Cycles: merge of cycles-x branch, a major update to the renderer

This includes much improved GPU rendering performance, viewport interactivity,
new shadow catcher, revamped sampling settings, subsurface scattering anisotropy,
new GPU volume sampling, improved PMJ sampling pattern, and more.

Some features have also been removed or changed, breaking backwards compatibility.
Including the removal of the OpenCL backend, for which alternatives are under
development.

Release notes and code docs:
https://wiki.blender.org/wiki/Reference/Release_Notes/3.0/Cycles
https://wiki.blender.org/wiki/Source/Render/Cycles

Credits:
* Sergey Sharybin
* Brecht Van Lommel
* Patrick Mours (OptiX backend)
* Christophe Hery (subsurface scattering anisotropy)
* William Leeson (PMJ sampling pattern)
* Alaska (various fixes and tweaks)
* Thomas Dinges (various fixes)

For the full commit history, see the cycles-x branch. This squashes together
all the changes since intermediate changes would often fail building or tests.

Ref T87839, T87837, T87836
Fixes T90734, T89353, T80267, T80267, T77185, T69800

1 files changed, 106 insertions, 208 deletions

diff --git a/intern/cycles/kernel/CMakeLists.txt b/intern/cycles/kernel/CMakeLists.txt
index 0ce33c51778..4196539a9b1 100644
--- a/intern/cycles/kernel/CMakeLists.txt
+++ b/intern/cycles/kernel/CMakeLists.txt
@@ -22,68 +22,22 @@ set(INC_SYS
 
 )
 
-set(SRC_CPU_KERNELS
-  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
+set(SRC_DEVICE_CPU
+  device/cpu/kernel.cpp
+  device/cpu/kernel_sse2.cpp
+  device/cpu/kernel_sse3.cpp
+  device/cpu/kernel_sse41.cpp
+  device/cpu/kernel_avx.cpp
+  device/cpu/kernel_avx2.cpp
 )
 
-set(SRC_CUDA_KERNELS
-  kernels/cuda/kernel.cu
-  kernels/cuda/kernel_split.cu
-  kernels/cuda/filter.cu
+set(SRC_DEVICE_CUDA
+  device/cuda/kernel.cu
 )
 
-set(SRC_OPENCL_KERNELS
-  kernels/opencl/kernel_adaptive_stopping.cl
-  kernels/opencl/kernel_adaptive_filter_x.cl
-  kernels/opencl/kernel_adaptive_filter_y.cl
-  kernels/opencl/kernel_adaptive_adjust_samples.cl
-  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
-)
-
-set(SRC_OPTIX_KERNELS
-  kernels/optix/kernel_optix.cu
+set(SRC_DEVICE_OPTIX
+  device/optix/kernel.cu
+  device/optix/kernel_shader_raytrace.cu
 )
 
 set(SRC_BVH_HEADERS
@@ -105,63 +59,56 @@ set(SRC_HEADERS
   kernel_bake.h
   kernel_camera.h
   kernel_color.h
-  kernel_compat_cpu.h
-  kernel_compat_cuda.h
-  kernel_compat_optix.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_light_background.h
   kernel_light_common.h
+  kernel_lookup_table.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_shadow_catcher.h
   kernel_textures.h
   kernel_types.h
-  kernel_volume.h
   kernel_work_stealing.h
   kernel_write_passes.h
 )
 
-set(SRC_KERNELS_CPU_HEADERS
-  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
+set(SRC_DEVICE_CPU_HEADERS
+  device/cpu/compat.h
+  device/cpu/image.h
+  device/cpu/globals.h
+  device/cpu/kernel.h
+  device/cpu/kernel_arch.h
+  device/cpu/kernel_arch_impl.h
 )
-
-set(SRC_KERNELS_CUDA_HEADERS
-  kernels/cuda/kernel_config.h
-  kernels/cuda/kernel_cuda_image.h
+set(SRC_DEVICE_GPU_HEADERS
+  device/gpu/image.h
+  device/gpu/kernel.h
+  device/gpu/parallel_active_index.h
+  device/gpu/parallel_prefix_sum.h
+  device/gpu/parallel_reduce.h
+  device/gpu/parallel_sorted_index.h
 )
 
