Tesla has released two new deep-dive presentations into its Dojo AI supercomputer ahead of its upcoming AI Day 2.
You can check the slides from both presentations below.
Dojo is Tesla’s own custom supercomputer platform built from the ground up for AI machine learning and more specifically for video training using the video data coming from its fleet of vehicles.
The automaker already has a large NVIDIA GPU-based supercomputer that is one of the most powerful in the world, but the new Dojo custom-built computer is using chips designed by Tesla.
The custom-built supercomputer is expected to elevate Tesla’s capacity to train neural nets using video data, which is critical to its computer vision technology powering its self-driving effort.
Last year, at Tesla’s AI Day, the company unveiled its Dojo supercomputer, but the company was still ramping up its effort at the time. It only had its first chip and training tiles, and it was still working on building a full Dojo cluster or “Exapod.”
Tesla’s AI Day 2 is coming up on September 30, and the automaker is expected to give an update on its Dojo program, but it has now given an update ahead of the event.
Tesla’s Dojo team gave two presentations at the Hot Chip event this week that are likely to preview some of what the company will discuss next month at AI Day.
Serve The Home shared the slides from the two presentations, which you can see below. The info is quite “inside baseball” on the system architecture side and requires some technical knowledge to appreciate, but I thought it would be worth sharing here for those who can appreciate it.
Tesla Dojo AI System Microarchitecture
Tesla Dojo Custom AI Supercomputer
I would assume that the main goal of those presentations is to recruit engineers for the Dojo program, and we might get a version more palatable to a wider audience next month at Tesla AI Day 2.
Subscribe to Electrek on YouTube for exclusive videos and subscribe to the podcast.