+- Inference (https://www.datasculpting.ai/inference)
+-- Forum: LLM Models (https://www.datasculpting.ai/inference/forumdisplay.php?fid=12)
+--- Forum: Open Weight (https://www.datasculpting.ai/inference/forumdisplay.php?fid=49)
+---- Forum: Local (https://www.datasculpting.ai/inference/forumdisplay.php?fid=51)
+---- Thread: simple local serving recipe on GB200 (/showthread.php?tid=1153)
Pages:
1
2
simple local serving recipe on GB200 - mlxgarden - 10-13-2026
Did a bit of testing and thought I'd post the result here.
I tried this with Llama using SGLang on Threadripper box. The headline result was better than I expected, but tokens/sec still ended up being the limiting factor.
Happy to post more details if anyone wants them.
Re: simple local serving recipe on GB200 - sparrowghost - 10-13-2026
What worked for me was pretty simple. On dual 3090, 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.
I'm happy to test one more variant if you want a second data point.
Re: simple local serving recipe on GB200 - rocmriley - 10-14-2026
Quote:What worked for me was pretty simple. On dual 3090, the easiest win is often to back off the aggressive settings and see whether the problem survives.
Ran a quick check on my box. My result was mixed: first run looked bad, second run looked normal. The surprise was that memory pressure showed up before throughput collapsed.
Re: simple local serving recipe on GB200 - batchshed - 10-15-2026
Quote:[quote]What worked for me was pretty simple. On dual 3090, the easiest win is often to back off the aggressive settings and see whether the problem su
My first guess would be this: Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at tokens/sec.
Re: simple local serving recipe on GB200 - modgarden - 10-15-2026
A couple things I'd check first. 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: simple local serving recipe on GB200 - batchscope72 - 10-15-2026
I'd start with the simpler stack.
Re: simple local serving recipe on GB200 - cachebench - 10-18-2026
What worked for me was pretty simple. Before rebuilding anything, I'd test one clean baseline with Phi and compare from there. Once you know the plain path works, then add the faster path back in piece by piece.
Re: simple local serving recipe on GB200 - audionode8 - 10-18-2026
Appreciate the detail. That matches what I saw on B200.
Re: simple local serving recipe on GB200 - batchsize - 10-19-2026
Maybe, but I had the opposite result. I've seen people blame KV cache pressure when the real issue was much more mundane. Worth testing the simpler explanation before adopting the more interesting one.
I'm happy to test one more variant if you want a second data point.
Re: simple local serving recipe on GB200 - deltagrid - 10-19-2026
A couple questions before I guess too hard. Is this local-only or are you routing anything externally?