-set(SRC_KERNELS_OPTIX_HEADERS
+set(SRC_DEVICE_CUDA_HEADERS
+  device/cuda/compat.h
+  device/cuda/config.h
+  device/cuda/globals.h
 )
 
-set(SRC_KERNELS_OPENCL_HEADERS
-  kernels/opencl/kernel_split_function.h
-  kernels/opencl/kernel_opencl_image.h
+set(SRC_DEVICE_OPTIX_HEADERS
+  device/optix/compat.h
+  device/optix/globals.h
 )
 
 set(SRC_CLOSURE_HEADERS
@@ -259,25 +206,32 @@ set(SRC_GEOM_HEADERS
   geom/geom_object.h
   geom/geom_patch.h
   geom/geom_primitive.h
+  geom/geom_shader_data.h
   geom/geom_subd_triangle.h
   geom/geom_triangle.h
   geom/geom_triangle_intersect.h
   geom/geom_volume.h
 )
 
-set(SRC_FILTER_HEADERS
-  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
+set(SRC_INTEGRATOR_HEADERS
+  integrator/integrator_init_from_bake.h
+  integrator/integrator_init_from_camera.h
+  integrator/integrator_intersect_closest.h
+  integrator/integrator_intersect_shadow.h
+  integrator/integrator_intersect_subsurface.h
+  integrator/integrator_intersect_volume_stack.h
+  integrator/integrator_megakernel.h
+  integrator/integrator_shade_background.h
+  integrator/integrator_shade_light.h
+  integrator/integrator_shade_shadow.h
+  integrator/integrator_shade_surface.h
+  integrator/integrator_shade_volume.h
+  integrator/integrator_state.h
+  integrator/integrator_state_flow.h
+  integrator/integrator_state_template.h
+  integrator/integrator_state_util.h
+  integrator/integrator_subsurface.h
+  integrator/integrator_volume_stack.h
 )
 
 set(SRC_UTIL_HEADERS
@@ -333,36 +287,6 @@ set(SRC_UTIL_HEADERS
   ../util/util_types_vector3_impl.h
 )
 
-set(SRC_SPLIT_HEADERS
-  split/kernel_adaptive_adjust_samples.h
-  split/kernel_adaptive_filter_x.h
-  split/kernel_adaptive_filter_y.h
-  split/kernel_adaptive_stopping.h
-  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
-)
-
 set(LIB
 
 )
@@ -393,21 +317,17 @@ if(WITH_CYCLES_CUDA_BINARIES)
   endif()
 
   # build for each arch
-  set(cuda_sources kernels/cuda/kernel.cu kernels/cuda/kernel_split.cu
+  set(cuda_sources device/cuda/kernel.cu
     ${SRC_HEADERS}
-    ${SRC_KERNELS_CUDA_HEADERS}
+    ${SRC_DEVICE_GPU_HEADERS}
+    ${SRC_DEVICE_CUDA_HEADERS}
     ${SRC_BVH_HEADERS}
     ${SRC_SVM_HEADERS}
     ${SRC_GEOM_HEADERS}
+    ${SRC_INTEGRATOR_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)
@@ -427,7 +347,7 @@ if(WITH_CYCLES_CUDA_BINARIES)
       endif()
     endif()
 
-    set(cuda_kernel_src "/kernels/cuda/${name}.cu")
+    set(cuda_kernel_src "/device/cuda/${name}.cu")
 
     set(cuda_flags ${flags}
       -D CCL_NAMESPACE_BEGIN=
@@ -435,7 +355,7 @@ if(WITH_CYCLES_CUDA_BINARIES)
       -D NVCC
       -m ${CUDA_BITS}
       -I ${CMAKE_CURRENT_SOURCE_DIR}/..
-      -I ${CMAKE_CURRENT_SOURCE_DIR}/kernels/cuda
+      -I ${CMAKE_CURRENT_SOURCE_DIR}/device/cuda
       --use_fast_math
       -o ${CMAKE_CURRENT_BINARY_DIR}/${cuda_file})
 
@@ -523,14 +443,8 @@ if(WITH_CYCLES_CUDA_BINARIES)
     endif()
     if(DEFINED cuda_nvcc_executable AND DEFINED cuda_toolkit_root_dir)
       # 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()
@@ -547,15 +461,15 @@ endif()
 # OptiX PTX modules
 
 if(WITH_CYCLES_DEVICE_OPTIX AND WITH_CYCLES_CUDA_BINARIES)
-  macro(CYCLES_OPTIX_KERNEL_ADD name flags)
-    set(input "kernels/optix/kernel_optix.cu")
+  macro(CYCLES_OPTIX_KERNEL_ADD name input flags)
     set(output "${CMAKE_CURRENT_BINARY_DIR}/${name}.ptx")
 
     set(cuda_flags ${flags}
       -I "${OPTIX_INCLUDE_DIR}"
       -I "${CMAKE_CURRENT_SOURCE_DIR}/.."
-      -I "${CMAKE_CURRENT_SOURCE_DIR}/kernels/cuda"
+      -I "${CMAKE_CURRENT_SOURCE_DIR}/device/cuda"
       --use_fast_math
+      -Wno-deprecated-gpu-targets
       -o ${output})
 
     if(WITH_NANOVDB)
@@ -580,11 +494,13 @@ if(WITH_CYCLES_DEVICE_OPTIX AND WITH_CYCLES_CUDA_BINARIES)
         DEPENDS
           ${input}
           ${SRC_HEADERS}
-          ${SRC_KERNELS_CUDA_HEADERS}
-          ${SRC_KERNELS_OPTIX_HEADERS}
+          ${SRC_DEVICE_GPU_HEADERS}
+          ${SRC_DEVICE_CUDA_HEADERS}
+          ${SRC_DEVICE_OPTIX_HEADERS}
           ${SRC_BVH_HEADERS}
           ${SRC_SVM_HEADERS}
           ${SRC_GEOM_HEADERS}
+          ${SRC_INTEGRATOR_HEADERS}
           ${SRC_CLOSURE_HEADERS}
           ${SRC_UTIL_HEADERS}
         COMMAND ${CUBIN_CC_ENV}
@@ -603,11 +519,13 @@ if(WITH_CYCLES_DEVICE_OPTIX AND WITH_CYCLES_CUDA_BINARIES)
         DEPENDS
           ${input}
           ${SRC_HEADERS}
-          ${SRC_KERNELS_CUDA_HEADERS}
-          ${SRC_KERNELS_OPTIX_HEADERS}
+          ${SRC_DEVICE_GPU_HEADERS}
+          ${SRC_DEVICE_CUDA_HEADERS}
+          ${SRC_DEVICE_OPTIX_HEADERS}
           ${SRC_BVH_HEADERS}
           ${SRC_SVM_HEADERS}
           ${SRC_GEOM_HEADERS}
+          ${SRC_INTEGRATOR_HEADERS}
           ${SRC_CLOSURE_HEADERS}
           ${SRC_UTIL_HEADERS}
         COMMAND
@@ -624,8 +542,14 @@ if(WITH_CYCLES_DEVICE_OPTIX AND WITH_CYCLES_CUDA_BINARIES)
     delayed_install("${CMAKE_CURRENT_BINARY_DIR}" "${output}" ${CYCLES_INSTALL_PATH}/lib)
   endmacro()
 
