2015-03-27 10:55:58 +00:00
|
|
|
remove_extra_strict_flags()
|
|
|
|
|
2011-11-08 20:27:37 +00:00
|
|
|
set(INC
|
2019-04-17 04:17:24 +00:00
|
|
|
..
|
2011-11-08 20:27:37 +00:00
|
|
|
)
|
2012-10-17 04:13:03 +00:00
|
|
|
|
2012-09-20 09:04:43 +00:00
|
|
|
set(INC_SYS
|
2012-10-17 04:13:03 +00:00
|
|
|
|
2012-09-20 09:04:43 +00:00
|
|
|
)
|
2011-11-08 20:27:37 +00:00
|
|
|
|
2017-10-07 12:08:02 +00:00
|
|
|
set(SRC_CPU_KERNELS
|
2019-04-17 04:17:24 +00:00
|
|
|
kernels/cpu/kernel.cpp
|
|
|
|
kernels/cpu/kernel_sse2.cpp
|
|
|
|
kernels/cpu/kernel_sse3.cpp
|
|
|
|
kernels/cpu/kernel_sse41.cpp
|
|
|
|
kernels/cpu/kernel_avx.cpp
|
|
|
|
kernels/cpu/kernel_avx2.cpp
|
|
|
|
kernels/cpu/kernel_split.cpp
|
|
|
|
kernels/cpu/kernel_split_sse2.cpp
|
|
|
|
kernels/cpu/kernel_split_sse3.cpp
|
|
|
|
kernels/cpu/kernel_split_sse41.cpp
|
|
|
|
kernels/cpu/kernel_split_avx.cpp
|
|
|
|
kernels/cpu/kernel_split_avx2.cpp
|
|
|
|
kernels/cpu/filter.cpp
|
|
|
|
kernels/cpu/filter_sse2.cpp
|
|
|
|
kernels/cpu/filter_sse3.cpp
|
|
|
|
kernels/cpu/filter_sse41.cpp
|
|
|
|
kernels/cpu/filter_avx.cpp
|
|
|
|
kernels/cpu/filter_avx2.cpp
|
2017-10-07 12:08:02 +00:00
|
|
|
)
|
|
|
|
|
|
|
|
set(SRC_CUDA_KERNELS
|
2019-04-17 04:17:24 +00:00
|
|
|
kernels/cuda/kernel.cu
|
|
|
|
kernels/cuda/kernel_split.cu
|
|
|
|
kernels/cuda/filter.cu
|
2017-10-07 12:08:02 +00:00
|
|
|
)
|
|
|
|
|
|
|
|
set(SRC_OPENCL_KERNELS
|
2019-04-17 04:17:24 +00:00
|
|
|
kernels/opencl/kernel_bake.cl
|
|
|
|
kernels/opencl/kernel_base.cl
|
|
|
|
kernels/opencl/kernel_displace.cl
|
|
|
|
kernels/opencl/kernel_background.cl
|
|
|
|
kernels/opencl/kernel_state_buffer_size.cl
|
|
|
|
kernels/opencl/kernel_split_bundle.cl
|
|
|
|
kernels/opencl/kernel_data_init.cl
|
|
|
|
kernels/opencl/kernel_path_init.cl
|
|
|
|
kernels/opencl/kernel_queue_enqueue.cl
|
|
|
|
kernels/opencl/kernel_scene_intersect.cl
|
|
|
|
kernels/opencl/kernel_lamp_emission.cl
|
|
|
|
kernels/opencl/kernel_do_volume.cl
|
|
|
|
kernels/opencl/kernel_indirect_background.cl
|
|
|
|
kernels/opencl/kernel_shader_setup.cl
|
|
|
|
kernels/opencl/kernel_shader_sort.cl
|
|
|
|
kernels/opencl/kernel_shader_eval.cl
|
|
|
|
kernels/opencl/kernel_holdout_emission_blurring_pathtermination_ao.cl
|
|
|
|
kernels/opencl/kernel_subsurface_scatter.cl
|
|
|
|
kernels/opencl/kernel_direct_lighting.cl
|
|
|
|
kernels/opencl/kernel_shadow_blocked_ao.cl
|
|
|
|
kernels/opencl/kernel_shadow_blocked_dl.cl
|
|
|
|
kernels/opencl/kernel_enqueue_inactive.cl
|
|
|
|
kernels/opencl/kernel_next_iteration_setup.cl
|
|
|
|
kernels/opencl/kernel_indirect_subsurface.cl
|
|
|
|
kernels/opencl/kernel_buffer_update.cl
|
|
|
|
kernels/opencl/filter.cl
|
2011-11-08 20:27:37 +00:00
|
|
|
)
|
2011-04-27 11:58:34 +00:00
|
|
|
|
2016-07-11 10:28:45 +00:00
|
|
|
set(SRC_BVH_HEADERS
|
2019-04-17 04:17:24 +00:00
|
|
|
bvh/bvh.h
|
|
|
|
bvh/bvh_nodes.h
|
|
|
|
bvh/bvh_shadow_all.h
|
|
|
|
bvh/bvh_local.h
|
|
|
|
bvh/bvh_traversal.h
|
|
|
|
bvh/bvh_types.h
|
|
|
|
bvh/bvh_volume.h
|
|
|
|
bvh/bvh_volume_all.h
|
|
|
|
bvh/qbvh_nodes.h
|
|
|
|
bvh/qbvh_shadow_all.h
|
|
|
|
bvh/qbvh_local.h
|
|
|
|
bvh/qbvh_traversal.h
|
|
|
|
bvh/qbvh_volume.h
|
|
|
|
bvh/qbvh_volume_all.h
|
|
|
|
bvh/obvh_nodes.h
|
|
|
|
bvh/obvh_shadow_all.h
|
|
|
|
