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
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
|
/*
* Copyright 2011-2013 Blender Foundation
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include <stdlib.h>
#include <string.h>
#include "device/device.h"
#include "device/device_intern.h"
#include "util/util_foreach.h"
#include "util/util_half.h"
#include "util/util_logging.h"
#include "util/util_math.h"
#include "util/util_opengl.h"
#include "util/util_string.h"
#include "util/util_system.h"
#include "util/util_time.h"
#include "util/util_types.h"
#include "util/util_vector.h"
CCL_NAMESPACE_BEGIN
bool Device::need_types_update = true;
bool Device::need_devices_update = true;
thread_mutex Device::device_mutex;
vector<DeviceInfo> Device::opencl_devices;
vector<DeviceInfo> Device::cuda_devices;
vector<DeviceInfo> Device::optix_devices;
vector<DeviceInfo> Device::cpu_devices;
vector<DeviceInfo> Device::network_devices;
uint Device::devices_initialized_mask = 0;
/* Device Requested Features */
std::ostream &operator<<(std::ostream &os, const DeviceRequestedFeatures &requested_features)
{
os << "Experimental features: " << (requested_features.experimental ? "On" : "Off") << std::endl;
os << "Max nodes group: " << requested_features.max_nodes_group << std::endl;
/* TODO(sergey): Decode bitflag into list of names. */
os << "Nodes features: " << requested_features.nodes_features << std::endl;
os << "Use Hair: " << string_from_bool(requested_features.use_hair) << std::endl;
os << "Use Object Motion: " << string_from_bool(requested_features.use_object_motion)
<< std::endl;
os << "Use Camera Motion: " << string_from_bool(requested_features.use_camera_motion)
<< std::endl;
os << "Use Baking: " << string_from_bool(requested_features.use_baking) << std::endl;
os << "Use Subsurface: " << string_from_bool(requested_features.use_subsurface) << std::endl;
os << "Use Volume: " << string_from_bool(requested_features.use_volume) << std::endl;
os << "Use Branched Integrator: " << string_from_bool(requested_features.use_integrator_branched)
<< std::endl;
os << "Use Patch Evaluation: " << string_from_bool(requested_features.use_patch_evaluation)
<< std::endl;
os << "Use Transparent Shadows: " << string_from_bool(requested_features.use_transparent)
<< std::endl;
os << "Use Principled BSDF: " << string_from_bool(requested_features.use_principled)
<< std::endl;
os << "Use Denoising: " << string_from_bool(requested_features.use_denoising) << std::endl;
os << "Use Displacement: " << string_from_bool(requested_features.use_true_displacement)
<< std::endl;
os << "Use Background Light: " << string_from_bool(requested_features.use_background_light)
<< std::endl;
return os;
}
/* Device */
Device::~Device()
{
if (!background) {
if (vertex_buffer != 0) {
glDeleteBuffers(1, &vertex_buffer);
}
if (fallback_shader_program != 0) {
glDeleteProgram(fallback_shader_program);
}
}
}
/* TODO move shaders to standalone .glsl file. */
const char *FALLBACK_VERTEX_SHADER =
"#version 330\n"
"uniform vec2 fullscreen;\n"
"in vec2 texCoord;\n"
"in vec2 pos;\n"
"out vec2 texCoord_interp;\n"
"\n"
"vec2 normalize_coordinates()\n"
