OpenCL and CUDA


I will be having an situation where I will be having two function(or library) in same process running OpenCL Kernel CUDA Kernel.

These two application will not run parallel but they will be sequential. My concern is how much of time delay if my first function CUDA between second function with OpenCL function starts?

Is it significant? is there any smart solution?

Thank you,

Not sure about NVIDIA drivers, but on AMD the dispatch of a kernel takes about several milliseconds.