bvh/obvh_local.h
|
|
|
|
bvh/obvh_traversal.h
|
|
|
|
bvh/obvh_volume.h
|
|
|
|
bvh/obvh_volume_all.h
|
|
|
|
bvh/bvh_embree.h
|
2016-07-11 10:28:45 +00:00
|
|
|
)
|
|
|
|
|
2011-11-08 20:27:37 +00:00
|
|
|
set(SRC_HEADERS
|
2019-04-17 04:17:24 +00:00
|
|
|
kernel_accumulate.h
|
|
|
|
kernel_bake.h
|
|
|
|
kernel_camera.h
|
|
|
|
kernel_color.h
|
|
|
|
kernel_compat_cpu.h
|
|
|
|
kernel_compat_cuda.h
|
|
|
|
kernel_compat_opencl.h
|
|
|
|
kernel_differential.h
|
|
|
|
kernel_emission.h
|
|
|
|
kernel_film.h
|
|
|
|
kernel_globals.h
|
|
|
|
kernel_id_passes.h
|
|
|
|
kernel_jitter.h
|
|
|
|
kernel_light.h
|
|
|
|
kernel_math.h
|
|
|
|
kernel_montecarlo.h
|
|
|
|
kernel_passes.h
|
|
|
|
kernel_path.h
|
|
|
|
kernel_path_branched.h
|
|
|
|
kernel_path_common.h
|
|
|
|
kernel_path_state.h
|
|
|
|
kernel_path_surface.h
|
|
|
|
kernel_path_subsurface.h
|
|
|
|
kernel_path_volume.h
|
|
|
|
kernel_profiling.h
|
|
|
|
kernel_projection.h
|
|
|
|
kernel_queues.h
|
|
|
|
kernel_random.h
|
|
|
|
kernel_shader.h
|
|
|
|
kernel_shadow.h
|
|
|
|
kernel_subsurface.h
|
|
|
|
kernel_textures.h
|
|
|
|
kernel_types.h
|
|
|
|
kernel_volume.h
|
|
|
|
kernel_work_stealing.h
|
2015-12-30 13:51:35 +00:00
|
|
|
)
|
2015-12-30 12:54:02 +00:00
|
|
|
|
2015-12-30 13:51:35 +00:00
|
|
|
set(SRC_KERNELS_CPU_HEADERS
|
2019-04-17 04:17:24 +00:00
|
|
|
kernel.h
|
|
|
|
kernels/cpu/kernel_cpu.h
|
|
|
|
kernels/cpu/kernel_cpu_impl.h
|
|
|
|
kernels/cpu/kernel_cpu_image.h
|
|
|
|
kernels/cpu/filter_cpu.h
|
|
|
|
kernels/cpu/filter_cpu_impl.h
|
2011-11-08 20:27:37 +00:00
|
|
|
)
|
2011-05-20 12:26:01 +00:00
|
|
|
|
2017-02-14 10:50:29 +00:00
|
|
|
set(SRC_KERNELS_CUDA_HEADERS
|
2019-04-17 04:17:24 +00:00
|
|
|
kernels/cuda/kernel_config.h
|
|
|
|
kernels/cuda/kernel_cuda_image.h
|
2017-02-14 10:50:29 +00:00
|
|
|
)
|
|
|
|
|
2017-05-11 23:23:49 +00:00
|
|
|
set(SRC_KERNELS_OPENCL_HEADERS
|
2019-04-17 04:17:24 +00:00
|
|
|
kernels/opencl/kernel_split_function.h
|
|
|
|
kernels/opencl/kernel_opencl_image.h
|
2017-05-11 23:23:49 +00:00
|
|
|
)
|
|
|
|
|
2012-10-20 14:08:49 +00:00
|
|
|
set(SRC_CLOSURE_HEADERS
|
2019-04-17 04:17:24 +00:00
|
|
|
closure/alloc.h
|
|
|
|
closure/bsdf.h
|
|
|
|
closure/bsdf_ashikhmin_velvet.h
|
|
|
|
closure/bsdf_diffuse.h
|
|
|
|
closure/bsdf_diffuse_ramp.h
|
|
|
|
closure/bsdf_microfacet.h
|
|
|
|
closure/bsdf_microfacet_multi.h
|
|
|
|
closure/bsdf_microfacet_multi_impl.h
|
|
|
|
closure/bsdf_oren_nayar.h
|
|
|
|
closure/bsdf_phong_ramp.h
|
|
|
|
closure/bsdf_reflection.h
|
|
|
|
closure/bsdf_refraction.h
|
|
|
|
closure/bsdf_toon.h
|
|
|
|
closure/bsdf_transparent.h
|
|
|
|
closure/bsdf_util.h
|
|
|
|
closure/bsdf_ashikhmin_shirley.h
|
|
|
|
closure/bsdf_hair.h
|
|
|
|
closure/bssrdf.h
|
|
|
|
closure/emissive.h
|
|
|
|
closure/volume.h
|
|
|
|
closure/bsdf_principled_diffuse.h
|
|
|
|
closure/bsdf_principled_sheen.h
|
2018-07-18 09:14:43 +00:00
|
|
|
closure/bsdf_hair_principled.h
|
2012-10-20 14:08:49 +00:00
|
|
|
)
|
Cycles: OpenCL kernel split
This commit contains all the work related on the AMD megakernel split work
which was mainly done by Varun Sundar, George Kyriazis and Lenny Wang, plus
some help from Sergey Sharybin, Martijn Berger, Thomas Dinges and likely
someone else which we're forgetting to mention.
