How to pronounce“tpu”in English

Performance out of my TPU or my iPhone accelerator or whatever right all these different things and how and so this is a layered problem with a lot of really

stuff with the TPU sort of there's a closer to the hardware that could be optimizations around the IROC closer to the instruction set that could be

But Google also has TPUs, tensor processing units.

Like Google selling TPUs or something like this?

you're talking to tpus you need you know ml or xlaa or one of these compilers that talks to tpus to build on top of okay and if you're talking to circuits

or "The tile size of a TPU is 128," like, these factoids, right?

batteries but the thing about the tesla tpu um is anybody can build that it's just a question of you know

and so the you know a TPU with a matrix multiply unit and a systolic array is much more efficient than having a scaler core doing multiplies and adds and

Google loves to rent you TPUs.

- And that was like when TPUs were waking up.

specialized for driving data it's just legit just tpu they want to inter go the apple way basically everything required in the chain is done

I don't know if it ran them GPUs or TPUs.

This is one of the things that powers Google TPUs.

And I was attracted to Google to help them with the TPUs, right?

you know uh you have maybe google working on tpus and so on you have all these other companies with asics

many projects but really challenging very varied compilers tpus Hardware

- So in infrastructure, you think TPU is giving an advantage?

- It's not a handful of TPU alternatives.

of the tensor system on a chip are the cpus the gpu isp tpu contacts hub and

and model using tensor flow and have it run on 1,24 nodes of a TPU yeah that's

today you can you can write a check and get access to 1,000 TPU cores and do really interesting large scale training and inference and things like

He's founder and editor-in-chief of the British comic book publisher TPub.

At Google, I worked on TPUs.

chris lattner for a second time and he was a key person about tpus and there's a lot of fun stuff going on in

shallow looking at it it seems like something like google did with tpus but specialized for uh driving data i don't think it's

of a specific thing, think like a GPU or TPU, a very specific type of device?

For example, tensor processing unit, TPU, is powered by Google.

No, the only ASIC that is remotely successful is Google's TPU.

optimization the compiled design stuff the llvm uh then there's he's also a google work that the tpu stuff

i'm interested what's what he's going to do next in the same kind of way yes so on either the tpu or maybe the

techniques and then Google buys them and throws a boatload of uh gpus at them and invents tpus for them

Like, you know, you look at Google with TPUs, massive, huge exif flops of compute strapped together into machines that researchers

this guy is ridiculous, is when I was at Google and we were bringing up TPUs and we had a whole team of people

like gpu wise is that something you can get into you know i know that google works with tpus and so on i mean for machine

now to train your options are nvidia or google so google is not even an option their tpus are only available in google

created llvm compiler infrastructure project the clang compiler the Swift programming language a lot of key contributions to tensor flown tpus as

Well, there are several that have-- of course, clearly the machine learning TPU stuff is pretty dramatic.

- So for auto-tuning, for like tile size, you mentioned 120f for TPU.

That uses, you know, these days, you know, tens of thousands, sometimes many tens of thousands of GPUs or TPUs

More recently, we've actually now started to see that we can replicate this on CPUs and then potentially even on TPUs and GPUs.

At Google, some of you will know that we've invested in a variety of different computational platforms-- CPUs, GPUs, now TPUs for the tensor processing

the Clang compiler, the Swift programming language, a lot of key contributions to TensorFlow and TPUs as part of Google.

And it's not just about CPUs, or GPUs, or TPUs, or NPUs, or IPUs, or whatever, all the PUs, right?

So I've worked on TensorFlow and TPUs and things like that.

In a data center, you now have multiple different machines, sometimes very specialized, sometimes with GPUs or TPUs in one node and sometimes with disks in other nodes.

- And Jeremy and one of his researchers crushed Google by not through sheer force of the amazing amount of compute and the number of TPUs and stuff like that,

doing anything that competes with google without google's prior written permission well okay i mean google is not a platform company uh i wouldn't i wouldn't touch tpus with