SEXP R_auto_cudaMemset2D(SEXP r_devPtr, SEXP r_pitch, SEXP r_value, SEXP r_width, SEXP r_height)
{
SEXP r_ans = R_NilValue;
void * devPtr = GET_REF(r_devPtr, void );
size_t pitch = REAL(r_pitch)[0];
int value = INTEGER(r_value)[0];
size_t width = REAL(r_width)[0];
size_t height = REAL(r_height)[0];
cudaError_t ans;
ans = cudaMemset2D(devPtr, pitch, value, width, height);
r_ans = Renum_convert_cudaError_t(ans) ;
return(r_ans);
}
cudaError_t WINAPI wine_cudaMemset2D( void *devPtr, size_t pitch, int value, size_t width, size_t height ) {
WINE_TRACE("\n");
return cudaMemset2D( devPtr, pitch, value, width, height );
}
void oz::gpu_image::clear() {
OZ_CUDA_SAFE_CALL(cudaMemset2D(ptr(), pitch(), 0, row_size(), h()));
}
void InitCudaLayers()
{
mmGridSizeX = sim_width/blockSizex;
mmGridSizeY = sim_height/blockSizey;
mmGridSize = mmGridSizeX*mmGridSizeY;
memset(mmGrid, 0, sizeof(mmGrid));
memset(mmYGGrid, 0, sizeof(mmYGGrid));
tempHostData = (float*)malloc(sim_width*sim_height*TEMP_HOST_ELEM*sizeof(float));
tempHostDataNoCuda = (float*)malloc(sim_width*sim_height*TEMP_HOST_ELEM*sizeof(float));
grid8ValTick = (float*)malloc(sim_width*sim_height*8*sizeof(float));
initColors();
memset(gCudaLayer, 0, sizeof(gCudaLayer));
memset(gCudaFuncLayer, 0, sizeof(gCudaFuncLayer));
memset(gPhysLayer, 0, sizeof(gPhysLayer));
memset(gStateLayer, 0, sizeof(gStateLayer));
srand(0);
int seed = rand();
const cudaChannelFormatDesc desc4 = cudaCreateChannelDesc<float4>();
cudaMallocArray(&gCudaVectArray, &desc4, sim_width, sim_height);
#if NFLAYERS ==2
const cudaChannelFormatDesc desc2 = cudaCreateChannelDesc<float2>();
#else if NFLAYERS ==4
const cudaChannelFormatDesc descF = desc4;
#endif
cudaMallocArray(&gCudaFlArray, &descF, sim_width, sim_height);
const cudaChannelFormatDesc desc = cudaCreateChannelDesc<float>();
cudaMallocArray(&gCudaFuncWavePack, &desc, sim_width);
cudaMallocArray(&gCudaFuncSmooth, &desc, sim_width);
cudaMallocArray(&(gCudaLayer[0]), &desc, sim_width, sim_height);
cudaMallocArray(&(gCudaLayer[1]), &desc, sim_width, sim_height);
cudaMallocArray(&(gCudaFuncLayer[0]), &desc, sim_width, sim_height);
cudaMalloc(&cuTempData, TEMP_SIZE*sizeof(float)*sim_width*sim_height);
cudaMalloc(&cuRandArr, sizeof(unsigned int)*sim_width*sim_height);
cudaMalloc(&gStateLayer[0], sim_rect*sizeof(float));
cudaMemset(gStateLayer[0], 0, sim_rect*sizeof(float));
cudaMalloc(&gStateLayer[1], sim_rect*sizeof(float));
cudaMemset(gStateLayer[1], 0, sim_rect*sizeof(float));
cudaMalloc(&gPhysLayer[0], sim_rect*sizeof(float));
cudaMemset(gPhysLayer[0], 0, sim_rect*sizeof(float));
cudaMalloc(&gPhysLayer[1], sim_rect*sizeof(float));
cudaMemset(gPhysLayer[1], 0, sim_rect*sizeof(float));
cudaMalloc(&gRedBlueField, NFLAYERS*sim_rect*sizeof(float));
cudaMemset(gRedBlueField, 0, NFLAYERS*sim_rect*sizeof(float));
size_t pitch = 4*sim_width*sizeof(float);
cudaMallocPitch((void**)&gVectorLayer, &pitch, 4*sim_width*sizeof(float), sim_height);
cudaMemset2D(gVectorLayer, 4*sim_width*sizeof(float), 0, 4*sim_width*sizeof(float), sim_height);
InitWavePack(32, 1.f, sim_width, sim_height, cuTempData, gCudaFuncWavePack);
InitSmooth(1, sim_width, cuTempData, gCudaFuncSmooth);
InitRnd2DInt(seed, cuRandArr, sim_width, sim_height);
InitFuncLayer(gCudaFuncLayer[0], cuTempData, sim_width, sim_height);
InitPhysLayer(gPhysLayer[0], gStateLayer[0], cuRandArr, sim_width, sim_height);
float* gridIni = cuTempData+3*sim_rect/2;
float* halfTemp = cuTempData + sim_rect;
float* out = cuTempData + 2*sim_rect;
cudaMemset(out, 0, sim_rect*sizeof(float));
seed = rand();
int gridx = INTERP_SIZEX;
int gridy = INTERP_SIZEX;
InitRnd2DF(seed, gridIni, gridx, gridy);
float scaleadd = .7f;
Spline2D(gridIni, gridx, gridy, halfTemp, scaleadd, out, sim_width, sim_height);
seed = rand();
gridx = (int)(gridx*2);
gridy = (int)(gridy*2);
InitRnd2DF(seed, gridIni, gridx, gridy);
scaleadd = .3f;
Spline2D(gridIni, gridx, gridy, halfTemp, scaleadd, out, sim_width, sim_height);
cudaMemcpyToArray(gCudaLayer[0], 0, 0, out, sizeof(float)*sim_rect, cudaMemcpyDeviceToDevice);
cudaMemset(out, 0, sim_rect*sizeof(float));
gridx = INTERP_SIZEX;
gridy = INTERP_SIZEX;
seed = rand();
InitRnd2DF(seed, gridIni, gridx, gridy);
scaleadd = .7f;
Spline2D(gridIni, gridx, gridy, halfTemp, scaleadd, out, sim_width, sim_height);
seed = rand();
gridx = (int)(gridx*1.5);
gridy = (int)(gridy*1.5);
InitRnd2DF(seed, gridIni, gridx, gridy);
scaleadd = .3f;
Spline2D(gridIni, gridx, gridy, halfTemp, scaleadd, out, sim_width, sim_height);
cudaMemcpyToArray(gCudaLayer[1], 0, 0, out, sizeof(float)*sim_rect, cudaMemcpyDeviceToDevice);
float2 pos0;
pos0.x = gObj0X;
pos0.y = gObj0Y;
float2 pos1;
pos1.x = gObj1X;
pos1.y = gObj1Y;
gObjInertia.Init(pos0, pos1);
LayerProc(sim_width, sim_height, gCudaLayer[0], gCudaFuncLayer[0], cuTempData, pos0.x , pos0.y, pos1.x , pos1.y);
ParticleStateInit(cuTempData, cuRandArr,
gStateLayer[0], gPhysLayer[0], gRedBlueField);
InitBhv();
}
