+- 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: prefix caching on repeated workflows on GB200 (/showthread.php?tid=490)
Pages:
1
2
prefix caching on repeated workflows on GB200 - cacheline - 09-25-2026
Spent some time on this today and wanted to compare notes.
I've gone back and forth between LM Studio and TensorRT-LLM for Mistral. The interesting difference for me has been prefix caching.
Trying to avoid rebuilding the whole thing twice.
Re: prefix caching on repeated workflows on GB200 - cacheline - 09-25-2026
Quote:Spent some time on this today and wanted to compare notes.
Can you fill in a bit more context? Is this local-only or are you routing anything externally?
Re: prefix caching on repeated workflows on GB200 - quietarch - 09-25-2026
Did a small comparison this afternoon. My result was mixed: first run looked bad, second run looked normal.
Re: prefix caching on repeated workflows on GB200 - rocmriley - 09-27-2026
that fixed it for me too.
Re: prefix caching on repeated workflows on GB200 - cacheline - 09-27-2026
Appreciate the detail. The pattern around MoE serving has been pretty similar for me. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.
Re: prefix caching on repeated workflows on GB200 - rocmcrate26 - 09-27-2026
Quote:Appreciate the detail. The pattern around MoE serving has been pretty similar for me. Worth rechecking your defaults though, because one bad assumptio
Need a little more detail here. Is this local-only or are you routing anything externally?
Re: prefix caching on repeated workflows on GB200 - cacheline - 09-30-2026
following this because I'm in the same boat.
Re: prefix caching on repeated workflows on GB200 - modrunner27 - 09-30-2026
I'd approach it in a boring order. On M3 Max, 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.
Re: prefix caching on repeated workflows on GB200 - cacheline - 09-30-2026
Interesting, I hadn't tried that. I don't think your conclusion is far off. The biggest change for me was simplifying the stack instead of adding one more optimization layer.
Re: prefix caching on repeated workflows on GB200 - vrambandit - 10-03-2026
That matches what I was seeing. The pattern around consumer GPUs 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. I keep promising myself this is the last rebuild.