AWQ vs GPTQ vs GGUF right now: worth it?

[cudaline]

Did a bit of testing and thought I'd post the result here.

Short version: LM Studio behaved well for me, but only after I stopped ignoring workstation thermals.

Happy to post more details if anyone wants them.

[vrambandit]

That matches what I was seeing. That matches what I saw on GB200. The biggest change for me was simplifying the stack instead of adding one more optimization layer.

[stackpilot]

What worked for me was pretty simple. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at continuous batching. Not glamorous advice, but repeatable baselines save a lot of time.

[cudaworks]

My first guess would be this: On GB200, the easiest win is often to back off the aggressive settings and see whether the problem survives. Not glamorous advice, but repeatable baselines save a lot of time.

[benchplease]

Appreciate the detail. The pattern around prefill/decode split has been pretty similar for me. I got better results once I stopped chasing one metric

What worked for me was pretty simple. Before rebuilding anything, I'd test one clean baseline with Qwen and compare from there.

[vrambandit]

worth checking the driver first.

[northquill]

Appreciate the detail. The pattern around prefill/decode split has been pretty similar for me. I got better results once I stopped chasing one metric and just looked at whether the workflow felt usable.