diff options
| author | Sergey Sharybin <sergey.vfx@gmail.com> | 2017-03-27 18:06:37 +0300 |
|---|---|---|
| committer | Sergey Sharybin <sergey.vfx@gmail.com> | 2017-03-28 18:26:47 +0300 |
| commit | 6ea54fe9ffe2b2514990fdf3489ca53d05ce449a (patch) | |
| tree | 3026ce049c43bf8e0d56fbad57e3cbb844444fc4 /intern/cycles/kernel/bvh/qbvh_volume.h | |
| parent | 69aa6577b3dfea5d8a6d915fad7fb7650d8b6278 (diff) | |
Cycles: Switch to reformulated Pluecker ray/triangle intersection
The intention of this commit it to address issues mentioned in the
reports T43865,T50164 and T50452.
The code is based on Embree code with some extra vectorization
to speed up single ray to single triangle intersection.
Unfortunately, such a fix is not coming for free. There is some
slowdown for AVX2 processors, mainly due to different vectorization
code, which caused different number of instructions to be executed
and different instructions-per-cycle counters. But on another hand
this commit makes pre-AVX2 platforms such as AVX and SSE4.1 a bit
faster. The prerformance goes as following:
2.78c AVX2 2.78c AVX Patch AVX2 Patch AVX
BMW 05:21.09 06:05.34 05:32.97 (+3.5%) 05:34.97 (-8.5%)
Classroom 16:55.36 18:24.51 17:10.41 (+1.4%) 17:15.87 (-6.3%)
Fishy Cat 08:08.49 08:36.26 08:09.19 (+0.2%) 08:12.25 (-4.7%
Koro 11:22.54 11:45.24 11:13.25 (-1.5%) 11:43.81 (-0.3%)
Barcelone 14:18.32 16:09.46 14:15.20 (-0.4%) 14:25.15 (-10.8%)
On GPU the performance is about 1.5-2% slower in my tests on GTX1080
but afraid we can't do much as a part of this chaneg here and
consider it a price to pay for more proper intersection check.
Made in collaboration with Maxym Dmytrychenko, big thanks to him!
Reviewers: brecht, juicyfruit, lukasstockner97, dingto
Differential Revision: https://developer.blender.org/D1574
Diffstat (limited to 'intern/cycles/kernel/bvh/qbvh_volume.h')
| -rw-r--r-- | intern/cycles/kernel/bvh/qbvh_volume.h | 11 |
1 files changed, 2 insertions, 9 deletions
diff --git a/intern/cycles/kernel/bvh/qbvh_volume.h b/intern/cycles/kernel/bvh/qbvh_volume.h index 7cfb2cf8c18..192ce009524 100644 --- a/intern/cycles/kernel/bvh/qbvh_volume.h +++ b/intern/cycles/kernel/bvh/qbvh_volume.h @@ -91,9 +91,6 @@ ccl_device bool BVH_FUNCTION_FULL_NAME(QBVH)(KernelGlobals *kg, &near_x, &near_y, &near_z, &far_x, &far_y, &far_z); - TriangleIsectPrecalc isect_precalc; - ray_triangle_intersect_precalc(dir, &isect_precalc); - /* Traversal loop. */ do { do { @@ -266,7 +263,7 @@ ccl_device bool BVH_FUNCTION_FULL_NAME(QBVH)(KernelGlobals *kg, continue; } /* Intersect ray against primitive. */ - triangle_intersect(kg, &isect_precalc, isect, P, visibility, object, prim_addr); + triangle_intersect(kg, isect, P, dir, visibility, object, prim_addr); } break; } @@ -281,7 +278,7 @@ ccl_device bool BVH_FUNCTION_FULL_NAME(QBVH)(KernelGlobals *kg, continue; } /* Intersect ray against primitive. */ - motion_triangle_intersect(kg, &isect_precalc, isect, P, ray->time, visibility, object, prim_addr); + motion_triangle_intersect(kg, isect, P, dir, ray->time, visibility, object, prim_addr); } break; } @@ -316,8 +313,6 @@ ccl_device bool BVH_FUNCTION_FULL_NAME(QBVH)(KernelGlobals *kg, org4 = sse3f(ssef(P.x), ssef(P.y), ssef(P.z)); # endif - ray_triangle_intersect_precalc(dir, &isect_precalc); - ++stack_ptr; kernel_assert(stack_ptr < BVH_QSTACK_SIZE); traversal_stack[stack_ptr].addr = ENTRYPOINT_SENTINEL; @@ -362,8 +357,6 @@ ccl_device bool BVH_FUNCTION_FULL_NAME(QBVH)(KernelGlobals *kg, org4 = sse3f(ssef(P.x), ssef(P.y), ssef(P.z)); # endif - ray_triangle_intersect_precalc(dir, &isect_precalc); - object = OBJECT_NONE; node_addr = traversal_stack[stack_ptr].addr; --stack_ptr; |