Whilst Tesla’s AI Day did spend a lot of time covering off on full self-driving vehicles, a new tesla robot (image below), they also announced a whole new processor.
The Tesla D1 Dojo is a purpose-built processor designed to conquer the challenges of AI computing in our time. It’s touted as a ‘neural network exaflop supercomputer’. Tesla’s senior director of Autopilot hardware and Dojo project lead said the 7nm D1 chip was designed in-house specifically for machine learning and to remove bandwidth bottlenecks.
All of D1’s 354 chip nodes reportedly have one teraflop (1,024 gflops) of computing power ready to meet the complex and demanding needs. Venkataramanan also stated that the entire chip was capable of up to 363 teraflops of computing and 10tbps of on-chip bandwidth/4tbps of off-chip bandwidth.
To house the chips, the company has developed what it calls ‘training tiles’ with which to build its computing systems. Each tile consists of 25 D1 chips in an integrated multi-chip module, with each tile providing nine petaflops of computing and 36tbps of off-tile bandwidth.
To create Dojo, the company plans to install two trays of six tiles in a single cabinet, for 100 petaflops of compute per cabinet. Once complete, the company will have a single ‘Exapod’ capable 1.1 exaflops of AI compute via 10 connected cabinets; the entire system will house 120 tiles, 3,000 D1 chips, and more than one million nodes.
According to Tesla CEO Elon Musk, the system will be launched next year. The company also plans to make Dojo available to other AI developers in the future.