Engineering and technology notes

Google Tensor processing units

Open main menu

Search Wikipedia
EditWatch this pageRead in another language
Tensor processing unit
Tensor processing units (or TPUs) are application-specific integrated circuits (ASIC) developed specifically for machine learning. Compared to graphics processing units (which as of 2016 are frequently used for the same tasks), they are designed explicitly for a higher volume of reduced precision computation with higher IOPS per watt (e.g. as little as 8-bit precision[1]), and lack hardware for rasterisation/texture mapping.[2] The chip has been specifically designed for Google’s TensorFlow framework, however Google still uses CPUs and GPUs for other machine learning.[3] Other AI accelerator designs are appearing from other vendors also and are aimed at embedded and robotics markets.

Google has stated that its proprietary tensor processing units were used in the AlphaGo versus Lee Sedol series of man-machine Go games.[2] Google has also used TPUs for Google Street View text processing, and was able to find all the text in the Street View database in less than five days. In Google Photos, an individual TPU can process over 100 million photos a day. It is also used in RankBrain which Google uses to provide search results.[4] The tensor processing unit was announced in 2016 at Google I/O, although the company stated that the TPU had been used inside their datacenter for over a year prior.[3][2]

The chip size can fit in a hard drive slot within a data center rack according to Google Distinguished Hardware Engineer Norm Jouppi.[3]

See also Edit

Vision processing unit a similar device specialised for vision processing.
TrueNorth a similar device simulating spiking neurons instead of low precision tensors.
Neural processing unit
References Edit

^ Armasu, Lucian (2016-05-19). “Google’s Big Chip Unveil For Machine Learning: Tensor Processing Unit With 10x Better Efficiency (Updated)”. Tom’s Hardware. Retrieved 2016-06-26.
^ a b c Jouppi, Norm (May 18, 2016). “Google supercharges machine learning tasks with TPU custom chip”. Google Cloud Platform Blog. Google. Retrieved 2017-01-22.
^ a b c “Google’s Tensor Processing Unit explained: this is what the future of computing looks like”. TechRadar. Retrieved 2017-01-19.
^ “Google’s Tensor Processing Unit could advance Moore’s Law 7 years into the future”. PCWorld. Retrieved 2017-01-19.

Last edited 12 days ago by PirateImpulse

Content is available under CC BY-SA 3.0 unless otherwise noted.
Terms of UsePrivacyDesktop

Mary Jackson (engineer)

Mary Winston Jackson (April 9, 1921 – February 11, 2005) was an African American mathematician and aerospace engineer at the National Advisory Committee for Aeronautics (NACA), which in 1958 was succeeded by theNational Aeronautics and Space Administration (NASA). She worked atLangley Research Center in Hampton, Virginia, for most of her career. She started as a computer at the segregated West Area Computing division. She took advanced engineering classes and in 1958 became NASA’s first black female engineer.

After 34 years at NASA, Jackson had earned the most senior engineering title available. She realized she could not earn further promotions without becoming a supervisor. She accepted a demotion to become a manager of both the Federal Women’s Program, in the NASA Office of Equal Opportunity Programs, and of the Affirmative Action Program. In this role, she worked to influence both the hiring and promotion of women in NASA’s science, engineering, and mathematics careers.

Jackson’s story features in the non-fiction book Hidden Figures: The Story of the African-American Women Who Helped Win the Space Race (2016). She is one of the three protagonists in Hidden Figures, the film adaptation released the same year.

https://en.m.wikipedia.org/wiki/Mary_Jackson_(engineer)

Dorothy Vaughan

Dorothy Johnson Vaughan (September 20, 1910 – November 10, 2008) was an African American mathematician who worked for the National Advisory Committee for Aeronautics (NACA), and NASA, at Langley Research Center inHampton, Virginia. In 1949, she became acting supervisor of the West Area Computers, the first African-American woman to supervise a staff at the center.

She later was promoted officially to this position. During her 28-year career, Vaughan prepared for the introduction of machine computers in the early 1960s by teaching herself and her staff the programming language of FORTRAN; she later headed the programming section of the Analysis and Computation Division (ACD) at Langley.

Vaughan is one of the women featured in Margot Lee Shetterly‘s history Hidden Figures: The Story of the African-American Women Who Helped Win the Space Race(2016). It was adapted as a biographical film of the same name, also released in 2016.

https://en.m.wikipedia.org/wiki/Dorothy_Vaughan

Katherine Johnson

Katherine Coleman Goble Johnson (born August 26, 1918) is an African American physicist and mathematician who made contributions to the United States’ aeronautics and space programs with the early application of digital electronic computers at NASA. Known for accuracy in computerized celestial navigation, she conducted technical work at NASA that spanned decades. During this time, she calculated the trajectories,launch windows, and emergency back-up return paths for many flights from Project Mercury, including the early NASA missions of John Glennand Alan Shepard, and the 1969 Apollo 11 flight to the Moon, through theSpace Shuttle program.[1][2] Her calculations were critical to the success of these missions.[1] Johnson also did calculations for plans for a mission to Mars.
https://en.m.wikipedia.org/wiki/Katherine_Johnson

Lua: getting started

Lua is a powerful and fast programming language that is easy to learn and use and to embed into your application. Lua is designed to be a lightweight embeddable scripting language and is used for all sorts of applications from games to web applications and image processing. See the about page for details and some reasons why you should choose Lua. See what Lua programs look and feel like in the live demo.

Source: Lua: getting started

App Inventor Tutorials and Examples: Manager Screen | Pura Vida Apps

The recommended method of switching screens in App InventorNote: Before starting to create another screen, first you should think about is it really necessary? See also Building apps with many screens and SteveJG’s post about advantages/disadvantages, because in only one screen you also can use vertical arrangements to simulate different screens, just set the arrangements to visible = true/false as needed…See also Martyn_HK’s example about how to use Tabs in App Inventor and another example from Cyd.Note from the Colored Dots tutorial:You can have many screens in an App Inventor app, but a screen always returns to the screen that opened it. On the other hand, you can get the effect of screens switching to arbitrary other screen by setting up a “manager screen” that is used for opening all the other screens. When a screen wants to switch, it returns to the manager with a value saying which screen to open next.This version demonstrates the use of a manager screen. If in a screen you want to go to another screen, you always go back to the manager screen telling him where you want to go. With this method, there only is one screen active in any moment. Additionally the back button will be catched with the Screen.BackPressed block.Note: if you only use the open another screen and never the close screen block, you will run out of memory after a while. Also reopening an already opened screen is a bad idea. For example if you open Screen1 twice, you also have to close your app twice!

Source: App Inventor Tutorials and Examples: Manager Screen | Pura Vida Apps

Creating new objects at runtime – Google Groups

Hi people!! I am loving the App Inventor, I have previously tried some droid dev using eclipse and the android SDK but I found it so hard to get to grips with. I am written a few apps and now am trying to write on that allows me to take pictures of my friends and assign their numbers to them… i want to be able to click my friends faces and call them from within the app but… is it possible to create new buttons at runtime? when i take a picture it should change the buttons image property to d

Source: Creating new objects at runtime – Google Groups