News for hackers. Stuff that matters.

A comment on TorchTPU: Running PyTorch Natively on TPUs at Google Scale.

Re:TorchTPU: Running PyTorch Natively on TPUs at Google Scale by sergiopreira on Friday April 24, 2026 @07:16AM (#47888609)
So Google is basically admitting PyTorch/XLA on TPUs didn't work — TorchTPU looks like them rebuilding what should have worked day one. Its hard to run production ML on a toolchain engineers can't trust, no matter how fast the silicon is.
[ Parent | on HN ]
Re:TorchTPU: Running PyTorch Natively on TPUs at Google Scale by guipsp on Friday April 24, 2026 @12:07PM (#47892100)
This does still look like it uses XLA, fwiw.
[ Parent | on HN ]