"{\n"
" return (vec2(2.0) * (pos / fullscreen)) - vec2(1.0);\n"
"}\n"
"\n"
"void main()\n"
"{\n"
" gl_Position = vec4(normalize_coordinates(), 0.0, 1.0);\n"
" texCoord_interp = texCoord;\n"
"}\n\0";
const char *FALLBACK_FRAGMENT_SHADER =
"#version 330\n"
"uniform sampler2D image_texture;\n"
"in vec2 texCoord_interp;\n"
"out vec4 fragColor;\n"
"\n"
"void main()\n"
"{\n"
" fragColor = texture(image_texture, texCoord_interp);\n"
"}\n\0";
static void shader_print_errors(const char *task, const char *log, const char *code)
{
LOG(ERROR) << "Shader: " << task << " error:";
LOG(ERROR) << "===== shader string ====";
stringstream stream(code);
string partial;
int line = 1;
while (getline(stream, partial, '\n')) {
if (line < 10) {
LOG(ERROR) << " " << line << " " << partial;
}
else {
LOG(ERROR) << line << " " << partial;
}
line++;
}
LOG(ERROR) << log;
}
static int bind_fallback_shader(void)
{
GLint status;
GLchar log[5000];
GLsizei length = 0;
GLuint program = 0;
struct Shader {
const char *source;
GLenum type;
} shaders[2] = {{FALLBACK_VERTEX_SHADER, GL_VERTEX_SHADER},
{FALLBACK_FRAGMENT_SHADER, GL_FRAGMENT_SHADER}};
program = glCreateProgram();
for (int i = 0; i < 2; i++) {
GLuint shader = glCreateShader(shaders[i].type);
string source_str = shaders[i].source;
const char *c_str = source_str.c_str();
glShaderSource(shader, 1, &c_str, NULL);
glCompileShader(shader);
glGetShaderiv(shader, GL_COMPILE_STATUS, &status);
if (!status) {
glGetShaderInfoLog(shader, sizeof(log), &length, log);
shader_print_errors("compile", log, c_str);
return 0;
}
glAttachShader(program, shader);
}
/* Link output. */
glBindFragDataLocation(program, 0, "fragColor");
/* Link and error check. */
glLinkProgram(program);
glGetProgramiv(program, GL_LINK_STATUS, &status);
if (!status) {
glGetShaderInfoLog(program, sizeof(log), &length, log);
shader_print_errors("linking", log, FALLBACK_VERTEX_SHADER);
shader_print_errors("linking", log, FALLBACK_FRAGMENT_SHADER);
return 0;
}
return program;
}
bool Device::bind_fallback_display_space_shader(const float width, const float height)
{
if (fallback_status == FALLBACK_SHADER_STATUS_ERROR) {
return false;
}
if (fallback_status == FALLBACK_SHADER_STATUS_NONE) {
fallback_shader_program = bind_fallback_shader();
fallback_status = FALLBACK_SHADER_STATUS_ERROR;
if (fallback_shader_program == 0) {
return false;
}
glUseProgram(fallback_shader_program);
image_texture_location = glGetUniformLocation(fallback_shader_program, "image_texture");
if (image_texture_location < 0) {
LOG(ERROR) << "Shader doesn't containt the 'image_texture' uniform.";
return false;
}
fullscreen_location = glGetUniformLocation(fallback_shader_program, "fullscreen");
if (fullscreen_location < 0) {
LOG(ERROR) << "Shader doesn't containt the 'fullscreen' uniform.";
return false;
}
fallback_status = FALLBACK_SHADER_STATUS_SUCCESS;
}
/* Run this every time. */
glUseProgram(fallback_shader_program);
glUniform1i(image_texture_location, 0);
glUniform2f(fullscreen_location, width, height);
return true;
}
void Device::draw_pixels(device_memory &rgba,
int y,
int w,
int h,
int width,
int height,
int dx,
int dy,
int dw,
int dh,
