memcpy from 3D array to devPtr of cudaGLMapBufferObject(devPtr)

  1. ptr3D = cudaMalloc3D() => this also returns a pitch.
  2. cudaGLMapBufferObject((void **) devPtr, pboId);

To check if my CUDA-GL interop is working, I write something to the devPtr using a cuda kernel and contents are what I expect. I just to need to know how to copy from ptr3D to devPtr since there is this pitch to ptr3D.

What is the memory alignment of devPtr ?
How to copy from ptr3D to devPtr ?