added Samurai Monastry wavefront .obj

added alternative batching kernel (slow)
tweaked controls a bit
added command-line options --selected_demo=<int> and --new_batching
started looking into parallel 3d sap

opencl/gpu_rigidbody/host/btConfig.h

@@ -18,7 +18,7 @@ struct	btConfig
 	int m_maxTriConvexPairCapacity;
 
 	btConfig()
-		:m_maxConvexBodies(64*1024),
+		:m_maxConvexBodies(32*1024),
 		m_maxConvexShapes(8192),
 		m_maxVerticesPerFace(64),
 		m_maxFacesPerShape(64),
@@ -26,7 +26,7 @@ struct	btConfig
 		m_maxConvexIndices(8192),
 		m_maxConvexUniqueEdges(8192),
 		m_maxCompoundChildShapes(8192),
-		m_maxTriConvexPairCapacity(64*1024)
+		m_maxTriConvexPairCapacity(512*1024)
 	{
 		m_maxBroadphasePairs = 16*m_maxConvexBodies;
 	}