YHK - yet another hacker news clone

Google open-sources TorchTPU, a native PyTorch back end for TPUs

54 points ilreb 1 hour ago 18 comments

Lucasoato 1 hour ago | parent

Are there TPUs outside Google Cloud?

sschueller 1 hour ago | parent

Google coral USB and m.2 devices are TPUs.

boulos 54 minutes ago | parent

Only in marketing name. They're not related otherwise.

westurner 1 hour ago | parent

TPU / neural compute accelerator / Math coprocessor;

"OpenTPU: Open-Source Reimplementation of Google Tensor Processing Unit (TPU)" https://news.ycombinator.com/item?id=44111452

From https://news.ycombinator.com/item?id=45848676 :

> How does Cerebras WSE-3 with 44GB of 'L2' on-chip SRAM compare to Google's TPUs, Tesla's TPUs, NorthPole, Groq LPU, Tenstorrent's, and AMD's NPU designs?

There are TPUs made out of silicon carbide with standard fab.

There is in-flight tensor processing with photons, but that's not quite a TPU.

hackernudes 3 minutes ago | parent

I think the commenter was asking if Google sells TPU chip hardware to use outside of Google's own infrastructure and cloud.

boulos 53 minutes ago | parent

Yes. Google is famously selling TPUs on prem to big customers (e.g., Anthropic) and Crux AI is now a venture to do just that.

goda90 6 minutes ago | parent

Maybe just marketing, but they call the NPUs in their Pixel phones "TPUs"

npalli 1 hour ago | parent

Interesting, Amazon doing the same for AWS Trainium.

https://github.com/aws-neuron/torch-neuronx

yipinwong 54 minutes ago | parent

Dumb question from non-AI engineer.

Is this somehow Google's answer to today's best DeepSeek open-sourcing inferencing engine?

bitpush 47 minutes ago | parent

I dont think big companies like Google react to news like that. Very likely this was meant to go out on Tuesday for weeks now. And it just happened to have coincided with another news.

Tanjreeve 42 minutes ago | parent

They do different things and sit in different parts of the stack (training vs inference). TorchTPU seems more like a play to get more people familiar with Googles TPU chips as they want to sell them to people outside of google.

mathisfun123 15 minutes ago | parent

lol no this is a replacement for https://github.com/pytorch/xla

JacobAsmuth 3 minutes ago | parent

Yeah they read the news this morning and whipped up a quick Torch->XLA translation layer and coordinated with four different inference engines (who also all wrote their own blog posts) and shoved this out the door quick to stay on top of the news cycle

23agt 34 minutes ago | parent

Google is selling TPU shovels via Cloud services and also physically to select partners. So this makes sense.

There will be an abundance of shovels. Nebius already crashed by almost half from its hype peak.

airza 28 minutes ago | parent

it doesn't seem like there's an abundance of shovels coming anytime soon

nelaggy 11 minutes ago | parent

so if the moat isn't the model and the moat isn't the inference layer, i guess the moat is the hardware now? what does this mean for frontier labs?

JacobAsmuth 4 minutes ago | parent

Products can also be a moat.