acm-header
Sign In

Communications of the ACM

ACM Opinion

Tesla's Dojo Is Impressive, but It Won't Transform Supercomputing


View as: Print Mobile App Share:
Tesla's Ganesh Venkataramanan holding a  9-petaflop training tile

Dojo is designed to do one particular thing: train AI. And Dojo chips don't operate individually, however; the smallest unit Tesla has built is what it calls Dojo training tiles.

During Tesla's recent AI Day, company officials revealed more about its Dojo supercomputer. Tesla claims that Dojo, built on its D1 chip, could end up breaking the exaflop barrier, something that no supercomputing company, university, or government has been capable of doing. The company's promises represent nothing short of a supercomputing breakthrough.

But to call Dojo a full-fledged supercomputer at this point in time is a bit generous. It hasn't been fully assembled yet, and its potential performance limits have yet to be tested. And claiming it will break the exaflop barrier may be a more difficult sell when everyone else is rating their systems using standards that are twice as complicated.

From TechRepublic
View Full Article


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account