Cycles: Cache split kernels in CUDA device
authorSergey Sharybin <sergey.vfx@gmail.com>
Tue, 2 May 2017 13:02:49 +0000 (15:02 +0200)
committerSergey Sharybin <sergey.vfx@gmail.com>
Tue, 2 May 2017 13:03:12 +0000 (15:03 +0200)
commit4174e533c001367b0ef391b72baa5b07cb517ce8
treea18112243e6a2c894d4421926cbc5394b54240a5
parentfc8f428224a8e77ba66394ad8238a406fa24ee3a
Cycles: Cache split kernels in CUDA device

This way we don't re-load kernels for every sample in the viewport.
Additionally, we don't risk global size changed inbetween of samples.
intern/cycles/device/device_cuda.cpp