Cr8-xyz/blender
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

Cycles: Restructure kernel files organization

Browse Source 
Since the kernel split work we're now having quite a few of new files, majority
of which are related on the kernel entry points. Keeping those files in the
root kernel folder will eventually make it really hard to follow which files are
actual implementation of Cycles kernel.

Those files are now moved to kernel/kernels/<device_type>. This way adding extra
entry points will be less noisy. It is also nice to have all device-specific
files grouped together.

Another change is in the way how split kernel invokes logic. Previously all the
logic was implemented directly in the .cl files, which makes it a bit tricky to
re-use the logic across other devices. Since we'll likely be looking into doing
same split work for CUDA devices eventually it makes sense to move logic from
.cl files to header files. Those files are stored in kernel/split. This does not
mean the header files will not give error messages when tried to be included
from other devices and their arguments will likely be changed, but having such
separation is a good start anyway.

There should be no functional changes.

Reviewers: juicyfruit, dingto

Differential Revision: https://developer.blender.org/D1314
This commit is contained in:
Sergey Sharybin
2015-05-21 17:40:04 +05:00
parent 7f4d5850fe
commit 2c503d8303
38 changed files with 1090 additions and 313 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
SConstruct
View File

@@ -1028,7 +1028,6 @@ if env['OURPLATFORM']!='darwin':
dir=os.path.join(env['BF_INSTALLDIR'], VERSION, 'scripts', 'addons','cycles', 'kernel')
source=os.listdir('intern/cycles/kernel')
if '__pycache__' in source: source.remove('__pycache__')
source.remove('kernel.cpp')
source.remove('CMakeLists.txt')
source.remove('SConscript')
source.remove('svm')
@@ -1036,6 +1035,7 @@ if env['OURPLATFORM']!='darwin':
source.remove('geom')
source.remove('shaders')
source.remove('osl')
source.remove('split')
source=['intern/cycles/kernel/'+s for s in source]
source.append('intern/cycles/util/util_atomic.h')
source.append('intern/cycles/util/util_color.h')
@@ -1063,6 +1063,12 @@ if env['OURPLATFORM']!='darwin':
if '__pycache__' in source: source.remove('__pycache__')
source=['intern/cycles/kernel/geom/'+s for s in source]
scriptinstall.append(env.Install(dir=dir,source=source))
# split
dir=os.path.join(env['BF_INSTALLDIR'], VERSION, 'scripts', 'addons','cycles', 'kernel', 'split')
source=os.listdir('intern/cycles/kernel/split')
if '__pycache__' in source: source.remove('__pycache__')
source=['intern/cycles/kernel/split/'+s for s in source]
scriptinstall.append(env.Install(dir=dir,source=source))
# licenses
dir=os.path.join(env['BF_INSTALLDIR'], VERSION, 'scripts', 'addons','cycles', 'license')

8
intern/cycles/SConscript
View File

@@ -34,12 +34,8 @@ cycles.Depends('../../source/blender/makesrna/intern/RNA_blender_cpp.h', 'makesr
sources = cycles.Glob('bvh/*.cpp') + cycles.Glob('device/*.cpp') + cycles.Glob('kernel/*.cpp') + cycles.Glob('render/*.cpp') + cycles.Glob('subd/*.cpp') + cycles.Glob('util/*.cpp') + cycles.Glob('blender/*.cpp')
sources.append(path.join('kernel', 'kernels', 'cpu', 'kernel.cpp'))
sources.remove(path.join('util', 'util_view.cpp'))
sources.remove(path.join('kernel', 'kernel_sse2.cpp'))
sources.remove(path.join('kernel', 'kernel_sse3.cpp'))
sources.remove(path.join('kernel', 'kernel_sse41.cpp'))
sources.remove(path.join('kernel', 'kernel_avx.cpp'))
sources.remove(path.join('kernel', 'kernel_avx2.cpp'))
incs = []
defs = []
@@ -146,7 +142,7 @@ for kernel_type in kernel_flags.keys():
defs.append('WITH_KERNEL_' + kernel_type.upper())
for kernel_type in kernel_flags.keys():
kernel_source = path.join('kernel', 'kernel_' + kernel_type + '.cpp')
kernel_source = path.join('kernel', 'kernels', 'cpu', 'kernel_' + kernel_type + '.cpp')
kernel_cxxflags = Split(env['CXXFLAGS'])
kernel_cxxflags.append(kernel_flags[kernel_type].split())
kernel_defs = defs[:]

2
intern/cycles/device/device_cuda.cpp
View File

@@ -266,7 +266,7 @@ public:
printf("CUDA version %d.%d detected, build may succeed but only CUDA 6.5 is officially supported.\n", cuda_version/10, cuda_version%10);
/* compile */
string kernel = path_join(kernel_path, "kernel.cu");
string kernel = path_join(kernel_path, path_join("kernels", path_join("cuda", "kernel.cu")));
string include = kernel_path;
const int machine = system_cpu_bits();

6
intern/cycles/device/device_opencl.cpp
View File

@@ -748,7 +748,7 @@ public:
}
else {
string init_kernel_source = "#include \"kernel.cl\" // " + kernel_md5 + "\n";
string init_kernel_source = "#include \"kernels/opencl/kernel.cl\" // " + kernel_md5 + "\n";
/* if does not exist or loading binary failed, compile kernel */
if(!compile_kernel(kernel_path, init_kernel_source, "", &cpProgram, debug_src))
@@ -1322,7 +1322,7 @@ public:
/* Kernel loaded from binary, nothing to do. */
}
else {
string init_kernel_source = "#include \"kernel.cl\" // " +
string init_kernel_source = "#include \"kernels/opencl/kernel.cl\" // " +
kernel_md5 + "\n";
/* If does not exist or loading binary failed, compile kernel. */
if(!compile_kernel(kernel_path,
@@ -1996,7 +1996,7 @@ public:
#define GLUE(a, b) a ## b
#define LOAD_KERNEL(name) \
do { \
kernel_init_source = "#include \"kernel_" #name ".cl\" // " + \
kernel_init_source = "#include \"kernels/opencl/kernel_" #name ".cl\" // " + \
kernel_md5 + "\n"; \
device_md5 = device_md5_hash(build_options); \
clbin = string_printf("cycles_kernel_%s_%s_" #name ".clbin", \