Currently only AMD cards are enabled for the new split kernel, but it is
possible to force split opencl kernel to be used by setting the following
environment variable: CYCLES_OPENCL_SPLIT_KERNEL_TEST=1.
Not all the features are supported yet, and that being said no motion blur,
camera blur, SSS and volumetrics for now. Also transparent shadows are
disabled on AMD device because of some compiler bug.
This kernel is also only implements regular path tracing and supporting
branched one will take a bit. Branched path tracing is exposed to the
interface still, which is a bit misleading and will be hidden there soon.
More feature will be enabled once they're ported to the split kernel and
tested.
Neither regular CPU nor CUDA has any difference, they're generating the
same exact code, which means no regressions/improvements there.
Based on the research paper:
https://research.nvidia.com/sites/default/files/publications/laine2013hpg_paper.pdf
Here's the documentation:
https://docs.google.com/document/d/1LuXW-CV-sVJkQaEGZlMJ86jZ8FmoPfecaMdR-oiWbUY/edit
Design discussion of the patch:
https://developer.blender.org/T44197
Differential Revision: https://developer.blender.org/D1200
2015-05-09 14:34:30 +00:00
|
|
|
|
2012-10-20 14:08:49 +00:00
|
|
|
set(SRC_SVM_HEADERS
|
2019-04-17 04:17:24 +00:00
|
|
|
svm/svm.h
|
|
|
|
svm/svm_ao.h
|
|
|
|
svm/svm_attribute.h
|
|
|
|
svm/svm_bevel.h
|
|
|
|
svm/svm_blackbody.h
|
|
|
|
svm/svm_bump.h
|
|
|
|
svm/svm_camera.h
|
|
|
|
svm/svm_closure.h
|
|
|
|
svm/svm_convert.h
|
|
|
|
svm/svm_checker.h
|
|
|
|
svm/svm_color_util.h
|
|
|
|
svm/svm_brick.h
|
|
|
|
svm/svm_displace.h
|
|
|
|
svm/svm_fresnel.h
|
|
|
|
svm/svm_wireframe.h
|
|
|
|
svm/svm_wavelength.h
|
|
|
|
svm/svm_gamma.h
|
|
|
|
svm/svm_brightness.h
|
|
|
|
svm/svm_geometry.h
|
|
|
|
svm/svm_gradient.h
|
|
|
|
svm/svm_hsv.h
|
|
|
|
svm/svm_ies.h
|
|
|
|
svm/svm_image.h
|
|
|
|
svm/svm_invert.h
|
|
|
|
svm/svm_light_path.h
|
|
|
|
svm/svm_magic.h
|
|
|
|
svm/svm_mapping.h
|
|
|
|
svm/svm_math.h
|
|
|
|
svm/svm_math_util.h
|
|
|
|
svm/svm_mix.h
|
|
|
|
svm/svm_musgrave.h
|
|
|
|
svm/svm_noise.h
|
|
|
|
svm/svm_noisetex.h
|
|
|
|
svm/svm_normal.h
|
|
|
|
svm/svm_ramp.h
|
|
|
|
svm/svm_ramp_util.h
|
|
|
|
svm/svm_sepcomb_hsv.h
|
|
|
|
svm/svm_sepcomb_vector.h
|
|
|
|
svm/svm_sky.h
|
|
|
|
svm/svm_tex_coord.h
|
|
|
|
svm/svm_texture.h
|
|
|
|
svm/svm_types.h
|
|
|
|
svm/svm_value.h
|
|
|
|
svm/svm_vector_transform.h
|
|
|
|
svm/svm_voronoi.h
|
|
|
|
svm/svm_voxel.h
|
|
|
|
svm/svm_wave.h
|
2011-11-08 20:27:37 +00:00
|
|
|
)
|
2011-04-27 11:58:34 +00:00
|
|
|
|
2014-03-29 12:03:45 +00:00
|
|
|
set(SRC_GEOM_HEADERS
|
2019-04-17 04:17:24 +00:00
|
|
|
geom/geom.h
|
|
|
|
geom/geom_attribute.h
|
|
|
|
geom/geom_curve.h
|
|
|
|
geom/geom_curve_intersect.h
|
|
|
|
geom/geom_motion_curve.h
|
|
|
|
geom/geom_motion_triangle.h
|
|
|
|
geom/geom_motion_triangle_intersect.h
