Difference between revisions of "Tesla Dojo"
Jump to navigation
Jump to search
↑ "Tesla Dojo Technology" (PDF). Tesla. Retrieved 29 October 2021.<templatestyles src="Module:Citation/CS1/styles.css"></templatestyles>
| Line 13: | Line 13: | ||
* [[Fugaku]] (supercomputer) | * [[Fugaku]] (supercomputer) | ||
* Float8 (CFloat8), an 8-bit floating point format | * Float8 (CFloat8), an 8-bit floating point format | ||
| + | * SHP, [[UHP]] | ||
== See also == | == See also == | ||
Revision as of 08:17, 6 December 2021
wikipedia:Tesla Dojo is an artificial intelligence (AI) neural network training supercomputer based on self developed chip.
The D1 chip, that would be used in Dojo supercomputer, according to Venkataramanan uses a 7-nanometer manufacturing process, with 362 teraflops of processing power.
Tesla Dojo whitepaper was released in September 2021[1] Configurable
Related
- Gartner research vice president Chirag Dekate
- Fugaku (supercomputer)
- Float8 (CFloat8), an 8-bit floating point format
- SHP, UHP
See also
- Tesla, LIB, Tesla Model 3, Tesla Cybertruck, Tesla Model S, Tesla Model X, Tesla Model Y, Tesla Autopilot, Tesla Megapack, Tesla Plaid trim, Andrej Karpathy, Elon Musk, Tesla Shanghai, Giga Berlin, Supercharger, OTA, Tesla Active Safety Features, Dojo (D1 chip), Tesla TRIP, Tesla Powerwall, 1280, Tesla roadside assistance, Tesla Supercharger, Tesla safety, Tesla Bot (Optimus), YouTube Tesla, SolarCity, GB/T, Tesla app, Tesla Solar Roof, The Boring Company, Robotaxi
Advertising:
