ruby-bullet 0.0.2-x86-mingw32

data/deps/include/bullet/BulletMultiThreaded/CMakeFiles/BulletMultiThreaded.dir/flags.make

@@ -0,0 +1,8 @@
+# CMAKE generated file: DO NOT EDIT!
+# Generated by "MSYS Makefiles" Generator, CMake Version 2.8
+
+# compile CXX with C:/MinGW/bin/g++.exe
+CXX_FLAGS = -O3 -DNDEBUG @CMakeFiles/BulletMultiThreaded.dir/includes_CXX.rsp  
+
+CXX_DEFINES = -DUSE_GRAPHICAL_BENCHMARK -D_IRR_STATIC_LIB_ -D_CRT_SECURE_NO_WARNINGS -D_CRT_SECURE_NO_DEPRECATE -D_SCL_SECURE_NO_WARNINGS
+

data/deps/include/bullet/BulletMultiThreaded/CMakeFiles/BulletMultiThreaded.dir/includes_CXX.rsp

@@ -0,0 +1 @@
+-IC:/Users/abex/Documents/GitHub/ruby-bullet/deps/src/bullet-2.80-rev2531/src 

data/deps/include/bullet/BulletMultiThreaded/CMakeFiles/BulletMultiThreaded.dir/link.txt

@@ -0,0 +1,2 @@
+C:/MinGW/bin/ar.exe cr ../../lib/libBulletMultiThreaded.a  CMakeFiles/BulletMultiThreaded.dir/SpuFakeDma.obj CMakeFiles/BulletMultiThreaded.dir/SpuLibspe2Support.obj CMakeFiles/BulletMultiThreaded.dir/btThreadSupportInterface.obj CMakeFiles/BulletMultiThreaded.dir/Win32ThreadSupport.obj CMakeFiles/BulletMultiThreaded.dir/PosixThreadSupport.obj CMakeFiles/BulletMultiThreaded.dir/SequentialThreadSupport.obj CMakeFiles/BulletMultiThreaded.dir/SpuSampleTaskProcess.obj CMakeFiles/BulletMultiThreaded.dir/SpuCollisionObjectWrapper.obj CMakeFiles/BulletMultiThreaded.dir/SpuCollisionTaskProcess.obj CMakeFiles/BulletMultiThreaded.dir/SpuGatheringCollisionDispatcher.obj CMakeFiles/BulletMultiThreaded.dir/SpuContactManifoldCollisionAlgorithm.obj CMakeFiles/BulletMultiThreaded.dir/btParallelConstraintSolver.obj CMakeFiles/BulletMultiThreaded.dir/SpuNarrowPhaseCollisionTask/boxBoxDistance.obj CMakeFiles/BulletMultiThreaded.dir/SpuNarrowPhaseCollisionTask/SpuContactResult.obj CMakeFiles/BulletMultiThreaded.dir/SpuNarrowPhaseCollisionTask/SpuMinkowskiPenetrationDepthSolver.obj CMakeFiles/BulletMultiThreaded.dir/SpuNarrowPhaseCollisionTask/SpuGatheringCollisionTask.obj CMakeFiles/BulletMultiThreaded.dir/SpuNarrowPhaseCollisionTask/SpuCollisionShapes.obj CMakeFiles/BulletMultiThreaded.dir/btGpu3DGridBroadphase.obj
+C:/MinGW/bin/ranlib.exe ../../lib/libBulletMultiThreaded.a

data/deps/include/bullet/BulletMultiThreaded/CMakeFiles/BulletMultiThreaded.dir/progress.make

@@ -0,0 +1,19 @@
+CMAKE_PROGRESS_1 = 
+CMAKE_PROGRESS_2 = 68
+CMAKE_PROGRESS_3 = 69
+CMAKE_PROGRESS_4 = 
+CMAKE_PROGRESS_5 = 70
+CMAKE_PROGRESS_6 = 
+CMAKE_PROGRESS_7 = 71
+CMAKE_PROGRESS_8 = 
+CMAKE_PROGRESS_9 = 72
+CMAKE_PROGRESS_10 = 73
+CMAKE_PROGRESS_11 = 
+CMAKE_PROGRESS_12 = 74
+CMAKE_PROGRESS_13 = 
+CMAKE_PROGRESS_14 = 75
+CMAKE_PROGRESS_15 = 76
+CMAKE_PROGRESS_16 = 
+CMAKE_PROGRESS_17 = 77
+CMAKE_PROGRESS_18 = 
+

data/deps/include/bullet/BulletMultiThreaded/CMakeFiles/CMakeDirectoryInformation.cmake

