Bart/bullet3
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

add re-usable createGraphicsSphere method in GpuDemo.

Browse Source 
introduce and use maxContactCapacity (needs to be fixed in various other contact kernels)
implement sphere versus trimesh
disable new/sequential GPU batching (only uses 1 thread in a warp, slow but works on NVIDIA/Apple OpenCL)
This commit is contained in:
erwin coumans
2013-04-04 17:54:45 -07:00
parent 733572e625
commit 358f4f97a2
15 changed files with 1008 additions and 139 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
opencl/gpu_sat/kernels/sat.cl
View File

@@ -1081,6 +1081,13 @@ __kernel void findConcaveSeparatingAxisKernel( __global int4* concavePairs,
int shapeIndexA = collidables[collidableIndexA].m_shapeIndex;
int shapeIndexB = collidables[collidableIndexB].m_shapeIndex;
if (collidables[collidableIndexB].m_shapeType!=SHAPE_CONVEX_HULL)
{
concavePairs[pairIdx].w = 0;
return;
}
int numFacesA = convexShapes[shapeIndexA].m_numFaces;
int numActualConcaveConvexTests = 0;