Cycles: Trace indirect subsurface rays by restarting the integrator loop

This gives much lower stack usage on GPU and reduces kernel memory size to
around 448MB on GTX560Ti (comparing to 652MB with previous commit and 946MB
with official release). There's also a barely measurable speedup of around
5%, but this is to be confirmed still.

At this stage we're using only ~3% for the experimental kernel and SSS
rendering seems to be faster by 40% and after some further testing we might
consider making SSS and CMJ official features and remove experimental
precompiled kernels.

1 files changed, 16 insertions, 1 deletions

diff --git a/intern/cycles/kernel/kernel_bake.h b/intern/cycles/kernel/kernel_bake.h
index 0f572b34ae0..3efd7ecca5f 100644
--- a/intern/cycles/kernel/kernel_bake.h
+++ b/intern/cycles/kernel/kernel_bake.h
@@ -74,7 +74,22 @@ ccl_device void compute_light_pass(KernelGlobals *kg, ShaderData *sd, PathRadian
 			                                  &throughput,
 			                                  &ss_indirect))
 			{
-				kernel_path_subsurface_scatter_indirect(kg, &L_sample, &state, &rng, &ray, &ss_indirect);
+				while(ss_indirect.num_rays) {
+					kernel_path_subsurface_setup_indirect(kg,
+					                                      &ss_indirect,
+					                                      &L_sample,
+					                                      &state,
+					                                      &ray,
+					                                      &ray,
+					                                      &throughput);
+					kernel_path_indirect(kg,
+					                     &rng,
+					                     &ray,
+					                     throughput,
+					                     state.num_samples,
+					                     &state,
+					                     &L_sample);
+				}
 				is_sss_sample = true;
 			}
 		}