diff --git a/Extras/BulletMultiThreaded/SpuSampleTask/SpuSampleTask.cpp b/Extras/BulletMultiThreaded/SpuSampleTask/SpuSampleTask.cpp
index 2ea1280ce..32909f9dc 100644
--- a/Extras/BulletMultiThreaded/SpuSampleTask/SpuSampleTask.cpp
+++ b/Extras/BulletMultiThreaded/SpuSampleTask/SpuSampleTask.cpp
@@ -19,16 +19,24 @@ subject to the following restrictions:
 #include "../PlatformDefinitions.h"
 #include "../SpuFakeDma.h"
 #include "LinearMath/btMinMax.h"
-#include <stdio.h>
 
+#ifdef __SPU__
+#include <spu_printf.h>
+#else
+#include <stdio.h>
+#define spu_printf printf
+#endif
 
 struct SampleTask_LocalStoreMemory
 {
+	ATTRIBUTE_ALIGNED16(char gLocalRigidBody [sizeof(btRigidBody)+16]);
+	ATTRIBUTE_ALIGNED16(void* gPointerArray[1024]); //at max upload 1024 pointers
 
 };
 
 
 
+
 //-- MAIN METHOD
 void processSampleTask(void* userPtr, void* lsMemory)
 {
@@ -43,11 +51,70 @@ void processSampleTask(void* userPtr, void* lsMemory)
 	{
 	case CMD_SAMPLE_INTEGRATE_BODIES:
 		{
+			btTransform predictedTrans;
+			btCollisionObject** eaPtr = (btCollisionObject**)taskDesc.m_mainMemoryPtr;
+
+			int batchSize = taskDesc.m_sampleValue;
+			int dmaArraySize = batchSize*sizeof(void*);
+
+			uint64_t ppuArrayAddress = reinterpret_cast<uint64_t>(eaPtr);
+			
+//			spu_printf("array location is at %llx, batchSize = %d, DMA size = %d\n",ppuArrayAddress,batchSize,dmaArraySize);
+
+			if (dmaArraySize>=16)
+			{
+				cellDmaLargeGet((void*)&localMemory->gPointerArray[0], ppuArrayAddress  , dmaArraySize, DMA_TAG(1), 0, 0);	
+				cellDmaWaitTagStatusAll(DMA_MASK(1));
+			} else
+			{
+				stallingUnalignedDmaSmallGet((void*)&localMemory->gPointerArray[0], ppuArrayAddress  , dmaArraySize);
+			}
+			
+			
+			for ( int i=0;i<batchSize;i++)
+			{
+				///DMA rigid body
+				
+				void* localPtr = &localMemory->gLocalRigidBody[0];
+				void* shortAdd = localMemory->gPointerArray[i];
+				uint64_t ppuRigidBodyAddress = reinterpret_cast<uint64_t>(shortAdd);
+				
+			//	spu_printf("cellDmaGet at CMD_SAMPLE_INTEGRATE_BODIES from %llx to %llx\n",ppuRigidBodyAddress,localPtr);
+
+				int dmaBodySize = sizeof(btRigidBody);
+			
+				cellDmaGet((void*)localPtr, ppuRigidBodyAddress  , dmaBodySize, DMA_TAG(1), 0, 0);	
+				cellDmaWaitTagStatusAll(DMA_MASK(1));
+
+					
+				float timeStep = 1.f/60.f;
+
+				btRigidBody* body = (btRigidBody*) localPtr;//btRigidBody::upcast(colObj);
+				if (body)
+				{
+					if (body->isActive() && (!body->isStaticOrKinematicObject()))
+					{
+						body->predictIntegratedTransform(timeStep, predictedTrans);
+						body->proceedToTransform( predictedTrans);
+						void* ptr = (void*)localPtr;
+					//	spu_printf("cellDmaLargePut from %llx to LS %llx\n",ptr,ppuRigidBodyAddress);
+
+						cellDmaLargePut(ptr, ppuRigidBodyAddress  , dmaBodySize, DMA_TAG(1), 0, 0);
+						cellDmaWaitTagStatusAll(DMA_MASK(1));
+
+					}
+				}
+
+			}
+
+			/*
 #ifdef __SPU__
 			spu_printf("hello SPU world\n");
 #else
 			printf("hello world\n");
 #endif
+			*/
+
 			break;
 		}
 	default:
diff --git a/Extras/BulletMultiThreaded/SpuSampleTask/SpuSampleTask.h b/Extras/BulletMultiThreaded/SpuSampleTask/SpuSampleTask.h
index 9c02401e5..e3ca4d2bc 100644
--- a/Extras/BulletMultiThreaded/SpuSampleTask/SpuSampleTask.h
+++ b/Extras/BulletMultiThreaded/SpuSampleTask/SpuSampleTask.h
@@ -38,9 +38,9 @@ ATTRIBUTE_ALIGNED16(struct) SpuSampleTaskDesc
 	uint32_t						m_sampleCommand;
 	uint32_t						m_taskId;
 
-	uint64_t 	m_mainMemoryPtr;
-	int			m_sampleValue;
-	
+	uint64_t 	m_mainMemoryPtr;
+	int			m_sampleValue;
+	
 
 };
 
diff --git a/src/LinearMath/btQuadWord.h b/src/LinearMath/btQuadWord.h
index c14481928..647b9eeeb 100644
--- a/src/LinearMath/btQuadWord.h
+++ b/src/LinearMath/btQuadWord.h
@@ -19,13 +19,7 @@ subject to the following restrictions:
 #include "btScalar.h"
 #include "btMinMax.h"
 #include <math.h>
-#ifdef __PPU__
-#include <altivec.h>
-#endif
 
-#ifdef __SPU__
-#include <spu_intrinsics.h>
-#endif
 
 //ATTRIBUTE_ALIGNED16(class) btQuadWordStorage
 //some issues under PS3 Linux with IBM 2.1 SDK, gcc compiler prevent from using aligned quadword. todo: look into this
@@ -33,34 +27,13 @@ subject to the following restrictions:
 class btQuadWordStorage
 {
 protected:
-#if defined __PPU__ || defined __SPU__
-	union {
-		vec_float4 m_Vec128;
-		struct	
-		{
-#endif
+
 	btScalar	m_x;
 	btScalar	m_y;
 	btScalar	m_z;
 	btScalar	m_unusedW;
-#if defined __PPU__ || defined __SPU__
-		};
-	};
-#endif
 
 public:
-#if defined __PPU__ || defined __SPU__
-
-	inline const vec_float4 get128( ) const
-	{
-		return m_Vec128;
-	}
-
-	inline vec_float4 get128( )
-	{
-		return m_Vec128;
-	}
-#endif //
 
 };
 
diff --git a/src/LinearMath/btVector3.h b/src/LinearMath/btVector3.h
index d68e0279f..96548c6ba 100644
--- a/src/LinearMath/btVector3.h
+++ b/src/LinearMath/btVector3.h
@@ -48,13 +48,7 @@ public:
 	SIMD_FORCE_INLINE btVector3& operator+=(const btVector3& v)
 	{
 
-#ifdef __PPU__
-		m_Vec128 = vec_add( m_Vec128, v.get128() );
-#elif defined __SPU__
-		m_Vec128 = spu_add( m_Vec128, v.get128() ); 
-#else
 		m_x += v.x(); m_y += v.y(); m_z += v.z();
-#endif
 		return *this;
 	}