reorder files, in preparation for Bullet 3 -> Bullet 2 merge
This commit is contained in:
164
src/Bullet3OpenCL/BroadphaseCollision/kernels/sapFastKernels.h
Normal file
164
src/Bullet3OpenCL/BroadphaseCollision/kernels/sapFastKernels.h
Normal file
@@ -0,0 +1,164 @@
|
||||
//this file is autogenerated using stringify.bat (premake --stringify) in the build folder of this project
|
||||
static const char* sapFastCL= \
|
||||
"/*\n"
|
||||
"Copyright (c) 2012 Advanced Micro Devices, Inc. \n"
|
||||
"\n"
|
||||
"This software is provided 'as-is', without any express or implied warranty.\n"
|
||||
"In no event will the authors be held liable for any damages arising from the use of this software.\n"
|
||||
"Permission is granted to anyone to use this software for any purpose, \n"
|
||||
"including commercial applications, and to alter it and redistribute it freely, \n"
|
||||
"subject to the following restrictions:\n"
|
||||
"\n"
|
||||
"1. The origin of this software must not be misrepresented; you must not claim that you wrote the original software. If you use this software in a product, an acknowledgment in the product documentation would be appreciated but is not required.\n"
|
||||
"2. Altered source versions must be plainly marked as such, and must not be misrepresented as being the original software.\n"
|
||||
"3. This notice may not be removed or altered from any source distribution.\n"
|
||||
"*/\n"
|
||||
"//Originally written by Erwin Coumans\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"typedef struct \n"
|
||||
"{\n"
|
||||
" union\n"
|
||||
" {\n"
|
||||
" float4 m_min;\n"
|
||||
" float m_minElems[4];\n"
|
||||
" int m_minIndices[4];\n"
|
||||
" };\n"
|
||||
" union\n"
|
||||
" {\n"
|
||||
" float4 m_max;\n"
|
||||
" float m_maxElems[4];\n"
|
||||
" int m_maxIndices[4];\n"
|
||||
" };\n"
|
||||
"} b3AabbCL;\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"/// conservative test for overlap between two aabbs\n"
|
||||
"bool TestAabbAgainstAabb2(const b3AabbCL* aabb1, __local const b3AabbCL* aabb2);\n"
|
||||
"bool TestAabbAgainstAabb2(const b3AabbCL* aabb1, __local const b3AabbCL* aabb2)\n"
|
||||
"{\n"
|
||||
"//skip pairs between static (mass=0) objects\n"
|
||||
" if ((aabb1->m_maxIndices[3]==0) && (aabb2->m_maxIndices[3] == 0))\n"
|
||||
" return false;\n"
|
||||
" \n"
|
||||
" bool overlap = true;\n"
|
||||
" overlap = (aabb1->m_min.x > aabb2->m_max.x || aabb1->m_max.x < aabb2->m_min.x) ? false : overlap;\n"
|
||||
" overlap = (aabb1->m_min.z > aabb2->m_max.z || aabb1->m_max.z < aabb2->m_min.z) ? false : overlap;\n"
|
||||
" overlap = (aabb1->m_min.y > aabb2->m_max.y || aabb1->m_max.y < aabb2->m_min.y) ? false : overlap;\n"
|
||||
" return overlap;\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"//computePairsKernelBatchWrite\n"
|
||||
"__kernel void computePairsKernel( __global const b3AabbCL* aabbs, volatile __global int2* pairsOut,volatile __global int* pairCount, int numObjects, int axis, int maxPairs)\n"
|
||||
"{\n"
|
||||
" int i = get_global_id(0);\n"
|
||||
" int localId = get_local_id(0);\n"
|
||||
"\n"
|
||||
" __local int numActiveWgItems[1];\n"
|
||||
" __local int breakRequest[1];\n"
|
||||
" __local b3AabbCL localAabbs[128];// = aabbs[i];\n"
