OpenCL on CUDA

I have installed the recent 190.28 NVIDIA driver and have the CUDA SDK v2.3 installed. Whenever I try to create a basic Hello World type OpenCL program and compile and run it, it gives me the CL_DEVICE_NOT_FOUND error. I am creating a Context using the CL_DEVICE_TYPE_CPU. Any help would be appreciated.

NVIDIA SDK is not supporting compiling for CPU yet, you should try ATI/AMD Stream SDK instead (but note that this one, on the contrary, doesn’t support building for GPU yet).