@@ -0,0 +1,16 @@
+# CMAKE generated file: DO NOT EDIT!
+# Generated by "MSYS Makefiles" Generator, CMake Version 2.8
+
+# Relative path conversion top directories.
+SET(CMAKE_RELATIVE_PATH_TOP_SOURCE "C:/Users/abex/Documents/GitHub/ruby-bullet/deps/src/bullet-2.80-rev2531")
+SET(CMAKE_RELATIVE_PATH_TOP_BINARY "C:/Users/abex/Documents/GitHub/ruby-bullet/deps/src/bullet-2.80-rev2531")
+
+# Force unix paths in dependencies.
+SET(CMAKE_FORCE_UNIX_PATHS 1)
+
+
+# The C and CXX include file regular expressions for this directory.
+SET(CMAKE_C_INCLUDE_REGEX_SCAN "^.*$")
+SET(CMAKE_C_INCLUDE_REGEX_COMPLAIN "^$")
+SET(CMAKE_CXX_INCLUDE_REGEX_SCAN ${CMAKE_C_INCLUDE_REGEX_SCAN})
+SET(CMAKE_CXX_INCLUDE_REGEX_COMPLAIN ${CMAKE_C_INCLUDE_REGEX_COMPLAIN})

data/deps/include/bullet/BulletMultiThreaded/CMakeFiles/progress.marks

@@ -0,0 +1 @@
+11

data/deps/include/bullet/BulletMultiThreaded/CMakeLists.txt

@@ -0,0 +1,123 @@
+INCLUDE_DIRECTORIES(
+	${BULLET_PHYSICS_SOURCE_DIR}/src
+	${VECTOR_MATH_INCLUDE}
+)
+
+SET(BulletMultiThreaded_SRCS
+	SpuFakeDma.cpp
+	SpuLibspe2Support.cpp
+	btThreadSupportInterface.cpp
+	Win32ThreadSupport.cpp
+	PosixThreadSupport.cpp
+	SequentialThreadSupport.cpp
+	SpuSampleTaskProcess.cpp
+	SpuCollisionObjectWrapper.cpp 
+	SpuCollisionTaskProcess.cpp
+	SpuGatheringCollisionDispatcher.cpp
+	SpuContactManifoldCollisionAlgorithm.cpp
+	btParallelConstraintSolver.cpp
+	
+	#SPURS_PEGatherScatterTask/SpuPEGatherScatterTask.cpp
+	#SpuPEGatherScatterTaskProcess.cpp
+
+	SpuNarrowPhaseCollisionTask/boxBoxDistance.cpp
+	SpuNarrowPhaseCollisionTask/SpuContactResult.cpp
+	SpuNarrowPhaseCollisionTask/SpuMinkowskiPenetrationDepthSolver.cpp
+	SpuNarrowPhaseCollisionTask/SpuGatheringCollisionTask.cpp
+	SpuNarrowPhaseCollisionTask/SpuCollisionShapes.cpp
+
+	#Some GPU related stuff, mainly CUDA and perhaps OpenCL
+	btGpu3DGridBroadphase.cpp
+)
+
+SET(Root_HDRS
+	PlatformDefinitions.h
+	PpuAddressSpace.h
+	SpuFakeDma.h
+	SpuDoubleBuffer.h
+	SpuLibspe2Support.h
+	btThreadSupportInterface.h
+	Win32ThreadSupport.h
+	PosixThreadSupport.h
+	SequentialThreadSupport.h
+	SpuSampleTaskProcess.h
+	SpuCollisionObjectWrapper.cpp 
+	SpuCollisionObjectWrapper.h 
+	SpuCollisionTaskProcess.h
+	SpuGatheringCollisionDispatcher.h
+	SpuContactManifoldCollisionAlgorithm.h
+	btParallelConstraintSolver.h
+
+	#SPURS_PEGatherScatterTask/SpuPEGatherScatterTask.h
+	#SpuPEGatherScatterTaskProcess.h
+
+	#Some GPU related stuff, mainly CUDA and perhaps OpenCL
+	btGpu3DGridBroadphase.h
+	btGpu3DGridBroadphaseSharedCode.h
+	btGpu3DGridBroadphaseSharedDefs.h
+	btGpu3DGridBroadphaseSharedTypes.h
+	btGpuDefines.h
+	btGpuUtilsSharedCode.h
+	btGpuUtilsSharedDefs.h
+)
+
+SET(SpuNarrowPhaseCollisionTask_HDRS
+	SpuNarrowPhaseCollisionTask/Box.h
+	SpuNarrowPhaseCollisionTask/boxBoxDistance.h
+	SpuNarrowPhaseCollisionTask/SpuContactResult.h
+	SpuNarrowPhaseCollisionTask/SpuMinkowskiPenetrationDepthSolver.h
+	SpuNarrowPhaseCollisionTask/SpuConvexPenetrationDepthSolver.h
+	SpuNarrowPhaseCollisionTask/SpuPreferredPenetrationDirections.h
