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

(hpcwire.com)

64 points | by alok-g 3 hours ago

12 comments

  • woctordho 1 hour ago
    There's nothing wrong to run CUDA on non-Nvidia hardware. CUDA has an interface that is reasonably well-designed, well-documented/reverse-engineered, and battle-tested for decades. What we need is not to invent another interface just under the name of 'open standard', but to implement the same interface. ROCm is exactly doing this, and so are other hardware SDKs such as MooreThread and Alibaba T-Head.
  • u1hcw9nx 16 minutes ago
    Alternatives exist, but little demand outside hyperscalers and special uses.

    Neocloud customers just want plug-and-play CUDA. It works, it's tested, it adapts faster, and has known performance. Alternatives give no significant benefits.

    Things can change, but they are not changing now.

  • pjmlp 2 hours 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...

    • mschuetz 1 hour ago
      > Ease of programming and a giant leap in performance is one of the key reasons for the CUDA platform’s widespread adoption

      This, so much. Other platforms continue to ignore developer UX, but it's one of the main things that get's new users onboard and keeps old users around.

    • msond 2 hours ago
      We're actually targeting all of it, and not just CUDA C++.
      • pjmlp 2 hours 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 1 hour ago
          This post has some serious peanut-gallery vibes.
          • pjmlp 1 hour ago
            Peanut-gallery is happily using CUDA, and needs actual sound reasons to move.
            • account42 19 minutes ago
              Then the peanut gallery has nothing to complain when Nvidia jacks up prices.
      • embedding-shape 2 hours ago
        Ambitious 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?

        • msond 1 hour ago
          A guess would be some time next year — since our public launch our focus has generally been on API coverage and increasingly recently, on performance.

          While performance improvements will always remain a target, we're soon at full coverage of the core CUDA APIs and will be shifting an increasing amount of effort towards developer tooling.

  • dachworker 21 minutes ago
    Why should I not just port my kernel to Triton? What's the appeal of Scale?
  • luciana1u 1 hour ago
    every CUDA alternative follows the same arc: bold launch, works for 3 operations, then a Discord server where the last message is 'any updates?' from 2024
  • puschkinfr 1 hour ago
    In this context AdaptiveCpp should also be mentioned. Started as a SYCL implementation, but recently-ish added a compiler for compiling a CUDA dialect to GPUs and CPUs from basically all vendors
  • asdaqopqkq 20 minutes ago
    aren't llms smart enough to directly write custom kernels for custom hardware from cuda code?
  • maxloh 2 hours ago
    There is also ZLUDA, which is open source and works on pre-compiled binaries.

    https://github.com/vosen/ZLUDA

  • lulzx 1 hour ago
    I have been trying for cuda -> metal, to run it on mac, https://github.com/lulzx/cuda-metal
  • DiabloD3 1 hour 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.

    • mschuetz 1 hour ago
      A couple of years ago I evaluated both Vulkan and Cuda as a choice for future projects. I couldnt get anything done after a week in Vulkan, but had the test prototype project working after just a day in Cuda.

      Needless to say, I'd never ever pick Vulkan for any project after that experience. It's just way to needlessly overengineered and bloated.

      • pjmlp 26 minutes ago
        I used to be big into Khronos API camp, even did my project thesis in OpenGL, up to the famous Long Peaks fail.

        Vulkan ended up being the same extension spaghetti as its predecessor, and Khronos was only able to come up with something thanks to AMD offering Mantle, C++ bindings and a GLSL successor only came to be thanks to NVidia (Vulkan-hpp and Slang started at NVidia).

        The "we build the specification", and then "the community builds the tools", leads to very poor experiences, and if it wasn't for LunarG own interests, there wouldn't even exist any kind of Vulkan SDK.

        What they have going is naturally the vendor independence, however we can achieve the same with middleware with the benefit of much better developer experience.

        • DiabloD3 20 minutes ago
          I love how people say things like "extension spaghetti", as if all other non-standard APIs have the same problem: hardware gets new features that people want to use from that API, API gains extension to use that hardware feature.

          CUDA is no different, in fact, often worse. Nvidia is bad at documenting which hardware does what things, and CUDA users often have to use third party tables to figure out what hardware can't do what and disappoint customers who unwisely invested into it.

      • DiabloD3 23 minutes ago
        Weird, most people have the exact opposite experience.

        Having to deal with closed source opaque poorly documented stacks sucks.

        • mschuetz 7 minutes ago
          They really don't, no. Vulkan: 50 lines to allocate device memory. Cuda: One single line. What kind of extensive documentation stack do you want for functionality that is trivial in Cuda? And that exact issue continues through every little step of the way to your first usable application. I know there is VMA, it is a very poor solution to a problem that shouldn't even exist, and it only poorly addresses one of 100 parts of the API where Cuda is vastly simpler than Vulkan. Cuda also doesnt force you to use queue families but you can optionally use streams. No ridiculous descriptor management and binding in cuda, just passing pointers and handles via launch arguments. No overengineered explicit syncing mechanis in cuda, everything is nicely implicitly synced until you explicitly opt in to parallel streams. etc.
    • sollycb 40 minutes ago
      Ports are very often incredibly difficult and very time consuming.

      One of the biggest complaints we hear from the industry is "we tried to port to X and we could never complete it".

      An established codebase can have years of refinement. It will take time to achieve the same with the port.

      And with our compiler, just using cuda is no longer putting urself inside the moat :)

      • DiabloD3 19 minutes ago
        Ironically, this is what people claim AI can do with a snap of the fingers.

        Should be real simple if the HN AI echochamber is right, right?

    • swerner 1 hour ago
      Unfortunately, Vulkan Compute doesn’t to all the things that OpenCL, SYCL, HIP or CUDA do.
      • binsquare 1 hour ago
        Yep, there are inference stacks where it just does not work without cuda in any meaningful performance
        • DiabloD3 26 minutes ago
          Weird, since the most used open source inference engine is faster on Vulkan on platforms that offer multiple options, with the sole exception being Nvidia, due to poor Nvidia driver quality (which I am forced to assume is intentional, Nvidia wishes to maintain their moat after all).
    • pjmlp 32 minutes ago
      Vulkan tooling is light years behind what CUDA offers in 2026, across programming languages, IDE tooling, graphical debuggers and libraries.
    • dannecodez 1 hour ago
      [dead]
  • tangsoupgallery 14 minutes ago
    [dead]
  • z0ltan 1 hour ago
    [dead]