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
|
/*
* 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.
*/
/* CUDA kernel entry points */
#include "../../kernel_compat_cuda.h"
#include "../../kernel_math.h"
#include "../../kernel_types.h"
#include "../../kernel_globals.h"
#include "../../kernel_film.h"
#include "../../kernel_path.h"
#include "../../kernel_path_branched.h"
#include "../../kernel_bake.h"
/* device data taken from CUDA occupancy calculator */
#ifdef __CUDA_ARCH__
/* 2.0 and 2.1 */
#if __CUDA_ARCH__ == 200 || __CUDA_ARCH__ == 210
# define CUDA_MULTIPRESSOR_MAX_REGISTERS 32768
# define CUDA_MULTIPROCESSOR_MAX_BLOCKS 8
# define CUDA_BLOCK_MAX_THREADS 1024
# define CUDA_THREAD_MAX_REGISTERS 63
/* tunable parameters */
# define CUDA_THREADS_BLOCK_WIDTH 16
# define CUDA_KERNEL_MAX_REGISTERS 32
# define CUDA_KERNEL_BRANCHED_MAX_REGISTERS 40
/* 3.0 and 3.5 */
#elif __CUDA_ARCH__ == 300 || __CUDA_ARCH__ == 350
# define CUDA_MULTIPRESSOR_MAX_REGISTERS 65536
# define CUDA_MULTIPROCESSOR_MAX_BLOCKS 16
# define CUDA_BLOCK_MAX_THREADS 1024
# define CUDA_THREAD_MAX_REGISTERS 63
/* tunable parameters */
# define CUDA_THREADS_BLOCK_WIDTH 16
# define CUDA_KERNEL_MAX_REGISTERS 63
# define CUDA_KERNEL_BRANCHED_MAX_REGISTERS 63
/* 3.2 */
#elif __CUDA_ARCH__ == 320
# define CUDA_MULTIPRESSOR_MAX_REGISTERS 32768
# define CUDA_MULTIPROCESSOR_MAX_BLOCKS 16
# define CUDA_BLOCK_MAX_THREADS 1024
# define CUDA_THREAD_MAX_REGISTERS 63
/* tunable parameters */
# define CUDA_THREADS_BLOCK_WIDTH 16
# define CUDA_KERNEL_MAX_REGISTERS 63
# define CUDA_KERNEL_BRANCHED_MAX_REGISTERS 63
/* 3.7 */
#elif __CUDA_ARCH__ == 370
# define CUDA_MULTIPRESSOR_MAX_REGISTERS 65536
# define CUDA_MULTIPROCESSOR_MAX_BLOCKS 16
# define CUDA_BLOCK_MAX_THREADS 1024
# define CUDA_THREAD_MAX_REGISTERS 255
/* tunable parameters */
# define CUDA_THREADS_BLOCK_WIDTH 16
# define CUDA_KERNEL_MAX_REGISTERS 63
# define CUDA_KERNEL_BRANCHED_MAX_REGISTERS 63
/* 5.0, 5.2, 5.3, 6.0, 6.1 */
#elif __CUDA_ARCH__ >= 500
# define CUDA_MULTIPRESSOR_MAX_REGISTERS 65536
# define CUDA_MULTIPROCESSOR_MAX_BLOCKS 32
# define CUDA_BLOCK_MAX_THREADS 1024
# define CUDA_THREAD_MAX_REGISTERS 255
/* tunable parameters */
# define CUDA_THREADS_BLOCK_WIDTH 16
# define CUDA_KERNEL_MAX_REGISTERS 48
# define CUDA_KERNEL_BRANCHED_MAX_REGISTERS 63
/* unknown architecture */
#else
# error "Unknown or unsupported CUDA architecture, can't determine launch bounds"
#endif
/* compute number of threads per block and minimum blocks per multiprocessor
* given the maximum number of registers per thread */
#define CUDA_LAUNCH_BOUNDS(threads_block_width, thread_num_registers) \
__launch_bounds__( \
threads_block_width*threads_block_width, \
CUDA_MULTIPRESSOR_MAX_REGISTERS/(threads_block_width*threads_block_width*thread_num_registers) \
)
/* sanity checks */
#if CUDA_THREADS_BLOCK_WIDTH*CUDA_THREADS_BLOCK_WIDTH > CUDA_BLOCK_MAX_THREADS
# error "Maximum number of threads per block exceeded"
#endif
