Hacker News Reader

Comment by DeathArrow | Hacker News Reader

Do they have a good CUDA alternative?

Can't speak to how it compares to CUDA however they are developing ROCm[0]

[0]: https://www.amd.com/en/products/software/rocm.html

Next to nothing is written that targets this. Not Stable Diffusion, not RVC, not Vall-E, not Tacotron, not Tortoise, ...

Maybe the LLM space is better about this, but the generative media side definitely isn't.

AMD has a market share of 0% here, and nobody publishes models with AMD support.

Could there be a compiler/transpiler from CUSA to whatever AMD is pushing ?

That's what https://github.com/ROCm/HIPIFY is (as a part of ROCm)

I think it's llama.cpp that simply #defines all cuda_ functions to rocm_ (99% name-name). Porting seems to be that trivial.

The things is you can actually run Stable Diffusion.

And I got PyTorch working on my AMD 7900 XT graphics card recently, though it was a bit of a hassle to do so.

You can also run Stable Diffusion in cpu mode, if you don't mind it being slower. I have an NVIDIA card but it's not powerful enough to run it. I'm on Ubuntu.

> though it was a bit of a hassle to do so.

Incredible understatement. And the diverse set of community tools also breaks down.

We're still a year or more out from proper AMD support in the ecosystem.

I can attest that it works really well on my 6900xt. Compiling CUDA kernels is merely a matter of a using a #define shim. Also, provided you download the ROCm pytorch (and force compatibility with the HSA_OVERRIDE_GFX_VERSION) everything just works.

it's called hip, and it's mostly the same

AMD have their own thrust gpu impls, so from a high level they are somewhat interchangeable

I don't know, but couldn't people use LLMs to drastically lower the cost of switching? Converting a codebase to use a different platform doesn't require creativity.

I think we shouldnt even try, because LLMs will simply design new hardware and software all by its own. All we need to do is sit and watch, while collecting UBI.

Reply on news.ycombinator.com