I am a beginner in GPU programming. I tried CUDA for a few weeks and now I am trying OpenCL.
In CUDA I used the functions cudaMemcpy2D, cudaMemset2D and cudaMallocPitch to have data aligned, and it signifcantly improved performance on 8800GT, less on Fermi (because of cache ?).
My problem is that I cannot find any similar functions in OpenCL.
Would you have any idea of how I can do the same thing ?
Any help would be appreciated.