Does AMD APP have some optimization on the AMD CPU when running OpenCL on CPU?
I have tried to evaluate the performance for OpenCL on CPU. In my test, I run OpenCL code with 1 workgroup and 2 work items. It seems on AMD CPU, the OpenCL can utilize the 2 CPU core, and on Intel CPU, it can only use 1 CPU core.
Does anyone know if there's optimize work done by AMD APP driver for AMD CPU?