Video Of Day

Breaking News

Nvidia Debuts Cloud Server Platform To Unify Ai In Addition To Hpc (Nvda)

This is a pretty large deal.
High Performance Computing has been a split as well as distinct cast of machines since the University of Manchester's Atlas reckoner went online inwards 1962.
And forthwith NVIDIA is blurring the business inwards a few unlike ways.

First off, existing as well as planned supercomputers usage graphics processing units to accelerate processing speeds at the real move yesteryear of the heap. Here's Jan 2018's "China prepare to launch its novel supercomputer"
NVIDIA watches, or as well as so commentary later the jump...

.... On the end Top500 list of the world's fastest supercomputers, 87 systems usage NVIDIA GPUs equally accelerators. Not the move yesteryear 2 however. You bring to become downwards to #3, Switzerland's Piz Daint which was upgraded amongst NVIDIA chips inwards 2016. mainland People's Republic of China is using an only unlike architecture which nosotros touched on inwards June 2016's "Milestone: mainland People's Republic of China Builds The (NEW) World's Fastest Supercomuter Using Only Chinese Components (and other news) INTC; NVDA; IBM".
Then at that topographic point was NVIDIA's ain supercomputer highlighted inwards Nov 2016:

Now they're but showing off.

The reckoner isn't going to hold out a production business or anything that generates immediate revenues but it puts the companionship inwards a real exclusive guild as well as may Pb to or as well as so in-house breakthroughs inwards fleck blueprint going forward.
The stock is upwardly $4.97 (+5.77%) at $91.16.

To hold out clear, this isn't someone using NVDA's graphics processors to speed upwardly their supercomputer equally the Swiss did amongst the i they permit CERN usage as well as which is currently the 8th fastest inwards the basis or the reckoner that's being built correct now at Oak Ridge National Laboratory as well as is planned to hold out the fastest inwards the basis (but may non brand it, China's Sunway TaihuLight is very, real fast)....
Before that they explained how to construct your ain (remember, this was pre-media-rapture, thus the explanatory tone) supercomputer, relayed inwards May 2015's:
Nvidia Wants to Be the Brains Of Your Autonomous Car (NVDA)
...Among the fastest processors inwards the concern are the one's originally developed for video games as well as known equally Graphics Processing Units or GPU's. Since Nvidia released their Tesla hardware inwards 2008 hobbyists (and others) bring used GPU's to construct personal supercomputers.
Here's Nvidias Build your Own page.
Or bring your tech guy construct i for you....
NVIDIA really started blurring the lines  amongst their mini-supercomputer for preparation AI inwards 2016:
Technology Review on NVIDIA's Pint-Sized Supercomputer (NVDA)
$129,000

In Apr 2018 the companionship had a novel offering that stunned reviewers into caveman-speak:
UPDATED—NVIDIA Wants to Be the Brains Behind the Surveillance State (NVDA)
The companionship but rolled out a $399,000 two-petaflop supercomputer that every picayune totalitarian as well as his blood brother is going to lust later to run their surveillance-city smart-city information slurping dreams.

The coming municipal information centers volition halt upwardly matching the NSA inwards full storage capacity as well as NVIDIA wants to hold out the i sifting through it all. More on this downwards the road, for forthwith here's the beast.

From Hot Hardware:
NVIDIA Unveils Beastly 2 Petaflop DGX-2 AI Supercomputer With 32GB Tesla V100 And NVSwitch Tech (Updated)...
Ummm, fauna fast.

For reference the #500 supercomputer inwards the basis is an HPE automobile that clocks inwards at a theoretical maximum speed of 712.9 teraflops at NASA/Goddard Space Flight Center's Climate Simulation Platform, piece of occupation of the 3.5 petaflop Discover system.

That way yous tin sack purchase an off-the-shelf supercomputer that is close iii times faster than the a 'puter on the Top500 list. Anyway, that's a much longer than commons introduction, here's the floor at Silicon Angle, May 30:

Nvidia debuts cloud server platform to unify AI as well as high-performance computing
Hoping to hold the high basis inwards artificial intelligence as well as high-performance computing, Nvidia Corp. belatedly Tuesday debuted a novel computing architecture that it claims volition unify both fast-growing areas of the industry.
The announcement of the HGX-2 cloud-server platform (pictured), made yesteryear Nvidia Chief Executive Jensen Huang at its GPU Technology Conference inwards Taipei, Taiwan, is aimed at many novel applications that combine AI as well as HPC.

“We believe the futurity requires a unified platform for AI as well as high-performance computing,” Paresh Kharya, production marketing director for Nvidia’s accelerated-computing group, said during a press telephone band Tuesday.

Others agree. “I think that Exascale Computers: Competing With mainland People's Republic of China Is About to Get Serious
...The 200 petaflop/second (quadrillions of calculations per second) Summit supercomputer beingness built at Oak Ridge National Laboratory volition blow yesteryear the electrical flow world's fastest, the 93 petaflop/s Sunway TaihuLight but this latest monster is planned to hold out 5 times faster still.

At that point, if yous were to usage it for, saaay, preparation artificial intelligence, yous are going places that the human heed literally can't covert much less forecast or, dream on, guide.... 

No comments