Extras/RigidBodyGpuPipeline/opencl/primitives/AdlPrimitives/Copy/CopyKernelsDX11.h

   1 static const char* copyKernelsDX11= \\r
   2 "/*\n"\r
   3 "               2011 Takahiro Harada\n"\r
   4 "*/\n"\r
   5 "\n"\r
   6 "typedef uint u32;\n"\r
   7 "\n"\r
   8 "#define GET_GROUP_IDX groupIdx.x\n"\r
   9 "#define GET_LOCAL_IDX localIdx.x\n"\r
  10 "#define GET_GLOBAL_IDX globalIdx.x\n"\r
  11 "#define GROUP_LDS_BARRIER GroupMemoryBarrierWithGroupSync()\n"\r
  12 "#define GROUP_MEM_FENCE\n"\r
  13 "#define DEFAULT_ARGS uint3 globalIdx : SV_DispatchThreadID, uint3 localIdx : SV_GroupThreadID, uint3 groupIdx : SV_GroupID\n"\r
  14 "#define AtomInc(x) InterlockedAdd(x, 1)\n"\r
  15 "#define AtomInc1(x, out) InterlockedAdd(x, 1, out)\n"\r
  16 "\n"\r
  17 "#define make_uint4 uint4\n"\r
  18 "#define make_uint2 uint2\n"\r
  19 "#define make_int2 int2\n"\r
  20 "\n"\r
  21 "#define WG_SIZE 64\n"\r
  22 "\n"\r
  23 "#define GET_GROUP_SIZE WG_SIZE\n"\r
  24 "\n"\r
  25 "\n"\r
  26 "\n"\r
  27 "cbuffer CB : register( b0 )\n"\r
  28 "{\n"\r
  29 "       int m_n;\n"\r
  30 "       int m_padding[3];\n"\r
  31 "};\n"\r
  32 "\n"\r
  33 "RWStructuredBuffer<float4> dst : register( u0 );\n"\r
  34 "StructuredBuffer<float4> src : register( t0 );\n"\r
  35 "\n"\r
  36 "[numthreads(WG_SIZE, 1, 1)]\n"\r
  37 "void Copy1F4Kernel( DEFAULT_ARGS )\n"\r
  38 "{\n"\r
  39 "       int gIdx = GET_GLOBAL_IDX;\n"\r
  40 "\n"\r
  41 "       if( gIdx < m_n )\n"\r
  42 "       {\n"\r
  43 "               float4 a0 = src[gIdx];\n"\r
  44 "\n"\r
  45 "               dst[ gIdx ] = a0;\n"\r
  46 "       }\n"\r
  47 "}\n"\r
  48 "\n"\r
  49 "[numthreads(WG_SIZE, 1, 1)]\n"\r
  50 "void Copy2F4Kernel( DEFAULT_ARGS )\n"\r
  51 "{\n"\r
  52 "       int gIdx = GET_GLOBAL_IDX;\n"\r
  53 "\n"\r
  54 "       if( 2*gIdx <= m_n )\n"\r
  55 "       {\n"\r
  56 "               float4 a0 = src[gIdx*2+0];\n"\r
  57 "               float4 a1 = src[gIdx*2+1];\n"\r
  58 "\n"\r
  59 "               dst[ gIdx*2+0 ] = a0;\n"\r
  60 "               dst[ gIdx*2+1 ] = a1;\n"\r
  61 "       }\n"\r
  62 "}\n"\r
  63 "\n"\r
  64 "[numthreads(WG_SIZE, 1, 1)]\n"\r
  65 "void Copy4F4Kernel( DEFAULT_ARGS )\n"\r
  66 "{\n"\r
  67 "       int gIdx = GET_GLOBAL_IDX;\n"\r
  68 "\n"\r
  69 "       if( 4*gIdx <= m_n )\n"\r
  70 "       {\n"\r
  71 "               int idx0 = gIdx*4+0;\n"\r
  72 "               int idx1 = gIdx*4+1;\n"\r
  73 "               int idx2 = gIdx*4+2;\n"\r
  74 "               int idx3 = gIdx*4+3;\n"\r
  75 "\n"\r
  76 "               float4 a0 = src[idx0];\n"\r
  77 "               float4 a1 = src[idx1];\n"\r
  78 "               float4 a2 = src[idx2];\n"\r
  79 "               float4 a3 = src[idx3];\n"\r
  80 "\n"\r
  81 "               dst[ idx0 ] = a0;\n"\r
  82 "               dst[ idx1 ] = a1;\n"\r
  83 "               dst[ idx2 ] = a2;\n"\r
  84 "               dst[ idx3 ] = a3;\n"\r
  85 "       }\n"\r
  86 "}\n"\r
  87 "\n"\r
  88 "RWStructuredBuffer<float> dstF1 : register( u0 );\n"\r
  89 "StructuredBuffer<float> srcF1 : register( t0 );\n"\r
  90 "\n"\r
  91 "[numthreads(WG_SIZE, 1, 1)]\n"\r
  92 "void CopyF1Kernel( DEFAULT_ARGS )\n"\r
  93 "{\n"\r
  94 "       int gIdx = GET_GLOBAL_IDX;\n"\r
  95 "\n"\r
  96 "       if( gIdx < m_n )\n"\r
  97 "       {\n"\r
  98 "               float a0 = srcF1[gIdx];\n"\r
  99 "\n"\r
 100 "               dstF1[ gIdx ] = a0;\n"\r
 101 "       }\n"\r
 102 "\n"\r
 103 "}\n"\r
 104 "\n"\r
 105 "RWStructuredBuffer<float2> dstF2 : register( u0 );\n"\r
 106 "StructuredBuffer<float2> srcF2 : register( t0 );\n"\r
 107 "\n"\r
 108 "[numthreads(WG_SIZE, 1, 1)]\n"\r
 109 "void CopyF2Kernel( DEFAULT_ARGS )\n"\r
 110 "{\n"\r
 111 "       int gIdx = GET_GLOBAL_IDX;\n"\r
 112 "\n"\r
 113 "       if( gIdx < m_n )\n"\r
 114 "       {\n"\r
 115 "               float2 a0 = srcF2[gIdx];\n"\r
 116 "\n"\r
 117 "               dstF2[ gIdx ] = a0;\n"\r
 118 "       }\n"\r
 119 "}\n"\r
 120 ;\r