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
|
/*
* 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.
*
* The Original Code is Copyright (C) 2017 by Blender Foundation.
* All rights reserved.
*/
/** \file
* \ingroup draw
*
* \brief Contains procedural GPU hair drawing methods.
*/
#include "DRW_render.h"
#include "BLI_utildefines.h"
#include "BLI_string_utils.h"
#include "DNA_modifier_types.h"
#include "DNA_particle_types.h"
#include "DNA_customdata_types.h"
#include "BKE_anim.h"
#include "GPU_batch.h"
#include "GPU_shader.h"
#include "draw_hair_private.h"
#ifndef __APPLE__
# define USE_TRANSFORM_FEEDBACK
#endif
typedef enum ParticleRefineShader {
PART_REFINE_CATMULL_ROM = 0,
PART_REFINE_MAX_SHADER,
} ParticleRefineShader;
#ifndef USE_TRANSFORM_FEEDBACK
typedef struct ParticleRefineCall {
struct ParticleRefineCall *next;
GPUVertBuf *vbo;
DRWShadingGroup *shgrp;
uint vert_len;
} ParticleRefineCall;
static ParticleRefineCall *g_tf_calls = NULL;
static int g_tf_id_offset;
static int g_tf_target_width;
static int g_tf_target_height;
#endif
static GPUShader *g_refine_shaders[PART_REFINE_MAX_SHADER] = {NULL};
static DRWPass *g_tf_pass; /* XXX can be a problem with multiple DRWManager in the future */
extern char datatoc_common_hair_lib_glsl[];
extern char datatoc_common_hair_refine_vert_glsl[];
extern char datatoc_gpu_shader_3D_smooth_color_frag_glsl[];
static GPUShader *hair_refine_shader_get(ParticleRefineShader sh)
{
if (g_refine_shaders[sh]) {
return g_refine_shaders[sh];
}
char *vert_with_lib = BLI_string_joinN(datatoc_common_hair_lib_glsl,
datatoc_common_hair_refine_vert_glsl);
#ifdef USE_TRANSFORM_FEEDBACK
const char *var_names[1] = {"finalColor"};
g_refine_shaders[sh] = DRW_shader_create_with_transform_feedback(
vert_with_lib, NULL, "#define HAIR_PHASE_SUBDIV\n", GPU_SHADER_TFB_POINTS, var_names, 1);
#else
g_refine_shaders[sh] = DRW_shader_create(vert_with_lib,
NULL,
datatoc_gpu_shader_3D_smooth_color_frag_glsl,
"#define HAIR_PHASE_SUBDIV\n"
"#define TF_WORKAROUND\n");
#endif
MEM_freeN(vert_with_lib);
return g_refine_shaders[sh];
}
void DRW_hair_init(void)
{
#ifdef USE_TRANSFORM_FEEDBACK
g_tf_pass = DRW_pass_create("Update Hair Pass", 0);
#else
g_tf_pass = DRW_pass_create("Update Hair Pass", DRW_STATE_WRITE_COLOR);
#endif
}
typedef struct DRWHairInstanceData {
DrawData dd;
float mat[4][4];
} DRWHairInstanceData;
static DRWShadingGroup *drw_shgroup_create_hair_procedural_ex(Object *object,
ParticleSystem *psys,
ModifierData *md,
DRWPass *hair_pass,
struct GPUMaterial *gpu_mat,
GPUShader *gpu_shader)
{
/* TODO(fclem): Pass the scene as parameter */
const DRWContextState *draw_ctx = DRW_context_state_get();
Scene *scene = draw_ctx->scene;
static float unit_mat[4][4] = {
{1, 0, 0, 0},
{0, 1, 0, 0},
{0, 0, 1, 0},
{0, 0, 0, 1},
};
float(*dupli_mat)[4];
Object *dupli_parent = DRW_object_get_dupli_parent(object);
DupliObject *dupli_object = DRW_object_get_dupli(object);
int subdiv = scene->r.hair_subdiv;
int thickness_res = (scene->r.hair_type == SCE_HAIR_SHAPE_STRAND) ? 1 : 2;
ParticleHairCache *hair_cache;
ParticleSettings *part = psys->part;
bool need_ft_update = particles_ensure_procedural_data(
