From 6353ecb996898b4ce2fe8065130ed1f5ea3b6989 Mon Sep 17 00:00:00 2001 From: Sergey Sharybin Date: Mon, 1 Aug 2016 15:40:46 +0200 Subject: Cycles: Tweaks to support CUDA 8 toolkit All the changes are mainly giving explicit tips on inlining functions, so they match how inlining worked with previous toolkit. This make kernel compiled by CUDA 8 render in average with same speed as previous kernels. Some scenes are somewhat faster, some of them are somewhat slower. But slowdown is within 1% so far. On a positive side it allows us to enable newer generation cards on buildbots (so GTX 10x0 will be officially supported soon). --- intern/cycles/kernel/kernel_path_surface.h | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) (limited to 'intern/cycles/kernel/kernel_path_surface.h') diff --git a/intern/cycles/kernel/kernel_path_surface.h b/intern/cycles/kernel/kernel_path_surface.h index 74b1ae0ca32..250b8e92a45 100644 --- a/intern/cycles/kernel/kernel_path_surface.h +++ b/intern/cycles/kernel/kernel_path_surface.h @@ -222,8 +222,13 @@ ccl_device_inline void kernel_path_surface_connect_light(KernelGlobals *kg, ccl_ #endif /* path tracing: bounce off or through surface to with new direction stored in ray */ -ccl_device_inline bool kernel_path_surface_bounce(KernelGlobals *kg, ccl_addr_space RNG *rng, - ShaderData *sd, ccl_addr_space float3 *throughput, ccl_addr_space PathState *state, PathRadiance *L, ccl_addr_space Ray *ray) +ccl_device bool kernel_path_surface_bounce(KernelGlobals *kg, + ccl_addr_space RNG *rng, + ShaderData *sd, + ccl_addr_space float3 *throughput, + ccl_addr_space PathState *state, + PathRadiance *L, + ccl_addr_space Ray *ray) { /* no BSDF? we can stop here */ if(ccl_fetch(sd, flag) & SD_BSDF) { -- cgit v1.2.3