96
intern/cycles/kernel/CMakeLists.txt
View File

@@ -12,20 +12,20 @@ set(INC_SYS
)
set(SRC
kernel.cpp
kernel.cl
kernel_data_init.cl
kernel_queue_enqueue.cl
kernel_scene_intersect.cl
kernel_lamp_emission.cl
kernel_background_buffer_update.cl
kernel_shader_eval.cl
kernel_holdout_emission_blurring_pathtermination_ao.cl
kernel_direct_lighting.cl
kernel_shadow_blocked.cl
kernel_next_iteration_setup.cl
kernel_sum_all_radiance.cl
kernel.cu
kernels/cpu/kernel.cpp
kernels/opencl/kernel.cl
kernels/opencl/kernel_data_init.cl
kernels/opencl/kernel_queue_enqueue.cl
kernels/opencl/kernel_scene_intersect.cl
kernels/opencl/kernel_lamp_emission.cl
kernels/opencl/kernel_background_buffer_update.cl
kernels/opencl/kernel_shader_eval.cl
kernels/opencl/kernel_holdout_emission_blurring_pathtermination_ao.cl
kernels/opencl/kernel_direct_lighting.cl
kernels/opencl/kernel_shadow_blocked.cl
kernels/opencl/kernel_next_iteration_setup.cl
kernels/opencl/kernel_sum_all_radiance.cl
kernels/cuda/kernel.cu
)
set(SRC_HEADERS
@@ -57,7 +57,6 @@ set(SRC_HEADERS
kernel_shader.h
kernel_shaderdata_vars.h
kernel_shadow.h
kernel_split.h
kernel_subsurface.h
kernel_textures.h
kernel_types.h
@@ -162,6 +161,22 @@ set(SRC_UTIL_HEADERS
../util/util_transform.h
../util/util_types.h
)
set(SRC_SPLIT_HEADERS
split/kernel_background_buffer_update.h
split/kernel_data_init.h
split/kernel_direct_lighting.h
split/kernel_holdout_emission_blurring_pathtermination_ao.h
split/kernel_lamp_emission.h
split/kernel_next_iteration_setup.h
split/kernel_queue_enqueue.h
split/kernel_scene_intersect.h
split/kernel_shader_eval.h
split/kernel_shadow_blocked.h
split/kernel_split_common.h
split/kernel_sum_all_radiance.h
)
# CUDA module
if(WITH_CYCLES_CUDA_BINARIES)
@@ -187,7 +202,7 @@ if(WITH_CYCLES_CUDA_BINARIES)
endif()
# build for each arch
set(cuda_sources kernel.cu ${SRC_HEADERS} ${SRC_SVM_HEADERS} ${SRC_GEOM_HEADERS} ${SRC_CLOSURE_HEADERS} ${SRC_UTIL_HEADERS})
set(cuda_sources kernels/cuda/kernel.cu ${SRC_HEADERS} ${SRC_SVM_HEADERS} ${SRC_GEOM_HEADERS} ${SRC_CLOSURE_HEADERS} ${SRC_UTIL_HEADERS})
set(cuda_cubins)
macro(CYCLES_CUDA_KERNEL_ADD arch experimental)
@@ -213,7 +228,7 @@ if(WITH_CYCLES_CUDA_BINARIES)
COMMAND ${CUDA_NVCC_EXECUTABLE}
-arch=${arch}
-m${CUDA_BITS}
--cubin ${CMAKE_CURRENT_SOURCE_DIR}/kernel.cu
--cubin ${CMAKE_CURRENT_SOURCE_DIR}/kernels/cuda/kernel.cu
-o ${CMAKE_CURRENT_BINARY_DIR}/${cuda_cubin}
--ptxas-options="-v"
${cuda_arch_flags}
@@ -261,28 +276,28 @@ include_directories(SYSTEM ${INC_SYS})
if(CXX_HAS_SSE)
list(APPEND SRC
kernel_sse2.cpp
kernel_sse3.cpp
kernel_sse41.cpp
kernels/cpu/kernel_sse2.cpp
kernels/cpu/kernel_sse3.cpp
kernels/cpu/kernel_sse41.cpp
)
set_source_files_properties(kernel_sse2.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_SSE2_KERNEL_FLAGS}")
set_source_files_properties(kernel_sse3.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_SSE3_KERNEL_FLAGS}")
set_source_files_properties(kernel_sse41.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_SSE41_KERNEL_FLAGS}")
set_source_files_properties(kernels/cpu/kernel_sse2.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_SSE2_KERNEL_FLAGS}")
set_source_files_properties(kernels/cpu/kernel_sse3.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_SSE3_KERNEL_FLAGS}")
set_source_files_properties(kernels/cpu/kernel_sse41.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_SSE41_KERNEL_FLAGS}")
endif()
if(CXX_HAS_AVX)
list(APPEND SRC
kernel_avx.cpp
kernels/cpu/kernel_avx.cpp
)
set_source_files_properties(kernel_avx.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_AVX_KERNEL_FLAGS}")
set_source_files_properties(kernels/cpu/kernel_avx.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_AVX_KERNEL_FLAGS}")
endif()
if(CXX_HAS_AVX2)
list(APPEND SRC
kernel_avx2.cpp
kernels/cpu/kernel_avx2.cpp
)
set_source_files_properties(kernel_avx2.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_AVX2_KERNEL_FLAGS}")
set_source_files_properties(kernels/cpu/kernel_avx2.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_AVX2_KERNEL_FLAGS}")
endif()
add_library(cycles_kernel ${SRC} ${SRC_HEADERS} ${SRC_CLOSURE_HEADERS} ${SRC_SVM_HEADERS} ${SRC_GEOM_HEADERS})
@@ -301,22 +316,23 @@ endif()
#add_custom_target(cycles_kernel_preprocess ALL DEPENDS ${KERNEL_PREPROCESSED})
#delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${KERNEL_PREPROCESSED}" ${CYCLES_INSTALL_PATH}/kernel)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernel.cl" ${CYCLES_INSTALL_PATH}/kernel)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernel_data_init.cl" ${CYCLES_INSTALL_PATH}/kernel)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernel_queue_enqueue.cl" ${CYCLES_INSTALL_PATH}/kernel)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernel_scene_intersect.cl" ${CYCLES_INSTALL_PATH}/kernel)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernel_lamp_emission.cl" ${CYCLES_INSTALL_PATH}/kernel)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernel_background_buffer_update.cl" ${CYCLES_INSTALL_PATH}/kernel)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernel_shader_eval.cl" ${CYCLES_INSTALL_PATH}/kernel)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernel_holdout_emission_blurring_pathtermination_ao.cl" ${CYCLES_INSTALL_PATH}/kernel)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernel_direct_lighting.cl" ${CYCLES_INSTALL_PATH}/kernel)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernel_shadow_blocked.cl" ${CYCLES_INSTALL_PATH}/kernel)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernel_next_iteration_setup.cl" ${CYCLES_INSTALL_PATH}/kernel)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernel_sum_all_radiance.cl" ${CYCLES_INSTALL_PATH}/kernel)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernel.cu" ${CYCLES_INSTALL_PATH}/kernel)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernels/opencl/kernel.cl" ${CYCLES_INSTALL_PATH}/kernel/kernels/opencl)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernels/opencl/kernel_data_init.cl" ${CYCLES_INSTALL_PATH}/kernel/kernels/opencl)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernels/opencl/kernel_queue_enqueue.cl" ${CYCLES_INSTALL_PATH}/kernel/kernels/opencl)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernels/opencl/kernel_scene_intersect.cl" ${CYCLES_INSTALL_PATH}/kernel/kernels/opencl)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernels/opencl/kernel_lamp_emission.cl" ${CYCLES_INSTALL_PATH}/kernel/kernels/opencl)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernels/opencl/kernel_background_buffer_update.cl" ${CYCLES_INSTALL_PATH}/kernel/kernels/opencl)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernels/opencl/kernel_shader_eval.cl" ${CYCLES_INSTALL_PATH}/kernel/kernels/opencl)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernels/opencl/kernel_holdout_emission_blurring_pathtermination_ao.cl" ${CYCLES_INSTALL_PATH}/kernel/kernels/opencl)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernels/opencl/kernel_direct_lighting.cl" ${CYCLES_INSTALL_PATH}/kernel/kernels/opencl)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernels/opencl/kernel_shadow_blocked.cl" ${CYCLES_INSTALL_PATH}/kernel/kernels/opencl)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernels/opencl/kernel_next_iteration_setup.cl" ${CYCLES_INSTALL_PATH}/kernel/kernels/opencl)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernels/opencl/kernel_sum_all_radiance.cl" ${CYCLES_INSTALL_PATH}/kernel/kernels/opencl)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "kernels/cuda/kernel.cu" ${CYCLES_INSTALL_PATH}/kernel/kernels/cuda)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_HEADERS}" ${CYCLES_INSTALL_PATH}/kernel)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_CLOSURE_HEADERS}" ${CYCLES_INSTALL_PATH}/kernel/closure)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_SVM_HEADERS}" ${CYCLES_INSTALL_PATH}/kernel/svm)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_GEOM_HEADERS}" ${CYCLES_INSTALL_PATH}/kernel/geom)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_UTIL_HEADERS}" ${CYCLES_INSTALL_PATH}/kernel)
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_SPLIT_HEADERS}" ${CYCLES_INSTALL_PATH}/kernel/split)

7
intern/cycles/kernel/SConscript
View File

@@ -57,8 +57,9 @@ if env['WITH_BF_CYCLES_CUDA_BINARIES']:
build_dir = os.path.join(root_build_dir, 'intern/cycles/kernel')
# source directories and files
kernel_file_rel = os.path.join("kernels", "cuda", "kernel.cu")
source_dir = Dir('.').srcnode().path
kernel_file = os.path.join(source_dir, "kernel.cu")
kernel_file = os.path.join(source_dir, kernel_file_rel)
util_dir = os.path.join(source_dir, "../util")
svm_dir = os.path.join(source_dir, "../svm")
geom_dir = os.path.join(source_dir, "../geom")
@@ -83,7 +84,7 @@ if env['WITH_BF_CYCLES_CUDA_BINARIES']:
nvcc_flags += " -D__KERNEL_DEBUG__"
# dependencies
dependencies = ['kernel.cu'] + kernel.Glob('*.h') + kernel.Glob('../util/*.h') + kernel.Glob('svm/*.h') + kernel.Glob('geom/*.h') + kernel.Glob('closure/*.h')
dependencies = [kernel_file_rel] + kernel.Glob('*.h') + kernel.Glob('../util/*.h') + kernel.Glob('svm/*.h') + kernel.Glob('geom/*.h') + kernel.Glob('closure/*.h')
last_cubin_file = None
configs = (("kernel_%s.cubin", ''),
@@ -105,7 +106,7 @@ if env['WITH_BF_CYCLES_CUDA_BINARIES']:
else:
command = "\"%s\" -arch=%s %s \"%s\" -o \"%s\"" % (nvcc, arch, current_flags, kernel_file, cubin_file)
kernel.Command(cubin_file, 'kernel.cu', command)
kernel.Command(cubin_file, kernel_file_rel, command)
kernel.Depends(cubin_file, dependencies)
kernel_binaries.append(cubin_file)

