Inference
PCIe lane weirdness with inference loads - Printable Version

+- Inference (https://www.datasculpting.ai/inference)
+-- Forum: Hardware (https://www.datasculpting.ai/inference/forumdisplay.php?fid=4)
+--- Forum: GPUs (https://www.datasculpting.ai/inference/forumdisplay.php?fid=28)
+--- Thread: PCIe lane weirdness with inference loads (/showthread.php?tid=1052)

Pages: 1 2


Re: PCIe lane weirdness with inference loads - amberhall - 10-17-2026

Tried to reproduce it just now. My result was mixed: first run looked bad, second run looked normal. The surprise was that memory pressure showed up before throughput collapsed.


Re: PCIe lane weirdness with inference loads - kernelmike - 10-17-2026

Quote:Tried to reproduce it just now. My result was mixed: first run looked bad, second run looked normal. The surprise was that memory pressure showed up b

Need a little more detail here. Do you see the same thing with a smaller batch or shorter prompt?