diff --git a/src/BulletMultiThreaded/Makefile.original b/src/BulletMultiThreaded/Makefile.original
index cff0cb336..ea134416c 100644
--- a/src/BulletMultiThreaded/Makefile.original
+++ b/src/BulletMultiThreaded/Makefile.original
@@ -28,6 +28,9 @@ ifeq "$(__ARCH_BITS__)" "64"
 else
   SPU_DEFFLAGS= -DUSE_LIBSPE2 -D__SPU__
 endif
+
+SPU_DEFFLAGS+=-DUSE_PE_BOX_BOX
+
 SPU_GCC=$(USE_CCACHE) /usr/bin/spu-gcc
 SPU_INCLUDEDIR= -Ivectormath/scalar/cpp -I. -I$(CELL_SYSROOT)/usr/spu/include -I../../src -I$(NARROWPHASEDIR)
 #SPU_CFLAGS= $(DEBUGFLAG) -W -Wall -Winline -Os -c -include spu_intrinsics.h -include stdbool.h
diff --git a/src/BulletMultiThreaded/PpuAddressSpace.h b/src/BulletMultiThreaded/PpuAddressSpace.h
index 0f453fcbf..93c83648c 100644
--- a/src/BulletMultiThreaded/PpuAddressSpace.h
+++ b/src/BulletMultiThreaded/PpuAddressSpace.h
@@ -1,6 +1,13 @@
 #ifndef __PPU_ADDRESS_SPACE_H
 #define __PPU_ADDRESS_SPACE_H
 
+
+#ifdef WIN32
+//stop those casting warnings until we have a better solution for ppu_address_t / void* / uint64 conversions
+#pragma warning (disable: 4311)
+#pragma warning (disable: 4312)
+#endif //WIN32
+
 #ifdef USE_ADDR64
 typedef uint64_t ppu_address_t;
 #else
diff --git a/src/BulletMultiThreaded/SpuNarrowPhaseCollisionTask/SpuGatheringCollisionTask.cpp b/src/BulletMultiThreaded/SpuNarrowPhaseCollisionTask/SpuGatheringCollisionTask.cpp
index 6d9b503e1..41a99c622 100644
--- a/src/BulletMultiThreaded/SpuNarrowPhaseCollisionTask/SpuGatheringCollisionTask.cpp
+++ b/src/BulletMultiThreaded/SpuNarrowPhaseCollisionTask/SpuGatheringCollisionTask.cpp
@@ -981,7 +981,9 @@ void	processCollisionTask(void* userPtr, void* lsMemPtr)
 #endif											
 										)
 										{
+//#define USE_PE_BOX_BOX 1
 #ifdef USE_PE_BOX_BOX
+											{
 
 												//getCollisionMargin0
 												btScalar margin0 = lsMem.getlocalCollisionAlgorithm()->getCollisionMargin0();
diff --git a/src/BulletMultiThreaded/SpuSolverTask/SpuParallellSolverTask.cpp b/src/BulletMultiThreaded/SpuSolverTask/SpuParallellSolverTask.cpp
index 2927812e8..24eba629d 100644
--- a/src/BulletMultiThreaded/SpuSolverTask/SpuParallellSolverTask.cpp
+++ b/src/BulletMultiThreaded/SpuSolverTask/SpuParallellSolverTask.cpp
@@ -491,7 +491,7 @@ static void solveFriction ( btSolverBody& bodyA, btSolverBody& bodyB, btSolverCo
 			j1 = constraint.m_appliedImpulse - oldTangentImpulse;
 
 		}
-		j1*=0.9;
+		
 		if (bodyA.m_invMass > 0)
 		{
 			bodyA.m_linearVelocity += constraint.m_contactNormal*bodyA.m_invMass*j1;
@@ -1134,7 +1134,7 @@ void processSolverTask(void* userPtr, void* lsMemory)
 									constraint.m_penetration *= (taskDesc.m_commandData.m_manifoldSetup.m_solverInfo.m_erp/taskDesc.m_commandData.m_manifoldSetup.m_solverInfo.m_timeStep);
 								
 									constraint.m_restitution = rest;
-									constraint.m_appliedImpulse = cp.m_appliedImpulse*0.85;
+									constraint.m_appliedImpulse = cp.m_appliedImpulse*taskDesc.m_commandData.m_manifoldSetup.m_solverInfo.m_warmstartingFactor;
 									if (constraint.m_appliedImpulse!= 0.f)
 									{
 										if (solverBodyA)