62
intern/cycles/kernel/kernel_split.h
View File

@@ -1,62 +0,0 @@
/*
* Copyright 2011-2015 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.
*/
#ifndef _KERNEL_SPLIT_H_
#define _KERNEL_SPLIT_H_
#include "kernel_compat_opencl.h"
#include "kernel_math.h"
#include "kernel_types.h"
#include "kernel_globals.h"
#include "util_atomic.h"
#include "kernel_random.h"
#include "kernel_projection.h"
#include "kernel_montecarlo.h"
#include "kernel_differential.h"
#include "kernel_camera.h"
#include "geom/geom.h"
#include "kernel_accumulate.h"
#include "kernel_shader.h"
#include "kernel_light.h"
#include "kernel_passes.h"
#ifdef __SUBSURFACE__
#include "kernel_subsurface.h"
#endif
#ifdef __VOLUME__
#include "kernel_volume.h"
#endif
#include "kernel_path_state.h"
#include "kernel_shadow.h"
#include "kernel_emission.h"
#include "kernel_path_common.h"
#include "kernel_path_surface.h"
#include "kernel_path_volume.h"
#ifdef __KERNEL_DEBUG__
#include "kernel_debug.h"
#endif
#include "kernel_queues.h"
#include "kernel_work_stealing.h"
#endif

0
intern/cycles/kernel/kernel.cpp → intern/cycles/kernel/kernels/cpu/kernel.cpp
View File

0
intern/cycles/kernel/kernel_avx.cpp → intern/cycles/kernel/kernels/cpu/kernel_avx.cpp
View File

0
intern/cycles/kernel/kernel_avx2.cpp → intern/cycles/kernel/kernels/cpu/kernel_avx2.cpp
View File

0
intern/cycles/kernel/kernel_sse2.cpp → intern/cycles/kernel/kernels/cpu/kernel_sse2.cpp
View File

0
intern/cycles/kernel/kernel_sse3.cpp → intern/cycles/kernel/kernels/cpu/kernel_sse3.cpp
View File

0
intern/cycles/kernel/kernel_sse41.cpp → intern/cycles/kernel/kernels/cpu/kernel_sse41.cpp
View File

14
intern/cycles/kernel/kernel.cu → intern/cycles/kernel/kernels/cuda/kernel.cu
View File

@@ -16,13 +16,13 @@
/* 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_bake.h"
#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_bake.h"
/* device data taken from CUDA occupancy calculator */

34
intern/cycles/kernel/kernel.cl → intern/cycles/kernel/kernels/opencl/kernel.cl
View File

@@ -16,14 +16,14 @@
/* OpenCL kernel entry points - unfinished */
#include "kernel_compat_opencl.h"
#include "kernel_math.h"
#include "kernel_types.h"
#include "kernel_globals.h"
#include "../../kernel_compat_opencl.h"
#include "../../kernel_math.h"
#include "../../kernel_types.h"
#include "../../kernel_globals.h"
#include "kernel_film.h"
#include "kernel_path.h"
#include "kernel_bake.h"
#include "../../kernel_film.h"
#include "../../kernel_path.h"
#include "../../kernel_bake.h"
#ifdef __COMPILE_ONLY_MEGAKERNEL__
@@ -34,7 +34,7 @@ __kernel void kernel_ocl_path_trace(
#define KERNEL_TEX(type, ttype, name) \
ccl_global type *name,
#include "kernel_textures.h"
#include "../../kernel_textures.h"
int sample,
int sx, int sy, int sw, int sh, int offset, int stride)
@@ -45,7 +45,7 @@ __kernel void kernel_ocl_path_trace(
#define KERNEL_TEX(type, ttype, name) \
kg->name = name;
#include "kernel_textures.h"
#include "../../kernel_textures.h"
int x = sx + get_global_id(0);
int y = sy + get_global_id(1);
@@ -63,7 +63,7 @@ __kernel void kernel_ocl_shader(
#define KERNEL_TEX(type, ttype, name) \
ccl_global type *name,
#include "kernel_textures.h"
#include "../../kernel_textures.h"
int type, int sx, int sw, int offset, int sample)
{
@@ -73,7 +73,7 @@ __kernel void kernel_ocl_shader(
#define KERNEL_TEX(type, ttype, name) \
kg->name = name;
#include "kernel_textures.h"
#include "../../kernel_textures.h"
int x = sx + get_global_id(0);
@@ -88,7 +88,7 @@ __kernel void kernel_ocl_bake(
#define KERNEL_TEX(type, ttype, name) \
ccl_global type *name,
#include "kernel_textures.h"
#include "../../kernel_textures.h"
int type, int sx, int sw, int offset, int sample)
{
@@ -98,7 +98,7 @@ __kernel void kernel_ocl_bake(
#define KERNEL_TEX(type, ttype, name) \
kg->name = name;
#include "kernel_textures.h"
#include "../../kernel_textures.h"
int x = sx + get_global_id(0);
@@ -124,7 +124,7 @@ __kernel void kernel_ocl_convert_to_byte(
#define KERNEL_TEX(type, ttype, name) \
ccl_global type *name,
#include "kernel_textures.h"
#include "../../kernel_textures.h"
float sample_scale,
int sx, int sy, int sw, int sh, int offset, int stride)
@@ -135,7 +135,7 @@ __kernel void kernel_ocl_convert_to_byte(
#define KERNEL_TEX(type, ttype, name) \
kg->name = name;
#include "kernel_textures.h"
#include "../../kernel_textures.h"
int x = sx + get_global_id(0);
int y = sy + get_global_id(1);
@@ -151,7 +151,7 @@ __kernel void kernel_ocl_convert_to_half_float(
#define KERNEL_TEX(type, ttype, name) \
ccl_global type *name,
#include "kernel_textures.h"
#include "../../kernel_textures.h"
float sample_scale,
int sx, int sy, int sw, int sh, int offset, int stride)
@@ -162,7 +162,7 @@ __kernel void kernel_ocl_convert_to_half_float(
#define KERNEL_TEX(type, ttype, name) \
kg->name = name;
#include "kernel_textures.h"
#include "../../kernel_textures.h"
int x = sx + get_global_id(0);
int y = sy + get_global_id(1);

81
intern/cycles/kernel/kernels/opencl/kernel_background_buffer_update.cl Normal file
View File

@@ -0,0 +1,81 @@
/*
* Copyright 2011-2015 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 "split/kernel_background_buffer_update.h"
__kernel void kernel_ocl_path_trace_background_buffer_update(
ccl_global char *globals,
ccl_constant KernelData *data,
ccl_global char *shader_data,
ccl_global float *per_sample_output_buffers,
ccl_global uint *rng_state,
ccl_global uint *rng_coop, /* Required for buffer Update */
ccl_global float3 *throughput_coop, /* Required for background hit processing */
PathRadiance *PathRadiance_coop, /* Required for background hit processing and buffer Update */
ccl_global Ray *Ray_coop, /* Required for background hit processing */
ccl_global PathState *PathState_coop, /* Required for background hit processing */
ccl_global float *L_transparent_coop, /* Required for background hit processing and buffer Update */
ccl_global char *ray_state, /* Stores information on the current state of a ray */
int sw, int sh, int sx, int sy, int stride,
int rng_state_offset_x,
int rng_state_offset_y,
int rng_state_stride,
ccl_global unsigned int *work_array, /* Denotes work of each ray */
ccl_global int *Queue_data, /* Queues memory */
ccl_global int *Queue_index, /* Tracks the number of elements in each queue */
int queuesize, /* Size (capacity) of each queue */
int end_sample,
int start_sample,
#ifdef __WORK_STEALING__
ccl_global unsigned int *work_pool_wgs,
unsigned int num_samples,
#endif
#ifdef __KERNEL_DEBUG__
DebugData *debugdata_coop,
#endif
int parallel_samples) /* Number of samples to be processed in parallel */
{
kernel_background_buffer_update(globals,
data,
shader_data,
per_sample_output_buffers,
rng_state,
rng_coop,
throughput_coop,
PathRadiance_coop,
Ray_coop,
PathState_coop,
L_transparent_coop,
ray_state,
sw, sh, sx, sy, stride,
rng_state_offset_x,
rng_state_offset_y,
rng_state_stride,
work_array,
Queue_data,
Queue_index,
queuesize,
end_sample,
start_sample,
#ifdef __WORK_STEALING__
work_pool_wgs,
num_samples,
#endif
#ifdef __KERNEL_DEBUG__
debugdata_coop,
#endif
parallel_samples);
}

