Hacker News Reader

Comment by curt15 | Hacker News Reader

Pretty much any modern NVIDIA GPU supports CUDA. You don't have to buy a datacenter-class unit to get your feet wet with CUDA programming. ROCm will count as "something" when the same is true for AMD GPUs.

View on HN

ROCm supports current gen consumer gpus officially and a decent chunk of recent gen consumer gpus unofficially. Not all of them of course but a decent chunk.

It's not ideal but I'm pretty sure CUDA didn't support everything from day 1. And ROCm is part of AMD's vendor part of the Windows AI stack so from upcoming gen on out basically anything that outputs video should support ROCm.

No, but CUDA at least supported the 8800 gt on release [1]. ROCm didn't support any consumer cards on release, looks like they didn't support any till last year? [2]

[1]https://www.gamesindustry.biz/nvidia-unveils-cuda-the-gpu-co...

[2]https://www.tomshardware.com/news/amd-rocm-comes-to-windows-...

I don't think AMD needs to support 5+ year old GPUs personally. And all the recent generations are already practically supported.

AMD only claims support for a select few GPUs, but in my testing I find all the GPUs work fine if the architecture is supported. I've tested rx6600, rx6700xt for example and even though they aren't officially supported, they work fine on ROCm.

> 5+ year old GPUs

AMD had a big architecture switchover exactly 5 years ago, and the full launch wasn't over until 4.5 years ago. I think that generation should have full support. Especially because it's not like they're cutting support now. They didn't support it at launch, and they didn't support it after 1, 2, 3, 4 years either.

The other way to look at things, I'd say that for a mid to high tier GPU to be obsolete based on performance, the replacement model needs to be over twice as fast. 7700XT is just over 50% faster than 5700XT.

I'm on a 5+ year old GPU, because I don't trust AMD to offer a compelling GPU that actually works. An RX 7 570 is good enough for the little gaming I do. It mostly acts as an oversized iGPU that has good Linux drivers, but since AMD is not supporting ROCm on this GPU, there is no need to hurry on upgrading to a better GPU or to get my feet wet on running things locally on the GPU like Stable Diffusion, LLMs, etc.

Here is the support list:

https://rocm.docs.amd.com/projects/install-on-linux/en/lates...

AMD's definition of "support" I think is different than what people expect, and pretty misleading - ROCm itself will run on almost anything, back as far as the RX 400/500 series:

https://en.wikipedia.org/wiki/ROCm#:~:text=GCN%205%20%2D%20V...

Stable Diffusion ran fine for me on RX 570 and RX 6600XT with nothing but distro packages.

I don't buy it. Even running things like llama.cpp on my RX 570 via Vulkan crashes the entire system.

The last time I checked, I was stuck with a pretty old kernel if I wanted to have the last version of ROCm available for my rx470. It's compatible at some point in time, but not kept compatible with recent kernels.

It's the responsibility of your distro to ship things that work together,

There are out-of-bounds writes in the BLAS libraries for gfx803 GPUs (such as the RX 570). That hardware might work fine for your use case, but there's a lot of failures in the test suites.

I agree that the official support list is very conservative, but I wouldn't recommend pre-Vega GPUs for use with ROCm. Stick to gfx900 and newer, if you can.

AMD should focus their efforts on competitive hardware offerings, because that is where the need and the money is. Sorry, I don't think the hobbyist should be a priority.

Huh? I've been running ROCm for SD and LLMs for over a year and a half on my puny consumer 6750X - not even latest gen.

Reply on news.ycombinator.com