+	SpuNarrowPhaseCollisionTask/SpuGatheringCollisionTask.h
+	SpuNarrowPhaseCollisionTask/SpuCollisionShapes.h
+)
+
+SET(BulletMultiThreaded_HDRS
+	${Root_HDRS}
+	${SpuNarrowPhaseCollisionTask_HDRS}
+)
+
+ADD_LIBRARY(BulletMultiThreaded ${BulletMultiThreaded_SRCS} ${BulletMultiThreaded_HDRS})
+SET_TARGET_PROPERTIES(BulletMultiThreaded PROPERTIES VERSION ${BULLET_VERSION})
+SET_TARGET_PROPERTIES(BulletMultiThreaded PROPERTIES SOVERSION ${BULLET_VERSION})
+
+
+SUBDIRS(GpuSoftBodySolvers)
+
+
+IF (BUILD_SHARED_LIBS)
+	IF (UNIX)
+		TARGET_LINK_LIBRARIES(BulletMultiThreaded BulletDynamics BulletCollision pthread)
+	ELSE()
+		TARGET_LINK_LIBRARIES(BulletMultiThreaded BulletDynamics BulletCollision)
+	ENDIF()
+ENDIF (BUILD_SHARED_LIBS)
+
+
+IF (INSTALL_LIBS)
+	IF (NOT INTERNAL_CREATE_DISTRIBUTABLE_MSVC_PROJECTFILES)
+		#INSTALL of other files requires CMake 2.6
+		IF (${CMAKE_MAJOR_VERSION}.${CMAKE_MINOR_VERSION} GREATER 2.5)
+#			IF(INSTALL_EXTRA_LIBS)
+				IF (APPLE AND BUILD_SHARED_LIBS AND FRAMEWORK)
+					INSTALL(TARGETS BulletMultiThreaded DESTINATION .)
+				ELSE (APPLE AND BUILD_SHARED_LIBS AND FRAMEWORK)
+					INSTALL(TARGETS BulletMultiThreaded DESTINATION lib${LIB_SUFFIX})
+					INSTALL(DIRECTORY
+${CMAKE_CURRENT_SOURCE_DIR} DESTINATION ${INCLUDE_INSTALL_DIR} FILES_MATCHING
+PATTERN "*.h"  PATTERN ".svn" EXCLUDE PATTERN "CMakeFiles" EXCLUDE)
+				ENDIF (APPLE AND BUILD_SHARED_LIBS AND FRAMEWORK)
+#			ENDIF (INSTALL_EXTRA_LIBS)
+		ENDIF (${CMAKE_MAJOR_VERSION}.${CMAKE_MINOR_VERSION} GREATER 2.5)
+
+		IF (APPLE AND BUILD_SHARED_LIBS AND FRAMEWORK)
+			SET_TARGET_PROPERTIES(BulletMultiThreaded PROPERTIES FRAMEWORK true)
+		
+			SET_TARGET_PROPERTIES(BulletMultiThreaded PROPERTIES PUBLIC_HEADER "${Root_HDRS}")
+			# Have to list out sub-directories manually:
+			SET_PROPERTY(SOURCE ${SpuNarrowPhaseCollisionTask_HDRS} PROPERTY MACOSX_PACKAGE_LOCATION Headers/SpuNarrowPhaseCollisionTask)
+		
+		ENDIF (APPLE AND BUILD_SHARED_LIBS AND FRAMEWORK)
+	ENDIF (NOT INTERNAL_CREATE_DISTRIBUTABLE_MSVC_PROJECTFILES)
+ENDIF (INSTALL_LIBS)
+

data/deps/include/bullet/BulletMultiThreaded/GpuSoftBodySolvers/CMakeFiles/CMakeDirectoryInformation.cmake

@@ -0,0 +1,16 @@
+# CMAKE generated file: DO NOT EDIT!
+# Generated by "MSYS Makefiles" Generator, CMake Version 2.8
+
+# Relative path conversion top directories.
+SET(CMAKE_RELATIVE_PATH_TOP_SOURCE "C:/Users/abex/Documents/GitHub/ruby-bullet/deps/src/bullet-2.80-rev2531")
+SET(CMAKE_RELATIVE_PATH_TOP_BINARY "C:/Users/abex/Documents/GitHub/ruby-bullet/deps/src/bullet-2.80-rev2531")
+
+# Force unix paths in dependencies.
+SET(CMAKE_FORCE_UNIX_PATHS 1)
+
+
+# The C and CXX include file regular expressions for this directory.
+SET(CMAKE_C_INCLUDE_REGEX_SCAN "^.*$")
+SET(CMAKE_C_INCLUDE_REGEX_COMPLAIN "^$")
+SET(CMAKE_CXX_INCLUDE_REGEX_SCAN ${CMAKE_C_INCLUDE_REGEX_SCAN})
+SET(CMAKE_CXX_INCLUDE_REGEX_COMPLAIN ${CMAKE_C_INCLUDE_REGEX_COMPLAIN})

data/deps/include/bullet/BulletMultiThreaded/GpuSoftBodySolvers/CMakeFiles/progress.marks

