Cycles: shadow function optimization for transparent shadows (CPU only).

Old algorithm:

Raytrace from one transparent surface to the next step by step. To minimize
overhead in cases where we don't need transparent shadows, we first trace a
regular shadow ray. We check if the hit primitive was potentially transparent,
and only in that case start marching. this gives extra ray cast for the cases
were we do want transparency.

New algorithm:

We trace a single ray. If it hits any opaque surface, or more than a given
number of transparent surfaces is hit, then we consider the geometry to be
entirely blocked. If not, all transparent surfaces will be recorded and we
will shade them one by one to determine how much light is blocked. This all
happens in one scene intersection function.

Recording all hits works well in some cases but may be slower in others. If
we have many semi-transparent hairs, one intersection may be faster because
you'd be reinteresecting the same hairs a lot with each step otherwise. If
however there is mostly binary transparency then we may be recording many
unnecessary intersections when one of the first surfaces blocks all light.

We found that this helps quite nicely in some scenes, on koro.blend this can
give a 50% reduction in render time, on the pabellon barcelona scene and a
forest scene with transparent leaves it was 30%. Some other files rendered
maybe 1% or 2% slower, but this seems a reasonable tradeoff.

Differential Revision: https://developer.blender.org/D473

1 files changed, 76 insertions, 0 deletions

diff --git a/intern/cycles/kernel/geom/geom_bvh.h b/intern/cycles/kernel/geom/geom_bvh.h
index 12ebc646c14..dd7c25d581d 100644
--- a/intern/cycles/kernel/geom/geom_bvh.h
+++ b/intern/cycles/kernel/geom/geom_bvh.h
@@ -93,6 +93,36 @@ CCL_NAMESPACE_BEGIN
 #include "geom_bvh_subsurface.h"
 #endif
 
+#if defined(__SHADOW_RECORD_ALL__)
+#define BVH_FUNCTION_NAME bvh_intersect_shadow_all
+#define BVH_FUNCTION_FEATURES 0
+#include "geom_bvh_shadow.h"
+#endif
+
+#if defined(__SUBSURFACE__) && defined(__INSTANCING__)
+#define BVH_FUNCTION_NAME bvh_intersect_shadow_all_instancing
+#define BVH_FUNCTION_FEATURES BVH_INSTANCING
+#include "geom_bvh_shadow.h"
+#endif
+
+#if defined(__SUBSURFACE__) && defined(__HAIR__)
+#define BVH_FUNCTION_NAME bvh_intersect_shadow_all_hair
+#define BVH_FUNCTION_FEATURES BVH_INSTANCING|BVH_HAIR
+#include "geom_bvh_shadow.h"
+#endif
+
+#if defined(__SUBSURFACE__) && defined(__OBJECT_MOTION__)
+#define BVH_FUNCTION_NAME bvh_intersect_shadow_all_motion
+#define BVH_FUNCTION_FEATURES BVH_INSTANCING|BVH_MOTION
+#include "geom_bvh_shadow.h"
+#endif
+
+#if defined(__SUBSURFACE__) && defined(__HAIR__) && defined(__OBJECT_MOTION__)
+#define BVH_FUNCTION_NAME bvh_intersect_shadow_all_hair_motion
+#define BVH_FUNCTION_FEATURES BVH_INSTANCING|BVH_HAIR|BVH_MOTION
+#include "geom_bvh_shadow.h"
+#endif
+
 /* to work around titan bug when using arrays instead of textures */
 #if !defined(__KERNEL_CUDA__) || defined(__KERNEL_CUDA_TEX_STORAGE__)
 ccl_device_inline
@@ -185,6 +215,52 @@ uint scene_intersect_subsurface(KernelGlobals *kg, const Ray *ray, Intersection
 }
 #endif
 
+/* to work around titan bug when using arrays instead of textures */
+#ifdef __SHADOW_RECORD_ALL__
+#if !defined(__KERNEL_CUDA__) || defined(__KERNEL_CUDA_TEX_STORAGE__)
+ccl_device_inline
+#else
+ccl_device_noinline
+#endif
+uint scene_intersect_shadow_all(KernelGlobals *kg, const Ray *ray, Intersection *isect, uint max_hits, uint *num_hits)
+{
+#ifdef __OBJECT_MOTION__
+	if(kernel_data.bvh.have_motion) {
+#ifdef __HAIR__
+		if(kernel_data.bvh.have_curves)
+			return bvh_intersect_shadow_all_hair_motion(kg, ray, isect, max_hits, num_hits);
+#endif /* __HAIR__ */
+
+		return bvh_intersect_shadow_all_motion(kg, ray, isect, max_hits, num_hits);
+	}
+#endif /* __OBJECT_MOTION__ */
+
+#ifdef __HAIR__ 
+	if(kernel_data.bvh.have_curves)
+		return bvh_intersect_shadow_all_hair(kg, ray, isect, max_hits, num_hits);
+#endif /* __HAIR__ */
+
+#ifdef __KERNEL_CPU__
+
+#ifdef __INSTANCING__
+	if(kernel_data.bvh.have_instancing)
+		return bvh_intersect_shadow_all_instancing(kg, ray, isect, max_hits, num_hits);
+#endif /* __INSTANCING__ */
+
+	return bvh_intersect_shadow_all(kg, ray, isect, max_hits, num_hits);
+#else /* __KERNEL_CPU__ */
+
+#ifdef __INSTANCING__
+	return bvh_intersect_shadow_all_instancing(kg, ray, isect, max_hits, num_hits);
+#else
+	return bvh_intersect_shadow_all(kg, ray, isect, max_hits, num_hits);
+#endif /* __INSTANCING__ */
+
+#endif /* __KERNEL_CPU__ */
+}
+#endif
+
+
 /* Ray offset to avoid self intersection.
  *
  * This function should be used to compute a modified ray start position for