|
|
|
|
geom/geom_motion_triangle_shader.h
|
|
|
|
geom/geom_object.h
|
|
|
|
geom/geom_patch.h
|
|
|
|
geom/geom_primitive.h
|
|
|
|
geom/geom_subd_triangle.h
|
|
|
|
geom/geom_triangle.h
|
|
|
|
geom/geom_triangle_intersect.h
|
|
|
|
geom/geom_volume.h
|
2014-03-29 12:03:45 +00:00
|
|
|
)
|
|
|
|
|
2017-05-07 12:40:58 +00:00
|
|
|
set(SRC_FILTER_HEADERS
|
2019-04-17 04:17:24 +00:00
|
|
|
filter/filter.h
|
|
|
|
filter/filter_defines.h
|
|
|
|
filter/filter_features.h
|
|
|
|
filter/filter_features_sse.h
|
|
|
|
filter/filter_kernel.h
|
|
|
|
filter/filter_nlm_cpu.h
|
|
|
|
filter/filter_nlm_gpu.h
|
|
|
|
filter/filter_prefilter.h
|
|
|
|
filter/filter_reconstruction.h
|
|
|
|
filter/filter_transform.h
|
|
|
|
filter/filter_transform_gpu.h
|
|
|
|
filter/filter_transform_sse.h
|
2017-05-07 12:40:58 +00:00
|
|
|
)
|
|
|
|
|
2011-11-08 20:27:37 +00:00
|
|
|
set(SRC_UTIL_HEADERS
|
2019-04-17 04:17:24 +00:00
|
|
|
../util/util_atomic.h
|
|
|
|
../util/util_color.h
|
|
|
|
../util/util_defines.h
|
|
|
|
../util/util_half.h
|
|
|
|
../util/util_hash.h
|
|
|
|
../util/util_math.h
|
|
|
|
../util/util_math_fast.h
|
|
|
|
../util/util_math_intersect.h
|
|
|
|
../util/util_math_float2.h
|
|
|
|
../util/util_math_float3.h
|
|
|
|
../util/util_math_float4.h
|
|
|
|
../util/util_math_int2.h
|
|
|
|
../util/util_math_int3.h
|
|
|
|
../util/util_math_int4.h
|
|
|
|
../util/util_math_matrix.h
|
|
|
|
../util/util_projection.h
|
|
|
|
../util/util_rect.h
|
|
|
|
../util/util_static_assert.h
|
|
|
|
../util/util_transform.h
|
|
|
|
../util/util_texture.h
|
|
|
|
../util/util_types.h
|
|
|
|
../util/util_types_float2.h
|
|
|
|
../util/util_types_float2_impl.h
|
|
|
|
../util/util_types_float3.h
|
|
|
|
../util/util_types_float3_impl.h
|
|
|
|
../util/util_types_float4.h
|
|
|
|
../util/util_types_float4_impl.h
|
|
|
|
../util/util_types_float8.h
|
|
|
|
../util/util_types_float8_impl.h
|
|
|
|
../util/util_types_int2.h
|
|
|
|
../util/util_types_int2_impl.h
|
|
|
|
../util/util_types_int3.h
|
|
|
|
../util/util_types_int3_impl.h
|
|
|
|
../util/util_types_int4.h
|
|
|
|
../util/util_types_int4_impl.h
|
|
|
|
../util/util_types_uchar2.h
|
|
|
|
../util/util_types_uchar2_impl.h
|
|
|
|
../util/util_types_uchar3.h
|
|
|
|
../util/util_types_uchar3_impl.h
|
|
|
|
../util/util_types_uchar4.h
|
|
|
|
../util/util_types_uchar4_impl.h
|
|
|
|
../util/util_types_uint2.h
|
|
|
|
../util/util_types_uint2_impl.h
|
|
|
|
../util/util_types_uint3.h
|
|
|
|
../util/util_types_uint3_impl.h
|
|
|
|
../util/util_types_uint4.h
|
|
|
|
../util/util_types_uint4_impl.h
|
|
|
|
../util/util_types_ushort4.h
|
|
|
|
../util/util_types_vector3.h
|
|
|
|
../util/util_types_vector3_impl.h
|
2011-11-08 20:27:37 +00:00
|
|
|
)
|
2015-05-21 12:40:04 +00:00
|
|
|
|
|
|
|
set(SRC_SPLIT_HEADERS
|
2019-04-17 04:17:24 +00:00
|
|
|
split/kernel_branched.h
|
|
|
|
split/kernel_buffer_update.h
