3 ms·

It now supports AMD GPUs since 3 weeks ago, check the latest commit at the repo: https://github.com/vosen/ZLUDA https://github.com/vosen/ZLUDA The article als

It now supports AMD GPUs since 3 weeks ago, check the latest commit at the repo: https://github.com/vosen/ZLUDA https://github.com/vosen/ZLUDA The article also mentions exactly this fact.

KeplerBoy 3y ago
That's exactly what the post from 3 weeks ago discussed. AMD dropped ZLUDA funding, the remains were opensourced and it won't ever be a viable option.
- bri3d 3y ago
  > it won't ever be a viable option For production workloads, I generally agree. It's an unsupported hack with a questionable future, I wouldn't do anything money-making with it. However, for tinkering and consumer workloads, it already works pretty well. Enough of cuDNN and cuBLAS work to run PyTorch and in turn, Stable Diffusion with https://github.com/lshqqytiger/ZLUDA https://github.com/lshqqytiger/ZLUDA - there's even a fairly user-friendly setup process already in https://github.com/vladmandic/automatic https://github.com/vladmandic/automatic . I was able to get a personal non-ML related project working on my AMD card in just a few minutes, which saved me a lot of development time before I then deployed the production workload on NV hardware (this is probably why AMD pulled the plug on the project - it's almost more of a boost to NV than anything else, AMD really need people to be writing code on ROCm to deploy on AMD datacenter hardware).
  - HarHarVeryFunny 3y ago
    Given that PyTorch supports AMD cards with ROCm support, I'm curious why would you be coding in CUDA (very low level!) as opposed to PyTorch ?
    - bri3d 3y ago
      > a personal non-ML related project Not everything is DNNs and tensors... As for the Stable Diffusion thing, a silly edge case - because MIOpen (and therefore PyTorch-on-ROCm) doesn't work on Windows yet (they're slated to ship it next month, I think).
      - HarHarVeryFunny 3y ago
        > Not everything is DNNs and tensors... Sure, but you could use PyTorch for cuBLAS/cuRAND etc type functionality too.

It now supports AMD GPUs since 3 weeks ago, check the latest commit at the repo: https://github.com/vosen/ZLUDA https://github.com/vosen/ZLUDA The article also mentions exactly this fact.

KeplerBoy 3y ago

That's exactly what the post from 3 weeks ago discussed. AMD dropped ZLUDA funding, the remains were opensourced and it won't ever be a viable option.

bri3d 3y ago
> it won't ever be a viable option For production workloads, I generally agree. It's an unsupported hack with a questionable future, I wouldn't do anything money-making with it. However, for tinkering and consumer workloads, it already works pretty well. Enough of cuDNN and cuBLAS work to run PyTorch and in turn, Stable Diffusion with https://github.com/lshqqytiger/ZLUDA https://github.com/lshqqytiger/ZLUDA - there's even a fairly user-friendly setup process already in https://github.com/vladmandic/automatic https://github.com/vladmandic/automatic . I was able to get a personal non-ML related project working on my AMD card in just a few minutes, which saved me a lot of development time before I then deployed the production workload on NV hardware (this is probably why AMD pulled the plug on the project - it's almost more of a boost to NV than anything else, AMD really need people to be writing code on ROCm to deploy on AMD datacenter hardware).
- HarHarVeryFunny 3y ago
  Given that PyTorch supports AMD cards with ROCm support, I'm curious why would you be coding in CUDA (very low level!) as opposed to PyTorch ?
  - bri3d 3y ago
    > a personal non-ML related project Not everything is DNNs and tensors... As for the Stable Diffusion thing, a silly edge case - because MIOpen (and therefore PyTorch-on-ROCm) doesn't work on Windows yet (they're slated to ship it next month, I think).
    - HarHarVeryFunny 3y ago
      > Not everything is DNNs and tensors... Sure, but you could use PyTorch for cuBLAS/cuRAND etc type functionality too.