Microsoft developed the supercomputer for OpenAI.

In 2019, Microsoft (NASDAQ: MSFT) entered a multi-year partnership with OpenAI to build a platform that OpenAI will use to create new AI technologies and deliver on the promise of artificial general intelligence. As part of the agreement, Microsoft agreed to invest $1 billion in the project. At its ‘Build 2020’ developers’ conference, the tech giant unveiled “one of the top five publicly disclosed supercomputers in the world, making new infrastructure available in Azure to train extremely large artificial intelligence models.” (sify.com)

Supercomputing monster

The single system machine comprises more than 285,000 CPU cores, 10,000 GPUs, and 400 gigabits per second of network connectivity for each GPU server.

According to hpcwire.com, though Microsoft did not reveal the throughput of the machine, the top-five moniker would place the machine as capable of greater than 23.5 (double-precision, Linpack) petaflops.

The system will be hosted on Microsoft’s Azure public cloud and is a part of Microsoft’s “AI at Scale” initiative. It will provide the massive infrastructure needed to support very large AI models. It will, therefore, be accessible to AI organizations, researchers, and developers.

“The exciting thing about these models is the breadth of things they’re going to enable,” said Microsoft CTO Kevin Scott. “This is about being able to do a hundred exciting things in natural language processing at once and a hundred exciting things in computer vision, and when you start to see combinations of these perceptual domains, you’re going to have new applications that are hard to even imagine right now.”

Dream system

OpenAI CEO Sam Altman said AI researchers have only now been able to specify the configuration of their “so-called” ultimate dream machine. Fortunately, Microsoft has been able to build it.

“We are seeing that larger-scale systems are an important component in training more powerful models,” Altman commented.

AI at Scale

The company released the Microsoft Turing model for natural language generation earlier this year. It is the largest AI language model in the world.

The company has deployed its family of large Turing AI models for solving many language processing tasks in its productivity products.

Ultimately, however, Microsoft intends to throw open its AI at Scale resources to all interested parties through Azure AI services and Github.

Related Story:   Novartis And Microsoft Team Up to Transform Global Medicine