CUDA中三维数组的创建,读取,拷贝。以下创建float4类型的width*height*depth大小的数组,在线性存储器中。
Host Code:
...
cudaPitchedPtr data;
cudaExtent extent = make_cudaExtent(width* sizeof(float4), height, depth);
cudaMalloc3D(&(d_data), extent);
...
Kernel Code:
__global__ kernel(char* data, size_t pitch, ...)
{
...
float4 element = *((float4*) (data + (x*sizeof(float4) + y*pitch + z*pitch*height)));
...
}
Host Code:
unsigned int size = width*height*depth*sizeof(float4);
float4 * h_volume=(float4*)malloc(size);
cudaMemcpy3DParms copyParams = {0};
copyParams.srcPtr = data;
copyParams.dstPtr = make_cudaPitchedPtr(h_volume,width*sizeof(float4),width,height);
copyParams.extent = extent;
copyParams.kind = cudaMemcpyDeviceToHost;
cutilSafeCall(cudaMemcpy3D(©Params));