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/cpu/kernel_cpu_impl.h | |
| 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/cpu/kernel_cpu_impl.h')
| -rw-r--r-- | intern/cycles/kernel/kernels/cpu/kernel_cpu_impl.h | 8 |
1 files changed, 7 insertions, 1 deletions
diff --git a/intern/cycles/kernel/kernels/cpu/kernel_cpu_impl.h b/intern/cycles/kernel/kernels/cpu/kernel_cpu_impl.h index 96f54bb427e..8c05dd1d9ef 100644 --- a/intern/cycles/kernel/kernels/cpu/kernel_cpu_impl.h +++ b/intern/cycles/kernel/kernels/cpu/kernel_cpu_impl.h @@ -44,6 +44,8 @@ # include "kernel/split/kernel_do_volume.h" # include "kernel/split/kernel_queue_enqueue.h" # include "kernel/split/kernel_indirect_background.h" +# include "kernel/split/kernel_shader_setup.h" +# include "kernel/split/kernel_shader_sort.h" # include "kernel/split/kernel_shader_eval.h" # include "kernel/split/kernel_holdout_emission_blurring_pathtermination_ao.h" # include "kernel/split/kernel_subsurface_scatter.h" @@ -181,7 +183,9 @@ DEFINE_SPLIT_KERNEL_FUNCTION(lamp_emission) DEFINE_SPLIT_KERNEL_FUNCTION(do_volume) DEFINE_SPLIT_KERNEL_FUNCTION_LOCALS(queue_enqueue, QueueEnqueueLocals) DEFINE_SPLIT_KERNEL_FUNCTION(indirect_background) -DEFINE_SPLIT_KERNEL_FUNCTION_LOCALS(shader_eval, uint) +DEFINE_SPLIT_KERNEL_FUNCTION_LOCALS(shader_setup, uint) +DEFINE_SPLIT_KERNEL_FUNCTION_LOCALS(shader_sort, ShaderSortLocals) +DEFINE_SPLIT_KERNEL_FUNCTION(shader_eval) DEFINE_SPLIT_KERNEL_FUNCTION_LOCALS(holdout_emission_blurring_pathtermination_ao, BackgroundAOLocals) DEFINE_SPLIT_KERNEL_FUNCTION(subsurface_scatter) DEFINE_SPLIT_KERNEL_FUNCTION_LOCALS(direct_lighting, uint) @@ -209,6 +213,8 @@ void KERNEL_FUNCTION_FULL_NAME(register_functions)(void(*reg)(const char* name, REGISTER(do_volume); REGISTER(queue_enqueue); REGISTER(indirect_background); + REGISTER(shader_setup); + REGISTER(shader_sort); REGISTER(shader_eval); REGISTER(holdout_emission_blurring_pathtermination_ao); REGISTER(subsurface_scatter); |