+- Inference (https://www.datasculpting.ai/inference)
+-- Forum: Inference Engines (https://www.datasculpting.ai/inference/forumdisplay.php?fid=9)
+--- Forum: vLLM / SGLang (https://www.datasculpting.ai/inference/forumdisplay.php?fid=36)
+--- Thread: router in front of vLLM for Phi (/showthread.php?tid=1256)
Pages:
1
2
router in front of vLLM for Phi - cacheline - 10-16-2026
Curious what people here are seeing.
My current setup is LM Studio with B200, and I keep running into MoE serving.
Not looking for the perfect answer, mostly what worked in real use.
For context, I'm not chasing leaderboard numbers. I mostly want the system to feel consistent day to day.
Re: router in front of vLLM for Phi - rocmrunner40 - 10-16-2026
worth checking the driver first.
Re: router in front of vLLM for Phi - signalcraft14 - 10-16-2026
Did a small comparison this afternoon. I couldn't reproduce the slowdown until I pushed context much higher. I expected the model choice to matter more than it did.
Re: router in front of vLLM for Phi - audioforge - 10-16-2026
I'd start with the simpler stack.
Re: router in front of vLLM for Phi - batchsize - 10-16-2026
What worked for me was pretty simple. Before rebuilding anything, I'd test one clean baseline with Mistral and compare from there. Once you know the plain path works, then add the faster path back in piece by piece.
Re: router in front of vLLM for Phi - tensorfield88 - 10-16-2026
Quote:What worked for me was pretty simple. Before rebuilding anything, I'd test one clean baseline with Mistral and compare from there. Once you know the p
following this because I'm in the same boat.
Re: router in front of vLLM for Phi - deltapoint17 - 10-16-2026
that fixed it for me too.
Re: router in front of vLLM for Phi - dockridge - 10-16-2026
A couple things I'd check first. On GB200, the easiest win is often to back off the aggressive settings and see whether the problem survives.
Re: router in front of vLLM for Phi - audioforge - 10-17-2026
What exact setup are you running? Is this local-only or are you routing anything externally?
Re: router in front of vLLM for Phi - signalcraft14 - 10-17-2026
Good point. The pattern around quant drift 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.