Being on the Khronos board means nothing really when it comes to actual implementation for the most part. Its just a standards committe… Microsoft is on the Khronos board, but they are not evangelizing OpenGL, they promote their api DirectX. As far as OpenCL goes, Nvidia have pretty much ‘*** the pooch’ ( sorry for the language, but it had to be said ) on this one. Of course they are going to promote CUDA, which from a business standpoint kind make sense. But when Ithink about it, it does really, since most user’s seems to prefer Nvidia GPU ( usually driver stability and a wider range of extensions and other feature ). So for ex. if a developer is dead set on using OpenCL for his/her Compute API, then Nvidia just lost a customer. This is all from a general point of view since I assume Nvidia has already done the market research to aid their decisions to pursue CUDA. In closing, in my eyes ( and most other engineers I interact with ), CUDA will always have its place in history as the first general purpose, easy to use compute api, so lets close that chapter and work to further OpenCL on all platforms.