+- Inference (https://www.datasculpting.ai/inference)
+-- Forum: Hardware (https://www.datasculpting.ai/inference/forumdisplay.php?fid=4)
+--- Forum: CPUs / NPUs / Accelerators (https://www.datasculpting.ai/inference/forumdisplay.php?fid=30)
+--- Thread: CPU-only fallback performance for Gemma (/showthread.php?tid=129)
Pages:
1
2
Re: CPU-only fallback performance for Gemma - macrotinker - 09-07-2026
Good point. 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: CPU-only fallback performance for Gemma - vramscope - 09-07-2026
What exact setup are you running? Is this local-only or are you routing anything externally?
Re: CPU-only fallback performance for Gemma - quietsupply11 - 09-08-2026
Quote:What exact setup are you running? Is this local-only or are you routing anything externally?
I'd approach it in a boring order. If you're on Ollama, I'd confirm the version first and then reduce variables until the behavior is obvious. Not glamorous advice, but repeatable baselines save a lot of time.
Re: CPU-only fallback performance for Gemma - circuitkite - 09-08-2026
Good point. That matches what I saw on dual 3090.
Re: CPU-only fallback performance for Gemma - vramridge - 09-09-2026
A couple things I'd check first. Before rebuilding anything, I'd test one clean baseline with Qwen and compare from there.