-  CYCLES_OPTIX_KERNEL_ADD(kernel_optix "-D __NO_SHADER_RAYTRACE__")
-  CYCLES_OPTIX_KERNEL_ADD(kernel_optix_shader_raytrace "--keep-device-functions")
+  CYCLES_OPTIX_KERNEL_ADD(
+    kernel_optix
+    "device/optix/kernel.cu"
+    "")
+  CYCLES_OPTIX_KERNEL_ADD(
+    kernel_optix_shader_raytrace
+    "device/optix/kernel_shader_raytrace.cu"
+    "--keep-device-functions")
 
   add_custom_target(cycles_kernel_optix ALL DEPENDS ${optix_ptx})
   cycles_set_solution_folder(cycles_kernel_optix)
@@ -659,62 +583,47 @@ if(WITH_COMPILER_ASAN)
   endif()
 endif()
 
-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}")
-set_source_files_properties(kernels/cpu/filter.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_KERNEL_FLAGS}")
+set_source_files_properties(device/cpu/kernel.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_KERNEL_FLAGS}")
 
 if(CXX_HAS_SSE)
-  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 "${CYCLES_SSE41_KERNEL_FLAGS}")
+  set_source_files_properties(device/cpu/kernel_sse2.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_SSE2_KERNEL_FLAGS}")
+  set_source_files_properties(device/cpu/kernel_sse3.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_SSE3_KERNEL_FLAGS}")
+  set_source_files_properties(device/cpu/kernel_sse41.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_SSE41_KERNEL_FLAGS}")
 endif()
 
 if(CXX_HAS_AVX)
-  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 "${CYCLES_AVX_KERNEL_FLAGS}")
+  set_source_files_properties(device/cpu/kernel_avx.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_AVX_KERNEL_FLAGS}")
 endif()
 
 if(CXX_HAS_AVX2)
-  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 "${CYCLES_AVX2_KERNEL_FLAGS}")
+  set_source_files_properties(device/cpu/kernel_avx2.cpp PROPERTIES COMPILE_FLAGS "${CYCLES_AVX2_KERNEL_FLAGS}")
 endif()
 
 cycles_add_library(cycles_kernel "${LIB}"
-  ${SRC_CPU_KERNELS}
-  ${SRC_CUDA_KERNELS}
-  ${SRC_OPTIX_KERNELS}
-  ${SRC_OPENCL_KERNELS}
+  ${SRC_DEVICE_CPU}
+  ${SRC_DEVICE_CUDA}
+  ${SRC_DEVICE_OPTIX}
   ${SRC_HEADERS}
-  ${SRC_KERNELS_CPU_HEADERS}
-  ${SRC_KERNELS_CUDA_HEADERS}
-  ${SRC_KERNELS_OPTIX_HEADERS}
-  ${SRC_KERNELS_OPENCL_HEADERS}
+  ${SRC_DEVICE_CPU_HEADERS}
+  ${SRC_DEVICE_GPU_HEADERS}
+  ${SRC_DEVICE_CUDA_HEADERS}
+  ${SRC_DEVICE_OPTIX_HEADERS}
   ${SRC_BVH_HEADERS}
   ${SRC_CLOSURE_HEADERS}
-  ${SRC_FILTER_HEADERS}
   ${SRC_SVM_HEADERS}
   ${SRC_GEOM_HEADERS}
-  ${SRC_SPLIT_HEADERS}
+  ${SRC_INTEGRATOR_HEADERS}
 )
 
