Monday, January 29, 2018

"China set to launch its new supercomputer"

NVIDIA watches, some commentary after the jump
From Asia Times:

Chinese supercomputers running on indigenous processing units are charging ahead; new model capable of 1 billion billion calculations per second
A prototype of China’s next generation “exascale” supercomputer capable of crunching data that’s equal to one billion billion calculations within a blink of an eye, is slated for trial later this year,  the National Supercomputer Centre in Tianjin said in a statement on Monday.

That number translates into 1,000,000,000,000,000,000 calculations per second.

Yet to be christened, the new supercomputer – co-developed with the People’s Liberation Army-affiliated National University of Defence Technology – will be 200 times faster and have 100 times more storage capacity than the 600 million yuan (US$95 million) Tianhe-1 (Sky River) supercomputer launched in Tianjin in 2010. The latter model was capable of a maximum range of 2.5 peta floating point operations per second (FLOPS), a measure of computer performance, and briefly retained the crown as the world’s fastest computer from October 2010 to June 2011, Xinhua reported.
The Tianhe-1 has been operating at full capacity as China’s military and commercial needs for supercomputing has shot up since it was launched.

The power of supercomputers is harnessed for tasks such as simulated nuclear tests and weaponry design, oil exploration, equipment modeling, biological medicine, animation design and weather forecasting.

These machines are in fact clusters of tens of thousands of CPUs – central processing units – and graphic processing units, as well as legions of servers that typically occupy hundreds of square meters....MORE
On the last Top500 list of the world's fastest supercomputers, 87 systems use NVIDIA GPUs as accelerators. Not the top two however. You have to go down to #3, Switzerland's Piz Daint which was upgraded with NVIDIA chips in 2016. China is using an entirely different architecture which we touched on in June 2016's "Milestone: China Builds The (NEW) World's Fastest Supercomuter Using Only Chinese Components (and other news) INTC; NVDA; IBM".

The different approaches have implications beyond simple bragging rights. As we said a couple weeks ago in  the outro from "'Can Chinese AI Chip Makers Compete with Nvidia?' (NVDA)":
One thing not mentioned in the Nanalyze report that is critical to understanding Chinese R&D is the importance of the People's Liberation Army (Navy) and the advantage the closed loop of academia, end-user and PLA contracted-and-government-owned companies conveys:...
You'll note the Asia Times story has the co-developer research lab as the PLA funded National University of Defence Technology.

The 200 petaflop/second (quadrillions of calculations per second) Summit supercomputer being built at Oak Ridge National Laboratory will blow past the current world's fastest, the 93 petaflop/s Sunway TaihuLight but this latest monster is planned to be five times faster still.

At that point, if you were to use it for, saaay, training artificial intelligence, you are going places that the human mind literally can't comprehend much less forecast or, dream on, guide.

If interested see also:
Jan. 11, 2018
With the Summit Supercomputer, U.S. Could Retake Computing’s Top Spot (NVDA)
Jan. 1, 2018
Military AI: China, Russia and the U.S. are Running Neck-and-Neck in an Arms Race
September 2017
"China Upgrading Milky Way 2 Supercomputer to 95 Petaflops"
September 2017
"The Astonishing Engineering Behind America's Latest, Greatest Supercomputer"