Name: 雲計算TPU（TensorFlow @ O'Reilly AI大會，18年舊金山）。 (Cloud TPUs (TensorFlow @ O’Reilly AI Conference, San Francisco '18))
Uploaded: 2021-01-14T10:05:14.000Z
Duration: 22 min 9 s
Description: 【看影片學英語】數萬部 YouTube 影片，搭配英漢字典即點即查，輕鬆掌握單字發音與用法，長久累積看電影不必再看字幕。

And I work on the Google Brain team working on TensorFlow.

And today for the first part of this talk,

I'm going to talk to you about accelerating machine learning

So the motivation question here is, why is Google

but if you look at the data, this has been--

for the past 10 or 15 years, where we don't really see

the 52% year-on-year growth in single-threaded performance

that we saw from the late 1980s through the early 2000s

anymore, where now single-threaded performance

for CPUs is really growing at a rate of about maybe 3% or 5%

wait 18 months for my machine learning models

You have people uploading hundreds and hundreds

of hours of video every minute to YouTube.

People are leaving product reviews on Amazon.

People are using chat systems, such as WhatsApp.

People are talking about personal assistance

So more data is generated than ever before.

equipped to make sense of them to use them properly.

And the third thread is that at the same time,

we have this sort of exponential increase

in the amount of compute needed by these machine learning

This is a very interesting blog post by OpenAI.

where deep learning was first becoming useful.

Dropout, which used a fair amount of computing power,

but not that much compared to in late 2017 where

DeepMind published the AlphaGo Zero and AlphaGo.

In the Alpha Zero paper, we see in about six, seven years,

we see the compute demand increase by 300,000 times.

The end of Moore's law plus this sort of exponential increase

in computer requirements means that we need a new approach

At the same time, of course, everyone still

wants to do compute, do machine learning,

So that's why Google is building specialized hardware.

Now, the second question you might be asking

is, what sort of accelerators is Google building?

know that Google is building a type of accelerator

that we call Tensor Processing Units, which are really

specialized ASICs designed for machine learning.

now called Cloud TPU version 2 that we introduced

can be combined into pods called Cloud TPU v2 Pods.

we introduced the third generation of cloud TPUs.

And of course, you can link a bunch of them

So what are the differences between these generations

So the first version of TPUs, it was really

The second generation of TPUs does both training

And again, it does training and inference.

And of course, we see the same sort of progress

And our new generation of pods does over 100 petaflops

And of course, the new generation of pods

字幕列表影片播放

雲計算TPU（TensorFlow @ O'Reilly AI大會，18年舊金山）。 (Cloud TPUs (TensorFlow @ O’Reilly AI Conference, San Francisco '18))

sort

strategy

scale

cortex