object, psys, md, &hair_cache, subdiv, thickness_res);
DRWShadingGroup *shgrp;
if (gpu_mat) {
shgrp = DRW_shgroup_material_create(gpu_mat, hair_pass);
}
else if (gpu_shader) {
shgrp = DRW_shgroup_create(gpu_shader, hair_pass);
}
else {
shgrp = NULL;
BLI_assert(0);
}
/* TODO optimize this. Only bind the ones GPUMaterial needs. */
for (int i = 0; i < hair_cache->num_uv_layers; i++) {
for (int n = 0; n < MAX_LAYER_NAME_CT && hair_cache->uv_layer_names[i][n][0] != '\0'; n++) {
DRW_shgroup_uniform_texture(shgrp, hair_cache->uv_layer_names[i][n], hair_cache->uv_tex[i]);
}
}
for (int i = 0; i < hair_cache->num_col_layers; i++) {
for (int n = 0; n < MAX_LAYER_NAME_CT && hair_cache->col_layer_names[i][n][0] != '\0'; n++) {
DRW_shgroup_uniform_texture(
shgrp, hair_cache->col_layer_names[i][n], hair_cache->col_tex[i]);
}
}
if ((dupli_parent != NULL) && (dupli_object != NULL)) {
DRWHairInstanceData *hair_inst_data = (DRWHairInstanceData *)DRW_drawdata_ensure(
&object->id,
(DrawEngineType *)&drw_shgroup_create_hair_procedural_ex,
sizeof(DRWHairInstanceData),
NULL,
NULL);
dupli_mat = hair_inst_data->mat;
if (dupli_object->type & OB_DUPLICOLLECTION) {
copy_m4_m4(dupli_mat, dupli_parent->obmat);
}
else {
copy_m4_m4(dupli_mat, dupli_object->ob->obmat);
invert_m4(dupli_mat);
mul_m4_m4m4(dupli_mat, object->obmat, dupli_mat);
}
}
else {
dupli_mat = unit_mat;
}
DRW_shgroup_uniform_texture(shgrp, "hairPointBuffer", hair_cache->final[subdiv].proc_tex);
DRW_shgroup_uniform_int(shgrp, "hairStrandsRes", &hair_cache->final[subdiv].strands_res, 1);
DRW_shgroup_uniform_int_copy(shgrp, "hairThicknessRes", thickness_res);
DRW_shgroup_uniform_float(shgrp, "hairRadShape", &part->shape, 1);
DRW_shgroup_uniform_mat4(shgrp, "hairDupliMatrix", dupli_mat);
DRW_shgroup_uniform_float_copy(shgrp, "hairRadRoot", part->rad_root * part->rad_scale * 0.5f);
DRW_shgroup_uniform_float_copy(shgrp, "hairRadTip", part->rad_tip * part->rad_scale * 0.5f);
DRW_shgroup_uniform_bool_copy(
shgrp, "hairCloseTip", (part->shape_flag & PART_SHAPE_CLOSE_TIP) != 0);
/* TODO(fclem): Until we have a better way to cull the hair and render with orco, bypass culling
* test. */
GPUBatch *geom = hair_cache->final[subdiv].proc_hairs[thickness_res - 1];
DRW_shgroup_call_no_cull(shgrp, geom, object);
/* Transform Feedback subdiv. */
if (need_ft_update) {
int final_points_len = hair_cache->final[subdiv].strands_res * hair_cache->strands_len;
if (final_points_len) {
GPUShader *tf_shader = hair_refine_shader_get(PART_REFINE_CATMULL_ROM);
#ifdef USE_TRANSFORM_FEEDBACK
DRWShadingGroup *tf_shgrp = DRW_shgroup_transform_feedback_create(
tf_shader, g_tf_pass, hair_cache->final[subdiv].proc_buf);
#else
DRWShadingGroup *tf_shgrp = DRW_shgroup_create(tf_shader, g_tf_pass);
ParticleRefineCall *pr_call = MEM_mallocN(sizeof(*pr_call), __func__);
pr_call->next = g_tf_calls;
pr_call->vbo = hair_cache->final[subdiv].proc_buf;
pr_call->shgrp = tf_shgrp;
pr_call->vert_len = final_points_len;
g_tf_calls = pr_call;
DRW_shgroup_uniform_int(tf_shgrp, "targetHeight", &g_tf_target_height, 1);
DRW_shgroup_uniform_int(tf_shgrp, "targetWidth", &g_tf_target_width, 1);
DRW_shgroup_uniform_int(tf_shgrp, "idOffset", &g_tf_id_offset, 1);
#endif
