1 |
742 |
jeremybenn |
// -*- C++ -*-
|
2 |
|
|
|
3 |
|
|
// Copyright (C) 2007, 2008, 2009, 2010, 2011 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/quicksort.h
|
26 |
|
|
* @brief Implementation of a unbalanced parallel quicksort (in-place).
|
27 |
|
|
* This file is a GNU parallel extension to the Standard C++ Library.
|
28 |
|
|
*/
|
29 |
|
|
|
30 |
|
|
// Written by Johannes Singler.
|
31 |
|
|
|
32 |
|
|
#ifndef _GLIBCXX_PARALLEL_QUICKSORT_H
|
33 |
|
|
#define _GLIBCXX_PARALLEL_QUICKSORT_H 1
|
34 |
|
|
|
35 |
|
|
#include <parallel/parallel.h>
|
36 |
|
|
#include <parallel/partition.h>
|
37 |
|
|
|
38 |
|
|
namespace __gnu_parallel
|
39 |
|
|
{
|
40 |
|
|
/** @brief Unbalanced quicksort divide step.
|
41 |
|
|
* @param __begin Begin iterator of subsequence.
|
42 |
|
|
* @param __end End iterator of subsequence.
|
43 |
|
|
* @param __comp Comparator.
|
44 |
|
|
* @param __pivot_rank Desired __rank of the pivot.
|
45 |
|
|
* @param __num_samples Choose pivot from that many samples.
|
46 |
|
|
* @param __num_threads Number of threads that are allowed to work on
|
47 |
|
|
* this part.
|
48 |
|
|
*/
|
49 |
|
|
template<typename _RAIter, typename _Compare>
|
50 |
|
|
typename std::iterator_traits<_RAIter>::difference_type
|
51 |
|
|
__parallel_sort_qs_divide(_RAIter __begin, _RAIter __end,
|
52 |
|
|
_Compare __comp, typename std::iterator_traits
|
53 |
|
|
<_RAIter>::difference_type __pivot_rank,
|
54 |
|
|
typename std::iterator_traits
|
55 |
|
|
<_RAIter>::difference_type
|
56 |
|
|
__num_samples, _ThreadIndex __num_threads)
|
57 |
|
|
{
|
58 |
|
|
typedef std::iterator_traits<_RAIter> _TraitsType;
|
59 |
|
|
typedef typename _TraitsType::value_type _ValueType;
|
60 |
|
|
typedef typename _TraitsType::difference_type _DifferenceType;
|
61 |
|
|
|
62 |
|
|
_DifferenceType __n = __end - __begin;
|
63 |
|
|
__num_samples = std::min(__num_samples, __n);
|
64 |
|
|
|
65 |
|
|
// Allocate uninitialized, to avoid default constructor.
|
66 |
|
|
_ValueType* __samples = static_cast<_ValueType*>
|
67 |
|
|
(::operator new(__num_samples * sizeof(_ValueType)));
|
68 |
|
|
|
69 |
|
|
for (_DifferenceType __s = 0; __s < __num_samples; ++__s)
|
70 |
|
|
{
|
71 |
|
|
const unsigned long long __index = static_cast<unsigned long long>
|
72 |
|
|
(__s) * __n / __num_samples;
|
73 |
|
|
::new(&(__samples[__s])) _ValueType(__begin[__index]);
|
74 |
|
|
}
|
75 |
|
|
|
76 |
|
|
__gnu_sequential::sort(__samples, __samples + __num_samples, __comp);
|
77 |
|
|
|
78 |
|
|
_ValueType& __pivot = __samples[__pivot_rank * __num_samples / __n];
|
79 |
|
|
|
80 |
|
|
__gnu_parallel::__binder2nd<_Compare, _ValueType, _ValueType, bool>
|
81 |
|
|
__pred(__comp, __pivot);
|
82 |
|
|
_DifferenceType __split = __parallel_partition(__begin, __end,
|
83 |
|
|
__pred, __num_threads);
|
84 |
|
|
|
85 |
|
|
for (_DifferenceType __s = 0; __s < __num_samples; ++__s)
|
86 |
|
|
__samples[__s].~_ValueType();
|
87 |
|
|
::operator delete(__samples);
|
88 |
|
|
|
89 |
|
|
return __split;
|
90 |
|
|
}
|
91 |
|
|
|
92 |
|
|
/** @brief Unbalanced quicksort conquer step.
|
93 |
|
|
* @param __begin Begin iterator of subsequence.
|
94 |
|
|
* @param __end End iterator of subsequence.