|
||||
" \n"
|
||||
" int2 myPairs[64];\n"
|
||||
" \n"
|
||||
" b3AabbCL myAabb;\n"
|
||||
" \n"
|
||||
" myAabb = (i<numObjects)? aabbs[i]:aabbs[0];\n"
|
||||
" float testValue = myAabb.m_maxElems[axis];\n"
|
||||
" \n"
|
||||
" if (localId==0)\n"
|
||||
" {\n"
|
||||
" numActiveWgItems[0] = 0;\n"
|
||||
" breakRequest[0] = 0;\n"
|
||||
" }\n"
|
||||
" int localCount=0;\n"
|
||||
" int block=0;\n"
|
||||
" localAabbs[localId] = (i+block)<numObjects? aabbs[i+block] : aabbs[0];\n"
|
||||
" localAabbs[localId+64] = (i+block+64)<numObjects? aabbs[i+block+64]: aabbs[0];\n"
|
||||
" \n"
|
||||
" barrier(CLK_LOCAL_MEM_FENCE);\n"
|
||||
" atomic_inc(numActiveWgItems);\n"
|
||||
" barrier(CLK_LOCAL_MEM_FENCE);\n"
|
||||
" int localBreak = 0;\n"
|
||||
" int curNumPairs = 0;\n"
|
||||
" \n"
|
||||
" int j=i+1;\n"
|
||||
" do\n"
|
||||
" {\n"
|
||||
" barrier(CLK_LOCAL_MEM_FENCE);\n"
|
||||
" \n"
|
||||
" if (j<numObjects)\n"
|
||||
" {\n"
|
||||
" if(testValue < (localAabbs[localCount+localId+1].m_minElems[axis])) \n"
|
||||
" {\n"
|
||||
" if (!localBreak)\n"
|
||||
" {\n"
|
||||
" atomic_inc(breakRequest);\n"
|
||||
" localBreak = 1;\n"
|
||||
" }\n"
|
||||
" }\n"
|
||||
" }\n"
|
||||
" \n"
|
||||
" barrier(CLK_LOCAL_MEM_FENCE);\n"
|
||||
" \n"
|
||||
" if (j>=numObjects && !localBreak)\n"
|
||||
" {\n"
|
||||
" atomic_inc(breakRequest);\n"
|
||||
" localBreak = 1;\n"
|
||||
" }\n"
|
||||
" barrier(CLK_LOCAL_MEM_FENCE);\n"
|
||||
" \n"
|
||||
" if (!localBreak)\n"
|
||||
" {\n"
|
||||
" if (TestAabbAgainstAabb2(&myAabb,&localAabbs[localCount+localId+1]))\n"
|
||||
" {\n"
|
||||
" int2 myPair;\n"
|
||||
" myPair.x = myAabb.m_minIndices[3];\n"
|
||||
" myPair.y = localAabbs[localCount+localId+1].m_minIndices[3];\n"
|
||||
" myPairs[curNumPairs] = myPair;\n"
|
||||
" curNumPairs++;\n"
|
||||
" if (curNumPairs==64)\n"
|
||||
" {\n"
|
||||
" int curPair = atomic_add(pairCount,curNumPairs);\n"
|
||||
" //avoid a buffer overrun\n"
|
||||
" if ((curPair+curNumPairs)<maxPairs)\n"
|
||||
" {\n"
|
||||
" for (int p=0;p<curNumPairs;p++)\n"
|
||||
" {\n"
|
||||
" pairsOut[curPair+p] = myPairs[p]; //flush to main memory\n"
|
||||
" }\n"
|
||||
" }\n"
|
||||
" curNumPairs = 0;\n"
|
||||
" }\n"
|
||||
" }\n"
|
||||
" }\n"
|
||||
" barrier(CLK_LOCAL_MEM_FENCE);\n"
|
||||
" \n"
|
||||
" localCount++;\n"
|
||||
" if (localCount==64)\n"
|
||||
" {\n"
|
||||
" localCount = 0;\n"
|
||||
" block+=64; \n"
|
||||
" localAabbs[localId] = ((i+block)<numObjects) ? aabbs[i+block] : aabbs[0];\n"
|
||||
" localAabbs[localId+64] = ((i+64+block)<numObjects) ? aabbs[i+block+64] : aabbs[0];\n"
|
||||
" }\n"
|
||||
" j++;\n"
|
||||
" \n"
|
||||
" } while (breakRequest[0]<numActiveWgItems[0]);\n"
|
||||
" \n"
|
||||
" \n"
|
||||
" if (curNumPairs>0)\n"
|
||||
" {\n"
|
||||
" //avoid a buffer overrun\n"
|
||||
" int curPair = atomic_add(pairCount,curNumPairs);\n"
|
||||
" if ((curPair+curNumPairs)<maxPairs)\n"
|
||||
" {\n"
|
||||
" for (int p=0;p<curNumPairs;p++)\n"
|
||||
" {\n"
|
||||
" pairsOut[curPair+p] = myPairs[p]; //flush to main memory\n"
|
||||
" }\n"
|
||||
" }\n"
|
||||
" curNumPairs = 0;\n"
|
||||
" }\n"
|
||||
"}\n"
|
||||
;
|
||||
Reference in New Issue
Block a user