I had a quick question about cudaMemcpy2D. If you are making a CP from host to device then what do you use for the source pitch since it was not allocated with cudaMallocPitch?
The data width.
pitch = width + padding;
In this case, padding is 0.
Coll, thanks for the help.