|
95 |
|
|
* @param __comp Comparator.
|
96 |
|
|
* @param __num_threads Number of threads that are allowed to work on
|
97 |
|
|
* this part.
|
98 |
|
|
*/
|
99 |
|
|
template<typename _RAIter, typename _Compare>
|
100 |
|
|
void
|
101 |
|
|
__parallel_sort_qs_conquer(_RAIter __begin, _RAIter __end,
|
102 |
|
|
_Compare __comp,
|
103 |
|
|
_ThreadIndex __num_threads)
|
104 |
|
|
{
|
105 |
|
|
typedef std::iterator_traits<_RAIter> _TraitsType;
|
106 |
|
|
typedef typename _TraitsType::value_type _ValueType;
|
107 |
|
|
typedef typename _TraitsType::difference_type _DifferenceType;
|
108 |
|
|
|
109 |
|
|
if (__num_threads <= 1)
|
110 |
|
|
{
|
111 |
|
|
__gnu_sequential::sort(__begin, __end, __comp);
|
112 |
|
|
return;
|
113 |
|
|
}
|
114 |
|
|
|
115 |
|
|
_DifferenceType __n = __end - __begin, __pivot_rank;
|
116 |
|
|
|
117 |
|
|
if (__n <= 1)
|
118 |
|
|
return;
|
119 |
|
|
|
120 |
|
|
_ThreadIndex __num_threads_left;
|
121 |
|
|
|
122 |
|
|
if ((__num_threads % 2) == 1)
|
123 |
|
|
__num_threads_left = __num_threads / 2 + 1;
|
124 |
|
|
else
|
125 |
|
|
__num_threads_left = __num_threads / 2;
|
126 |
|
|
|
127 |
|
|
__pivot_rank = __n * __num_threads_left / __num_threads;
|
128 |
|
|
|
129 |
|
|
_DifferenceType __split = __parallel_sort_qs_divide
|
130 |
|
|
(__begin, __end, __comp, __pivot_rank,
|
131 |
|
|
_Settings::get().sort_qs_num_samples_preset, __num_threads);
|
132 |
|
|
|
133 |
|
|
#pragma omp parallel sections num_threads(2)
|
134 |
|
|
{
|
135 |
|
|
#pragma omp section
|
136 |
|
|
__parallel_sort_qs_conquer(__begin, __begin + __split,
|
137 |
|
|
__comp, __num_threads_left);
|
138 |
|
|
#pragma omp section
|
139 |
|
|
__parallel_sort_qs_conquer(__begin + __split, __end,
|
140 |
|
|
__comp, __num_threads - __num_threads_left);
|
141 |
|
|
}
|
142 |
|
|
}
|
143 |
|
|
|
144 |
|
|
|
145 |
|
|
/** @brief Unbalanced quicksort main call.
|
146 |
|
|
* @param __begin Begin iterator of input sequence.
|
147 |
|
|
* @param __end End iterator input sequence, ignored.
|
148 |
|
|
* @param __comp Comparator.
|
149 |
|
|
* @param __num_threads Number of threads that are allowed to work on
|
150 |
|
|
* this part.
|
151 |
|
|
*/
|
152 |
|
|
template<typename _RAIter, typename _Compare>
|
153 |
|
|
void
|
154 |
|
|
__parallel_sort_qs(_RAIter __begin, _RAIter __end,
|
155 |
|
|
_Compare __comp,
|
156 |
|
|
_ThreadIndex __num_threads)
|
157 |
|
|
{
|
158 |
|
|
_GLIBCXX_CALL(__n)
|
159 |
|
|
|
160 |
|
|
typedef std::iterator_traits<_RAIter> _TraitsType;
|
161 |
|
|
typedef typename _TraitsType::value_type _ValueType;
|
162 |
|
|
typedef typename _TraitsType::difference_type _DifferenceType;
|
163 |
|
|
|
164 |
|
|
_DifferenceType __n = __end - __begin;
|
165 |
|
|
|
166 |
|
|
// At least one element per processor.
|
167 |
|
|
if (__num_threads > __n)
|
168 |
|
|
__num_threads = static_cast<_ThreadIndex>(__n);
|
169 |
|
|
|
170 |
|
|
__parallel_sort_qs_conquer(
|
171 |
|
|
__begin, __begin + __n, __comp, __num_threads);
|
172 |
|
|
}
|
173 |
|
|
|
174 |
|
|
} //namespace __gnu_parallel
|
175 |
|
|
|
176 |
|
|
#endif /* _GLIBCXX_PARALLEL_QUICKSORT_H */
|