diff options
| author | Brecht Van Lommel <brechtvanlommel@gmail.com> | 2017-09-27 02:38:19 +0300 |
|---|---|---|
| committer | Brecht Van Lommel <brechtvanlommel@gmail.com> | 2017-10-04 22:58:47 +0300 |
| commit | 6da6f8d33f65b427162d0c8b13a5a5f5043bc8a5 (patch) | |
| tree | be10ed89869fd67aa7d430897301f6ecbb2d577a /intern/cycles/kernel/kernels | |
| parent | 77f300e2a9289af026278171b51103bf485297e4 (diff) | |
Cycles: CUDA faster rendering of small tiles, using multiple samples like OpenCL.
The work size is still very conservative, and this doesn't help for progressive
refine. For that we will need to render multiple tiles at the same time. But this
should already help for denoising renders that require too much memory with big
tiles, and just generally soften the performance dropoff with small tiles.
Differential Revision: https://developer.blender.org/D2856
Diffstat (limited to 'intern/cycles/kernel/kernels')
| -rw-r--r-- | intern/cycles/kernel/kernels/cuda/kernel.cu | 2 |
1 files changed, 2 insertions, 0 deletions
diff --git a/intern/cycles/kernel/kernels/cuda/kernel.cu b/intern/cycles/kernel/kernels/cuda/kernel.cu index 799cd587fcf..e72edfa7bdf 100644 --- a/intern/cycles/kernel/kernels/cuda/kernel.cu +++ b/intern/cycles/kernel/kernels/cuda/kernel.cu @@ -21,6 +21,8 @@ #include "kernel/kernel_compat_cuda.h" #include "kernel_config.h" +#include "util/util_atomic.h" + #include "kernel/kernel_math.h" #include "kernel/kernel_types.h" #include "kernel/kernel_globals.h" |