Comment by hangonhn

Is CUDA tied very closely to the Nvidia hardware and architecture so that all the abstraction would not make sense on other platforms? I know very little about hardware and low level software.

Thanks