DRW_shgroup_uniform_texture(tf_shgrp, "hairPointBuffer", hair_cache->point_tex);
DRW_shgroup_uniform_texture(tf_shgrp, "hairStrandBuffer", hair_cache->strand_tex);
DRW_shgroup_uniform_texture(tf_shgrp, "hairStrandSegBuffer", hair_cache->strand_seg_tex);
DRW_shgroup_uniform_int(
tf_shgrp, "hairStrandsRes", &hair_cache->final[subdiv].strands_res, 1);
DRW_shgroup_call_procedural_points(tf_shgrp, NULL, final_points_len);
}
}
return shgrp;
}
DRWShadingGroup *DRW_shgroup_hair_create(
Object *object, ParticleSystem *psys, ModifierData *md, DRWPass *hair_pass, GPUShader *shader)
{
return drw_shgroup_create_hair_procedural_ex(object, psys, md, hair_pass, NULL, shader);
}
DRWShadingGroup *DRW_shgroup_material_hair_create(Object *object,
ParticleSystem *psys,
ModifierData *md,
DRWPass *hair_pass,
struct GPUMaterial *material)
{
return drw_shgroup_create_hair_procedural_ex(object, psys, md, hair_pass, material, NULL);
}
void DRW_hair_update(void)
{
#ifndef USE_TRANSFORM_FEEDBACK
/**
* Workaround to transform feedback not working on mac.
* On some system it crashes (see T58489) and on some other it renders garbage (see T60171).
*
* So instead of using transform feedback we render to a texture,
* read back the result to system memory and re-upload as VBO data.
* It is really not ideal performance wise, but it is the simplest
* and the most local workaround that still uses the power of the GPU.
*/
if (g_tf_calls == NULL) {
return;
}
/* Search ideal buffer size. */
uint max_size = 0;
for (ParticleRefineCall *pr_call = g_tf_calls; pr_call; pr_call = pr_call->next) {
max_size = max_ii(max_size, pr_call->vert_len);
}
/* Create target Texture / Framebuffer */
/* Don't use max size as it can be really heavy and fail.
* Do chunks of maximum 2048 * 2048 hair points. */
int width = 2048;
int height = min_ii(width, 1 + max_size / width);
GPUTexture *tex = DRW_texture_pool_query_2d(width, height, GPU_RGBA32F, (void *)DRW_hair_update);
g_tf_target_height = height;
g_tf_target_width = width;
GPUFrameBuffer *fb = NULL;
GPU_framebuffer_ensure_config(&fb,
{
GPU_ATTACHMENT_NONE,
GPU_ATTACHMENT_TEXTURE(tex),
});
float *data = MEM_mallocN(sizeof(float) * 4 * width * height, "tf fallback buffer");
GPU_framebuffer_bind(fb);
while (g_tf_calls != NULL) {
ParticleRefineCall *pr_call = g_tf_calls;
g_tf_calls = g_tf_calls->next;
g_tf_id_offset = 0;
while (pr_call->vert_len > 0) {
int max_read_px_len = min_ii(width * height, pr_call->vert_len);
DRW_draw_pass_subset(g_tf_pass, pr_call->shgrp, pr_call->shgrp);
/* Readback result to main memory. */
GPU_framebuffer_read_color(fb, 0, 0, width, height, 4, 0, data);
/* Upload back to VBO. */
GPU_vertbuf_use(pr_call->vbo);
glBufferSubData(GL_ARRAY_BUFFER,
sizeof(float) * 4 * g_tf_id_offset,
sizeof(float) * 4 * max_read_px_len,
data);
g_tf_id_offset += max_read_px_len;
pr_call->vert_len -= max_read_px_len;
}
MEM_freeN(pr_call);
}
MEM_freeN(data);
GPU_framebuffer_free(fb);
#else
/* TODO(fclem): replace by compute shader. */
/* Just render using transform feedback. */
DRW_draw_pass(g_tf_pass);
#endif
}
void DRW_hair_free(void)
{
for (int i = 0; i < PART_REFINE_MAX_SHADER; i++) {
DRW_SHADER_FREE_SAFE(g_refine_shaders[i]);
}
}
|