@@ -0,0 +1 @@
+1

data/deps/include/bullet/BulletMultiThreaded/GpuSoftBodySolvers/CMakeLists.txt

@@ -0,0 +1,13 @@
+
+INCLUDE_DIRECTORIES(
+${BULLET_PHYSICS_SOURCE_DIR}/src
+)
+
+
+SUBDIRS ( 
+	OpenCL
+)
+
+IF( USE_DX11 )
+	SUBDIRS( DX11 )
+ENDIF( USE_DX11 )

data/deps/include/bullet/BulletMultiThreaded/GpuSoftBodySolvers/DX11/CMakeLists.txt

@@ -0,0 +1,83 @@
+
+INCLUDE_DIRECTORIES(
+${BULLET_PHYSICS_SOURCE_DIR}/src
+)
+
+SET(DXSDK_DIR $ENV{DXSDK_DIR})
+SET(DX11_INCLUDE_PATH  "${DIRECTX_SDK_BASE_DIR}/Include" CACHE DOCSTRING "Microsoft directX SDK include path")
+
+
+INCLUDE_DIRECTORIES(
+${DX11_INCLUDE_PATH} "../Shared/"
+${VECTOR_MATH_INCLUDE}
+)
+
+SET(BulletSoftBodyDX11Solvers_SRCS
+	btSoftBodySolver_DX11.cpp
+	btSoftBodySolver_DX11SIMDAware.cpp
+)
+
+SET(BulletSoftBodyDX11Solvers_HDRS
+	btSoftBodySolver_DX11.h
+	btSoftBodySolver_DX11SIMDAware.h
+	../Shared/btSoftBodySolverData.h
+	btSoftBodySolverVertexData_DX11.h
+	btSoftBodySolverTriangleData_DX11.h
+	btSoftBodySolverLinkData_DX11.h
+	btSoftBodySolverLinkData_DX11SIMDAware.h
+	btSoftBodySolverBuffer_DX11.h
+	btSoftBodySolverVertexBuffer_DX11.h
+
+)
+
+# OpenCL and HLSL Shaders.
+# Build rules generated to stringify these into headers
+# which are needed by some of the sources
+SET(BulletSoftBodyDX11Solvers_Shaders
+	OutputToVertexArray
+	UpdateNormals
+	Integrate
+	UpdatePositions
+	UpdateNodes
+	ComputeBounds
+	SolvePositions
+	SolvePositionsSIMDBatched
+	SolveCollisionsAndUpdateVelocities
+	SolveCollisionsAndUpdateVelocitiesSIMDBatched
+	UpdatePositionsFromVelocities
+	ApplyForces
+	PrepareLinks
+	VSolveLinks
+)
+
+foreach(f ${BulletSoftBodyDX11Solvers_Shaders})
+    LIST(APPEND BulletSoftBodyDX11Solvers_HLSL "HLSL/${f}.hlsl")
+endforeach(f) 
+
+
+
+ADD_LIBRARY(BulletSoftBodySolvers_DX11  ${BulletSoftBodyDX11Solvers_SRCS} ${BulletSoftBodyDX11Solvers_HDRS} ${BulletSoftBodyDX11Solvers_HLSL})
+SET_TARGET_PROPERTIES(BulletSoftBodySolvers_DX11 PROPERTIES VERSION ${BULLET_VERSION})
+SET_TARGET_PROPERTIES(BulletSoftBodySolvers_DX11 PROPERTIES SOVERSION ${BULLET_VERSION})
+IF (BUILD_SHARED_LIBS)
+	TARGET_LINK_LIBRARIES(BulletSoftBodySolvers_DX11 BulletSoftBody BulletDynamics)
+ENDIF (BUILD_SHARED_LIBS)
+
+
+IF (INSTALL_LIBS)
+	IF (NOT INTERNAL_CREATE_DISTRIBUTABLE_MSVC_PROJECTFILES)
+		IF (${CMAKE_MAJOR_VERSION}.${CMAKE_MINOR_VERSION} GREATER 2.5)
+			IF (APPLE AND BUILD_SHARED_LIBS AND FRAMEWORK)
+				INSTALL(TARGETS BulletSoftBodySolvers_DX11 DESTINATION .)
+			ELSE (APPLE AND BUILD_SHARED_LIBS AND FRAMEWORK)
+			INSTALL(TARGETS BulletSoftBodySolvers_DX11 DESTINATION lib${LIB_SUFFIX})
+#headers are already installed by BulletMultiThreaded library
+			ENDIF (APPLE AND BUILD_SHARED_LIBS AND FRAMEWORK)
+		ENDIF (${CMAKE_MAJOR_VERSION}.${CMAKE_MINOR_VERSION} GREATER 2.5)
+
+		IF (APPLE AND BUILD_SHARED_LIBS AND FRAMEWORK)
+			SET_TARGET_PROPERTIES(BulletSoftBodySolvers_DX11 PROPERTIES FRAMEWORK true)
+			SET_TARGET_PROPERTIES(BulletSoftBodySolvers_DX11 PROPERTIES PUBLIC_HEADER "${BulletSoftBodyDX11Solvers_HDRS}")
+		ENDIF (APPLE AND BUILD_SHARED_LIBS AND FRAMEWORK)
+	ENDIF (NOT INTERNAL_CREATE_DISTRIBUTABLE_MSVC_PROJECTFILES)
+ENDIF (INSTALL_LIBS)

