[SlavikCA]

2x CPU system may be slower for LLM than 1x CPU system.

Because in 2x CPU system, the model may have to be passed via NUMA, which has 10% - 30% of memory bandwidth bandwidth

[ryao]

It would be interesting to see the performance if someone built a single socket version. I have a parts list here if anyone wants to try it:

https://news.ycombinator.com/item?id=42868360