+- Inference (https://www.datasculpting.ai/inference)
+-- Forum: Drivers / Frameworks / Toolkits (https://www.datasculpting.ai/inference/forumdisplay.php?fid=8)
+--- Forum: NVIDIA / CUDA (https://www.datasculpting.ai/inference/forumdisplay.php?fid=31)
+--- Thread: CUDA update broke my throughput with Ollama (/showthread.php?tid=438)
Pages:
1
2
CUDA update broke my throughput with Ollama - kernelmike - 09-23-2026
Did a bit of testing and thought I'd post the result here.
I tried this with DeepSeek using Ollama on 4090. The headline result was better than I expected, but INT4 still ended up being the limiting factor.
Happy to post more details if anyone wants them.
This is mostly for local work rather than public serving.
Re: CUDA update broke my throughput with Ollama - rasterhall - 09-23-2026
Maybe, but I had the opposite result. I wouldn't optimize around that first. On my side the bigger swing came from the runtime and not the model. Not saying you're wrong, just that I wouldn't lock that in as the lesson yet.
Re: CUDA update broke my throughput with Ollama - northgarden37 - 09-23-2026
A couple questions before I guess too hard. Is this local-only or are you routing anything externally?
Re: CUDA update broke my throughput with Ollama - dockember - 09-24-2026
worth checking the driver first.
Re: CUDA update broke my throughput with Ollama - sparrowkite53 - 09-24-2026
My first guess would be this: If you're on MLX, I'd confirm the version first and then reduce variables until the behavior is obvious.
Re: CUDA update broke my throughput with Ollama - edgecoast - 09-24-2026
Interesting, I hadn't tried that. The pattern around consumer GPUs has been pretty similar for me. The biggest change for me was simplifying the stack instead of adding one more optimization layer.
Re: CUDA update broke my throughput with Ollama - thetapath - 09-24-2026
I hit almost the same issue last night.
Re: CUDA update broke my throughput with Ollama - kernelmike - 09-24-2026
Not sure I agree there. I get why that seems right, but on M3 Max the bottleneck felt elsewhere. Could be a workload difference, but I wouldn't generalize too fast from one setup.
Re: CUDA update broke my throughput with Ollama - audiopath66 - 09-24-2026
Ran a quick check on my box. The issue went away once I changed one setting and restarted the stack cleanly. The surprise was that memory pressure showed up before throughput collapsed.
Re: CUDA update broke my throughput with Ollama - promptspark58 - 09-27-2026
A couple things I'd check first. 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. I keep promising myself this is the last rebuild.