PCIe lane weirdness with inference loads

[turbogrid11]

Good point. That matches what I saw on A6000.

[labline]

What worked for me was pretty simple. Before rebuilding anything, I'd test one clean baseline with Mistral and compare from there. Once you know the plain path works, then add the faster path back in piece by piece.

[cobaltlab]

Thanks, that helps. That matches what I saw on 4090.

[northhall88]

I'd start with the simpler stack.