运行与可靠性

Meta的Prometheus超算集群中数万台Nvidia GPU闲置,只待故障发生

Tens of thousands of Nvidia GPUs sit idle in Meta's Prometheus supercluster, just waiting for a failure

来源 The Register

报道

  1. Meta的Prometheus超算集群中数万台Nvidia GPU闲置,只待故障发生 Tens of thousands of Nvidia GPUs sit idle in Meta's Prometheus supercluster, just waiting for a failure The Register · · 首发