Home | History | Annotate | Download | only in parallel 
      1 // -*- C++ -*-
      2 
      3 // Copyright (C) 2007-2013 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/par_loop.h
     26  *  @brief Parallelization of embarrassingly parallel execution by
     27  *  means of equal splitting.
     28  *  This file is a GNU parallel extension to the Standard C++ Library.
     29  */
     30 
     31 // Written by Felix Putze.
     32 
     33 #ifndef _GLIBCXX_PARALLEL_PAR_LOOP_H
     34 #define _GLIBCXX_PARALLEL_PAR_LOOP_H 1
     35 
     36 #include <omp.h>
     37 #include <parallel/settings.h>
     38 #include <parallel/base.h>
     39 #include <parallel/equally_split.h>
     40 
     41 namespace __gnu_parallel
     42 {
     43   /** @brief Embarrassingly parallel algorithm for random access
     44    * iterators, using hand-crafted parallelization by equal splitting
     45    * the work.
     46    *
     47    *  @param __begin Begin iterator of element sequence.
     48    *  @param __end End iterator of element sequence.
     49    *  @param __o User-supplied functor (comparator, predicate, adding
     50    *  functor, ...)
     51    *  @param __f Functor to "process" an element with __op (depends on
     52    *  desired functionality, e. g. for std::for_each(), ...).
     53    *  @param __r Functor to "add" a single __result to the already
     54    *  processed elements (depends on functionality).
     55    *  @param __base Base value for reduction.
     56    *  @param __output Pointer to position where final result is written to
     57    *  @param __bound Maximum number of elements processed (e. g. for
     58    *  std::count_n()).
     59    *  @return User-supplied functor (that may contain a part of the result).
     60    */
     61   template<typename _RAIter,
     62 	   typename _Op,
     63 	   typename _Fu,
     64 	   typename _Red,
     65 	   typename _Result>
     66     _Op
     67     __for_each_template_random_access_ed(_RAIter __begin, _RAIter __end,
     68 					 _Op __o, _Fu& __f, _Red __r,
     69 					 _Result __base, _Result& __output,
     70       typename std::iterator_traits<_RAIter>::difference_type __bound)
     71     {
     72       typedef std::iterator_traits<_RAIter> _TraitsType;
     73       typedef typename _TraitsType::difference_type _DifferenceType;
     74       const _DifferenceType __length = __end - __begin;
     75       _Result *__thread_results;
     76       bool* __constructed;
     77 
     78       _ThreadIndex __num_threads = __gnu_parallel::min<_DifferenceType>
     79 	(__get_max_threads(), __length);
     80 
     81 #     pragma omp parallel num_threads(__num_threads)
     82       {
     83 #       pragma omp single
     84 	{
     85 	  __num_threads = omp_get_num_threads();
     86 	  __thread_results = static_cast<_Result*>
     87 	    (::operator new(__num_threads * sizeof(_Result)));
     88 	  __constructed = new bool[__num_threads];
     89 	}
     90 
     91 	_ThreadIndex __iam = omp_get_thread_num();
     92 
     93 	// Neutral element.
     94 	_Result* __reduct;
     95 
     96 	_DifferenceType
     97 	  __start = __equally_split_point(__length, __num_threads, __iam),
     98 	  __stop = __equally_split_point(__length, __num_threads, __iam + 1);
     99 
    100 	if (__start < __stop)
    101 	  {
    102 	    __reduct = new _Result(__f(__o, __begin + __start));
    103 	    ++__start;
    104 	    __constructed[__iam] = true;
    105 	  }
    106 	else
    107 	  __constructed[__iam] = false;
    108 
    109 	for (; __start < __stop; ++__start)
    110 	  *__reduct = __r(*__reduct, __f(__o, __begin + __start));
    111 
    112 	if (__constructed[__iam])
    113 	  {
    114 	    ::new(&__thread_results[__iam]) _Result(*__reduct);
    115 	    delete __reduct;
    116 	  }
    117       } //parallel
    118 
    119       for (_ThreadIndex __i = 0; __i < __num_threads; ++__i)
    120 	if (__constructed[__i])
    121 	  {
    122 	    __output = __r(__output, __thread_results[__i]);
    123 	    __thread_results[__i].~_Result();
    124 	  }
    125 
    126       // Points to last element processed (needed as return value for
    127       // some algorithms like transform).
    128       __f._M_finish_iterator = __begin + __length;
    129 
    130       ::operator delete(__thread_results);
    131 
    132       delete[] __constructed;
    133 
    134       return __o;
    135     }
    136 
    137 } // end namespace
    138 
    139 #endif /* _GLIBCXX_PARALLEL_PAR_LOOP_H */
    140