Hi there. A few days ago I started writing some code in OpenCL (never used it before) and I have the following problem. I have a kernel which when executed in the CPU works fine but when it runs in the GPU it gives me false results and leads to a segmentation fault after a while.
You can see the kernel here: http://pastebin.com/7N6WMw5k
What especially gives me the problem is the last small loop which updates the values. Why can this happen? What am I doing wrong?