
Meta (formerly Facebook) announced the creation of the world’s largest AI Research SuperCluster (RSC) for machine learning research.
Meta is announcing the AI Research SuperCluster (RSC), our latest AI supercomputer 💻 for AI research. RSC will allow our researchers to do new, groundbreaking experiments in #AI. Learn more about RSC and the important role it will play: https://t.co/l9CcQuFLyM pic.twitter.com/gD8Ve74ZqQ
— MetaAI (@MetaAI) January 24, 2022
Today, the RSC consists of 760 Nvidia DGX 100 systems, which contain 6800 Nvidia A100 video accelerators. This allowed the Meta system to take fifth place in the ranking of the fastest supercomputers in the world.

The completed RSC will house 16,000 Nvidia A100 graphics chips and 4,000 AMD Epyc Rome 7742 processors in 2,000 clusters, the company said. The peak performance of the supercomputer will be 5 exaflops. The setup will be able to transfer up to 16 terabytes of training information per second, Meta said.
“We expect this stepwise change in computing capabilities will allow us not only to create more accurate AI models for our existing services, but also provide a completely new user experience, especially in the metaverse,” said Meta engineers Kevin Lee and Shubho Sengupta.
According to them, the supercomputer is intended for the company’s research projects and should not expect any products based on it in the near future.
The company added that the supercomputer will enable the creation of new and better AI models that can learn from trillions of examples, analyze text, images and videos together, and develop augmented reality tools.
“We hope that RSC will help us develop completely new artificial intelligence systems that can, for example, provide real-time voice translation or augmented reality cooperative play,” said Li and Sengupta.
Meta plans to complete the supercomputer by mid-2022. The cost of the project, as well as the physical location of RSC in the company, were not disclosed.
Recall that in November 2021, Sber introduced the Christofari Neo supercomputer to accelerate artificial intelligence tasks. The performance of the installation reaches 11.95 petaflops.
In May, American scientists unveiled the Perlmutter AI supercomputer, which will help build the largest ever 3D map of the visible universe to study dark energy.
Subscribe to Cryplogger news in Telegram: Cryplogger AI – all the news from the world of AI!
Found a mistake in the text? Select it and press CTRL+ENTER