#if CUDA_MULTIPRESSOR_MAX_REGISTERS/(CUDA_THREADS_BLOCK_WIDTH*CUDA_THREADS_BLOCK_WIDTH*CUDA_KERNEL_MAX_REGISTERS) > CUDA_MULTIPROCESSOR_MAX_BLOCKS
# error "Maximum number of blocks per multiprocessor exceeded"
#endif
#if CUDA_KERNEL_MAX_REGISTERS > CUDA_THREAD_MAX_REGISTERS
# error "Maximum number of registers per thread exceeded"
#endif
#if CUDA_KERNEL_BRANCHED_MAX_REGISTERS > CUDA_THREAD_MAX_REGISTERS
# error "Maximum number of registers per thread exceeded"
#endif
/* kernels */
extern "C" __global__ void
CUDA_LAUNCH_BOUNDS(CUDA_THREADS_BLOCK_WIDTH, CUDA_KERNEL_MAX_REGISTERS)
kernel_cuda_path_trace(float *buffer, uint *rng_state, int sample, int sx, int sy, int sw, int sh, int offset, int stride)
{
int x = sx + blockDim.x*blockIdx.x + threadIdx.x;
int y = sy + blockDim.y*blockIdx.y + threadIdx.y;
if(x < sx + sw && y < sy + sh)
kernel_path_trace(NULL, buffer, rng_state, sample, x, y, offset, stride);
}
#ifdef __BRANCHED_PATH__
extern "C" __global__ void
CUDA_LAUNCH_BOUNDS(CUDA_THREADS_BLOCK_WIDTH, CUDA_KERNEL_BRANCHED_MAX_REGISTERS)
kernel_cuda_branched_path_trace(float *buffer, uint *rng_state, int sample, int sx, int sy, int sw, int sh, int offset, int stride)
{
int x = sx + blockDim.x*blockIdx.x + threadIdx.x;
int y = sy + blockDim.y*blockIdx.y + threadIdx.y;
if(x < sx + sw && y < sy + sh)
kernel_branched_path_trace(NULL, buffer, rng_state, sample, x, y, offset, stride);
}
#endif
extern "C" __global__ void
CUDA_LAUNCH_BOUNDS(CUDA_THREADS_BLOCK_WIDTH, CUDA_KERNEL_MAX_REGISTERS)
kernel_cuda_convert_to_byte(uchar4 *rgba, float *buffer, float sample_scale, int sx, int sy, int sw, int sh, int offset, int stride)
{
int x = sx + blockDim.x*blockIdx.x + threadIdx.x;
int y = sy + blockDim.y*blockIdx.y + threadIdx.y;
if(x < sx + sw && y < sy + sh)
kernel_film_convert_to_byte(NULL, rgba, buffer, sample_scale, x, y, offset, stride);
}
extern "C" __global__ void
CUDA_LAUNCH_BOUNDS(CUDA_THREADS_BLOCK_WIDTH, CUDA_KERNEL_MAX_REGISTERS)
kernel_cuda_convert_to_half_float(uchar4 *rgba, float *buffer, float sample_scale, int sx, int sy, int sw, int sh, int offset, int stride)
{
int x = sx + blockDim.x*blockIdx.x + threadIdx.x;
int y = sy + blockDim.y*blockIdx.y + threadIdx.y;
if(x < sx + sw && y < sy + sh)
kernel_film_convert_to_half_float(NULL, rgba, buffer, sample_scale, x, y, offset, stride);
}
extern "C" __global__ void
CUDA_LAUNCH_BOUNDS(CUDA_THREADS_BLOCK_WIDTH, CUDA_KERNEL_MAX_REGISTERS)
kernel_cuda_shader(uint4 *input,
float4 *output,
float *output_luma,
int type,
int sx,
int sw,
int offset,
int sample)
{
int x = sx + blockDim.x*blockIdx.x + threadIdx.x;
if(x < sx + sw) {
kernel_shader_evaluate(NULL,
input,
output,
output_luma,
(ShaderEvalType)type,
x,
sample);
}
}
#ifdef __BAKING__
extern "C" __global__ void
CUDA_LAUNCH_BOUNDS(CUDA_THREADS_BLOCK_WIDTH, CUDA_KERNEL_MAX_REGISTERS)
kernel_cuda_bake(uint4 *input, float4 *output, int type, int filter, int sx, int sw, int offset, int sample)
{
int x = sx + blockDim.x*blockIdx.x + threadIdx.x;
if(x < sx + sw)
kernel_bake_evaluate(NULL, input, output, (ShaderEvalType)type, filter, x, offset, sample);
}
#endif
#endif
|