data/deps/include/bullet/BulletMultiThreaded/GpuSoftBodySolvers/DX11/HLSL/ApplyForces.hlsl

@@ -0,0 +1,95 @@
+MSTRINGIFY(
+
+cbuffer ApplyForcesCB : register( b0 )
+{
+	unsigned int numNodes;
+	float solverdt;
+	float epsilon;
+	int padding3;
+};
+
+
+StructuredBuffer<int> g_vertexClothIdentifier : register( t0 );
+StructuredBuffer<float4> g_vertexNormal : register( t1 );
+StructuredBuffer<float> g_vertexArea : register( t2 );
+StructuredBuffer<float> g_vertexInverseMass : register( t3 );
+// TODO: These could be combined into a lift/drag factor array along with medium density
+StructuredBuffer<float> g_clothLiftFactor : register( t4 );
+StructuredBuffer<float> g_clothDragFactor : register( t5 );
+StructuredBuffer<float4> g_clothWindVelocity : register( t6 );
+StructuredBuffer<float4> g_clothAcceleration : register( t7 );
+StructuredBuffer<float> g_clothMediumDensity : register( t8 );
+
+RWStructuredBuffer<float4> g_vertexForceAccumulator : register( u0 );
+RWStructuredBuffer<float4> g_vertexVelocity : register( u1 );
+
+float3 projectOnAxis( float3 v, float3 a )
+{
+	return (a*dot(v, a));
+}
+
+[numthreads(128, 1, 1)]
+void 
+ApplyForcesKernel( uint3 Gid : SV_GroupID, uint3 DTid : SV_DispatchThreadID, uint3 GTid : SV_GroupThreadID, uint GI : SV_GroupIndex )
+{
+	unsigned int nodeID = DTid.x;
+	if( nodeID < numNodes )
+	{		
+		int clothId = g_vertexClothIdentifier[nodeID];
+		float nodeIM = g_vertexInverseMass[nodeID];
+		
+		if( nodeIM > 0.0f )
+		{
+			float3 nodeV = g_vertexVelocity[nodeID].xyz;
+			float3 normal = g_vertexNormal[nodeID].xyz;
+			float area = g_vertexArea[nodeID];
+			float3 nodeF = g_vertexForceAccumulator[nodeID].xyz;
+			
+			// Read per-cloth values
+			float3 clothAcceleration = g_clothAcceleration[clothId].xyz;
+			float3 clothWindVelocity = g_clothWindVelocity[clothId].xyz;
+			float liftFactor = g_clothLiftFactor[clothId];
+			float dragFactor = g_clothDragFactor[clothId];
+			float mediumDensity = g_clothMediumDensity[clothId];
+		
+			// Apply the acceleration to the cloth rather than do this via a force
+			nodeV += (clothAcceleration*solverdt);
+
+			g_vertexVelocity[nodeID] = float4(nodeV, 0.f);
+
+			float3 relativeWindVelocity = nodeV - clothWindVelocity;
+			float relativeSpeedSquared = dot(relativeWindVelocity, relativeWindVelocity);
+			
+			if( relativeSpeedSquared > epsilon )
+			{
+				// Correct direction of normal relative to wind direction and get dot product
+				normal = normal * (dot(normal, relativeWindVelocity) < 0 ? -1.f : 1.f);
+				float dvNormal = dot(normal, relativeWindVelocity);
+				if( dvNormal > 0 )
+				{
+					float3 force = float3(0.f, 0.f, 0.f);
+					float c0 = area * dvNormal * relativeSpeedSquared / 2.f;
+					float c1 = c0 * mediumDensity;
+					force += normal * (-c1 * liftFactor);
+					force += normalize(relativeWindVelocity)*(-c1 * dragFactor);
+					
+					float dtim = solverdt * nodeIM;
+					float3 forceDTIM = force * dtim;
+					
+					float3 nodeFPlusForce = nodeF + force;
+					
+					// m_nodesf[i] -= ProjectOnAxis(m_nodesv[i], force.normalized())/dtim;	
+					float3 nodeFMinus = nodeF - (projectOnAxis(nodeV, normalize(force))/dtim);
+					
+					nodeF = nodeFPlusForce;
+					if( dot(forceDTIM, forceDTIM) > dot(nodeV, nodeV) )
+						nodeF = nodeFMinus;
+									
+					g_vertexForceAccumulator[nodeID] = float4(nodeF, 0.0f);	
+				}
+			}
+		}
+	}
+}
+
+);

