+- Inference (https://www.datasculpting.ai/inference)
+-- Forum: Inference Engines (https://www.datasculpting.ai/inference/forumdisplay.php?fid=9)
+--- Forum: Ollama / LM Studio / Desktop Runtimes (https://www.datasculpting.ai/inference/forumdisplay.php?fid=38)
+--- Thread: Ollama for power users (/showthread.php?tid=626)
Ollama for power users - fusedhall - 09-30-2026
Spent some time on this today and wanted to compare notes.
I'm comparing two approaches here: keep it simple on B200, or squeeze more out of the stack and deal with the complexity.
What would you pick if you had to live with it for a few months?
This is mostly for local work rather than public serving.
Re: Ollama for power users - batchscope72 - 09-30-2026
Tried to reproduce it just now. The issue went away once I changed one setting and restarted the stack cleanly.
Re: Ollama for power users - oxidemixer - 09-30-2026
Thanks, that helps. That matches what I saw on A6000. I got better results once I stopped chasing one metric and just looked at whether the workflow felt usable.
Re: Ollama for power users - pixelspark - 09-30-2026
What exact setup are you running? Are you testing this on GB200 or something else?
Re: Ollama for power users - latentscope35 - 10-01-2026
A couple questions before I guess too hard. Do you see the same thing with a smaller batch or shorter prompt?
Re: Ollama for power users - orbitline - 10-02-2026
subscribed, curious where this lands.
Re: Ollama for power users - prompttinker44 - 10-04-2026
Tried to reproduce it just now. The issue went away once I changed one setting and restarted the stack cleanly. The main thing that moved was latency, not quality.
Re: Ollama for power users - amberhall - 10-04-2026
My first guess would be this: If you're on LM Studio, 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.
Re: Ollama for power users - edgeriver - 10-05-2026
Tried to reproduce it just now. On A6000 with SGLang, I saw basically the same behavior. I expected the model choice to matter more than it did.
Re: Ollama for power users - pixelspark - 10-05-2026
I tested this tonight. My result was mixed: first run looked bad, second run looked normal. I expected the model choice to matter more than it did.