242
intern/cycles/kernel/kernels/opencl/kernel_data_init.cl Normal file
View File

@@ -0,0 +1,242 @@
/*
* Copyright 2011-2015 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 "split/kernel_data_init.h"
__kernel void kernel_ocl_path_trace_data_init(
ccl_global char *globals,
ccl_global char *shader_data_sd, /* Arguments related to ShaderData */
ccl_global char *shader_data_sd_DL_shadow, /* Arguments related to ShaderData */
ccl_global float3 *P_sd,
ccl_global float3 *P_sd_DL_shadow,
ccl_global float3 *N_sd,
ccl_global float3 *N_sd_DL_shadow,
ccl_global float3 *Ng_sd,
ccl_global float3 *Ng_sd_DL_shadow,
ccl_global float3 *I_sd,
ccl_global float3 *I_sd_DL_shadow,
ccl_global int *shader_sd,
ccl_global int *shader_sd_DL_shadow,
ccl_global int *flag_sd,
ccl_global int *flag_sd_DL_shadow,
ccl_global int *prim_sd,
ccl_global int *prim_sd_DL_shadow,
ccl_global int *type_sd,
ccl_global int *type_sd_DL_shadow,
ccl_global float *u_sd,
ccl_global float *u_sd_DL_shadow,
ccl_global float *v_sd,
ccl_global float *v_sd_DL_shadow,
ccl_global int *object_sd,
ccl_global int *object_sd_DL_shadow,
ccl_global float *time_sd,
ccl_global float *time_sd_DL_shadow,
ccl_global float *ray_length_sd,
ccl_global float *ray_length_sd_DL_shadow,
ccl_global int *ray_depth_sd,
ccl_global int *ray_depth_sd_DL_shadow,
ccl_global int *transparent_depth_sd,
ccl_global int *transparent_depth_sd_DL_shadow,
/* Ray differentials. */
ccl_global differential3 *dP_sd,
ccl_global differential3 *dP_sd_DL_shadow,
ccl_global differential3 *dI_sd,
ccl_global differential3 *dI_sd_DL_shadow,
ccl_global differential *du_sd,
ccl_global differential *du_sd_DL_shadow,
ccl_global differential *dv_sd,
ccl_global differential *dv_sd_DL_shadow,
/* Dp/Du */
ccl_global float3 *dPdu_sd,
ccl_global float3 *dPdu_sd_DL_shadow,
ccl_global float3 *dPdv_sd,
ccl_global float3 *dPdv_sd_DL_shadow,
/* Object motion. */
ccl_global Transform *ob_tfm_sd,
ccl_global Transform *ob_tfm_sd_DL_shadow,
ccl_global Transform *ob_itfm_sd,
ccl_global Transform *ob_itfm_sd_DL_shadow,
ShaderClosure *closure_sd,
ShaderClosure *closure_sd_DL_shadow,
ccl_global int *num_closure_sd,
ccl_global int *num_closure_sd_DL_shadow,
ccl_global float *randb_closure_sd,
ccl_global float *randb_closure_sd_DL_shadow,
ccl_global float3 *ray_P_sd,
ccl_global float3 *ray_P_sd_DL_shadow,
ccl_global differential3 *ray_dP_sd,
ccl_global differential3 *ray_dP_sd_DL_shadow,
ccl_constant KernelData *data,
ccl_global float *per_sample_output_buffers,
ccl_global uint *rng_state,
ccl_global uint *rng_coop, /* rng array to store rng values for all rays */
ccl_global float3 *throughput_coop, /* throughput array to store throughput values for all rays */
ccl_global float *L_transparent_coop, /* L_transparent array to store L_transparent values for all rays */
PathRadiance *PathRadiance_coop, /* PathRadiance array to store PathRadiance values for all rays */
ccl_global Ray *Ray_coop, /* Ray array to store Ray information for all rays */
ccl_global PathState *PathState_coop, /* PathState array to store PathState information for all rays */
ccl_global char *ray_state, /* Stores information on current state of a ray */
#define KERNEL_TEX(type, ttype, name) \
ccl_global type *name,
#include "../../kernel_textures.h"
int start_sample, int sx, int sy, int sw, int sh, int offset, int stride,
int rng_state_offset_x,
int rng_state_offset_y,
int rng_state_stride,
ccl_global int *Queue_data, /* Memory for queues */
ccl_global int *Queue_index, /* Tracks the number of elements in queues */
int queuesize, /* size (capacity) of the queue */
ccl_global char *use_queues_flag, /* flag to decide if scene-intersect kernel should use queues to fetch ray index */
ccl_global unsigned int *work_array, /* work array to store which work each ray belongs to */
#ifdef __WORK_STEALING__
ccl_global unsigned int *work_pool_wgs, /* Work pool for each work group */
unsigned int num_samples, /* Total number of samples per pixel */
#endif
#ifdef __KERNEL_DEBUG__
DebugData *debugdata_coop,
#endif
int parallel_samples /* Number of samples to be processed in parallel */
)
{
kernel_data_init(globals,
shader_data_sd,
shader_data_sd_DL_shadow,
P_sd,
P_sd_DL_shadow,
N_sd,
N_sd_DL_shadow,
Ng_sd,
Ng_sd_DL_shadow,
I_sd,
I_sd_DL_shadow,
shader_sd,
shader_sd_DL_shadow,
flag_sd,
flag_sd_DL_shadow,
prim_sd,
prim_sd_DL_shadow,
type_sd,
type_sd_DL_shadow,
u_sd,
u_sd_DL_shadow,
v_sd,
v_sd_DL_shadow,
object_sd,
object_sd_DL_shadow,
time_sd,
time_sd_DL_shadow,
ray_length_sd,
ray_length_sd_DL_shadow,
ray_depth_sd,
ray_depth_sd_DL_shadow,
transparent_depth_sd,
transparent_depth_sd_DL_shadow,
/* Ray differentials. */
dP_sd,
dP_sd_DL_shadow,
dI_sd,
dI_sd_DL_shadow,
du_sd,
du_sd_DL_shadow,
dv_sd,
dv_sd_DL_shadow,
/* Dp/Du */
dPdu_sd,
dPdu_sd_DL_shadow,
dPdv_sd,
dPdv_sd_DL_shadow,
/* Object motion. */
ob_tfm_sd,
ob_tfm_sd_DL_shadow,
ob_itfm_sd,
ob_itfm_sd_DL_shadow,
closure_sd,
closure_sd_DL_shadow,
num_closure_sd,
num_closure_sd_DL_shadow,
randb_closure_sd,
randb_closure_sd_DL_shadow,
ray_P_sd,
ray_P_sd_DL_shadow,
ray_dP_sd,
ray_dP_sd_DL_shadow,
data,
per_sample_output_buffers,
rng_state,
rng_coop,
throughput_coop,
L_transparent_coop,
PathRadiance_coop,
Ray_coop,
PathState_coop,
ray_state,
#define KERNEL_TEX(type, ttype, name) name,
#include "../../kernel_textures.h"
start_sample, sx, sy, sw, sh, offset, stride,
rng_state_offset_x,
rng_state_offset_y,
rng_state_stride,
Queue_data,
Queue_index,
queuesize,
use_queues_flag,
work_array,
#ifdef __WORK_STEALING__
work_pool_wgs,
num_samples,
#endif
#ifdef __KERNEL_DEBUG__
debugdata_coop,
#endif
parallel_samples);
}

47
intern/cycles/kernel/kernels/opencl/kernel_direct_lighting.cl Normal file
View File

