+- Inference (https://www.datasculpting.ai/inference)
+-- Forum: Drivers / Frameworks / Toolkits (https://www.datasculpting.ai/inference/forumdisplay.php?fid=8)
+--- Forum: AMD / ROCm (https://www.datasculpting.ai/inference/forumdisplay.php?fid=32)
+--- Thread: 7900 XTX local stack (/showthread.php?tid=140)
Pages:
1
2
7900 XTX local stack - relaynest82 - 09-05-2026
Thought I'd post a small recipe because a lot of threads ask similar things.
This isn't revolutionary, but the combination of LM Studio and dual 3090 felt surprisingly usable once I tuned around tool calling.
If someone wants the exact settings I can clean them up and post them.
Re: 7900 XTX local stack - phasetrack - 09-07-2026
Quote:Thought I'd post a small recipe because a lot of threads ask similar things.
I'd approach it in a boring order. Before rebuilding anything, I'd test one clean baseline with FLUX and compare from there. Once you know the plain path works, then add the faster path back in piece by piece.
Re: 7900 XTX local stack - mintquill - 09-07-2026
That matches what I was seeing. That matches what I saw on M3 Max. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.
Re: 7900 XTX local stack - thetaarch43 - 09-07-2026
I'd approach it in a boring order. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at NVFP4.
Re: 7900 XTX local stack - dockworks23 - 09-08-2026
Ran a quick check on my box. On M3 Max with llama.cpp, I saw basically the same behavior. The main thing that moved was latency, not quality.
Re: 7900 XTX local stack - quantshed - 09-10-2026
Did a small comparison this afternoon. On 7900 XTX with LM Studio, I saw basically the same behavior.
Re: 7900 XTX local stack - thetaember38 - 09-10-2026
Quote:Did a small comparison this afternoon. On 7900 XTX with LM Studio, I saw basically the same behavior.
What worked for me was pretty simple. If you're on PyTorch, I'd confirm the version first and then reduce variables until the behavior is obvious.
Re: 7900 XTX local stack - routecraft - 09-13-2026
Quote:Quote:Did a small comparison this afternoon. On 7900 XTX with LM Studio, I saw basically the same behavior.
A couple things I'd check first. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at KV cache pressure. Once you know the plain path works, then add the faster path back in piece by piece.
Re: 7900 XTX local stack - macropoint - 09-13-2026
Not sure I agree there. I've seen people blame tokens/sec when the real issue was much more mundane.
Re: 7900 XTX local stack - mintquill - 09-13-2026
Can you fill in a bit more context? Are you testing this on 5090 or something else?