News for hackers. Stuff that matters.
Re:AWS acquires Talen's 960MW nuclear data center campus in ... by zrm on Tuesday March 05, 2024 @04:52AM (#39601330)
> This is getting worse with ML clusters.

Is it? What causes there to be unused capacity when there is currently so much demand for training?

Re:AWS acquires Talen's 960MW nuclear data center campus in ... by deelowe on Tuesday March 05, 2024 @03:45PM (#39608957)
Training workloads are highly variable and mostly consistent across the cluster (all systems ramp up and down at the same time).