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
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
|
/*
* This program 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 Foundation; either version 2
* of the License, or (at your option) any later version.
*
* This program 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.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software Foundation,
* Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
*/
/** \file
* \ingroup bli
*
* Task parallel range functions.
*/
#include <stdlib.h>
#include "MEM_guardedalloc.h"
#include "DNA_listBase.h"
#include "BLI_task.h"
#include "BLI_threads.h"
#include "atomic_ops.h"
#ifdef WITH_TBB
/* Quiet top level deprecation message, unrelated to API usage here. */
# define TBB_SUPPRESS_DEPRECATED_MESSAGES 1
# include <tbb/tbb.h>
#endif
#ifdef WITH_TBB
/* Functor for running TBB parallel_for and parallel_reduce. */
struct RangeTask {
TaskParallelRangeFunc func;
void *userdata;
const TaskParallelSettings *settings;
void *userdata_chunk;
/* Root constructor. */
RangeTask(TaskParallelRangeFunc func, void *userdata, const TaskParallelSettings *settings)
: func(func), userdata(userdata), settings(settings)
{
init_chunk(settings->userdata_chunk);
}
/* Copy constructor. */
RangeTask(const RangeTask &other)
: func(other.func), userdata(other.userdata), settings(other.settings)
{
init_chunk(settings->userdata_chunk);
}
/* Splitting constructor for parallel reduce. */
RangeTask(RangeTask &other, tbb::split)
: func(other.func), userdata(other.userdata), settings(other.settings)
{
init_chunk(settings->userdata_chunk);
}
~RangeTask()
{
if (settings->func_free != NULL) {
settings->func_free(userdata, userdata_chunk);
}
MEM_SAFE_FREE(userdata_chunk);
}
void init_chunk(void *from_chunk)
{
if (from_chunk) {
userdata_chunk = MEM_mallocN(settings->userdata_chunk_size, "RangeTask");
memcpy(userdata_chunk, from_chunk, settings->userdata_chunk_size);
}
else {
userdata_chunk = NULL;
}
}
void operator()(const tbb::blocked_range<int> &r) const
{
tbb::this_task_arena::isolate([this, r] {
TaskParallelTLS tls;
tls.userdata_chunk = userdata_chunk;
for (int i = r.begin(); i != r.end(); ++i) {
func(userdata, i, &tls);
}
});
}
void join(const RangeTask &other)
{
settings->func_reduce(userdata, userdata_chunk, other.userdata_chunk);
}
};
#endif
void BLI_task_parallel_range(const int start,
const int stop,
void *userdata,
TaskParallelRangeFunc func,
const TaskParallelSettings *settings)
{
#ifdef WITH_TBB
/* Multithreading. */
if (settings->use_threading && BLI_task_scheduler_num_threads() > 1) {
RangeTask task(func, userdata, settings);
const size_t grainsize = MAX2(settings->min_iter_per_thread, 1);
const tbb::blocked_range<int> range(start, stop, grainsize);
if (settings->func_reduce) {
parallel_reduce(range, task);
if (settings->userdata_chunk) {
memcpy(settings->userdata_chunk, task.userdata_chunk, settings->userdata_chunk_size);
}
}
else {
parallel_for(range, task);
}
return;
}
#endif
/* Single threaded. Nothing to reduce as everything is accumulated into the
* main userdata chunk directly. */
TaskParallelTLS tls;
tls.userdata_chunk = settings->userdata_chunk;
for (int i = start; i < stop; i++) {
func(userdata, i, &tls);
}
if (settings->func_free != NULL) {
settings->func_free(userdata, settings->userdata_chunk);
}
}
int BLI_task_parallel_thread_id(const TaskParallelTLS *UNUSED(tls))
{
#ifdef WITH_TBB
/* Get a unique thread ID for texture nodes. In the future we should get rid
* of the thread ID and change texture evaluation to not require per-thread
* storage that can't be efficiently allocated on the stack. */
static tbb::enumerable_thread_specific<int> tbb_thread_id(-1);
static int tbb_thread_id_counter = 0;
int &thread_id = tbb_thread_id.local();
if (thread_id == -1) {
thread_id = atomic_fetch_and_add_int32(&tbb_thread_id_counter, 1);
if (thread_id >= BLENDER_MAX_THREADS) {
BLI_assert(!"Maximum number of threads exceeded for sculpting");
thread_id = thread_id % BLENDER_MAX_THREADS;
}
}
return thread_id;
#else
return 0;
#endif
}
|
