2007-09-12 00:32:51 +02:00
|
|
|
// -*- C++ -*-
|
|
|
|
|
2009-04-09 17:00:19 +02:00
|
|
|
// Copyright (C) 2007, 2008, 2009 Free Software Foundation, Inc.
|
2007-09-12 00:32:51 +02:00
|
|
|
//
|
|
|
|
// This file is part of the GNU ISO C++ Library. This library is free
|
|
|
|
// software; you can redistribute it and/or modify it under the terms
|
|
|
|
// of the GNU General Public License as published by the Free Software
|
2009-04-09 17:00:19 +02:00
|
|
|
// Foundation; either version 3, or (at your option) any later
|
2007-09-12 00:32:51 +02:00
|
|
|
// 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
|
|
|
|
// General Public License for more details.
|
|
|
|
|
2009-04-09 17:00:19 +02:00
|
|
|
// Under Section 7 of GPL version 3, you are granted additional
|
|
|
|
// permissions described in the GCC Runtime Library Exception, version
|
|
|
|
// 3.1, as published by the Free Software Foundation.
|
|
|
|
|
|
|
|
// You should have received a copy of the GNU General Public License and
|
|
|
|
// a copy of the GCC Runtime Library Exception along with this program;
|
|
|
|
// see the files COPYING3 and COPYING.RUNTIME respectively. If not, see
|
|
|
|
// <http://www.gnu.org/licenses/>.
|
2007-09-12 00:32:51 +02:00
|
|
|
|
|
|
|
/** @file parallel/omp_loop_static.h
|
|
|
|
* @brief Parallelization of embarrassingly parallel execution by
|
|
|
|
* means of an OpenMP for loop with static scheduling.
|
|
|
|
* This file is a GNU parallel extension to the Standard C++ Library.
|
|
|
|
*/
|
|
|
|
|
|
|
|
// Written by Felix Putze.
|
|
|
|
|
|
|
|
#ifndef _GLIBCXX_PARALLEL_OMP_LOOP_STATIC_H
|
|
|
|
#define _GLIBCXX_PARALLEL_OMP_LOOP_STATIC_H 1
|
|
|
|
|
|
|
|
#include <omp.h>
|
|
|
|
|
|
|
|
#include <parallel/settings.h>
|
|
|
|
#include <parallel/basic_iterator.h>
|
|
|
|
|
|
|
|
namespace __gnu_parallel
|
|
|
|
{
|
|
|
|
|
|
|
|
/** @brief Embarrassingly parallel algorithm for random access
|
|
|
|
* iterators, using an OpenMP for loop with static scheduling.
|
|
|
|
*
|
2009-09-16 11:47:25 +02:00
|
|
|
* @param __begin Begin iterator of element __sequence.
|
|
|
|
* @param __end End iterator of element __sequence.
|
|
|
|
* @param __o User-supplied functor (comparator, predicate, adding
|
2007-09-12 00:32:51 +02:00
|
|
|
* functor, ...).
|
2009-09-16 11:47:25 +02:00
|
|
|
* @param __f Functor to "process" an element with __op (depends on
|
2007-09-12 00:32:51 +02:00
|
|
|
* desired functionality, e. g. for std::for_each(), ...).
|
2009-09-16 11:47:25 +02:00
|
|
|
* @param __r Functor to "add" a single __result to the already processed
|
|
|
|
* __elements (depends on functionality).
|
|
|
|
* @param __base Base value for reduction.
|
|
|
|
* @param __output Pointer to position where final result is written to
|
|
|
|
* @param __bound Maximum number of elements processed (e. g. for
|
2007-09-12 00:32:51 +02:00
|
|
|
* std::count_n()).
|
|
|
|
* @return User-supplied functor (that may contain a part of the result).