|
|
|
|
split/kernel_data_init.h
|
|
|
|
split/kernel_direct_lighting.h
|
|
|
|
split/kernel_do_volume.h
|
|
|
|
split/kernel_enqueue_inactive.h
|
|
|
|
split/kernel_holdout_emission_blurring_pathtermination_ao.h
|
|
|
|
split/kernel_indirect_background.h
|
|
|
|
split/kernel_indirect_subsurface.h
|
|
|
|
split/kernel_lamp_emission.h
|
|
|
|
split/kernel_next_iteration_setup.h
|
|
|
|
split/kernel_path_init.h
|
|
|
|
split/kernel_queue_enqueue.h
|
|
|
|
split/kernel_scene_intersect.h
|
|
|
|
split/kernel_shader_setup.h
|
|
|
|
split/kernel_shader_sort.h
|
|
|
|
split/kernel_shader_eval.h
|
|
|
|
split/kernel_shadow_blocked_ao.h
|
|
|
|
split/kernel_shadow_blocked_dl.h
|
|
|
|
split/kernel_split_common.h
|
|
|
|
split/kernel_split_data.h
|
|
|
|
split/kernel_split_data_types.h
|
|
|
|
split/kernel_subsurface_scatter.h
|
2015-05-21 12:40:04 +00:00
|
|
|
)
|
|
|
|
|
2019-04-16 04:11:27 +00:00
|
|
|
set(LIB
|
|
|
|
|
|
|
|
)
|
|
|
|
|
2011-11-10 12:52:17 +00:00
|
|
|
# CUDA module
|
|
|
|
|
|
|
|
if(WITH_CYCLES_CUDA_BINARIES)
|
2019-04-17 04:17:24 +00:00
|
|
|
# 64 bit only
|
|
|
|
set(CUDA_BITS 64)
|
|
|
|
|
|
|
|
# CUDA version
|
|
|
|
execute_process(COMMAND ${CUDA_NVCC_EXECUTABLE} "--version" OUTPUT_VARIABLE NVCC_OUT)
|
|
|
|
string(REGEX REPLACE ".*release ([0-9]+)\\.([0-9]+).*" "\\1" CUDA_VERSION_MAJOR "${NVCC_OUT}")
|
|
|
|
string(REGEX REPLACE ".*release ([0-9]+)\\.([0-9]+).*" "\\2" CUDA_VERSION_MINOR "${NVCC_OUT}")
|
|
|
|
set(CUDA_VERSION "${CUDA_VERSION_MAJOR}${CUDA_VERSION_MINOR}")
|
|
|
|
|
|
|
|
# warn for other versions
|
|
|
|
if(CUDA_VERSION MATCHES "101")
|
|
|
|
else()
|
|
|
|
message(WARNING
|
|
|
|
"CUDA version ${CUDA_VERSION_MAJOR}.${CUDA_VERSION_MINOR} detected, "
|
|
|
|
"build may succeed but only CUDA 10.1 is officially supported")
|
|
|
|
endif()
|
|
|
|
|
|
|
|
# build for each arch
|
|
|
|
set(cuda_sources kernels/cuda/kernel.cu kernels/cuda/kernel_split.cu
|
|
|
|
${SRC_HEADERS}
|
|
|
|
${SRC_KERNELS_CUDA_HEADERS}
|
|
|
|
${SRC_BVH_HEADERS}
|
|
|
|
${SRC_SVM_HEADERS}
|
|
|
|
${SRC_GEOM_HEADERS}
|
|
|
|
${SRC_CLOSURE_HEADERS}
|
|
|
|
${SRC_UTIL_HEADERS}
|
|
|
|
)
|
|
|
|
set(cuda_filter_sources kernels/cuda/filter.cu
|
|
|
|
${SRC_HEADERS}
|
|
|
|
${SRC_KERNELS_CUDA_HEADERS}
|
|
|
|
${SRC_FILTER_HEADERS}
|
|
|
|
${SRC_UTIL_HEADERS}
|
|
|
|
)
|
|
|
|
set(cuda_cubins)
|
|
|
|
|
|
|
|
macro(CYCLES_CUDA_KERNEL_ADD arch prev_arch name flags sources experimental)
|
|
|
|
set(cuda_cubin ${name}_${arch}.cubin)
|
|
|
|
|
|
|
|
set(kernel_sources ${sources})
|
|
|
|
if(NOT ${prev_arch} STREQUAL "none")
|
|
|
|
set(kernel_sources ${kernel_sources} ${name}_${prev_arch}.cubin)
|
|
|
|
endif()
|
|
|
|
|
|
|
|
set(cuda_kernel_src "/kernels/cuda/${name}.cu")
|
|
|
|
|
|
|
|
set(cuda_flags
|
|
|
|
-D CCL_NAMESPACE_BEGIN=
|
|
|
|
-D CCL_NAMESPACE_END=
|
|
|
|
-D NVCC
|
|
|
|
-m ${CUDA_BITS}
|
|
|
|
-I ${CMAKE_CURRENT_SOURCE_DIR}/..
