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

Some performance improvements and fixes related to btVector3 being aligned on SPU.

Browse Source 
btQuantizedBvh has a version number, memory layout might be different now (due to aligned btVector3)
reorganized some data members of some classes, to reduce memory footprint
This commit is contained in:
erwin.coumans
2008-10-29 02:44:50 +00:00
parent ce5df7cf47
commit 55b29e2355
10 changed files with 241 additions and 110 deletions
Download Patch File Download Diff File Expand all files Collapse all files

74
src/BulletMultiThreaded/SpuNarrowPhaseCollisionTask/SpuCollisionShapes.cpp
View File

@@ -16,6 +16,17 @@ subject to the following restrictions:
#include "SpuCollisionShapes.h"
#ifdef __SPU__
#include <spu_intrinsics.h>
static inline vec_float4 vec_dot3( vec_float4 vec0, vec_float4 vec1 )
{
vec_float4 result;
result = spu_mul( vec0, vec1 );
result = spu_madd( spu_rlqwbyte( vec0, 4 ), spu_rlqwbyte( vec1, 4 ), result );
return spu_madd( spu_rlqwbyte( vec0, 8 ), spu_rlqwbyte( vec1, 8 ), result );
}
#endif //__SPU__
btVector3 localGetSupportingVertexWithoutMargin(int shapeType, void* shape, const btVector3& localDir,struct SpuConvexPolyhedronVertexData* convexVertexData)//, int *featureIndex)
{
switch (shapeType)
@@ -170,7 +181,45 @@ btVector3 localGetSupportingVertexWithoutMargin(int shapeType, void* shape, cons
{
//spu_printf("SPU: todo: getSupport CONVEX_HULL_SHAPE_PROXYTYPE\n");
#if defined (__SPU__)
vec_float4 v_distMax = {-FLT_MAX,0,0,0};
vec_int4 v_idxMax = {-999,0,0,0};
int v=0;
int numverts = convexVertexData->gNumConvexPoints;
btVector3* points = convexVertexData->gConvexPoints;
for(;v<(int)numverts-4;v+=4) {
vec_float4 p0 = vec_dot3(points[v ].get128(),localDir.get128());
vec_float4 p1 = vec_dot3(points[v+1].get128(),localDir.get128());
vec_float4 p2 = vec_dot3(points[v+2].get128(),localDir.get128());
vec_float4 p3 = vec_dot3(points[v+3].get128(),localDir.get128());
const vec_int4 i0 = {v ,0,0,0};
const vec_int4 i1 = {v+1,0,0,0};
const vec_int4 i2 = {v+2,0,0,0};
const vec_int4 i3 = {v+3,0,0,0};
vec_uint4 retGt01 = spu_cmpgt(p0,p1);
vec_float4 pmax01 = spu_sel(p1,p0,retGt01);
vec_int4 imax01 = spu_sel(i1,i0,retGt01);
vec_uint4 retGt23 = spu_cmpgt(p2,p3);
vec_float4 pmax23 = spu_sel(p3,p2,retGt23);
vec_int4 imax23 = spu_sel(i3,i2,retGt23);
vec_uint4 retGt0123 = spu_cmpgt(pmax01,pmax23);
vec_float4 pmax0123 = spu_sel(pmax23,pmax01,retGt0123);
vec_int4 imax0123 = spu_sel(imax23,imax01,retGt0123);
vec_uint4 retGtMax = spu_cmpgt(v_distMax,pmax0123);
v_distMax = spu_sel(pmax0123,v_distMax,retGtMax);
v_idxMax = spu_sel(imax0123,v_idxMax,retGtMax);
}
for(;v<(int)numverts;v++) {
vec_float4 p = vec_dot3(points[v].get128(),localDir.get128());
const vec_int4 i = {v,0,0,0};
vec_uint4 retGtMax = spu_cmpgt(v_distMax,p);
v_distMax = spu_sel(p,v_distMax,retGtMax);
v_idxMax = spu_sel(i,v_idxMax,retGtMax);
}
int ptIndex = spu_extract(v_idxMax,0);
const btVector3& supVec= points[ptIndex];
#else
btVector3* points = 0;
int numPoints = 0;
@@ -179,7 +228,7 @@ btVector3 localGetSupportingVertexWithoutMargin(int shapeType, void* shape, cons
// spu_printf("numPoints = %d\n",numPoints);
btVector3 supVec(btScalar(0.),btScalar(0.),btScalar(0.));
int ptIndex = 0;
btScalar newDot,maxDot = btScalar(-1e30);
btVector3 vec0(localDir.getX(),localDir.getY(),localDir.getZ());
@@ -197,15 +246,18 @@ btVector3 localGetSupportingVertexWithoutMargin(int shapeType, void* shape, cons
