Blender Git Loki
Git Commits -> Revision 6bf4115
Revision 6bf4115 by Hristo Gueorguiev (master) May 3, 2017, 13:30 (GMT) |
Cycles: Split kernel - sort shaders Reduce thread divergence in kernel_shader_eval. Rays are sorted in blocks of 2048 according to shader->id. On R9 290 Classroom is ~30% faster, and Pabellon Barcelone is ~8% faster. No sorting for CUDA split kernel. Reviewers: sergey, maiself Reviewed By: maiself Differential Revision: https://developer.blender.org/D2598 |
Commit Details:
Full Hash: 6bf4115c13962c99d1cdc97f2be92c4922f3fd33
Parent Commit: 6f9c839
Lines Changed: +294, -37
4 Added Paths:
/intern/cycles/kernel/kernels/opencl/kernel_shader_setup.cl (+27, -0) (View)
/intern/cycles/kernel/kernels/opencl/kernel_shader_sort.cl (+28, -0) (View)
/intern/cycles/kernel/split/kernel_shader_setup.h (+70, -0) (View)
/intern/cycles/kernel/split/kernel_shader_sort.h (+97, -0) (View)
/intern/cycles/kernel/kernels/opencl/kernel_shader_sort.cl (+28, -0) (View)
/intern/cycles/kernel/split/kernel_shader_setup.h (+70, -0) (View)
/intern/cycles/kernel/split/kernel_shader_sort.h (+97, -0) (View)
11 Modified Paths:
/intern/cycles/device/device_split_kernel.cpp (+8, -0) (Diff)
/intern/cycles/device/device_split_kernel.h (+2, -0) (Diff)
/intern/cycles/kernel/CMakeLists.txt (+6, -0) (Diff)
/intern/cycles/kernel/kernels/cpu/kernel_cpu.h (+2, -0) (Diff)
/intern/cycles/kernel/kernels/cpu/kernel_cpu_impl.h (+7, -1) (Diff)
/intern/cycles/kernel/kernels/cuda/kernel_split.cu (+5, -1) (Diff)
/intern/cycles/kernel/kernels/opencl/kernel_shader_eval.cl (+1, -2) (Diff)
/intern/cycles/kernel/kernels/opencl/kernel_split.cl (+2, -0) (Diff)
/intern/cycles/kernel/kernel_types.h (+15, -0) (Diff)
/intern/cycles/kernel/split/kernel_shader_eval.h (+19, -33) (Diff)
/intern/cycles/kernel/split/kernel_split_data_types.h (+5, -0) (Diff)
/intern/cycles/device/device_split_kernel.h (+2, -0) (Diff)
/intern/cycles/kernel/CMakeLists.txt (+6, -0) (Diff)
/intern/cycles/kernel/kernels/cpu/kernel_cpu.h (+2, -0) (Diff)
/intern/cycles/kernel/kernels/cpu/kernel_cpu_impl.h (+7, -1) (Diff)
/intern/cycles/kernel/kernels/cuda/kernel_split.cu (+5, -1) (Diff)
/intern/cycles/kernel/kernels/opencl/kernel_shader_eval.cl (+1, -2) (Diff)
/intern/cycles/kernel/kernels/opencl/kernel_split.cl (+2, -0) (Diff)
/intern/cycles/kernel/kernel_types.h (+15, -0) (Diff)
/intern/cycles/kernel/split/kernel_shader_eval.h (+19, -33) (Diff)
/intern/cycles/kernel/split/kernel_split_data_types.h (+5, -0) (Diff)