data/deps/include/bullet/BulletMultiThreaded/GpuSoftBodySolvers/DX11/HLSL/ComputeBounds.hlsl

@@ -0,0 +1,83 @@
+MSTRINGIFY(
+
+cbuffer ComputeBoundsCB : register( b0 )
+{
+	int numNodes;
+	int numSoftBodies;
+	int padding1;
+	int padding2;
+};
+
+// Node indices for each link
+StructuredBuffer<int> g_vertexClothIdentifier : register( t0 );
+StructuredBuffer<float4> g_vertexPositions : register( t1 );
+
+RWStructuredBuffer<uint4> g_clothMinBounds : register( u0 );
+RWStructuredBuffer<uint4> g_clothMaxBounds : register( u1 );
+
+groupshared uint4 clothMinBounds[256];
+groupshared uint4 clothMaxBounds[256];
+
+[numthreads(128, 1, 1)]
+void 
+ComputeBoundsKernel( uint3 Gid : SV_GroupID, uint3 DTid : SV_DispatchThreadID, uint3 GTid : SV_GroupThreadID, uint GI : SV_GroupIndex )
+{
+	const unsigned int UINT_MAX = 0xffffffff;
+
+	// Init min and max bounds arrays
+	if( GTid.x < numSoftBodies )
+	{
+		clothMinBounds[GTid.x] = uint4(UINT_MAX, UINT_MAX, UINT_MAX, UINT_MAX);
+		clothMaxBounds[GTid.x] = uint4(0,0,0,0);
+	}
+
+	AllMemoryBarrierWithGroupSync();
+
+	int nodeID = DTid.x;
+	if( nodeID < numNodes )
+	{	
+		int clothIdentifier = g_vertexClothIdentifier[nodeID];
+		if( clothIdentifier >= 0 )
+		{
+			float3 position = g_vertexPositions[nodeID].xyz;
+
+			// Reinterpret position as uint
+			uint3 positionUInt = uint3(asuint(position.x), asuint(position.y), asuint(position.z));
+		
+			// Invert sign bit of positives and whole of negatives to allow comparison as unsigned ints
+			//positionUInt.x ^= uint((-int(positionUInt.x >> 31) | 0x80000000));
+			//positionUInt.y ^= uint((-int(positionUInt.y >> 31) | 0x80000000));
+			//positionUInt.z ^= uint((-int(positionUInt.z >> 31) | 0x80000000));
+			positionUInt.x ^= (1+~(positionUInt.x >> 31) | 0x80000000);
+			positionUInt.y ^= (1+~(positionUInt.y >> 31) | 0x80000000);		
+			positionUInt.z ^= (1+~(positionUInt.z >> 31) | 0x80000000);
+		
+			// Min/max with the LDS values
+			InterlockedMin(clothMinBounds[clothIdentifier].x, positionUInt.x);
+			InterlockedMin(clothMinBounds[clothIdentifier].y, positionUInt.y);
+			InterlockedMin(clothMinBounds[clothIdentifier].z, positionUInt.z);
+
+			InterlockedMax(clothMaxBounds[clothIdentifier].x, positionUInt.x);
+			InterlockedMax(clothMaxBounds[clothIdentifier].y, positionUInt.y);
+			InterlockedMax(clothMaxBounds[clothIdentifier].z, positionUInt.z);
+		}
+	}
+	
+	AllMemoryBarrierWithGroupSync();
+
+
+	// Use global atomics to update the global versions of the data
+	if( GTid.x < numSoftBodies )
+	{
+		InterlockedMin(g_clothMinBounds[GTid.x].x, clothMinBounds[GTid.x].x);
+		InterlockedMin(g_clothMinBounds[GTid.x].y, clothMinBounds[GTid.x].y);
+		InterlockedMin(g_clothMinBounds[GTid.x].z, clothMinBounds[GTid.x].z);
+
+		InterlockedMax(g_clothMaxBounds[GTid.x].x, clothMaxBounds[GTid.x].x);		
+		InterlockedMax(g_clothMaxBounds[GTid.x].y, clothMaxBounds[GTid.x].y);
+		InterlockedMax(g_clothMaxBounds[GTid.x].z, clothMaxBounds[GTid.x].z);
+	}
+}
+
+
+);

data/deps/include/bullet/BulletMultiThreaded/GpuSoftBodySolvers/DX11/HLSL/Integrate.hlsl