@@ -0,0 +1,47 @@
/*
* Copyright 2011-2015 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 "split/kernel_direct_lighting.h"
__kernel void kernel_ocl_path_trace_direct_lighting(
ccl_global char *globals,
ccl_constant KernelData *data,
ccl_global char *shader_data, /* Required for direct lighting */
ccl_global char *shader_DL, /* Required for direct lighting */
ccl_global uint *rng_coop, /* Required for direct lighting */
ccl_global PathState *PathState_coop, /* Required for direct lighting */
ccl_global int *ISLamp_coop, /* Required for direct lighting */
ccl_global Ray *LightRay_coop, /* Required for direct lighting */
ccl_global BsdfEval *BSDFEval_coop, /* Required for direct lighting */
ccl_global char *ray_state, /* Denotes the state of each ray */
ccl_global int *Queue_data, /* Queue memory */
ccl_global int *Queue_index, /* Tracks the number of elements in each queue */
int queuesize) /* Size (capacity) of each queue */
{
kernel_direct_lighting(globals,
data,
shader_data,
shader_DL,
rng_coop,
PathState_coop,
ISLamp_coop,
LightRay_coop,
BSDFEval_coop,
ray_state,
Queue_data,
Queue_index,
queuesize);
}

67
intern/cycles/kernel/kernels/opencl/kernel_holdout_emission_blurring_pathtermination_ao.cl Normal file
View File

@@ -0,0 +1,67 @@
/*
* Copyright 2011-2015 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 "split/kernel_holdout_emission_blurring_pathtermination_ao.h"
__kernel void kernel_ocl_path_trace_holdout_emission_blurring_pathtermination_ao(
ccl_global char *globals,
ccl_constant KernelData *data,
ccl_global char *shader_data, /* Required throughout the kernel except probabilistic path termination and AO */
ccl_global float *per_sample_output_buffers,
ccl_global uint *rng_coop, /* Required for "kernel_write_data_passes" and AO */
ccl_global float3 *throughput_coop, /* Required for handling holdout material and AO */
ccl_global float *L_transparent_coop, /* Required for handling holdout material */
PathRadiance *PathRadiance_coop, /* Required for "kernel_write_data_passes" and indirect primitive emission */
ccl_global PathState *PathState_coop, /* Required throughout the kernel and AO */
Intersection *Intersection_coop, /* Required for indirect primitive emission */
ccl_global float3 *AOAlpha_coop, /* Required for AO */
ccl_global float3 *AOBSDF_coop, /* Required for AO */
ccl_global Ray *AOLightRay_coop, /* Required for AO */
int sw, int sh, int sx, int sy, int stride,
ccl_global char *ray_state, /* Denotes the state of each ray */
ccl_global unsigned int *work_array, /* Denotes the work that each ray belongs to */
ccl_global int *Queue_data, /* Queue memory */
ccl_global int *Queue_index, /* Tracks the number of elements in each queue */
int queuesize, /* Size (capacity) of each queue */
#ifdef __WORK_STEALING__
unsigned int start_sample,
#endif
int parallel_samples) /* Number of samples to be processed in parallel */
{
kernel_holdout_emission_blurring_pathtermination_ao(globals,
data,
shader_data,
per_sample_output_buffers,
rng_coop,
throughput_coop,
L_transparent_coop,
PathRadiance_coop,
PathState_coop,
Intersection_coop,
AOAlpha_coop,
AOBSDF_coop,
AOLightRay_coop,
sw, sh, sx, sy, stride,
ray_state,
work_array,
Queue_data,
Queue_index,
queuesize,
#ifdef __WORK_STEALING__
start_sample,
#endif
parallel_samples);
}

52
intern/cycles/kernel/kernels/opencl/kernel_lamp_emission.cl Normal file
View File

@@ -0,0 +1,52 @@
/*
* Copyright 2011-2015 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 "split/kernel_lamp_emission.h"
__kernel void kernel_ocl_path_trace_lamp_emission(
ccl_global char *globals,
ccl_constant KernelData *data,
ccl_global char *shader_data, /* Required for lamp emission */
ccl_global float3 *throughput_coop, /* Required for lamp emission */
PathRadiance *PathRadiance_coop, /* Required for lamp emission */
ccl_global Ray *Ray_coop, /* Required for lamp emission */
ccl_global PathState *PathState_coop, /* Required for lamp emission */
Intersection *Intersection_coop, /* Required for lamp emission */
ccl_global char *ray_state, /* Denotes the state of each ray */
int sw, int sh,
ccl_global int *Queue_data, /* Memory for queues */
ccl_global int *Queue_index, /* Tracks the number of elements in queues */
int queuesize, /* Size (capacity) of queues */
ccl_global char *use_queues_flag, /* used to decide if this kernel should use queues to fetch ray index */
int parallel_samples /* Number of samples to be processed in parallel */
)
{
kernel_lamp_emission(globals,
data,
shader_data,
throughput_coop,
PathRadiance_coop,
Ray_coop,
PathState_coop,
Intersection_coop,
ray_state,
sw, sh,
Queue_data,
Queue_index,
queuesize,
use_queues_flag,
parallel_samples);
}

59
intern/cycles/kernel/kernels/opencl/kernel_next_iteration_setup.cl Normal file
View File

@@ -0,0 +1,59 @@
/*
* Copyright 2011-2015 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 "split/kernel_next_iteration_setup.h"
__kernel void kernel_ocl_path_trace_next_iteration_setup(
ccl_global char *globals,
ccl_constant KernelData *data,
ccl_global char *shader_data, /* Required for setting up ray for next iteration */
ccl_global uint *rng_coop, /* Required for setting up ray for next iteration */
ccl_global float3 *throughput_coop, /* Required for setting up ray for next iteration */
PathRadiance *PathRadiance_coop, /* Required for setting up ray for next iteration */
ccl_global Ray *Ray_coop, /* Required for setting up ray for next iteration */
ccl_global PathState *PathState_coop, /* Required for setting up ray for next iteration */
ccl_global Ray *LightRay_dl_coop, /* Required for radiance update - direct lighting */
ccl_global int *ISLamp_coop, /* Required for radiance update - direct lighting */
ccl_global BsdfEval *BSDFEval_coop, /* Required for radiance update - direct lighting */
ccl_global Ray *LightRay_ao_coop, /* Required for radiance update - AO */
ccl_global float3 *AOBSDF_coop, /* Required for radiance update - AO */
ccl_global float3 *AOAlpha_coop, /* Required for radiance update - AO */
ccl_global char *ray_state, /* Denotes the state of each ray */
ccl_global int *Queue_data, /* Queue memory */
ccl_global int *Queue_index, /* Tracks the number of elements in each queue */
int queuesize, /* Size (capacity) of each queue */
ccl_global char *use_queues_flag) /* flag to decide if scene_intersect kernel should use queues to fetch ray index */
{
kernel_next_iteration_setup(globals,
data,
shader_data,
rng_coop,
throughput_coop,
PathRadiance_coop,
Ray_coop,
PathState_coop,
LightRay_dl_coop,
ISLamp_coop,
BSDFEval_coop,
LightRay_ao_coop,
AOBSDF_coop,
AOAlpha_coop,
ray_state,
Queue_data,
Queue_index,
queuesize,
use_queues_flag);
}

29
intern/cycles/kernel/kernels/opencl/kernel_queue_enqueue.cl Normal file
View File

@@ -0,0 +1,29 @@
/*
* Copyright 2011-2015 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 "split/kernel_queue_enqueue.h"
__kernel void kernel_ocl_path_trace_queue_enqueue(
ccl_global int *Queue_data, /* Queue memory */
ccl_global int *Queue_index, /* Tracks the number of elements in each queue */
ccl_global char *ray_state, /* Denotes the state of each ray */
int queuesize) /* Size (capacity) of each queue */
{
kernel_queue_enqueue(Queue_data,
Queue_index,
ray_state,
queuesize);
}

53
intern/cycles/kernel/kernels/opencl/kernel_scene_intersect.cl Normal file
View File