bool transparent,
const DeviceDrawParams &draw_params)
{
const bool use_fallback_shader = (draw_params.bind_display_space_shader_cb == NULL);
assert(rgba.type == MEM_PIXELS);
mem_copy_from(rgba, y, w, h, rgba.memory_elements_size(1));
GLuint texid;
glActiveTexture(GL_TEXTURE0);
glGenTextures(1, &texid);
glBindTexture(GL_TEXTURE_2D, texid);
if (rgba.data_type == TYPE_HALF) {
GLhalf *data_pointer = (GLhalf *)rgba.host_pointer;
data_pointer += 4 * y * w;
glTexImage2D(GL_TEXTURE_2D, 0, GL_RGBA16F, w, h, 0, GL_RGBA, GL_HALF_FLOAT, data_pointer);
}
else {
uint8_t *data_pointer = (uint8_t *)rgba.host_pointer;
data_pointer += 4 * y * w;
glTexImage2D(GL_TEXTURE_2D, 0, GL_RGBA8, w, h, 0, GL_RGBA, GL_UNSIGNED_BYTE, data_pointer);
}
glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
if (transparent) {
glEnable(GL_BLEND);
glBlendFunc(GL_ONE, GL_ONE_MINUS_SRC_ALPHA);
}
GLint shader_program;
if (use_fallback_shader) {
if (!bind_fallback_display_space_shader(dw, dh)) {
return;
}
shader_program = fallback_shader_program;
}
else {
draw_params.bind_display_space_shader_cb();
glGetIntegerv(GL_CURRENT_PROGRAM, &shader_program);
}
if (!vertex_buffer) {
glGenBuffers(1, &vertex_buffer);
}
glBindBuffer(GL_ARRAY_BUFFER, vertex_buffer);
/* invalidate old contents - avoids stalling if buffer is still waiting in queue to be rendered
*/
glBufferData(GL_ARRAY_BUFFER, 16 * sizeof(float), NULL, GL_STREAM_DRAW);
float *vpointer = (float *)glMapBuffer(GL_ARRAY_BUFFER, GL_WRITE_ONLY);
if (vpointer) {
/* texture coordinate - vertex pair */
vpointer[0] = 0.0f;
vpointer[1] = 0.0f;
vpointer[2] = dx;
vpointer[3] = dy;
vpointer[4] = 1.0f;
vpointer[5] = 0.0f;
vpointer[6] = (float)width + dx;
vpointer[7] = dy;
vpointer[8] = 1.0f;
vpointer[9] = 1.0f;
vpointer[10] = (float)width + dx;
vpointer[11] = (float)height + dy;
vpointer[12] = 0.0f;
vpointer[13] = 1.0f;
vpointer[14] = dx;
vpointer[15] = (float)height + dy;
if (vertex_buffer) {
glUnmapBuffer(GL_ARRAY_BUFFER);
}
}
GLuint vertex_array_object;
GLuint position_attribute, texcoord_attribute;
glGenVertexArrays(1, &vertex_array_object);
glBindVertexArray(vertex_array_object);
texcoord_attribute = glGetAttribLocation(shader_program, "texCoord");
position_attribute = glGetAttribLocation(shader_program, "pos");
glEnableVertexAttribArray(texcoord_attribute);
glEnableVertexAttribArray(position_attribute);
glVertexAttribPointer(
texcoord_attribute, 2, GL_FLOAT, GL_FALSE, 4 * sizeof(float), (const GLvoid *)0);
glVertexAttribPointer(position_attribute,
2,
GL_FLOAT,
GL_FALSE,
4 * sizeof(float),
(const GLvoid *)(sizeof(float) * 2));
glDrawArrays(GL_TRIANGLE_FAN, 0, 4);
if (vertex_buffer) {
glBindBuffer(GL_ARRAY_BUFFER, 0);
}
if (use_fallback_shader) {
glUseProgram(0);
}
else {
draw_params.unbind_display_space_shader_cb();
}
glDeleteVertexArrays(1, &vertex_array_object);
glBindTexture(GL_TEXTURE_2D, 0);
glDeleteTextures(1, &texid);
if (transparent) {
glDisable(GL_BLEND);
}
}
Device *Device::create(DeviceInfo &info, Stats &stats, Profiler &profiler, bool background)
{
#ifdef WITH_MULTI
if (!info.multi_devices.empty()) {
/* Always create a multi device when info contains multiple devices.
