diff options
| author | Brecht Van Lommel <brechtvanlommel@pandora.be> | 2013-06-18 13:36:06 +0400 |
|---|---|---|
| committer | Brecht Van Lommel <brechtvanlommel@pandora.be> | 2013-06-18 13:36:06 +0400 |
| commit | d57c6748c4ebb37246caf25d4900ef6d5c16c0fe (patch) | |
| tree | 08491bec3d7310f7df1e2171c8fb44a68d508a90 /intern/cycles/bvh | |
| parent | 9131adca9f748f794c18c71d36f830a961c218b4 (diff) | |
Cycles: optimization for BVH traveral on CPU's with SSE3, using code from Embree.
On the BMW scene, this gives roughly a 10% speedup overall with clang/gcc, and 30%
speedup with visual studio (2008). It turns out visual studio was optimizing the
existing code quite poorly compared to pretty good autovectorization by clang/gcc,
but hand written SSE code also gives a smaller speed boost there.
This code isn't enabled when using the hair minimum width feature yet, need to
make that work with the SSE code still.
Diffstat (limited to 'intern/cycles/bvh')
| -rw-r--r-- | intern/cycles/bvh/bvh.cpp | 6 |
1 files changed, 3 insertions, 3 deletions
diff --git a/intern/cycles/bvh/bvh.cpp b/intern/cycles/bvh/bvh.cpp index 82a444bda76..69ccf2588c9 100644 --- a/intern/cycles/bvh/bvh.cpp +++ b/intern/cycles/bvh/bvh.cpp @@ -552,9 +552,9 @@ void RegularBVH::pack_node(int idx, const BoundBox& b0, const BoundBox& b1, int { int4 data[BVH_NODE_SIZE] = { - make_int4(__float_as_int(b0.min.x), __float_as_int(b0.max.x), __float_as_int(b0.min.y), __float_as_int(b0.max.y)), - make_int4(__float_as_int(b1.min.x), __float_as_int(b1.max.x), __float_as_int(b1.min.y), __float_as_int(b1.max.y)), - make_int4(__float_as_int(b0.min.z), __float_as_int(b0.max.z), __float_as_int(b1.min.z), __float_as_int(b1.max.z)), + make_int4(__float_as_int(b0.min.x), __float_as_int(b1.min.x), __float_as_int(b0.max.x), __float_as_int(b1.max.x)), + make_int4(__float_as_int(b0.min.y), __float_as_int(b1.min.y), __float_as_int(b0.max.y), __float_as_int(b1.max.y)), + make_int4(__float_as_int(b0.min.z), __float_as_int(b1.min.z), __float_as_int(b0.max.z), __float_as_int(b1.max.z)), make_int4(c0, c1, visibility0, visibility1) }; |