@@ -0,0 +1,53 @@
/*
* Copyright 2011-2015 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 "split/kernel_scene_intersect.h"
__kernel void kernel_ocl_path_trace_scene_intersect(
ccl_global char *globals,
ccl_constant KernelData *data,
ccl_global uint *rng_coop,
ccl_global Ray *Ray_coop, /* Required for scene_intersect */
ccl_global PathState *PathState_coop, /* Required for scene_intersect */
Intersection *Intersection_coop, /* Required for scene_intersect */
ccl_global char *ray_state, /* Denotes the state of each ray */
int sw, int sh,
ccl_global int *Queue_data, /* Memory for queues */
ccl_global int *Queue_index, /* Tracks the number of elements in queues */
int queuesize, /* Size (capacity) of queues */
ccl_global char *use_queues_flag, /* used to decide if this kernel should use queues to fetch ray index */
#ifdef __KERNEL_DEBUG__
DebugData *debugdata_coop,
#endif
int parallel_samples) /* Number of samples to be processed in parallel */
{
kernel_scene_intersect(globals,
data,
rng_coop,
Ray_coop,
PathState_coop,
Intersection_coop,
ray_state,
sw, sh,
Queue_data,
Queue_index,
queuesize,
use_queues_flag,
#ifdef __KERNEL_DEBUG__
debugdata_coop,
#endif
parallel_samples);
}

43
intern/cycles/kernel/kernels/opencl/kernel_shader_eval.cl Normal file
View File

@@ -0,0 +1,43 @@
/*
* Copyright 2011-2015 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 "split/kernel_shader_eval.h"
__kernel void kernel_ocl_path_trace_shader_eval(
ccl_global char *globals,
ccl_constant KernelData *data,
ccl_global char *shader_data, /* Output ShaderData structure to be filled */
ccl_global uint *rng_coop, /* Required for rbsdf calculation */
ccl_global Ray *Ray_coop, /* Required for setting up shader from ray */
ccl_global PathState *PathState_coop, /* Required for all functions in this kernel */
Intersection *Intersection_coop, /* Required for setting up shader from ray */
ccl_global char *ray_state, /* Denotes the state of each ray */
ccl_global int *Queue_data, /* queue memory */
ccl_global int *Queue_index, /* Tracks the number of elements in each queue */
int queuesize) /* Size (capacity) of each queue */
{
kernel_shader_eval(globals,
data,
shader_data,
rng_coop,
Ray_coop,
PathState_coop,
Intersection_coop,
ray_state,
Queue_data,
Queue_index,
queuesize);
}

47
intern/cycles/kernel/kernels/opencl/kernel_shadow_blocked.cl Normal file
View File

@@ -0,0 +1,47 @@
/*
* Copyright 2011-2015 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 "split/kernel_shadow_blocked.h"
__kernel void kernel_ocl_path_trace_shadow_blocked(
ccl_global char *globals,
ccl_constant KernelData *data,
ccl_global char *shader_shadow, /* Required for shadow blocked */
ccl_global PathState *PathState_coop, /* Required for shadow blocked */
ccl_global Ray *LightRay_dl_coop, /* Required for direct lighting's shadow blocked */
ccl_global Ray *LightRay_ao_coop, /* Required for AO's shadow blocked */
Intersection *Intersection_coop_AO,
Intersection *Intersection_coop_DL,
ccl_global char *ray_state,
ccl_global int *Queue_data, /* Queue memory */
ccl_global int *Queue_index, /* Tracks the number of elements in each queue */
int queuesize, /* Size (capacity) of each queue */
int total_num_rays)
{
kernel_shadow_blocked(globals,
data,
shader_shadow,
PathState_coop,
LightRay_dl_coop,
LightRay_ao_coop,
Intersection_coop_AO,
Intersection_coop_DL,
ray_state,
Queue_data,
Queue_index,
queuesize,
total_num_rays);
}

38
intern/cycles/kernel/kernels/opencl/kernel_sum_all_radiance.cl Normal file
View File

@@ -0,0 +1,38 @@
/*
* Copyright 2011-2015 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 "split/kernel_sum_all_radiance.h"
__kernel void kernel_ocl_path_trace_sum_all_radiance(
ccl_constant KernelData *data, /* To get pass_stride to offet into buffer */
ccl_global float *buffer, /* Output buffer of RenderTile */
ccl_global float *per_sample_output_buffer, /* Radiance contributed by all samples */
int parallel_samples, int sw, int sh, int stride,
int buffer_offset_x,
int buffer_offset_y,
int buffer_stride,
int start_sample)
{
kernel_sum_all_radiance(data,
buffer,
per_sample_output_buffer,
parallel_samples,
sw, sh, stride,
buffer_offset_x,
buffer_offset_y,
buffer_stride,
start_sample);
}

8
intern/cycles/kernel/kernel_background_buffer_update.cl → intern/cycles/kernel/split/kernel_background_buffer_update.h
View File

@@ -14,10 +14,10 @@
* limitations under the License.
*/
#include "kernel_split.h"
#include "kernel_split_common.h"
/*
* Note on kernel_ocl_path_trace_background_buffer_update kernel.
* Note on kernel_background_buffer_update kernel.
* This is the fourth kernel in the ray tracing logic, and the third
* of the path iteration kernels. This kernel takes care of rays that hit
* the background (sceneintersect kernel), and for the rays of
@@ -33,7 +33,7 @@
*
* The input and output are as follows,
*
* rng_coop ---------------------------------------------|--- kernel_ocl_path_trace_background_buffer_update --|--- PathRadiance_coop
* rng_coop ---------------------------------------------|--- kernel_background_buffer_update --|--- PathRadiance_coop
* throughput_coop --------------------------------------| |--- L_transparent_coop
* per_sample_output_buffers ----------------------------| |--- per_sample_output_buffers
* Ray_coop ---------------------------------------------| |--- ray_state
@@ -70,7 +70,7 @@
* QUEUE_ACTIVE_AND_REGENERATED_RAYS will be filled with RAY_ACTIVE and RAY_REGENERATED rays
* QUEUE_HITBG_BUFF_UPDATE_TOREGEN_RAYS will be empty
*/
__kernel void kernel_ocl_path_trace_background_buffer_update(
ccl_device void kernel_background_buffer_update(
ccl_global char *globals,
ccl_constant KernelData *data,
ccl_global char *shader_data,

12
intern/cycles/kernel/kernel_data_init.cl → intern/cycles/kernel/split/kernel_data_init.h
View File

@@ -14,10 +14,10 @@
* limitations under the License.
*/
#include "kernel_split.h"
#include "kernel_split_common.h"
/*
* Note on kernel_ocl_path_trace_data_initialization kernel
* Note on kernel_data_initialization kernel
* This kernel Initializes structures needed in path-iteration kernels.
* This is the first kernel in ray-tracing logic.
*
@@ -25,7 +25,7 @@
*
* Its input and output are as follows,
*
* Un-initialized rng---------------|--- kernel_ocl_path_trace_data_initialization ---|--- Initialized rng
* Un-initialized rng---------------|--- kernel_data_initialization ---|--- Initialized rng
* Un-initialized throughput -------| |--- Initialized throughput
* Un-initialized L_transparent ----| |--- Initialized L_transparent
* Un-initialized PathRadiance -----| |--- Initialized PathRadiance
@@ -51,7 +51,7 @@
* All slots in queues are initialized to queue empty slot;
* The number of elements in the queues is initialized to 0;
*/
__kernel void kernel_ocl_path_trace_data_init(
ccl_device void kernel_data_init(
ccl_global char *globals,
ccl_global char *shader_data_sd, /* Arguments related to ShaderData */
ccl_global char *shader_data_sd_DL_shadow, /* Arguments related to ShaderData */
@@ -156,7 +156,7 @@ __kernel void kernel_ocl_path_trace_data_init(
#define KERNEL_TEX(type, ttype, name) \
ccl_global type *name,
#include "kernel_textures.h"
#include "../kernel_textures.h"
int start_sample, int sx, int sy, int sw, int sh, int offset, int stride,
int rng_state_offset_x,
@@ -184,7 +184,7 @@ __kernel void kernel_ocl_path_trace_data_init(
kg->data = data;
#define KERNEL_TEX(type, ttype, name) \
kg->name = name;
#include "kernel_textures.h"
#include "../kernel_textures.h"
/* Load ShaderData structure */
ShaderData *sd = (ShaderData *)shader_data_sd;

8
intern/cycles/kernel/kernel_direct_lighting.cl → intern/cycles/kernel/split/kernel_direct_lighting.h
View File

