Google Cloud Platform Blog
Product updates, customer stories, and tips and tricks on Google Cloud Platform
Google supercharges machine learning tasks with TPU custom chip
Wednesday, May 18, 2016
Posted by Norm Jouppi, Distinguished Hardware Engineer, Google
Machine learning provides the underlying oomph to many of Google’s most-loved applications. In fact, more than 100 teams are currently using machine learning at Google today, from Street View, to Inbox Smart Reply, to voice search.
But one thing we know to be true at Google: great software shines brightest with great hardware underneath. That’s why we started a stealthy project at Google several years ago to see what we could accomplish with our own custom accelerators for machine learning applications.
The result is called a Tensor Processing Unit (TPU), a custom ASIC we built specifically for machine learning — and tailored for
TensorFlow
. We’ve been running TPUs inside our data centers for more than a year, and have found them to deliver an order of magnitude better-optimized performance per watt for machine learning. This is roughly equivalent to fast-forwarding technology about seven years into the future (three generations of Moore’s Law).
TPU is tailored to machine learning applications, allowing the chip to be more tolerant of reduced computational precision, which means it requires fewer transistors per operation. Because of this, we can squeeze more operations per second into the silicon, use more sophisticated and powerful machine learning models and apply these models more quickly, so users get more intelligent results more rapidly. A board with a TPU fits into a hard disk drive slot in our data center racks.
Tensor Processing Unit board
TPU is an example of how fast we turn research into practice — from first tested silicon, the team had them up and running applications at speed in our data centers within 22 days.
TPUs already power many applications at Google, including
RankBrain
, used to improve the relevancy of search results and
Street View
, to improve the accuracy and quality of our maps and navigation. AlphaGo was powered by TPUs in the
matches against Go world champion, Lee Sedol
, enabling it to "think" much faster and look farther ahead between moves.
Server racks with TPUs used in the AlphaGo matches with Lee Sedol
Our goal is to lead the industry on machine learning and make that innovation available to our customers. Building TPUs into our infrastructure stack will allow us to bring the power of Google to developers across software like
TensorFlow
and
Cloud Machine Learning
with advanced acceleration capabilities. Machine Learning is transforming how developers build intelligent applications that benefit customers and consumers, and we're excited to see the possibilities come to life.
Free Trial
GCP Blogs
Big Data & Machine Learning
Labels
Android
1
Announcement
15
Announcement Partners Technical Customers Compute Networking Storage Big Data & Analytics Developers Compute Engine Cloud Storage Cloud SQL Cloud BigTable
1
Announcements
1
api
2
app engine
50
Atmosphere Live
1
Big Data & Analytics
7
bigquery
15
BigTable
2
CDN
1
Cloud Console
2
Cloud Dataflow
5
Cloud Datastore
7
cloud endpoints
1
Cloud Pub/Sub
2
Cloud SDK
1
cloud sql
12
cloud storage
27
Cloudera
1
Compute
5
Compute Engine
56
container cluster
1
Container Engine
1
Container Registry
1
customer
59
Customers
4
Dataflow
4
DataLab
1
Dev Tools
1
developer tools
5
developer-insights
6
Developers
2
Developers Console
2
devfests
4
Disaster Recovery
1
Encryption Keys
1
ESG
1
Event
4
events
11
GA
1
Gaming
1
Go Client
1
Google App Engine
5
Google Apps
1
Google BigQuery
8
Google Cloud Deployment Manager
1
Google Cloud Networking
2
Google Cloud Platform
8
Google Cloud Storage
7
Google Compute Engine
9
Google Container Engine
1
gRPC
1
hadoop
3
Hardware
1
Helium
1
how to
2
IO2013
3
iOS
1
Kubernetes
15
Levyx
1
Local SSD
2
Logging
1
mapreduce
1
Media
3
Mobile
1
Nearline
1
networking
3
open source
98
PaaS Solution
1
Partner
12
Partners
2
Pricing
4
Products
15
Pub/Sub
2
Research
1
round-up
8
Server
1
Siggraph
1
solutions
4
Startup
1
Storage
2
Tableau
1
TCO
1
Technical
23
Windows
1
Wowza
1
Zync
3
Archive
2016
May
Apr
Mar
Feb
Jan
2015
Dec
Nov
Oct
Sep
Aug
Jul
Jun
May
Apr
Mar
Feb
Jan
2014
Dec
Nov
Oct
Sep
Aug
Jul
Jun
May
Apr
Mar
Feb
Jan
2013
Dec
Nov
Oct
Sep
Aug
Jul
Jun
May
Apr
Mar
Feb
Jan
2012
Dec
Nov
Oct
Sep
Aug
Jul
Jun
May
Apr
Mar
Feb
Jan
2011
Dec
Nov
Oct
Sep
Aug
Jul
Jun
May
Apr
Mar
Feb
Jan
2010
Dec
Oct
Sep
Aug
Jul
Jun
May
Apr
Mar
Feb
Jan
2009
Dec
Nov
Oct
Sep
Aug
Jul
Jun
May
Apr
Mar
Feb
Jan
2008
Dec
Nov
Oct
Sep
Aug
Jul
Jun
May
Apr
Feed
Technical questions? Check us out on
Stack Overflow
.
Subscribe to
our monthly newsletter
.
Follow @googlecloud