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

Added preparation for GPU hardware accelerated solvers for BulletSoftBody (OpenCL and DirectCompute backends will follow)

Browse Source 
Added assert to check for positive masses in btCompoundShape::calculatePrincipalAxisTransform, see Issue 399
Fixes for LLVM/GCC compilation issue in btSequentialImpulseConstraintSolver
(Untested) fix for Linux 64bit compilation Issue 409
This commit is contained in:
erwin.coumans
2010-07-16 23:26:25 +00:00
parent 13d9441f30
commit b3f081fc85
21 changed files with 834 additions and 175 deletions
Download Patch File Download Diff File Expand all files Collapse all files

54
Demos/Benchmarks/BenchmarkDemo.cpp
View File

@@ -41,12 +41,40 @@ subject to the following restrictions:
#ifdef USE_PARALLEL_DISPATCHER_BENCHMARK
#include "BulletMultiThreaded/Win32ThreadSupport.h"
#include "BulletMultiThreaded/SpuGatheringCollisionDispatcher.h"
#include "BulletMultiThreaded/btParallelConstraintSolver.h"
btThreadSupportInterface* createSolverThreadSupport(int maxNumThreads)
{
//#define SEQUENTIAL
#ifdef SEQUENTIAL
SequentialThreadSupport::SequentialThreadConstructionInfo tci("solverThreads",SolverThreadFunc,SolverlsMemoryFunc);
SequentialThreadSupport* threadSupport = new SequentialThreadSupport(tci);
threadSupport->startSPU();
#else
#ifdef _WIN32
Win32ThreadSupport::Win32ThreadConstructionInfo threadConstructionInfo("solverThreads",SolverThreadFunc,SolverlsMemoryFunc,maxNumThreads);
Win32ThreadSupport* threadSupport = new Win32ThreadSupport(threadConstructionInfo);
threadSupport->startSPU();
#elif defined (USE_PTHREADS)
PosixThreadSupport::ThreadConstructionInfo solverConstructionInfo("solver", SolverThreadFunc,
SolverlsMemoryFunc, maxNumThreads);
PosixThreadSupport* threadSupport = new PosixThreadSupport(solverConstructionInfo);
#else
SequentialThreadSupport::SequentialThreadConstructionInfo tci("solverThreads",SolverThreadFunc,SolverlsMemoryFunc);
SequentialThreadSupport* threadSupport = new SequentialThreadSupport(tci);
threadSupport->startSPU();
#endif
#endif
#undef USE_PARALLEL_SOLVER_BENCHMARK
#ifdef USE_PARALLEL_SOLVER_BENCHMARK
#include "BulletMultiThreaded/btParallelConstraintSolver.h"
return threadSupport;
}
#endif
class btRaycastBar2
@@ -319,8 +347,10 @@ void BenchmarkDemo::initPhysics()
///the default constraint solver. For parallel processing you can use a different solver (see Extras/BulletMultiThreaded)
#ifdef USE_PARALLEL_SOLVER_BENCHMARK
btConstraintSolver* sol = new btParallelConstraintSolver;
#ifdef USE_PARALLEL_DISPATCHER_BENCHMARK
btThreadSupportInterface* thread = createSolverThreadSupport(4);
btConstraintSolver* sol = new btParallelConstraintSolver(thread);
#else
btSequentialImpulseConstraintSolver* sol = new btSequentialImpulseConstraintSolver;
#endif //USE_PARALLEL_DISPATCHER_BENCHMARK
@@ -331,9 +361,13 @@ void BenchmarkDemo::initPhysics()
btDiscreteDynamicsWorld* dynamicsWorld;
m_dynamicsWorld = dynamicsWorld = new btDiscreteDynamicsWorld(m_dispatcher,m_overlappingPairCache,m_solver,m_collisionConfiguration);
#ifdef USE_PARALLEL_DISPATCHER_BENCHMARK
dynamicsWorld->getSimulationIslandManager()->setSplitIslands(false);
#endif //USE_PARALLEL_DISPATCHER_BENCHMARK
///the following 3 lines increase the performance dramatically, with a little bit of loss of quality
m_dynamicsWorld->getSolverInfo().m_solverMode |=SOLVER_ENABLE_FRICTION_DIRECTION_CACHING; //don't recalculate friction values each frame
dynamicsWorld->getSolverInfo().m_numIterations = 4; //few solver iterations
dynamicsWorld->getSolverInfo().m_numIterations = 5; //few solver iterations
m_defaultContactProcessingThreshold = 0.f;//used when creating bodies: body->setContactProcessingThreshold(...);
@@ -494,8 +528,8 @@ void BenchmarkDemo::createWall(const btVector3& offsetPosition,int stackSize,con
void BenchmarkDemo::createPyramid(const btVector3& offsetPosition,int stackSize,const btVector3& boxSize)
{
btScalar space = 0.0001f;
btScalar space = 0.000f;
btVector3 pos(0.0f, boxSize[1], 0.0f);
btBoxShape* blockShape = new btBoxShape(btVector3(boxSize[0]-COLLISION_RADIUS,boxSize[1]-COLLISION_RADIUS,boxSize[2]-COLLISION_RADIUS));
@@ -507,9 +541,9 @@ void BenchmarkDemo::createPyramid(const btVector3& offsetPosition,int stackSize,
blockShape->calculateLocalInertia(mass,localInertia);
btScalar diffX = boxSize[0];
btScalar diffY = boxSize[1];
btScalar diffZ = boxSize[2];
btScalar diffX = boxSize[0]*1.02f;
btScalar diffY = boxSize[1]*1.02f;
btScalar diffZ = boxSize[2]*1.02f;
btScalar offsetX = -stackSize * (diffX * 2.0f + space) * 0.5f;
btScalar offsetZ = -stackSize * (diffZ * 2.0f + space) * 0.5f;