Microsoft builds huge cloud-based AI supercomputer for OpenAI

Microsoft introduced on Tuesday that it has constructed a enormous supercomputer for OpenAI, a company with a mission to establish an synthetic typical intelligence that will advantage humanity as a entire.

The announcement arrived at Microsoft’s Create 2020 developer meeting, which is staying hosted online because of to the coronavirus pandemic.

Microsoft’s new supercomputer is hosted in the Azure cloud and is claimed to be among the top 5 supercomputers in the globe. It functions 285,000 processor cores and ten,000 GPUs and can present 400 gigabits for each next of community connectivity for every GPU server, Microsoft says.

Nevertheless, Microsoft did not share real performance figures for its machine.

Supercomputers throughout the globe are ranked by processing velocity two times a year by specialists at the Top500 venture. The machines receive scores centered on how rapid they can conduct Linpack take a look at. IBM’s Ability Method-centered Summit is at the moment ranked as the world’s most strong supercomputer, reaching above 148,000 teraflops of velocity.

In accordance to Microsoft, its Azure machine is excellent for AI that learns from analysing billions of internet pages of publicly readily available textual content.

Microsoft says that the supercomputer’s means, as very well as AI models and schooling instruments, will be readily available to experts, clients and builders via Azure AI and GitHub.

In July 2019, Microsoft introduced its prepare to spend $1bn in OpenAI to develop AI systems for supercomputers.

OpenAI is a non-gain organisation co-launched by Elon Musk in 2015. It gets funding from undertaking capitalist Peter Thiel, LinkedIn co-founder Reid Hoffman and Y Combinator’s Jessica Livingston, among some others. It also has corporate ties to Amazon World wide web Products and services and IT solutions organization Infosys.

When OpenAI was introduced, Musk said that the notion of this NPO was to develop AI systems that would be utilised for superior causes, reducing the risk of harm, by distributing AI systems as commonly as achievable.

Microsoft officials also rolled out the next variation of open up supply DeepSpeed deep finding out library for Pytorch, which Microsoft said would reduce the computing power required for schooling models.

Microsoft also said that it would quickly begin open up-sourcing Turing models for AI as very well as “recipes for schooling them in Azure Machine Learning”. Microsoft’s Turing models refer to a family of large AI models the company utilizes to enhance language comprehending throughout Place of work, Bing, Dynamics and other solutions.