+- 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: Recipes & Configurations (https://www.datasculpting.ai/inference/forumdisplay.php?fid=52)
+---- Thread: routing between local and hosted (/showthread.php?tid=1356)
Pages:
1
2
routing between local and hosted - fastbench - 10-16-2026
This keeps coming up in conversation so I figured I'd start a thread.
The more I work with local setups, the more it seems like structured output decides everything downstream.
Curious whether that lines up with what other people are seeing.
Re: routing between local and hosted - cudapoint - 10-16-2026
Did a small comparison this afternoon. My result was mixed: first run looked bad, second run looked normal. The surprise was that memory pressure showed up before throughput collapsed.
Re: routing between local and hosted - pixelspark - 10-17-2026
Tried to reproduce it just now. On Threadripper box with Ollama, I saw basically the same behavior. The surprise was that memory pressure showed up before throughput collapsed.
Re: routing between local and hosted - stackarch - 10-17-2026
Interesting, I hadn't tried that. I don't think your conclusion is far off. I got better results once I stopped chasing one metric and just looked at whether the workflow felt usable.
Re: routing between local and hosted - prompttinker44 - 10-18-2026
That matches what I was seeing. The pattern around reranking has been pretty similar for me. The biggest change for me was simplifying the stack instead of adding one more optimization layer.
Re: routing between local and hosted - vectorshed43 - 10-18-2026
Good point. I ended up in about the same place using llama.cpp. The biggest change for me was simplifying the stack instead of adding one more optimization layer.
Re: routing between local and hosted - kernelrunner - 10-18-2026
A couple questions before I guess too hard. Is this local-only or are you routing anything externally?
Re: routing between local and hosted - turboscope - 10-19-2026
Quote:Good point. I ended up in about the same place using llama.cpp. The biggest change for me was simplifying the stack instead of adding one more optimiz
following this because I'm in the same boat.
Re: routing between local and hosted - visionforge - 10-19-2026
I'd approach it in a boring order. On 7900 XTX, the easiest win is often to back off the aggressive settings and see whether the problem survives. Once you know the plain path works, then add the faster path back in piece by piece.
Re: routing between local and hosted - nanobrook - 10-19-2026
What worked for me was pretty simple. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at prefill/decode split. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.