|
|
|
|
-I ${CMAKE_CURRENT_SOURCE_DIR}/kernels/cuda
|
|
|
|
--use_fast_math
|
|
|
|
-o ${CMAKE_CURRENT_BINARY_DIR}/${cuda_cubin})
|
|
|
|
|
|
|
|
if(${experimental})
|
|
|
|
set(cuda_flags ${cuda_flags} -D __KERNEL_EXPERIMENTAL__)
|
|
|
|
set(name ${name}_experimental)
|
|
|
|
endif()
|
|
|
|
|
|
|
|
if(WITH_CYCLES_DEBUG)
|
|
|
|
set(cuda_flags ${cuda_flags} -D __KERNEL_DEBUG__)
|
|
|
|
endif()
|
|
|
|
|
|
|
|
if(WITH_CYCLES_CUBIN_COMPILER)
|
|
|
|
string(SUBSTRING ${arch} 3 -1 CUDA_ARCH)
|
|
|
|
|
|
|
|
# Needed to find libnvrtc-builtins.so. Can't do it from inside
|
|
|
|
# cycles_cubin_cc since the env variable is read before main()
|
|
|
|
if(APPLE)
|
|
|
|
set(CUBIN_CC_ENV ${CMAKE_COMMAND}
|
|
|
|
-E env DYLD_LIBRARY_PATH="${CUDA_TOOLKIT_ROOT_DIR}/lib")
|
|
|
|
elseif(UNIX)
|
|
|
|
set(CUBIN_CC_ENV ${CMAKE_COMMAND}
|
|
|
|
-E env LD_LIBRARY_PATH="${CUDA_TOOLKIT_ROOT_DIR}/lib64")
|
|
|
|
endif()
|
|
|
|
|
|
|
|
add_custom_command(
|
|
|
|
OUTPUT ${cuda_cubin}
|
|
|
|
COMMAND ${CUBIN_CC_ENV}
|
|
|
|
"$<TARGET_FILE:cycles_cubin_cc>"
|
|
|
|
-target ${CUDA_ARCH}
|
|
|
|
-i ${CMAKE_CURRENT_SOURCE_DIR}${cuda_kernel_src}
|
|
|
|
${cuda_flags}
|
|
|
|
-v
|
|
|
|
-cuda-toolkit-dir "${CUDA_TOOLKIT_ROOT_DIR}"
|
|
|
|
DEPENDS ${kernel_sources} cycles_cubin_cc)
|
|
|
|
else()
|
|
|
|
add_custom_command(
|
|
|
|
OUTPUT ${cuda_cubin}
|
|
|
|
COMMAND ${CUDA_NVCC_EXECUTABLE}
|
|
|
|
-arch=${arch}
|
|
|
|
${CUDA_NVCC_FLAGS}
|
|
|
|
--cubin
|
|
|
|
${CMAKE_CURRENT_SOURCE_DIR}${cuda_kernel_src}
|
|
|
|
--ptxas-options="-v"
|
|
|
|
${cuda_flags}
|
|
|
|
DEPENDS ${kernel_sources})
|
|
|
|
endif()
|
|
|
|
delayed_install("${CMAKE_CURRENT_BINARY_DIR}" "${cuda_cubin}" ${CYCLES_INSTALL_PATH}/lib)
|
|
|
|
list(APPEND cuda_cubins ${cuda_cubin})
|
|
|
|
|
|
|
|
unset(cuda_debug_flags)
|
|
|
|
endmacro()
|
|
|
|
|
|
|
|
set(prev_arch "none")
|
|
|
|
foreach(arch ${CYCLES_CUDA_BINARIES_ARCH})
|
|
|
|
if(${arch} MATCHES "sm_2.")
|
|
|
|
message(STATUS "CUDA binaries for ${arch} are no longer supported, skipped.")
|
|
|
|
elseif(${arch} MATCHES "sm_7." AND ${CUDA_VERSION} LESS 100)
|
|
|
|
message(STATUS "CUDA binaries for ${arch} require CUDA 10.0+, skipped.")
|
|
|
|
else()
|
|
|
|
# Compile regular kernel
|
|
|
|
CYCLES_CUDA_KERNEL_ADD(${arch} ${prev_arch} filter "" "${cuda_filter_sources}" FALSE)
|
|
|
|
CYCLES_CUDA_KERNEL_ADD(${arch} ${prev_arch} kernel "" "${cuda_sources}" FALSE)
|
|
|
|
|
|
|
|
if(WITH_CYCLES_CUDA_SPLIT_KERNEL_BINARIES)
|
|
|
|
# Compile split kernel
|
|
|
|
CYCLES_CUDA_KERNEL_ADD(${arch} ${prev_arch} kernel_split "-D __SPLIT__" "${cuda_sources}" FALSE)
|
|
|
|
endif()
|
|
|
|
|
|
|
|
if(WITH_CYCLES_CUDA_BUILD_SERIAL)
|
|
|
|
set(prev_arch ${arch})
|
|
|
|
endif()
|
|
|
|
endif()
|
|
|
|
endforeach()
|
|
|
|
|
|
|
|
add_custom_target(cycles_kernel_cuda ALL DEPENDS ${cuda_cubins})
|
|
|
|
cycles_set_solution_folder(cycles_kernel_cuda)
|
2011-11-10 12:52:17 +00:00
|
|
|
endif()
|
2011-08-11 12:36:08 +00:00
|
|
|
|
2011-04-27 11:58:34 +00:00
|
|
|
# OSL module
|
|
|
|
|
2011-08-16 16:15:34 +00:00
|
|
|
if(WITH_CYCLES_OSL)
|
2019-04-17 04:17:24 +00:00
|
|
|
list(APPEND LIB
|
|
|
|
cycles_kernel_osl
|
|
|
|
)
|
|
|
|
add_subdirectory(osl)
|
|
|
|
add_subdirectory(shaders)
|
2011-08-16 16:15:34 +00:00
|
|
|
endif()
|
2011-04-27 11:58:34 +00:00
|
|
|
|
|
|
|
# CPU module
|
|
|
|
|
2011-11-08 20:27:37 +00:00
|
|
|
include_directories(${INC})
|
2012-09-20 09:04:43 +00:00
|
|
|
include_directories(SYSTEM ${INC_SYS})
|
2011-04-27 11:58:34 +00:00
|
|
|
|
2017-03-30 09:23:26 +00:00
|
|
|
set_source_files_properties(kernels/cpu/kernel.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_KERNEL_FLAGS}")
|
|
|
|
set_source_files_properties(kernels/cpu/kernel_split.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_KERNEL_FLAGS}")
|
2017-05-07 12:40:58 +00:00
|
|
|
set_source_files_properties(kernels/cpu/filter.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_KERNEL_FLAGS}")
|
2017-03-30 09:23:26 +00:00
|
|
|
|
2017-05-16 19:27:57 +00:00
|
|
|
if(CXX_HAS_SSE)
|
2019-04-17 04:17:24 +00:00
|
|
|
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 "${CYCLES_SSE41_KERNEL_FLAGS}")
|
|
|
|
