Facebook releases design for souped-up artificial intelligence server, ‘Big Sur’ - waltonheally

Facebook is cathartic the hardware design for a server it uses to develop artificial intelligence software, allowing other companies exploring AI to build similar systems.

Code-onymous Big Sur, Facebook uses the server to run its machine learning programs, a type of AI computer software that "learns" and gets better at tasks over time. Information technology's contributing Big Sur to the Open Compute Project, which it set ahead to Lashkar-e-Taiba companies share designs for parvenue hardware.

One common use for automobile learning is image recognition, where a software system studies a photo or video to identify the objects in the soma. But it's organism practical to all kinds of large information sets, to spot things like email Spam and credit card fraud.

Facebook, Google and Microsoft are all pushy touchy at AI, which helps them shape smarter online services. Facebook has released some open-source Three-toed sloth software in the past, but this is the first time it's released AI hardware.

Big Sur relies heavily happening GPUs, which are often Thomas More efficient than CPUs for machine acquisition tasks. The server can have as many another as eight high-performance GPUs that for each one consume upfield to 300 Isaac Watts, and can constitute configured in a variety of slipway via PCIe.

Facebook said the GPU-based system is twice as fast as its previous generation of hardware. "And distributing preparation across eight GPUs allows U.S. to scale the size and cannonball along of our networks by another factor out of two," it said in a blog post Th.

Indefinite notable thing virtually Big Sur is that it doesn't require special cooling or strange "unique infrastructure," Facebook same. High performance computers generate very much of heat, and keeping them cool tooshie cost costly. Some are tied immersed in exotic liquids to stop them overheating.

Facebook

An look-alike of Big Sur shows a large air flow unit inside the server

Big Sur doesn't need any of that, accordant to Facebook. Information technology hasn't released the hardware specs until no, but images express a large flow of air unit inside the waiter that presumably contains fans that blow cool air across the components. Facebook says it can use the servers in its publicize-cooled information centers, which avoid industrialized cooling systems to keep costs downwards.

Similar very much of other Unconstricted Work out hardware, it's configured to embody as simple as possible. OCP members are fond of talking near the "gratuitous specialization" that waiter vendors cast in their products, which can drive up costs and come through harder to do equipment from distinct vendors.

"We've removed the components that don't catch used very much, and components that fail relatively frequently — so much as catchy drives and DIMMs — behind today embody removed and replaced in a few seconds," Facebook same. All the handles and levers that technicians are supposed to touch are colored green, so the machines can live serviced quickly, and even the motherboard can be removed within a careful. "In fact, Big Sur is just about entirely tool-fewer –the CPU warmth sinks are the entirely things you need a screwdriver for" Facebook says.

It's not sharing the design to be altruistic: Facebook hopes others will try out the hardware and suggest improvements. And if other enceinte companies ask server makers to build their have Big Sur systems, the economies of scale should help drive costs down for Facebook.

Machine learning has come to the fore lately for few reasons. One is that large data sets old to train the systems have get in public procurable. The other is that powerful computers have gotten affordable enough to do some impressive AI work.

Facebook pointed to software it developed already that can interpret stories, serve questions about an image, play games, and learn tasks by observing examples. "Simply we realized that truly tackling these problems at scurf would ask USA to design our have systems," it said.

Big Sur, named afterward a stretch of colorful California coastline, uses GPUs from Nvidia, including its Tesla Expedited Computing Platform.

Facebook said it testament to triad its investing in GPUs so that information technology can bring machine learnedness to to a greater extent of its services.

"Big Sur is twice as hurried as our previous generation, which substance we stern train twice A fast and explore networks doubly equally large," it said. "And distributing training across eight GPUs allows us to plate the size and speed of our networks aside other factor of two."

Google is also rolling out machine learning crossways more of its services. "Machine learning is a nucleus, transformative way by which we're rethinking everything we're doing," Google CEO Sundar Pichai aforesaid in October.

Facebook didn't say when IT would release the specifications for Big Sur. The next OCP Acme in the U.S. takes seat in March, thusly it might say more around the organisation more then.

Source: https://www.pcworld.com/article/418743/facebook-makes-its-big-sur-ai-server-design-available-to-anyone.html

Posted by: waltonheally.blogspot.com

Share this post