← Back to context

Comment by geremiiah

6 hours ago

TPUs are systolic arrays right? So does that mean that Google is using a hetreogenous cluster compromising both GPUs and TPUs, for workloads that don't map well or at all on TPUs?

I can't speak to what every team at Google does, but there are machines with Nvidia GPUs in Borg. However Google charges orgs internally for cpu/memory/gpu/tpu usage and TPUs are *way* more efficient in terms of FLOPS/$ than Nvidia GPUs, so there is a *huge* incentive for teams to use TPUs if they can, especially for teams operating large products.