set_source_files_properties(kernels/cpu/kernel_split_sse2.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_SSE2_KERNEL_FLAGS}")
|
|
|
|
set_source_files_properties(kernels/cpu/kernel_split_sse3.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_SSE3_KERNEL_FLAGS}")
|
|
|
|
set_source_files_properties(kernels/cpu/kernel_split_sse41.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_SSE41_KERNEL_FLAGS}")
|
|
|
|
set_source_files_properties(kernels/cpu/filter_sse2.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_SSE2_KERNEL_FLAGS}")
|
|
|
|
set_source_files_properties(kernels/cpu/filter_sse3.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_SSE3_KERNEL_FLAGS}")
|
|
|
|
set_source_files_properties(kernels/cpu/filter_sse41.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_SSE41_KERNEL_FLAGS}")
|
2017-05-16 19:27:57 +00:00
|
|
|
endif()
|
2014-03-26 23:15:36 +00:00
|
|
|
|
2017-05-16 19:27:57 +00:00
|
|
|
if(CXX_HAS_AVX)
|
2019-04-17 04:17:24 +00:00
|
|
|
set_source_files_properties(kernels/cpu/kernel_avx.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_AVX_KERNEL_FLAGS}")
|
|
|
|
set_source_files_properties(kernels/cpu/kernel_split_avx.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_AVX_KERNEL_FLAGS}")
|
|
|
|
set_source_files_properties(kernels/cpu/filter_avx.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_AVX_KERNEL_FLAGS}")
|
2017-05-16 19:27:57 +00:00
|
|
|
endif()
|
2017-05-15 16:37:33 +00:00
|
|
|
|
2017-05-16 19:27:57 +00:00
|
|
|
if(CXX_HAS_AVX2)
|
2019-04-17 04:17:24 +00:00
|
|
|
set_source_files_properties(kernels/cpu/kernel_avx2.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_AVX2_KERNEL_FLAGS}")
|
|
|
|
set_source_files_properties(kernels/cpu/kernel_split_avx2.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_AVX2_KERNEL_FLAGS}")
|
|
|
|
set_source_files_properties(kernels/cpu/filter_avx2.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_AVX2_KERNEL_FLAGS}")
|
2014-12-07 10:22:05 +00:00
|
|
|
endif()
|
2013-11-22 13:16:47 +00:00
|
|
|
|
2019-04-16 04:11:27 +00:00
|
|
|
cycles_add_library(cycles_kernel "${LIB}"
|
2019-04-17 04:17:24 +00:00
|
|
|
${SRC_CPU_KERNELS}
|
|
|
|
${SRC_CUDA_KERNELS}
|
|
|
|
${SRC_OPENCL_KERNELS}
|
|
|
|
${SRC_HEADERS}
|
|
|
|
${SRC_KERNELS_CPU_HEADERS}
|
|
|
|
${SRC_KERNELS_CUDA_HEADERS}
|
|
|
|
${SRC_KERNELS_OPENCL_HEADERS}
|
|
|
|
${SRC_BVH_HEADERS}
|
|
|
|
${SRC_CLOSURE_HEADERS}
|
|
|
|
${SRC_FILTER_HEADERS}
|
|
|
|
${SRC_SVM_HEADERS}
|
|
|
|
${SRC_GEOM_HEADERS}
|
|
|
|
${SRC_SPLIT_HEADERS}
|
2015-12-30 13:51:35 +00:00
|
|
|
)
|
2013-11-22 13:16:47 +00:00
|
|
|
|
2011-08-16 16:15:34 +00:00
|
|
|
if(WITH_CYCLES_CUDA)
|
2019-04-17 04:17:24 +00:00
|
|
|
add_dependencies(cycles_kernel cycles_kernel_cuda)
|
2011-08-16 16:15:34 +00:00
|
|
|
endif()
|
2011-04-27 11:58:34 +00:00
|
|
|
|
2011-11-15 15:13:38 +00:00
|
|
|
# OpenCL kernel
|
2011-05-20 12:26:01 +00:00
|
|
|
|
2011-11-08 20:27:37 +00:00
|
|
|
#set(KERNEL_PREPROCESSED ${CMAKE_CURRENT_BINARY_DIR}/kernel_preprocessed.cl)
|
2011-09-02 14:55:06 +00:00
|
|
|
#add_custom_command(
|
2019-04-17 04:17:24 +00:00
|
|
|
# OUTPUT ${KERNEL_PREPROCESSED}
|
|
|
|
# COMMAND gcc -x c++ -E ${CMAKE_CURRENT_SOURCE_DIR}/kernel.cl -I ${CMAKE_CURRENT_SOURCE_DIR}/../util/ -DCCL_NAMESPACE_BEGIN= -DCCL_NAMESPACE_END= -o ${KERNEL_PREPROCESSED}
|
|
|
|
# DEPENDS ${SRC_KERNEL} ${SRC_UTIL_HEADERS})
|
2011-11-08 20:27:37 +00:00
|
|
|
#add_custom_target(cycles_kernel_preprocess ALL DEPENDS ${KERNEL_PREPROCESSED})
|
|
|
|
#delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${KERNEL_PREPROCESSED}" ${CYCLES_INSTALL_PATH}/kernel)
|
2011-09-02 14:55:06 +00:00
|
|
|
|
2017-10-07 12:08:02 +00:00
|
|
|
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_OPENCL_KERNELS}" ${CYCLES_INSTALL_PATH}/source/kernel/kernels/opencl)
|
|
|
|
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_CUDA_KERNELS}" ${CYCLES_INSTALL_PATH}/source/kernel/kernels/cuda)
|
Cycles: Make all #include statements relative to cycles source directory
The idea is to make include statements more explicit and obvious where the
file is coming from, additionally reducing chance of wrong header being
picked up.
