8B vs 14B sweet spot

[deepseeked]

Posting this in case it saves someone else a little time.

I'm not trying to build a perfect architecture diagram. I just want a sane starting point around llama.cpp and GB200.

Simple and boring is a valid answer. At this point the real bottleneck is my patience.

[circuitkite]

I'd approach it in a boring order. If you're on PyTorch, I'd confirm the version first and then reduce variables until the behavior is obvious.

[localtinker]

I'd approach it in a boring order. Before rebuilding anything, I'd test one clean baseline with Phi and compare from there. Not glamorous advice, but repeatable baselines save a lot of time.

[rasterbench]

My first guess would be this: Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at speculative decoding.

[sanescope]

My first guess would be this: Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at speculative decoding.

same here on my side.

[benchplease]

Good point. That matches what I saw on dual 3090.

[kernelghost]

Good point. That matches what I saw on dual 3090.

Interesting, I hadn't tried that. That matches what I saw on 4090. The biggest change for me was simplifying the stack instead of adding one more optimization layer. Turns out 'just one more test' was a lie.

[batchwire]

My first guess would be this: Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at batch size.

[thetagrid15]

Can you fill in a bit more context? Are you testing this on mini PC or something else?

[fusedhall56]

worth checking the driver first.