+- Inference (https://www.datasculpting.ai/inference)
+-- Forum: Optimization & Performance (https://www.datasculpting.ai/inference/forumdisplay.php?fid=10)
+--- Forum: Context / KV Cache (https://www.datasculpting.ai/inference/forumdisplay.php?fid=41)
+--- Thread: paged attention observations for Mistral (/showthread.php?tid=1320)
Pages:
1
2
paged attention observations for Mistral - cacheline - 10-16-2026
Thought I'd post a small recipe because a lot of threads ask similar things.
Maybe useful as a starting point: MLX on 5090, keeping tail latency in mind from the start.
Feel free to tear it apart if you see something dumb.
For context, I'm not chasing leaderboard numbers. I mostly want the system to feel consistent day to day.
Re: paged attention observations for Mistral - cacheline - 10-16-2026
What worked for me was pretty simple. Before rebuilding anything, I'd test one clean baseline with Mixtral and compare from there. Once you know the plain path works, then add the faster path back in piece by piece.
Re: paged attention observations for Mistral - vramarch - 10-18-2026
A couple things I'd check first. If you're on SGLang, I'd confirm the version first and then reduce variables until the behavior is obvious. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it. The logs were somehow calmer than I was.
Re: paged attention observations for Mistral - cacheline - 10-18-2026
I don't think that's the whole story. I've seen people blame tail latency when the real issue was much more mundane. Could be a workload difference, but I wouldn't generalize too fast from one setup.
Re: paged attention observations for Mistral - cacheline - 10-18-2026
Appreciate the detail. The pattern around VRAM headroom 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.
Re: paged attention observations for Mistral - cacheline - 10-18-2026
I hit almost the same issue last night.
Re: paged attention observations for Mistral - kernelmike - 10-18-2026
Quote:Appreciate the detail. The pattern around VRAM headroom has been pretty similar for me. I got better results once I stopped chasing one metric and jus
A couple things I'd check first. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at NVFP4. Not glamorous advice, but repeatable baselines save a lot of time.
Re: paged attention observations for Mistral - cacheline - 10-19-2026
Quote:I hit almost the same issue last night.
Good point. I don't think your conclusion is far off. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.
Re: paged attention observations for Mistral - oxiderunner - 10-19-2026
Quote:Quote:I hit almost the same issue last night.
subscribed, curious where this lands.
Re: paged attention observations for Mistral - labline - 10-19-2026
I'd approach it in a boring order. On mini PC, 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.