Some simple OpenCL tests

I wrote some simple benchmarks to test performance of misc things like data transfers, compilation speed, parallel data transfers etc:

You can either clone the repository using mercurial, or if you don’t have mercurial, then you can see the “get source” option on the upper right of the page and get it as a zip file etc.
Tested on Linux. Uses Boost 1.48. I know thats a big dependency but boost’s installation process is pretty simple.
Also uses OpenCL C++ bindings, AFAIK Nvidia doesn’t provide them by default but you can get the cl.hpp from Khronos and modify a few lines to be compatible with Nvidia.

Actually, just for convenience, I have added a download marked v0.1 on the “downloads” page.