[lostmsu]

You are talking about a specific setup:

> Here we are going to utilize an 8 node cluster (64 H100 GPUs)