@@ -0,0 +1,41 @@
+MSTRINGIFY(
+
+cbuffer IntegrateCB : register( b0 )
+{
+	int numNodes;
+	float solverdt;
+	int padding1;
+	int padding2;
+};
+
+// Node indices for each link
+StructuredBuffer<float> g_vertexInverseMasses : register( t0 );
+
+RWStructuredBuffer<float4> g_vertexPositions : register( u0 );
+RWStructuredBuffer<float4> g_vertexVelocity : register( u1 );
+RWStructuredBuffer<float4> g_vertexPreviousPositions : register( u2 );
+RWStructuredBuffer<float4> g_vertexForceAccumulator : register( u3 );
+
+[numthreads(128, 1, 1)]
+void 
+IntegrateKernel( uint3 Gid : SV_GroupID, uint3 DTid : SV_DispatchThreadID, uint3 GTid : SV_GroupThreadID, uint GI : SV_GroupIndex )
+{
+	int nodeID = DTid.x;
+	if( nodeID < numNodes )
+	{	
+		float3 position = g_vertexPositions[nodeID].xyz;
+		float3 velocity = g_vertexVelocity[nodeID].xyz;
+		float3 force = g_vertexForceAccumulator[nodeID].xyz;
+		float inverseMass = g_vertexInverseMasses[nodeID];
+		
+		g_vertexPreviousPositions[nodeID] = float4(position, 0.f);
+		velocity += force * inverseMass * solverdt;
+		position += velocity * solverdt;
+		
+		g_vertexForceAccumulator[nodeID] = float4(0.f, 0.f, 0.f, 0.0f);
+		g_vertexPositions[nodeID] = float4(position, 0.f);
+		g_vertexVelocity[nodeID] = float4(velocity, 0.f);	
+	}
+}
+
+);

data/deps/include/bullet/BulletMultiThreaded/GpuSoftBodySolvers/DX11/HLSL/OutputToVertexArray.hlsl

@@ -0,0 +1,63 @@
+MSTRINGIFY(
+
+cbuffer OutputToVertexArrayCB : register( b0 )
+{
+	int startNode;
+	int numNodes;
+	int positionOffset;
+	int positionStride;
+	
+	int normalOffset;	
+	int normalStride;
+	int padding1;
+	int padding2;
+};
+
+
+StructuredBuffer<float4> g_vertexPositions : register( t0 );
+StructuredBuffer<float4> g_vertexNormals : register( t1 );
+
+RWBuffer<float> g_vertexBuffer : register( u0 );
+
+
+[numthreads(128, 1, 1)]
+void 
+OutputToVertexArrayWithNormalsKernel( uint3 Gid : SV_GroupID, uint3 DTid : SV_DispatchThreadID, uint3 GTid : SV_GroupThreadID, uint GI : SV_GroupIndex )
+{
+	int nodeID = DTid.x;
+	if( nodeID < numNodes )
+	{			
+		float4 position = g_vertexPositions[nodeID + startNode];
+		float4 normal = g_vertexNormals[nodeID + startNode];
+		
+		// Stride should account for the float->float4 conversion
+		int positionDestination = nodeID * positionStride + positionOffset;		
+		g_vertexBuffer[positionDestination] = position.x;
+		g_vertexBuffer[positionDestination+1] = position.y;
+		g_vertexBuffer[positionDestination+2] = position.z;
+		
+		int normalDestination = nodeID * normalStride + normalOffset;
+		g_vertexBuffer[normalDestination] = normal.x;
+		g_vertexBuffer[normalDestination+1] = normal.y;
+		g_vertexBuffer[normalDestination+2] = normal.z;		
+	}
+}
+
+[numthreads(128, 1, 1)]
+void 
+OutputToVertexArrayWithoutNormalsKernel( uint3 Gid : SV_GroupID, uint3 DTid : SV_DispatchThreadID, uint3 GTid : SV_GroupThreadID, uint GI : SV_GroupIndex )
+{
+	int nodeID = DTid.x;
+	if( nodeID < numNodes )
+	{			
+		float4 position = g_vertexPositions[nodeID + startNode];
+		float4 normal = g_vertexNormals[nodeID + startNode];
+		
+		// Stride should account for the float->float4 conversion
+		int positionDestination = nodeID * positionStride + positionOffset;		
+		g_vertexBuffer[positionDestination] = position.x;
+		g_vertexBuffer[positionDestination+1] = position.y;
+		g_vertexBuffer[positionDestination+2] = position.z;		
+	}
+}
+);

data/deps/include/bullet/BulletMultiThreaded/GpuSoftBodySolvers/DX11/HLSL/PrepareLinks.hlsl

