Blender Git Loki
Git Commits -> Revision 3844b8f
Revision 3844b8f by Thomas Dinges (soc-2014-cycles) May 14, 2014, 19:03 (GMT) |
Cycles: Use some dedicated FMA intrinsics in the AVX2 kernel. This gives me a small speedup of 2% in bmw.blend, and 3% in hair.blend. Could only test on my Macbook with clang though, no idea how gcc or msvc performs here. Thanks to Lockal for some input on this! :) |
Commit Details:
Full Hash: 3844b8f85c7dd849a10b80c5b6b92fe968a19ecf
Parent Commit: ac908f6
Lines Changed: +34, -2
8 Modified Paths:
/intern/cycles/CMakeLists.txt (+1, -1) (Diff)
/intern/cycles/kernel/geom/geom_bvh_shadow.h (+6, -0) (Diff)
/intern/cycles/kernel/geom/geom_bvh_subsurface.h (+6, -0) (Diff)
/intern/cycles/kernel/geom/geom_bvh_traversal.h (+6, -0) (Diff)
/intern/cycles/kernel/kernel_avx2.cpp (+2, -0) (Diff)
/intern/cycles/SConscript (+1, -1) (Diff)
/intern/cycles/util/util_optimization.h (+4, -0) (Diff)
/intern/cycles/util/util_simd.h (+8, -0) (Diff)
/intern/cycles/kernel/geom/geom_bvh_shadow.h (+6, -0) (Diff)
/intern/cycles/kernel/geom/geom_bvh_subsurface.h (+6, -0) (Diff)
/intern/cycles/kernel/geom/geom_bvh_traversal.h (+6, -0) (Diff)
/intern/cycles/kernel/kernel_avx2.cpp (+2, -0) (Diff)
/intern/cycles/SConscript (+1, -1) (Diff)
/intern/cycles/util/util_optimization.h (+4, -0) (Diff)
/intern/cycles/util/util_simd.h (+8, -0) (Diff)