Sunday, June 26, 2022
News

Our AI supercomputer will be world's fastest this year: Meta

   SocialTwist Tell-a-Friend    Print this Page   COMMENT

San Francisco | Tuesday, 2022 5:15:08 PM IST
Meta (formerly Facebook) has built an AI supercomputer that, it claims, will be the fastest in the world when it's fully ready in mid-2022.

Called the AI Research SuperCluster (RSC), the machine is already being used by Meta researchers to train large models in natural language processing (NLP) and computer vision for research, with the aim of training models with trillions of parameters in the near future.

"RSC will help Meta's AI researchers build new and better AI models that can learn from trillions of examples; work across hundreds of different languages; seamlessly analyse text, images, and video together; develop new augmented reality tools; and much more," Meta engineers Kevin Lee and Shubho Sengupta said in a statement late on Monday.

The first generation of this infrastructure, designed in 2017, has 22,000 NVIDIA V100 Tensor Core GPUs in a single cluster that performs 35,000 training jobs a day.

"We wanted this infrastructure to be able to train models with more than a trillion parameters on data sets as large as an exabyte -- which, to provide a sense of scale, is the equivalent of 36,000 years of high-quality video," said Meta researchers.

Early benchmarks on RSC, compared with Meta's legacy production and research infrastructure, have shown that it runs computer vision workflows up to 20 times faster and trains large-scale NLP models three times faster.

That means a model with tens of billions of parameters can finish training in three weeks, compared with nine weeks before.

"RSC is up and running today, but its development is ongoing. Once we complete phase two of building out RSC, we believe it will be the fastest AI supercomputer in the world, performing at nearly 5 exaflops of mixed precision compute," said Meta.

Through 2022, Meta will work to increase the number of GPUs from 6,080 to 16,000, which will increase AI training performance by more than 2.5x.

The storage system will have a target delivery bandwidth of 16 TB/s and exabyte-scale capacity to meet increased demand, the company added.

--IANS na/svn/dpb

( 351 Words)

2022-01-25-11:12:01 (IANS)

 
  LATEST COMMENTS ()
POST YOUR COMMENT
Comments Not Available
 
POST YOUR COMMENT
 
 
TRENDING TOPICS
 
 
CITY NEWS
MORE CITIES
 
 
 
MORE SCIENCE NEWS
Researchers find climate change negative...
Can body-checking reduce risk of injury ...
Research reveals people respond to each ...
Study: Elliptical craters shed light on ...
Research suggests unnoticed eye movement...
Post-traumatic stress has worse impact o...
More...
 
INDIA WORLD ASIA
Ensure no child is left behind, Sisodia ...
PM Modi to address 90th edition of Mann ...
Tripura by-election results to be announ...
Two AAP MLAs get extortion calls, police...
Jharkhand man held for killing fellow vi...
3 held for creating ruckus after consumi...
More...    
 
 Top Stories
'Rage Against the Machine' to donat... 
Can West counter terrorist threats ... 
Do you know long-term weight loss p... 
Pakistan: Sindh local body polls ma... 
Sidhu Moose Wala's last song 'SYL' ... 
Can body-checking reduce risk of in... 
Pilgrimages offer spiritual connect... 
Researchers find climate change neg...