1 // -*- C++ -*-
2 
3 // Copyright (C) 2007-2017 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
__for_each_template_random_access_ed(_RAIter __begin,_RAIter __end,_Op __o,_Fu & __f,_Red __r,_Result __base,_Result & __output,typename std::iterator_traits<_RAIter>::difference_type __bound)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