@@ -14,10 +14,10 @@
* limitations under the License.
*/
#include "kernel_split.h"
#include "kernel_split_common.h"
/*
* Note on kernel_ocl_path_trace_direct_lighting kernel.
* Note on kernel_direct_lighting kernel.
* This is the eighth kernel in the ray tracing logic. This is the seventh
* of the path iteration kernels. This kernel takes care of direct lighting
* logic. However, the "shadow ray cast" part of direct lighting is handled
@@ -29,7 +29,7 @@
*
* The input and output are as follows,
*
* rng_coop -----------------------------------------|--- kernel_ocl_path_trace_direct_lighting --|--- BSDFEval_coop
* rng_coop -----------------------------------------|--- kernel_direct_lighting --|--- BSDFEval_coop
* PathState_coop -----------------------------------| |--- ISLamp_coop
* shader_data --------------------------------------| |--- LightRay_coop
* ray_state ----------------------------------------| |--- ray_state
@@ -49,7 +49,7 @@
* QUEUE_SHADOW_RAY_CAST_DL_RAYS queue will be filled with rays for which a shadow_blocked function must be executed, after this
* kernel call. Before this kernel call the QUEUE_SHADOW_RAY_CAST_DL_RAYS will be empty.
*/
__kernel void kernel_ocl_path_trace_direct_lighting(
ccl_device void kernel_direct_lighting(
ccl_global char *globals,
ccl_constant KernelData *data,
ccl_global char *shader_data, /* Required for direct lighting */

8
intern/cycles/kernel/kernel_holdout_emission_blurring_pathtermination_ao.cl → intern/cycles/kernel/split/kernel_holdout_emission_blurring_pathtermination_ao.h
View File

@@ -14,10 +14,10 @@
* limitations under the License.
*/
#include "kernel_split.h"
#include "kernel_split_common.h"
/*
* Note on kernel_ocl_path_trace_holdout_emission_blurring_pathtermination_ao kernel.
* Note on kernel_holdout_emission_blurring_pathtermination_ao kernel.
* This is the sixth kernel in the ray tracing logic. This is the fifth
* of the path iteration kernels. This kernel takes care of the logic to process
* "material of type holdout", indirect primitive emission, bsdf blurring,
@@ -31,7 +31,7 @@
*
* The input and output are as follows,
*
* rng_coop ---------------------------------------------|--- kernel_ocl_path_trace_holdout_emission_blurring_pathtermination_ao ---|--- Queue_index (QUEUE_HITBG_BUFF_UPDATE_TOREGEN_RAYS)
* rng_coop ---------------------------------------------|--- kernel_holdout_emission_blurring_pathtermination_ao ---|--- Queue_index (QUEUE_HITBG_BUFF_UPDATE_TOREGEN_RAYS)
* throughput_coop --------------------------------------| |--- PathState_coop
* PathRadiance_coop ------------------------------------| |--- throughput_coop
* Intersection_coop ------------------------------------| |--- L_transparent_coop
@@ -72,7 +72,7 @@
* QUEUE_SHADOW_RAY_CAST_AO_RAYS will be filled with rays marked with flag RAY_SHADOW_RAY_CAST_AO
*/
__kernel void kernel_ocl_path_trace_holdout_emission_blurring_pathtermination_ao(
ccl_device void kernel_holdout_emission_blurring_pathtermination_ao(
ccl_global char *globals,
ccl_constant KernelData *data,
ccl_global char *shader_data, /* Required throughout the kernel except probabilistic path termination and AO */

10
intern/cycles/kernel/kernel_lamp_emission.cl → intern/cycles/kernel/split/kernel_lamp_emission.h
View File

@@ -14,17 +14,17 @@
* limitations under the License.
*/
#include "kernel_split.h"
#include "kernel_split_common.h"
/*
* Note on kernel_ocl_path_trace_lamp_emission
* Note on kernel_lamp_emission
* This is the 3rd kernel in the ray-tracing logic. This is the second of the
* path-iteration kernels. This kernel takes care of the indirect lamp emission logic.
* This kernel operates on QUEUE_ACTIVE_AND_REGENERATED_RAYS. It processes rays of state RAY_ACTIVE
* and RAY_HIT_BACKGROUND.
* We will empty QUEUE_ACTIVE_AND_REGENERATED_RAYS queue in this kernel.
* The input/output of the kernel is as follows,
* Throughput_coop ------------------------------------|--- kernel_ocl_path_trace_lamp_emission --|--- PathRadiance_coop
* Throughput_coop ------------------------------------|--- kernel_lamp_emission --|--- PathRadiance_coop
* Ray_coop -------------------------------------------| |--- Queue_data(QUEUE_ACTIVE_AND_REGENERATED_RAYS)
* PathState_coop -------------------------------------| |--- Queue_index(QUEUE_ACTIVE_AND_REGENERATED_RAYS)
* kg (globals + data) --------------------------------| |
@@ -38,9 +38,9 @@
* sh -------------------------------------------------| |
* parallel_samples -----------------------------------| |
*
* note : shader_data is neither input nor output. Its just filled and consumed in the same, kernel_ocl_path_trace_lamp_emission, kernel.
* note : shader_data is neither input nor output. Its just filled and consumed in the same, kernel_lamp_emission, kernel.
*/
__kernel void kernel_ocl_path_trace_lamp_emission(
ccl_device void kernel_lamp_emission(
ccl_global char *globals,
ccl_constant KernelData *data,
ccl_global char *shader_data, /* Required for lamp emission */

8
intern/cycles/kernel/kernel_next_iteration_setup.cl → intern/cycles/kernel/split/kernel_next_iteration_setup.h
View File

@@ -14,10 +14,10 @@
* limitations under the License.
*/
#include "kernel_split.h"
#include "kernel_split_common.h"
/*
* Note on kernel_ocl_path_trace_setup_next_iteration kernel.
* Note on kernel_setup_next_iteration kernel.
* This is the tenth kernel in the ray tracing logic. This is the ninth
* of the path iteration kernels. This kernel takes care of setting up
* Ray for the next iteration of path-iteration and accumulating radiance
@@ -27,7 +27,7 @@
*
* The input and output are as follows,
*
* rng_coop ---------------------------------------------|--- kernel_ocl_path_trace_next_iteration_setup -|--- Queue_index (QUEUE_HITBG_BUFF_UPDATE_TOREGEN_RAYS)
* rng_coop ---------------------------------------------|--- kernel_next_iteration_setup -|--- Queue_index (QUEUE_HITBG_BUFF_UPDATE_TOREGEN_RAYS)
* throughput_coop --------------------------------------| |--- Queue_data (QUEUE_HITBF_BUFF_UPDATE_TOREGEN_RAYS)
* PathRadiance_coop ------------------------------------| |--- throughput_coop
* PathState_coop ---------------------------------------| |--- PathRadiance_coop
@@ -61,7 +61,7 @@
* QUEUE_HITBG_BUFF_UPDATE_TOREGEN_RAYS will be filled with RAY_TO_REGENERATE and more RAY_UPDATE_BUFFER rays
*/
__kernel void kernel_ocl_path_trace_next_iteration_setup(
ccl_device void kernel_next_iteration_setup(
ccl_global char *globals,
ccl_constant KernelData *data,
ccl_global char *shader_data, /* Required for setting up ray for next iteration */

18
intern/cycles/kernel/kernel_queue_enqueue.cl → intern/cycles/kernel/split/kernel_queue_enqueue.h
View File

@@ -14,23 +14,23 @@
* limitations under the License.
*/
#include "kernel_compat_opencl.h"
#include "kernel_math.h"
#include "kernel_types.h"
#include "kernel_globals.h"
#include "kernel_queues.h"
#include "../kernel_compat_opencl.h"
#include "../kernel_math.h"
#include "../kernel_types.h"
#include "../kernel_globals.h"
#include "../kernel_queues.h"
/*
* The kernel "kernel_ocl_path_trace_queue_enqueue" enqueues rays of
* The kernel "kernel_queue_enqueue" enqueues rays of
* different ray state into their appropriate Queues;
* 1. Rays that have been determined to hit the background from the
* "kernel_ocl_path_trace_scene_intersect" kernel
* "kernel_scene_intersect" kernel
* are enqueued in QUEUE_HITBG_BUFF_UPDATE_TOREGEN_RAYS;
* 2. Rays that have been determined to be actively participating in path-iteration will be enqueued into QUEUE_ACTIVE_AND_REGENERATED_RAYS.
*
* The input and output of the kernel is as follows,
*
* ray_state -------------------------------------------|--- kernel_ocl_path_trace_queue_enqueue --|--- Queue_data (QUEUE_ACTIVE_AND_REGENERATED_RAYS & QUEUE_HITBF_BUFF_UPDATE_TOREGEN_RAYS)
* ray_state -------------------------------------------|--- kernel_queue_enqueue --|--- Queue_data (QUEUE_ACTIVE_AND_REGENERATED_RAYS & QUEUE_HITBF_BUFF_UPDATE_TOREGEN_RAYS)
* Queue_index(QUEUE_ACTIVE_AND_REGENERATED_RAYS) ------| |--- Queue_index (QUEUE_ACTIVE_AND_REGENERATED_RAYS & QUEUE_HITBF_BUFF_UPDATE_TOREGEN_RAYS)
* Queue_index(QUEUE_HITBG_BUFF_UPDATE_TOREGEN_RAYS) ---| |
* queuesize -------------------------------------------| |
@@ -52,7 +52,7 @@
* QUEUE_HITBG_BUFF_UPDATE_TOREGEN_RAYS will be filled with RAY_TO_REGENERATE, RAY_UPDATE_BUFFER, RAY_HIT_BACKGROUND rays.
*/
__kernel void kernel_ocl_path_trace_queue_enqueue(
ccl_device void kernel_queue_enqueue(
ccl_global int *Queue_data, /* Queue memory */
ccl_global int *Queue_index, /* Tracks the number of elements in each queue */
ccl_global char *ray_state, /* Denotes the state of each ray */

