::difference_type _DifferenceType; 98 99 if (__begin == __end) 100 return __result; 101 102 _ThreadIndex __num_threads = 103 std::min<_DifferenceType>(__get_max_threads(), __n - 1); 104 105 if (__num_threads < 2) 106 { 107 *__result = *__begin; 108 return __parallel_partial_sum_basecase(__begin + 1, __end, 109 __result + 1, __bin_op, 110 *__begin); 111 } 112 113 _DifferenceType* __borders; 114 _ValueType* __sums; 115 116 const _Settings& __s = _Settings::get(); 117 118 # pragma omp parallel num_threads(__num_threads) 119 { 120 # pragma omp single 121 { 122 __num_threads = omp_get_num_threads(); 123 124 __borders = new _DifferenceType[__num_threads + 2]; 125 126 if (__s.partial_sum_dilation == 1.0f) 127 equally_split(__n, __num_threads + 1, __borders); 128 else 129 { 130 _DifferenceType __chunk_length = 131 ((double)__n 132 / ((double)__num_threads + __s.partial_sum_dilation)), 133 __borderstart = __n - __num_threads * __chunk_length; 134 __borders[0] = 0; 135 for (_ThreadIndex __i = 1; __i < (__num_threads + 1); ++__i) 136 { 137 __borders[__i] = __borderstart; 138 __borderstart += __chunk_length; 139 } 140 __borders[__num_threads + 1] = __n; 141 } 142 143 __sums = static_cast<_ValueType*>(::operator new(sizeof(_ValueType) 144 * __num_threads)); 145 _OutputIterator __target_end; 146 } //single 147 148 _ThreadIndex __iam = omp_get_thread_num(); 149 if (__iam == 0) 150 { 151 *__result = *__begin; 152 __parallel_partial_sum_basecase(__begin + 1, 153 __begin + __borders[1], 154 __result + 1, 155 __bin_op, *__begin); 156 ::new(&(__sums[__iam])) _ValueType(*(__result + __borders[1] - 1)); 157 } 158 else 159 { 160 ::new(&(__sums[__iam])) 161 _ValueType(__gnu_parallel::accumulate( 162 __begin + __borders[__iam] + 1, 163 __begin + __borders[__iam + 1], 164 *(__begin + __borders[__iam]), 165 __bin_op, 166 __gnu_parallel::sequential_tag())); 167 } 168 169 # pragma omp barrier 170 171 # pragma omp single 172 __parallel_partial_sum_basecase(__sums + 1, __sums + __num_threads, 173 __sums + 1, __bin_op, __sums[0]); 174 175 # pragma omp barrier 176 177 // Still same team. 178 __parallel_partial_sum_basecase(__begin + __borders[__iam + 1], 179 __begin + __borders[__iam + 2], 180 __result + __borders[__iam + 1], 181 __bin_op, __sums[__iam]); 182 } //parallel 183 184 ::operator delete(__sums); 185 delete[] __borders; 186 187 return __result + __n; 188 } 189 190 /** @brief Parallel partial sum front-__end. 191 * @param __begin Begin iterator of input sequence. 192 * @param __end End iterator of input sequence. 193 * @param __result Begin iterator of output sequence. 194 * @param __bin_op Associative binary function. 195 * @return End iterator of output sequence. */ 196 template<typename _IIter, 197 typename _OutputIterator, 198 typename _BinaryOperation> 199 _OutputIterator 200 __parallel_partial_sum(_IIter __begin, _IIter __end, 201 _OutputIterator __result, _BinaryOperation __bin_op) 202 { 203 _GLIBCXX_CALL(__begin - __end) 204 205 typedef std::iterator_traits<_IIter> _TraitsType; 206 typedef typename _TraitsType::value_type _ValueType; 207 typedef typename _TraitsType::difference_type _DifferenceType; 208 209 _DifferenceType __n = __end - __begin; 210 211 switch (_Settings::get().partial_sum_algorithm) 212 { 213 case LINEAR: 214 // Need an initial offset. 215 return __parallel_partial_sum_linear(__begin, __end, __result, 216 __bin_op, __n); 217 default: 218 // Partial_sum algorithm not implemented. 219 _GLIBCXX_PARALLEL_ASSERT(0); 220 return __result + __n; 221 } 222 } 223 } 224 225 #endif /* _GLIBCXX_PARALLEL_PARTIAL_SUM_H */ 226