[GeekyBear]

It just goes to show that NVidia thinks supporting CUDA everywhere is very much in their competitive interest, while creating and supporting an OpenGL implementation simply is not.