@@ -0,0 +1,44 @@
+MSTRINGIFY(
+
+cbuffer PrepareLinksCB : register( b0 )
+{
+	int numLinks;
+	int padding0;
+	int padding1;
+	int padding2;
+};
+
+// Node indices for each link
+StructuredBuffer<int2> g_linksVertexIndices : register( t0 );
+StructuredBuffer<float> g_linksMassLSC : register( t1 );
+StructuredBuffer<float4> g_nodesPreviousPosition : register( t2 );
+
+RWStructuredBuffer<float> g_linksLengthRatio : register( u0 );
+RWStructuredBuffer<float4> g_linksCurrentLength : register( u1 );
+
+[numthreads(128, 1, 1)]
+void 
+PrepareLinksKernel( uint3 Gid : SV_GroupID, uint3 DTid : SV_DispatchThreadID, uint3 GTid : SV_GroupThreadID, uint GI : SV_GroupIndex )
+{
+	int linkID = DTid.x;
+	if( linkID < numLinks )
+	{	
+		int2 nodeIndices = g_linksVertexIndices[linkID];
+		int node0 = nodeIndices.x;
+		int node1 = nodeIndices.y;
+		
+		float4 nodePreviousPosition0 = g_nodesPreviousPosition[node0];
+		float4 nodePreviousPosition1 = g_nodesPreviousPosition[node1];
+
+		float massLSC = g_linksMassLSC[linkID];
+		
+		float4 linkCurrentLength = nodePreviousPosition1 - nodePreviousPosition0;
+		
+		float linkLengthRatio = dot(linkCurrentLength, linkCurrentLength)*massLSC;
+		linkLengthRatio = 1./linkLengthRatio;
+		
+		g_linksCurrentLength[linkID] = linkCurrentLength;
+		g_linksLengthRatio[linkID] = linkLengthRatio;		
+	}
+}
+);

data/deps/include/bullet/BulletMultiThreaded/GpuSoftBodySolvers/DX11/HLSL/SolvePositions.hlsl

@@ -0,0 +1,55 @@
+MSTRINGIFY(
+
+cbuffer SolvePositionsFromLinksKernelCB : register( b0 )
+{
+	int startLink;
+	int numLinks;
+	float kst;
+	float ti;
+};
+
+// Node indices for each link
+StructuredBuffer<int2> g_linksVertexIndices : register( t0 );
+
+StructuredBuffer<float> g_linksMassLSC : register( t1 );
+StructuredBuffer<float> g_linksRestLengthSquared : register( t2 );
+StructuredBuffer<float> g_verticesInverseMass : register( t3 );
+
+RWStructuredBuffer<float4> g_vertexPositions : register( u0 );
+
+[numthreads(128, 1, 1)]
+void 
+SolvePositionsFromLinksKernel( uint3 Gid : SV_GroupID, uint3 DTid : SV_DispatchThreadID, uint3 GTid : SV_GroupThreadID, uint GI : SV_GroupIndex )
+{
+	int linkID = DTid.x + startLink;
+	if( DTid.x < numLinks )
+	{	
+		float massLSC = g_linksMassLSC[linkID];
+		float restLengthSquared = g_linksRestLengthSquared[linkID];
+		
+		if( massLSC > 0.0f )
+		{		
+			int2 nodeIndices = g_linksVertexIndices[linkID];
+			int node0 = nodeIndices.x;
+			int node1 = nodeIndices.y;
+			
+			float3 position0 = g_vertexPositions[node0].xyz;
+			float3 position1 = g_vertexPositions[node1].xyz;
+
+			float inverseMass0 = g_verticesInverseMass[node0];
+			float inverseMass1 = g_verticesInverseMass[node1]; 
+
+			float3 del = position1 - position0;
+			float len = dot(del, del);
+			float k = ((restLengthSquared - len)/(massLSC*(restLengthSquared+len)))*kst;
+			position0 = position0 - del*(k*inverseMass0);
+			position1 = position1 + del*(k*inverseMass1);
+
+			g_vertexPositions[node0] = float4(position0, 0.f);
+			g_vertexPositions[node1] = float4(position1, 0.f);
+
+		}
+	}
+}
+
+);