* This is done so that the type can still be e.g. DEVICE_CPU to indicate
* that it is a homogeneous collection of devices, which simplifies checks. */
return device_multi_create(info, stats, profiler, background);
}
#endif
Device *device;
switch (info.type) {
case DEVICE_CPU:
device = device_cpu_create(info, stats, profiler, background);
break;
#ifdef WITH_CUDA
case DEVICE_CUDA:
if (device_cuda_init())
device = device_cuda_create(info, stats, profiler, background);
else
device = NULL;
break;
#endif
#ifdef WITH_OPTIX
case DEVICE_OPTIX:
if (device_optix_init())
device = device_optix_create(info, stats, profiler, background);
else
device = NULL;
break;
#endif
#ifdef WITH_NETWORK
case DEVICE_NETWORK:
device = device_network_create(info, stats, profiler, "127.0.0.1");
break;
#endif
#ifdef WITH_OPENCL
case DEVICE_OPENCL:
if (device_opencl_init())
device = device_opencl_create(info, stats, profiler, background);
else
device = NULL;
break;
#endif
default:
return NULL;
}
return device;
}
DeviceType Device::type_from_string(const char *name)
{
if (strcmp(name, "CPU") == 0)
return DEVICE_CPU;
else if (strcmp(name, "CUDA") == 0)
return DEVICE_CUDA;
else if (strcmp(name, "OPTIX") == 0)
return DEVICE_OPTIX;
else if (strcmp(name, "OPENCL") == 0)
return DEVICE_OPENCL;
else if (strcmp(name, "NETWORK") == 0)
return DEVICE_NETWORK;
else if (strcmp(name, "MULTI") == 0)
return DEVICE_MULTI;
return DEVICE_NONE;
}
string Device::string_from_type(DeviceType type)
{
if (type == DEVICE_CPU)
return "CPU";
else if (type == DEVICE_CUDA)
return "CUDA";
else if (type == DEVICE_OPTIX)
return "OPTIX";
else if (type == DEVICE_OPENCL)
return "OPENCL";
else if (type == DEVICE_NETWORK)
return "NETWORK";
else if (type == DEVICE_MULTI)
return "MULTI";
return "";
}
vector<DeviceType> Device::available_types()
{
vector<DeviceType> types;
types.push_back(DEVICE_CPU);
#ifdef WITH_CUDA
types.push_back(DEVICE_CUDA);
#endif
#ifdef WITH_OPTIX
types.push_back(DEVICE_OPTIX);
#endif
#ifdef WITH_OPENCL
types.push_back(DEVICE_OPENCL);
#endif
#ifdef WITH_NETWORK
types.push_back(DEVICE_NETWORK);
#endif
return types;
}
vector<DeviceInfo> Device::available_devices(uint mask)
{
/* Lazy initialize devices. On some platforms OpenCL or CUDA drivers can
* be broken and cause crashes when only trying to get device info, so
* we don't want to do any initialization until the user chooses to. */
thread_scoped_lock lock(device_mutex);
vector<DeviceInfo> devices;
#ifdef WITH_OPENCL
if (mask & DEVICE_MASK_OPENCL) {
if (!(devices_initialized_mask & DEVICE_MASK_OPENCL)) {
if (device_opencl_init()) {
device_opencl_info(opencl_devices);
}
devices_initialized_mask |= DEVICE_MASK_OPENCL;
}
foreach (DeviceInfo &info, opencl_devices) {
devices.push_back(info);
}
}
#endif
#ifdef WITH_CUDA
if (mask & DEVICE_MASK_CUDA) {
if (!(devices_initialized_mask & DEVICE_MASK_CUDA)) {
if (device_cuda_init()) {
device_cuda_info(cuda_devices);
}
devices_initialized_mask |= DEVICE_MASK_CUDA;
}
foreach (DeviceInfo &info, cuda_devices) {
devices.push_back(info);
}
}
#endif
#ifdef WITH_OPTIX
if (mask & DEVICE_MASK_OPTIX) {
if (!(devices_initialized_mask & DEVICE_MASK_OPTIX)) {
if (device_optix_init()) {
device_optix_info(optix_devices);
}
devices_initialized_mask |= DEVICE_MASK_OPTIX;
}