for (int i=0;i<numPoints;i++)
{
btVector3 vtx = points[i];// * m_localScaling;
const btVector3& vtx = points[i];// * m_localScaling;
newDot = vec.dot(vtx);
if (newDot > maxDot)
{
maxDot = newDot;
supVec = vtx;
ptIndex = i;
}
}
const btVector3& supVec= points[ptIndex];
#endif
return btVector3(supVec.getX(),supVec.getY(),supVec.getZ());
break;
@@ -223,7 +275,7 @@ btVector3 localGetSupportingVertexWithoutMargin(int shapeType, void* shape, cons
}
}
void computeAabb (btVector3& aabbMin, btVector3& aabbMax, btConvexInternalShape* convexShape, ppu_address_t convexShapePtr, int shapeType, btTransform xform)
void computeAabb (btVector3& aabbMin, btVector3& aabbMax, btConvexInternalShape* convexShape, ppu_address_t convexShapePtr, int shapeType, const btTransform& xform)
{
//calculate the aabb, given the types...
switch (shapeType)
@@ -235,7 +287,7 @@ void computeAabb (btVector3& aabbMin, btVector3& aabbMax, btConvexInternalShape*
float margin=convexShape->getMarginNV();
btVector3 halfExtents = convexShape->getImplicitShapeDimensions();
halfExtents += btVector3(margin,margin,margin);
btTransform& t = xform;
const btTransform& t = xform;
btMatrix3x3 abs_b = t.getBasis().absolute();
btVector3 center = t.getOrigin();
btVector3 extent = btVector3(abs_b[0].dot(halfExtents),abs_b[1].dot(halfExtents),abs_b[2].dot(halfExtents));
@@ -258,7 +310,7 @@ void computeAabb (btVector3& aabbMin, btVector3& aabbMax, btConvexInternalShape*
btScalar radius = convexShape->getRadius();
halfExtents[capsuleUpAxis] = radius + halfHeight;
#endif
btTransform& t = xform;
const btTransform& t = xform;
btMatrix3x3 abs_b = t.getBasis().absolute();
btVector3 center = t.getOrigin();
btVector3 extent = btVector3(abs_b[0].dot(halfExtents),abs_b[1].dot(halfExtents),abs_b[2].dot(halfExtents));
@@ -271,7 +323,7 @@ void computeAabb (btVector3& aabbMin, btVector3& aabbMax, btConvexInternalShape*
{
float radius = convexShape->getImplicitShapeDimensions().getX();// * convexShape->getLocalScaling().getX();
float margin = radius + convexShape->getMarginNV();
btTransform& t = xform;
const btTransform& t = xform;
const btVector3& center = t.getOrigin();
btVector3 extent(margin,margin,margin);
aabbMin = center - extent;
@@ -284,7 +336,7 @@ void computeAabb (btVector3& aabbMin, btVector3& aabbMax, btConvexInternalShape*
cellDmaGet(&convexHullShape0, convexShapePtr , sizeof(btConvexHullShape), DMA_TAG(1), 0, 0);
cellDmaWaitTagStatusAll(DMA_MASK(1));
btConvexHullShape* localPtr = (btConvexHullShape*)&convexHullShape0;
btTransform& t = xform;
const btTransform& t = xform;
btScalar margin = convexShape->getMarginNV();
localPtr->getNonvirtualAabb(t,aabbMin,aabbMax,margin);
//spu_printf("SPU convex aabbMin=%f,%f,%f=\n",aabbMin.getX(),aabbMin.getY(),aabbMin.getZ());
@@ -450,7 +502,9 @@ void spuWalkStacklessQuantizedTree(btNodeOverlapCallback* nodeCallback,unsigned
int curIndex = startNodeIndex;
int walkIterations = 0;
#ifdef BT_DEBUG
int subTreeSize = endNodeIndex - startNodeIndex;
#endif
int escapeIndex;
@@ -459,7 +513,7 @@ void spuWalkStacklessQuantizedTree(btNodeOverlapCallback* nodeCallback,unsigned
while (curIndex < endNodeIndex)
{
//catch bugs in tree data
assert (walkIterations < subTreeSize);
btAssert (walkIterations < subTreeSize);
walkIterations++;
aabbOverlap = spuTestQuantizedAabbAgainstQuantizedAabb(quantizedQueryAabbMin,quantizedQueryAabbMax,rootNode->m_quantizedAabbMin,rootNode->m_quantizedAabbMax);