|
|
|
|
*/
|
2009-09-16 11:47:25 +02:00
|
|
|
template<typename _RAIter,
|
2009-09-17 10:42:39 +02:00
|
|
|
typename _Op,
|
|
|
|
typename _Fu,
|
|
|
|
typename _Red,
|
|
|
|
typename _Result>
|
2009-09-16 11:47:25 +02:00
|
|
|
_Op
|
2009-09-17 10:42:39 +02:00
|
|
|
for_each_template_random_access_omp_loop_static(
|
|
|
|
_RAIter __begin, _RAIter __end, _Op __o, _Fu& __f, _Red __r,
|
|
|
|
_Result __base, _Result& __output,
|
|
|
|
typename std::iterator_traits<_RAIter>::difference_type __bound)
|
2007-09-12 00:32:51 +02:00
|
|
|
{
|
2007-11-22 11:13:08 +01:00
|
|
|
typedef typename
|
2009-09-16 11:47:25 +02:00
|
|
|
std::iterator_traits<_RAIter>::difference_type
|
|
|
|
_DifferenceType;
|
2007-09-12 00:32:51 +02:00
|
|
|
|
2009-09-16 11:47:25 +02:00
|
|
|
_DifferenceType __length = __end - __begin;
|
|
|
|
_ThreadIndex __num_threads =
|
|
|
|
std::min<_DifferenceType>(__get_max_threads(), __length);
|
2007-09-12 00:32:51 +02:00
|
|
|
|
2009-09-16 11:47:25 +02:00
|
|
|
_Result *__thread_results;
|
2007-11-22 11:13:08 +01:00
|
|
|
|
2009-09-16 11:47:25 +02:00
|
|
|
# pragma omp parallel num_threads(__num_threads)
|
2007-09-12 00:32:51 +02:00
|
|
|
{
|
2007-11-22 11:13:08 +01:00
|
|
|
# pragma omp single
|
|
|
|
{
|
2009-09-16 11:47:25 +02:00
|
|
|
__num_threads = omp_get_num_threads();
|
|
|
|
__thread_results = new _Result[__num_threads];
|
2007-11-22 11:13:08 +01:00
|
|
|
|
2009-09-16 11:47:25 +02:00
|
|
|
for (_ThreadIndex __i = 0; __i < __num_threads; ++__i)
|
|
|
|
__thread_results[__i] = _Result();
|
2007-11-22 11:13:08 +01:00
|
|
|
}
|
|
|
|
|
2009-09-16 11:47:25 +02:00
|
|
|
_ThreadIndex __iam = omp_get_thread_num();
|
2007-11-22 11:13:08 +01:00
|
|
|
|
2009-09-17 10:42:39 +02:00
|
|
|
#pragma omp for schedule(static, _Settings::get().workstealing_chunk_size)
|
2009-09-16 11:47:25 +02:00
|
|
|
for (_DifferenceType __pos = 0; __pos < __length; ++__pos)
|
2009-09-17 10:42:39 +02:00
|
|
|
__thread_results[__iam] = __r(__thread_results[__iam],
|
|
|
|
__f(__o, __begin+__pos));
|
2007-11-22 11:13:08 +01:00
|
|
|
} //parallel
|
2007-09-12 00:32:51 +02:00
|
|
|
|
2009-09-16 11:47:25 +02:00
|
|
|
for (_ThreadIndex __i = 0; __i < __num_threads; ++__i)
|
|
|
|
__output = __r(__output, __thread_results[__i]);
|
2007-09-12 00:32:51 +02:00
|
|
|
|
2009-09-16 11:47:25 +02:00
|
|
|
delete [] __thread_results;
|
2007-09-12 00:32:51 +02:00
|
|
|
|
|
|
|
// Points to last element processed (needed as return value for
|
|
|
|
// some algorithms like transform).
|
2009-09-16 11:47:25 +02:00
|
|
|
__f.finish_iterator = __begin + __length;
|
2007-09-12 00:32:51 +02:00
|
|
|
|
2009-09-16 11:47:25 +02:00
|
|
|
return __o;
|
2007-09-12 00:32:51 +02:00
|
|
|
}
|
2007-11-22 11:13:08 +01:00
|
|
|
|
2007-09-12 00:32:51 +02:00
|
|
|
} // end namespace
|
|
|
|
|
2008-07-15 12:06:58 +02:00
|
|
|
#endif /* _GLIBCXX_PARALLEL_OMP_LOOP_STATIC_H */
|