Voxel Operations with CUDA Using cudaMalloc3D


When I allocate the device memory using CudaMalloc3D and the cudaPitchedPtr, is it acceptable to use a normal cudaMemcpy for copying it to the device memory, using a pointer to the disk buffer in which I have stored reading from the file.

Also, please explain abt how typecasting to the void* ptr of ‘cudaPitchedPtr’ is done!!

Awaiting your reply.