diff options
author | Hristo Gueorguiev <prem.nirved@gmail.com> | 2017-05-03 16:30:45 +0300 |
---|---|---|
committer | Hristo Gueorguiev <prem.nirved@gmail.com> | 2017-05-03 16:30:45 +0300 |
commit | 6bf4115c13962c99d1cdc97f2be92c4922f3fd33 (patch) | |
tree | 569c512a242caf2ea4465f2eef561933ed937a2f /intern/cycles/kernel/kernels/opencl/kernel_split.cl | |
parent | 6f9c839f444f92c4b0c336a6f5e31cb9660d7dbc (diff) |
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
Diffstat (limited to 'intern/cycles/kernel/kernels/opencl/kernel_split.cl')
-rw-r--r-- | intern/cycles/kernel/kernels/opencl/kernel_split.cl | 2 |
1 files changed, 2 insertions, 0 deletions
diff --git a/intern/cycles/kernel/kernels/opencl/kernel_split.cl b/intern/cycles/kernel/kernels/opencl/kernel_split.cl index 732cda30115..8de82db7afe 100644 --- a/intern/cycles/kernel/kernels/opencl/kernel_split.cl +++ b/intern/cycles/kernel/kernels/opencl/kernel_split.cl @@ -23,6 +23,8 @@ #include "kernel/kernels/opencl/kernel_do_volume.cl" #include "kernel/kernels/opencl/kernel_indirect_background.cl" #include "kernel/kernels/opencl/kernel_queue_enqueue.cl" +#include "kernel/kernels/opencl/kernel_shader_setup.cl" +#include "kernel/kernels/opencl/kernel_shader_sort.cl" #include "kernel/kernels/opencl/kernel_shader_eval.cl" #include "kernel/kernels/opencl/kernel_holdout_emission_blurring_pathtermination_ao.cl" #include "kernel/kernels/opencl/kernel_subsurface_scatter.cl" |