Meta is building a new AI supercomputer, AI Research SuperCluster
At the end of last year, the metaverse became a very popular concept in the world. Its essence is a virtual world that is linked and created by scientific and technological means, a virtual world that maps and interacts with the real world, and a digital living space with a new social system. Many big-name companies have also entered the game, the most representative being Mark Zuckerberg’s Facebook. At the Facebook Connect 2021 augmented reality and virtual reality conference, they changed their name to “Meta” and made a full-scale foray into the metaverse.
“QCDOC Supercomputer” by Brookhaven National Laboratory is licensed under CC BY-NC-ND 2.0
After Meta, there are many actions. According to the Tomshardware reports, Meta recently is building a new AI supercomputer named AI Research SuperCluster, or RSC for short. It is a high-speed computer designed specifically for training machine learning systems, and although it is already operational, it is still in development. According to Meta’s official statement, if it can be completed in 2022, it will become the world’s fastest AI supercomputer.
In a blog, Meta writes:
“We expect such a step function change in compute capability to enable us not only to create more accurate AI models for our existing services, but also to enable completely new user experiences, especially in the metaverse. Our long-term investments in self-supervised learning and in building next-generation AI infrastructure with RSC are helping us create the foundational technologies that will power the metaverse and advance the broader AI community as well.”
The specs of the RSC supercomputer are also worth a look. It is understood that this AI supercomputer is currently in the first stage of operation, using 760 NVIDIA GGX A100 systems, including 6080 connected GPUs, which can provide excellent graphics processing capabilities. The second phase is expected to be installed in July 2022, when the machine will be upgraded with multiple 16,000 NVIDIA GPUs and an additional 1,240 DGX nodes.
It can be seen that Meta has built this AI supercomputer, which has put scientific research AI at the core of the enterprise and has also taken a big step towards the metaverse.
