Alternative(s) to run CUDA on non-Nvidia hardware

 (hpcwire.com)

31 points | by alok-g 1 hour ago

4 comments

  • DiabloD3 5 minutes ago
    Its easier to just get rid of your legacy code entirely and use Vulkan for compute, or have your compiler emit SPIR-V directly.

    No reason to tie yourself to Nvidia's moat.

  • maxloh 16 minutes ago
    There is also ZLUDA, which is open source and works on pre-compiled binaries.

    https://github.com/vosen/ZLUDA

  • pjmlp 1 hour ago
    Most of these "alternatives" focus on CUDA C++, and overlook what actually makes CUDA interesting.

    Already in 2020,

    https://developer.nvidia.com/blog/cuda-refresher-the-gpu-com...

    [-]
    • msond 47 minutes ago
      We're actually targeting all of it, and not just CUDA C++.
      [-]
      • pjmlp 40 minutes ago
        Including stuff like Fortran, Haskell, Java, .NET via PTX, Python JIT, IDE tooling integration with major IDEs, graphical GPU debugging and profiling, libraries and co?

        Then I guess all the best.

        [-]
        • zorked 4 minutes ago
          This post has some serious peanut-gallery vibes.
      • embedding-shape 32 minutes ago
        Ambitions but neat, good luck if nothing else :)

        If you were to guess, when do you think your Nsight Compute alternative might be ready with your own toolchain?

  • lulzx 4 minutes ago
    I have been trying for cuda -> metal, to run it on mac, https://github.com/lulzx/cuda-metal