comparing models without fooling yourself

[kernelarch]

Maybe I'm missing something obvious.

I'm trying to decide whether to keep the current stack or change course. Main question is how people here are handling HBM constraints.

Not looking for the perfect answer, mostly what worked in real use.

[quiethouse]

Not sure I agree there. I get why that seems right, but on dual 3090 the bottleneck felt elsewhere. Worth testing the simpler explanation before adopting the more interesting one.

[batchpeak]

Thanks, that helps. That matches what I saw on B200. I got better results once I stopped chasing one metric and just looked at whether the workflow felt usable.