Microsoft built a new AI supercomputer in Azure


Microsoft built a new AI supercomputer in Azure


The partnership that Microsoft announced last year with the synthetic intelligence startup OpenAI has been fruitful, to mention the smallest amount , because, at this year’s Build event, they announced that they need built the fifth most powerful AI computer within the world.

What are the tech specs of the Microsoft AI supercomputer?

The supercomputer developed along side OpenAI may be a single system with quite 285,000 CPU cores, 10,000 GPUs, and 400 gigabits per second of network connectivity for every GPU server.

Microsoft writes during a blog post that compared with other machines listed on the TOP500 supercomputers within the world, their system ranks within the top five.

this is often about having the ability to try to to 100 exciting things in tongue processing directly and 100 exciting things in computer vision, and once you start to ascertain combinations of those perceptual domains, you’re getting to have new applications that are hard to even imagine immediately .

, said OpenAI CEO Sam Altman.

What is the Microsoft AI supercomputer getting to be used for?
Microsoft announced that it might soon start to supply its Microsoft Turing models in open-source, and recipes for training them in Azure Machine Learning.

If you’re a developer, this may offer you access to an equivalent family of language models that Microsoft is using to enhance language understanding across all its products.

They will also use the supercomputer with the remake of DeepSpeed.

For people who aren't conversant in it, DeepSpeed is an open-source deep-learning library for PyTorch, that reduces the quantity of computing power needed for giant distributed model training.

The update may be a lot better than the version released three months ago. Now, it allows training models quite 15 times larger and 10 times faster than doing an equivalent thing without DeepSpeed on an equivalent infrastructure.





Post a Comment

0 Comments