I am thinking about using GPU programming with Mathematica using a MacBook Pro with a NVIDIA GPU. At this page, the Mathematica documentation says "Programming OpenCL in the Wolfram Language is simple since the user need not write C wrapper code -- which can be quite verbose, difficult to understand, and hard to debug. Using OpenCLLink also guarantees compatibility as new versions of the standard are released." So what advantages might there be to programming in CUDA instead of OpenCL?

  • $\begingroup$ This is a useful read. In Mathematica, CUDALink has more useful functions that OpenCLLink. Both the CUDALink and the OpenCLLink demos work fine for me on a 2012 MacBook Pro, but I didn't go deep into the technologies. I think using them from Mathematica will be quite similar. $\endgroup$ – Szabolcs Aug 18 '14 at 20:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.