Revision history [back]

I want to run a gpu moudule function (gpu::detectMultiScale for instance) with multiple threads on GPU. Someting like that:

detectMultiScale_routine<<< gridDiDim, blockm >>>(args);

and detectMultiScale_routine calls gpu::detectMultiScale in it.

There is an actual way to do that?

Bottom line: I want to run multiple threads of gpu::detectMultiScale on GPU.