foreach (DeviceInfo &info, optix_devices) {
devices.push_back(info);
}
}
#endif
if (mask & DEVICE_MASK_CPU) {
if (!(devices_initialized_mask & DEVICE_MASK_CPU)) {
device_cpu_info(cpu_devices);
devices_initialized_mask |= DEVICE_MASK_CPU;
}
foreach (DeviceInfo &info, cpu_devices) {
devices.push_back(info);
}
}
#ifdef WITH_NETWORK
if (mask & DEVICE_MASK_NETWORK) {
if (!(devices_initialized_mask & DEVICE_MASK_NETWORK)) {
device_network_info(network_devices);
devices_initialized_mask |= DEVICE_MASK_NETWORK;
}
foreach (DeviceInfo &info, network_devices) {
devices.push_back(info);
}
}
#endif
return devices;
}
string Device::device_capabilities(uint mask)
{
thread_scoped_lock lock(device_mutex);
string capabilities = "";
if (mask & DEVICE_MASK_CPU) {
capabilities += "\nCPU device capabilities: ";
capabilities += device_cpu_capabilities() + "\n";
}
#ifdef WITH_OPENCL
if (mask & DEVICE_MASK_OPENCL) {
if (device_opencl_init()) {
capabilities += "\nOpenCL device capabilities:\n";
capabilities += device_opencl_capabilities();
}
}
#endif
#ifdef WITH_CUDA
if (mask & DEVICE_MASK_CUDA) {
if (device_cuda_init()) {
capabilities += "\nCUDA device capabilities:\n";
capabilities += device_cuda_capabilities();
}
}
#endif
return capabilities;
}
DeviceInfo Device::get_multi_device(const vector<DeviceInfo> &subdevices,
int threads,
bool background)
{
assert(subdevices.size() > 0);
if (subdevices.size() == 1) {
/* No multi device needed. */
return subdevices.front();
}
DeviceInfo info;
info.type = subdevices.front().type;
info.id = "MULTI";
info.description = "Multi Device";
info.num = 0;
info.has_half_images = true;
info.has_volume_decoupled = true;
info.has_adaptive_stop_per_sample = true;
info.has_osl = true;
info.has_profiling = true;
foreach (const DeviceInfo &device, subdevices) {
/* Ensure CPU device does not slow down GPU. */
if (device.type == DEVICE_CPU && subdevices.size() > 1) {
if (background) {
int orig_cpu_threads = (threads) ? threads : system_cpu_thread_count();
int cpu_threads = max(orig_cpu_threads - (subdevices.size() - 1), 0);
VLOG(1) << "CPU render threads reduced from " << orig_cpu_threads << " to " << cpu_threads
<< ", to dedicate to GPU.";
if (cpu_threads >= 1) {
DeviceInfo cpu_device = device;
cpu_device.cpu_threads = cpu_threads;
info.multi_devices.push_back(cpu_device);
}
else {
continue;
}
}
else {
VLOG(1) << "CPU render threads disabled for interactive render.";
continue;
}
}
else {
info.multi_devices.push_back(device);
}
/* Create unique ID for this combination of devices. */
info.id += device.id;
/* Set device type to MULTI if subdevices are not of a common type. */
if (device.type != info.type) {
info.type = DEVICE_MULTI;
}
/* Accumulate device info. */
info.has_half_images &= device.has_half_images;
info.has_volume_decoupled &= device.has_volume_decoupled;
info.has_adaptive_stop_per_sample &= device.has_adaptive_stop_per_sample;
info.has_osl &= device.has_osl;
info.has_profiling &= device.has_profiling;
}
return info;
}
void Device::tag_update()
{
free_memory();
}
void Device::free_memory()
{
devices_initialized_mask = 0;
cuda_devices.free_memory();
optix_devices.free_memory();
opencl_devices.free_memory();
cpu_devices.free_memory();
network_devices.free_memory();
}
CCL_NAMESPACE_END
|