For example, it was not obvious whether bvh.h was refferring to builder
or traversal, whenter node.h is a generic graph node or a shader node
and cases like that.
Surely this might look obvious for the active developers, but after some
time of not touching the code it becomes less obvious where file is coming
from.
This was briefly mentioned in T50824 and seems @brecht is fine with such
explicitness, but need to agree with all active developers before committing
this.
Please note that this patch is lacking changes related on GPU/OpenCL
support. This will be solved if/when we all agree this is a good idea to move
forward.
Reviewers: brecht, lukasstockner97, maiself, nirved, dingto, juicyfruit, swerner
Reviewed By: lukasstockner97, maiself, nirved, dingto
Subscribers: brecht
Differential Revision: https://developer.blender.org/D2586
2017-03-28 18:39:14 +00:00
|
|
|
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_HEADERS}" ${CYCLES_INSTALL_PATH}/source/kernel)
|
2017-10-07 12:08:02 +00:00
|
|
|
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_KERNELS_OPENCL_HEADERS}" ${CYCLES_INSTALL_PATH}/source/kernel/kernels/opencl)
|
Cycles: Make all #include statements relative to cycles source directory
The idea is to make include statements more explicit and obvious where the
file is coming from, additionally reducing chance of wrong header being
picked up.
For example, it was not obvious whether bvh.h was refferring to builder
or traversal, whenter node.h is a generic graph node or a shader node
and cases like that.
Surely this might look obvious for the active developers, but after some
time of not touching the code it becomes less obvious where file is coming
from.
This was briefly mentioned in T50824 and seems @brecht is fine with such
explicitness, but need to agree with all active developers before committing
this.
Please note that this patch is lacking changes related on GPU/OpenCL
support. This will be solved if/when we all agree this is a good idea to move
forward.
Reviewers: brecht, lukasstockner97, maiself, nirved, dingto, juicyfruit, swerner
Reviewed By: lukasstockner97, maiself, nirved, dingto
Subscribers: brecht
Differential Revision: https://developer.blender.org/D2586
2017-03-28 18:39:14 +00:00
|
|
|
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_KERNELS_CUDA_HEADERS}" ${CYCLES_INSTALL_PATH}/source/kernel/kernels/cuda)
|
|
|
|
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_BVH_HEADERS}" ${CYCLES_INSTALL_PATH}/source/kernel/bvh)
|
|
|
|
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_CLOSURE_HEADERS}" ${CYCLES_INSTALL_PATH}/source/kernel/closure)
|
2017-05-07 12:40:58 +00:00
|
|
|
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_FILTER_HEADERS}" ${CYCLES_INSTALL_PATH}/source/kernel/filter)
|
Cycles: Make all #include statements relative to cycles source directory
The idea is to make include statements more explicit and obvious where the
file is coming from, additionally reducing chance of wrong header being
picked up.
For example, it was not obvious whether bvh.h was refferring to builder
or traversal, whenter node.h is a generic graph node or a shader node
and cases like that.
Surely this might look obvious for the active developers, but after some
time of not touching the code it becomes less obvious where file is coming
from.
This was briefly mentioned in T50824 and seems @brecht is fine with such
explicitness, but need to agree with all active developers before committing
this.
Please note that this patch is lacking changes related on GPU/OpenCL
support. This will be solved if/when we all agree this is a good idea to move
forward.
Reviewers: brecht, lukasstockner97, maiself, nirved, dingto, juicyfruit, swerner
Reviewed By: lukasstockner97, maiself, nirved, dingto
Subscribers: brecht
Differential Revision: https://developer.blender.org/D2586
2017-03-28 18:39:14 +00:00
|
|
|
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_SVM_HEADERS}" ${CYCLES_INSTALL_PATH}/source/kernel/svm)
|
|
|
|
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_GEOM_HEADERS}" ${CYCLES_INSTALL_PATH}/source/kernel/geom)
|
|
|
|
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_UTIL_HEADERS}" ${CYCLES_INSTALL_PATH}/source/util)
|
|
|
|
delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_SPLIT_HEADERS}" ${CYCLES_INSTALL_PATH}/source/kernel/split)
|