you use it to train your autonomous driving AI

Supercomputers are generally in academic or government organizations and are used for scientific research or predictions such as weather. However, Tesla seems to be here to break with that. They have presented their supercomputer, which they say is the fifth most powerful in the world right now.

Andrej Karpathy, Tesla’s head of artificial intelligence, has unveiled the supercomputer at the 2021 conference on Computer Vision and Pattern Recognition. During the conference he mentioned the supercomputer and showed some images of it. Its about Tesla’s third large computer cluster. Together they make up what is estimated to be the fifth most powerful supercomputer in existence.

If the specifications that Andrej Karpathy has shown are correct, we are facing a real beast. Is about a 720-node 8x A100 80GB system with 5760 GPUs in total. This computer can output 1.8 EFLOPS power with 640 Tbps switching capacity.

The long-awaited autonomous driving

Why a supercomputer of such a caliber? For autonomous driving. For the past few years Tesla has put its efforts into computing both inside and outside of its vehicles.The FSD inside every Tesla is in charge of analyzing and calculating in real time all the parameters. Meanwhile, it needs supercomputers to learn and train the AI ​​that is then mounted on the vehicles.

Screen Shot 2021 06 21 At 6 07 20 Am

But Tesla apparently isn’t going to stay here. Has been developing Dojo for a while Y Elon Musk has spoken about it on occasion. If the rumors and estimates are true, it will be one of the most powerful supercomputers in the world, if not the most. Dojo will use its own optimized architecture to train neural networks, it will not be a GPU cluster like the latter. Expected have it ready next year.

At the moment the most powerful supercomputer in the world it is the japanese fugaku. However, if we focus solely on artificial intelligence and neural networks, NVIDIA recently presented uno con 6.000 GPUs A100.

Via | Interesting Engineering


Source: Xataka by feeds.weblogssl.com.

*The article has been translated based on the content of Xataka by feeds.weblogssl.com. If there is any problem regarding the content, copyright, please leave a report below the article. We will try to process as quickly as possible to protect the rights of the author. Thank you very much!

*We just want readers to access information more quickly and easily with other multilingual content, instead of information only available in a certain language.

*We always respect the copyright of the content of the author and always include the original link of the source article.If the author disagrees, just leave the report below the article, the article will be edited or deleted at the request of the author. Thanks very much! Best regards!