10
intern/cycles/kernel/kernel_scene_intersect.cl → intern/cycles/kernel/split/kernel_scene_intersect.h
View File

@@ -14,10 +14,10 @@
* limitations under the License.
*/
#include "kernel_split.h"
#include "kernel_split_common.h"
/*
* Note on kernel_ocl_path_trace_scene_intersect kernel.
* Note on kernel_scene_intersect kernel.
* This is the second kernel in the ray tracing logic. This is the first
* of the path iteration kernels. This kernel takes care of scene_intersect function.
*
@@ -27,7 +27,7 @@
*
* The input and output are as follows,
*
* Ray_coop ---------------------------------------|--------- kernel_ocl_path_trace_scene_intersect----------|--- PathState
* Ray_coop ---------------------------------------|--------- kernel_scene_intersect----------|--- PathState
* PathState_coop ---------------------------------| |--- Intersection
* ray_state --------------------------------------| |--- ray_state
* use_queues_flag --------------------------------| |
@@ -40,7 +40,7 @@
* queuesize --------------------------------------| |
*
* Note on Queues :
* Ideally we would want kernel_ocl_path_trace_scene_intersect to work on queues.
* Ideally we would want kernel_scene_intersect to work on queues.
* But during the very first time, the queues wil be empty and hence we perform a direct mapping
* between ray-index and thread-index; From the next time onward, the queue will be filled and
* we may start operating on queues.
@@ -63,7 +63,7 @@
* QUEUE_HITBF_BUFF_UPDATE_TOREGEN_RAYS - no change
*/
__kernel void kernel_ocl_path_trace_scene_intersect(
ccl_device void kernel_scene_intersect(
ccl_global char *globals,
ccl_constant KernelData *data,
ccl_global uint *rng_coop,

8
intern/cycles/kernel/kernel_shader_eval.cl → intern/cycles/kernel/split/kernel_shader_eval.h
View File

@@ -14,17 +14,17 @@
* limitations under the License.
*/
#include "kernel_split.h"
#include "kernel_split_common.h"
/*
* Note on kernel_ocl_path_trace_shader_eval kernel
* Note on kernel_shader_eval kernel
* This kernel is the 5th kernel in the ray tracing logic. This is
* the 4rd kernel in path iteration. This kernel sets up the ShaderData
* structure from the values computed by the previous kernels. It also identifies
* the rays of state RAY_TO_REGENERATE and enqueues them in QUEUE_HITBG_BUFF_UPDATE_TOREGEN_RAYS queue.
*
* The input and output of the kernel is as follows,
* rng_coop -------------------------------------------|--- kernel_ocl_path_trace_shader_eval --|--- shader_data
* rng_coop -------------------------------------------|--- kernel_shader_eval --|--- shader_data
* Ray_coop -------------------------------------------| |--- Queue_data (QUEUE_HITBG_BUFF_UPDATE_TOREGEN_RAYS)
* PathState_coop -------------------------------------| |--- Queue_index (QUEUE_HITBG_BUFF_UPDATE_TOREGEN_RAYS)
* Intersection_coop ----------------------------------| |
@@ -46,7 +46,7 @@
* QUEUE_HITBG_BUFF_UPDATE_TOREGEN_RAYS will be filled with RAY_TO_REGENERATE rays
*/
__kernel void kernel_ocl_path_trace_shader_eval(
ccl_device void kernel_shader_eval(
ccl_global char *globals,
ccl_constant KernelData *data,
ccl_global char *shader_data, /* Output ShaderData structure to be filled */

8
intern/cycles/kernel/kernel_shadow_blocked.cl → intern/cycles/kernel/split/kernel_shadow_blocked.h
View File

@@ -14,17 +14,17 @@
* limitations under the License.
*/
#include "kernel_split.h"
#include "kernel_split_common.h"
/*
* Note on kernel_ocl_path_trace_shadow_blocked kernel.
* Note on kernel_shadow_blocked kernel.
* This is the ninth kernel in the ray tracing logic. This is the eighth
* of the path iteration kernels. This kernel takes care of "shadow ray cast"
* logic of the direct lighting and AO part of ray tracing.
*
* The input and output are as follows,
*
* PathState_coop ----------------------------------|--- kernel_ocl_path_trace_shadow_blocked --|
* PathState_coop ----------------------------------|--- kernel_shadow_blocked --|
* LightRay_dl_coop --------------------------------| |--- LightRay_dl_coop
* LightRay_ao_coop --------------------------------| |--- LightRay_ao_coop
* ray_state ---------------------------------------| |--- ray_state
@@ -47,7 +47,7 @@
* QUEUE_SHADOW_RAY_CAST_AO_RAYS and QUEUE_SHADOW_RAY_CAST_DL_RAYS will be empty at kernel exit.
*/
__kernel void kernel_ocl_path_trace_shadow_blocked(
ccl_device void kernel_shadow_blocked(
ccl_global char *globals,
ccl_constant KernelData *data,
ccl_global char *shader_shadow, /* Required for shadow blocked */

62
intern/cycles/kernel/split/kernel_split_common.h Normal file
View File

@@ -0,0 +1,62 @@
/*
* Copyright 2011-2015 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.
*/
#ifndef _KERNEL_SPLIT_H_
#define _KERNEL_SPLIT_H_
#include "../kernel_compat_opencl.h"
#include "../kernel_math.h"
#include "../kernel_types.h"
#include "../kernel_globals.h"
#include "../util_atomic.h"
#include "../kernel_random.h"
#include "../kernel_projection.h"
#include "../kernel_montecarlo.h"
#include "../kernel_differential.h"
#include "../kernel_camera.h"
#include "../geom/geom.h"
#include "../kernel_accumulate.h"
#include "../kernel_shader.h"
#include "../kernel_light.h"
#include "../kernel_passes.h"
#ifdef __SUBSURFACE__
#include "../kernel_subsurface.h"
#endif
#ifdef __VOLUME__
#include "../kernel_volume.h"
#endif
#include "../kernel_path_state.h"
#include "../kernel_shadow.h"
#include "../kernel_emission.h"
#include "../kernel_path_common.h"
#include "../kernel_path_surface.h"
#include "../kernel_path_volume.h"
#ifdef __KERNEL_DEBUG__
#include "../kernel_debug.h"
#endif
#include "../kernel_queues.h"
#include "../kernel_work_stealing.h"
#endif

10
intern/cycles/kernel/kernel_sum_all_radiance.cl → intern/cycles/kernel/split/kernel_sum_all_radiance.h
View File

@@ -14,10 +14,10 @@
* limitations under the License.
*/
#include "kernel_compat_opencl.h"
#include "kernel_math.h"
#include "kernel_types.h"
#include "kernel_globals.h"
#include "../kernel_compat_opencl.h"
#include "../kernel_math.h"
#include "../kernel_types.h"
#include "../kernel_globals.h"
/*
* Since we process various samples in parallel; The output radiance of different samples
@@ -25,7 +25,7 @@
* by all different samples and stores them in the RenderTile's output buffer.
*/
__kernel void kernel_ocl_path_trace_sum_all_radiance(
ccl_device void kernel_sum_all_radiance(
ccl_constant KernelData *data, /* To get pass_stride to offet into buffer */
ccl_global float *buffer, /* Output buffer of RenderTile */
ccl_global float *per_sample_output_buffer, /* Radiance contributed by all samples */