1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
|
/* SPDX-License-Identifier: GPL-2.0-or-later */
#pragma once
/** \file
* \ingroup bli
*/
#ifdef WITH_TBB
/* Quiet top level deprecation message, unrelated to API usage here. */
# if defined(WIN32) && !defined(NOMINMAX)
/* TBB includes Windows.h which will define min/max macros causing issues
* when we try to use std::min and std::max later on. */
# define NOMINMAX
# define TBB_MIN_MAX_CLEANUP
# endif
# include <tbb/blocked_range.h>
# include <tbb/parallel_for.h>
# include <tbb/parallel_for_each.h>
# include <tbb/parallel_invoke.h>
# include <tbb/parallel_reduce.h>
# include <tbb/task_arena.h>
# ifdef WIN32
/* We cannot keep this defined, since other parts of the code deal with this on their own, leading
* to multiple define warnings unless we un-define this, however we can only undefine this if we
* were the ones that made the definition earlier. */
# ifdef TBB_MIN_MAX_CLEANUP
# undef NOMINMAX
# endif
# endif
#endif
#include "BLI_index_range.hh"
#include "BLI_lazy_threading.hh"
#include "BLI_utildefines.h"
namespace blender::threading {
template<typename Range, typename Function>
void parallel_for_each(Range &range, const Function &function)
{
#ifdef WITH_TBB
tbb::parallel_for_each(range, function);
#else
for (auto &value : range) {
function(value);
}
#endif
}
template<typename Function>
void parallel_for(IndexRange range, int64_t grain_size, const Function &function)
{
if (range.size() == 0) {
return;
}
#ifdef WITH_TBB
/* Invoking tbb for small workloads has a large overhead. */
if (range.size() >= grain_size) {
lazy_threading::send_hint();
tbb::parallel_for(
tbb::blocked_range<int64_t>(range.first(), range.one_after_last(), grain_size),
[&](const tbb::blocked_range<int64_t> &subrange) {
function(IndexRange(subrange.begin(), subrange.size()));
});
return;
}
#else
UNUSED_VARS(grain_size);
#endif
function(range);
}
template<typename Value, typename Function, typename Reduction>
Value parallel_reduce(IndexRange range,
int64_t grain_size,
const Value &identity,
const Function &function,
const Reduction &reduction)
{
#ifdef WITH_TBB
if (range.size() >= grain_size) {
lazy_threading::send_hint();
return tbb::parallel_reduce(
tbb::blocked_range<int64_t>(range.first(), range.one_after_last(), grain_size),
identity,
[&](const tbb::blocked_range<int64_t> &subrange, const Value &ident) {
return function(IndexRange(subrange.begin(), subrange.size()), ident);
},
reduction);
}
#else
UNUSED_VARS(grain_size, reduction);
#endif
return function(range, identity);
}
/**
* Execute all of the provided functions. The functions might be executed in parallel or in serial
* or some combination of both.
*/
template<typename... Functions> void parallel_invoke(Functions &&...functions)
{
#ifdef WITH_TBB
tbb::parallel_invoke(std::forward<Functions>(functions)...);
#else
(functions(), ...);
#endif
}
/**
* Same #parallel_invoke, but allows disabling threading dynamically. This is useful because when
* the individual functions do very little work, there is a lot of overhead from starting parallel
* tasks.
*/
template<typename... Functions>
void parallel_invoke(const bool use_threading, Functions &&...functions)
{
if (use_threading) {
lazy_threading::send_hint();
parallel_invoke(std::forward<Functions>(functions)...);
}
else {
(functions(), ...);
}
}
/** See #BLI_task_isolate for a description of what isolating a task means. */
template<typename Function> void isolate_task(const Function &function)
{
#ifdef WITH_TBB
tbb::this_task_arena::isolate(function);
#else
function();
#endif
}
} // namespace blender::threading
|