 source_group("bvh" FILES ${SRC_BVH_HEADERS})
 source_group("closure" FILES ${SRC_CLOSURE_HEADERS})
-source_group("filter" FILES ${SRC_FILTER_HEADERS})
 source_group("geom" FILES ${SRC_GEOM_HEADERS})
+source_group("integrator" FILES ${SRC_INTEGRATOR_HEADERS})
 source_group("kernel" FILES ${SRC_HEADERS})
-source_group("kernel\\split" FILES ${SRC_SPLIT_HEADERS})
-source_group("kernels\\cpu" FILES ${SRC_CPU_KERNELS} ${SRC_KERNELS_CPU_HEADERS})
-source_group("kernels\\cuda" FILES ${SRC_CUDA_KERNELS} ${SRC_KERNELS_CUDA_HEADERS})
-source_group("kernels\\opencl" FILES ${SRC_OPENCL_KERNELS} ${SRC_KERNELS_OPENCL_HEADERS})
-source_group("kernels\\optix" FILES ${SRC_OPTIX_KERNELS} ${SRC_KERNELS_OPTIX_HEADERS})
+source_group("device\\cpu" FILES ${SRC_DEVICE_CPU} ${SRC_DEVICE_CPU_HEADERS})
+source_group("device\\gpu" FILES ${SRC_DEVICE_GPU_HEADERS})
+source_group("device\\cuda" FILES ${SRC_DEVICE_CUDA} ${SRC_DEVICE_CUDA_HEADERS})
+source_group("device\\optix" FILES ${SRC_DEVICE_OPTIX} ${SRC_DEVICE_OPTIX_HEADERS})
 source_group("svm" FILES ${SRC_SVM_HEADERS})
 
 if(WITH_CYCLES_CUDA)
@@ -724,31 +633,20 @@ if(WITH_CYCLES_DEVICE_OPTIX AND WITH_CYCLES_CUDA_BINARIES)
   add_dependencies(cycles_kernel cycles_kernel_optix)
 endif()
 
-# OpenCL kernel
-
-# set(KERNEL_PREPROCESSED ${CMAKE_CURRENT_BINARY_DIR}/kernel_preprocessed.cl)
-# add_custom_command(
-#    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})
-# add_custom_target(cycles_kernel_preprocess ALL DEPENDS ${KERNEL_PREPROCESSED})
-# delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${KERNEL_PREPROCESSED}" ${CYCLES_INSTALL_PATH}/kernel)
+# Install kernel source for runtime compilation
 
-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}" ${CYCLES_INSTALL_PATH}/source/kernel/kernels/cuda)
-delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_OPTIX_KERNELS}" ${CYCLES_INSTALL_PATH}/source/kernel/kernels/optix)
+delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_DEVICE_CUDA}" ${CYCLES_INSTALL_PATH}/source/kernel/device/cuda)
+delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_DEVICE_OPTIX}" ${CYCLES_INSTALL_PATH}/source/kernel/device/optix)
 delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_HEADERS}" ${CYCLES_INSTALL_PATH}/source/kernel)
-delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_KERNELS_OPENCL_HEADERS}" ${CYCLES_INSTALL_PATH}/source/kernel/kernels/opencl)
-delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_KERNELS_CUDA_HEADERS}" ${CYCLES_INSTALL_PATH}/source/kernel/kernels/cuda)
-delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_KERNELS_OPTIX_HEADERS}" ${CYCLES_INSTALL_PATH}/source/kernel/kernels/optix)
+delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_DEVICE_GPU_HEADERS}" ${CYCLES_INSTALL_PATH}/source/kernel/device/gpu)
+delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_DEVICE_CUDA_HEADERS}" ${CYCLES_INSTALL_PATH}/source/kernel/device/cuda)
+delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_DEVICE_OPTIX_HEADERS}" ${CYCLES_INSTALL_PATH}/source/kernel/device/optix)
 delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_BVH_HEADERS}" ${CYCLES_INSTALL_PATH}/source/kernel/bvh)
 delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_CLOSURE_HEADERS}" ${CYCLES_INSTALL_PATH}/source/kernel/closure)
-delayed_install(${CMAKE_CURRENT_SOURCE_DIR} "${SRC_FILTER_HEADERS}" ${CYCLES_INSTALL_PATH}/source/kernel/filter)
 delayed_install(${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_INTEGRATOR_HEADERS}" ${CYCLES_INSTALL_PATH}/source/kernel/integrator)
 delayed_install(${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)
-
 
 if(WITH_NANOVDB)
   set(SRC_NANOVDB_HEADERS