+- Inference (https://www.datasculpting.ai/inference)
+-- Forum: Optimization & Performance (https://www.datasculpting.ai/inference/forumdisplay.php?fid=10)
+--- Forum: Performance Tuning (https://www.datasculpting.ai/inference/forumdisplay.php?fid=42)
+--- Thread: consumer GPU serving economics: anyone testing this? (/showthread.php?tid=1408)
Pages:
1
2
consumer GPU serving economics: anyone testing this? - cudapeak - 10-17-2026
Wanted to compare notes with people doing similar things.
I'm mostly working with DeepSeek in LM Studio on 5090. The part I'm trying to understand is NVFP4.
Not looking for the perfect answer, mostly what worked in real use.
Re: consumer GPU serving economics: anyone testing this? - turboscope5 - 10-17-2026
Quote:Wanted to compare notes with people doing similar things.
that fixed it for me too.
Re: consumer GPU serving economics: anyone testing this? - cudapeak - 10-18-2026
A couple things I'd check first. If you're on MLX, I'd confirm the version first and then reduce variables until the behavior is obvious.
Re: consumer GPU serving economics: anyone testing this? - vramport - 10-18-2026
What worked for me was pretty simple. Before rebuilding anything, I'd test one clean baseline with Mixtral and compare from there. Once you know the plain path works, then add the faster path back in piece by piece.
Re: consumer GPU serving economics: anyone testing this? - signalcraft14 - 10-18-2026
Need a little more detail here. Are you testing this on 4090 or something else?
Re: consumer GPU serving economics: anyone testing this? - stackarch - 10-18-2026
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. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.
Re: consumer GPU serving economics: anyone testing this? - cudapeak - 10-18-2026
Thanks, that helps. I ended up in about the same place using LM Studio.
Re: consumer GPU serving economics: anyone testing this? - localpatch78 - 10-18-2026
My first guess would be this: If you're on vLLM, 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: consumer GPU serving economics: anyone testing this? - southden - 10-18-2026
Can you fill in a bit more context? What model, what runtime, and roughly what context size?
Re: consumer GPU serving economics: anyone testing this? - prompttinker44 - 10-18-2026
What exact setup are you running? Do you see the same thing with a smaller batch or shorter prompt?