diff options
| author | Thomas Dinges <blender@dingto.org> | 2014-01-16 20:04:11 +0400 |
|---|---|---|
| committer | Thomas Dinges <blender@dingto.org> | 2014-01-16 20:04:11 +0400 |
| commit | de28a4d4b2c9397c5233a5ee1dbf1400f450a15c (patch) | |
| tree | 1fe23de963e206af3fb2ff2d9e2e3393cd89149c /intern/cycles/util/util_optimization.h | |
| parent | 7c6d52eb07c4bd8142a95eca1dbdc794063859b8 (diff) | |
Cycles: Add an AVX kernel for CPU rendering.
* AVX is available on Intel Sandy Bridge and newer and AMD Bulldozer and newer.
* We don't use dedicated AVX intrinsics yet, but gcc auto vectorization gives a 3% performance improvement for Caminandes. Tested on an i5-3570, Linux x64.
* No change for Windows yet, MSVC 2008 does not support AVX.
Reviewed by: brecht
Differential Revision: https://developer.blender.org/D216
Diffstat (limited to 'intern/cycles/util/util_optimization.h')
| -rw-r--r-- | intern/cycles/util/util_optimization.h | 4 |
1 files changed, 3 insertions, 1 deletions
diff --git a/intern/cycles/util/util_optimization.h b/intern/cycles/util/util_optimization.h index b7a2506c950..ac94d43c998 100644 --- a/intern/cycles/util/util_optimization.h +++ b/intern/cycles/util/util_optimization.h @@ -42,10 +42,12 @@ /* no SSE2 kernel on x86-64, part of regular kernel */ #define WITH_CYCLES_OPTIMIZED_KERNEL_SSE3 #define WITH_CYCLES_OPTIMIZED_KERNEL_SSE41 +#define WITH_CYCLES_OPTIMIZED_KERNEL_AVX -/* VC2008 is not ready for sse41, probably broken blendv intrinsic... */ +/* MSVC 2008, no SSE41 (broken blendv intrinsic) and no AVX support */ #if defined(_MSC_VER) && (_MSC_VER < 1700) #undef WITH_CYCLES_OPTIMIZED_KERNEL_SSE41 +#undef WITH_CYCLES_OPTIMIZED_KERNEL_AVX #endif #endif |