Electric vehicle company Tesla has released a new white paper regarding a new standard for its Dojo supercomputing platform.

CEO Elon Musk called the newspaper “bigger than it looks,” Electrek reports. In the abstract, the automaker describes a new standard for working with its computing platform.

“This standard specifies Tesla arithmetic formats and methods for the novel 8-bit and 16-bit floating-point binary arithmetic in computer programming environments for deep learning of neural networks,” according to the report.

This standard also specifies exception conditions and their status indicators. An implementation of a floating point system conforming to this standard can be done entirely in software, entirely in hardware, or in any combination of software and material, ”he added.

For years, Tesla has teased the development of a new internal supercomputer optimized for video neural network training. Tesla manages an insane amount of video data from its fleet of more than one million vehicles. He uses this video data to train his neural networks.

For the past two years, Musk has teased the development of Tesla’s own supercomputer called the “Dojo”. Last year, he even announced that Tesla’s Dojo would have a capacity of more than one exaflop, or one quintillion (1018) floating point operations per second or 1000 petaFLOPS.

The automaker already has its Dojo chip and slab, but it is still working on building its full rack to create the supercomputer.

Read also :

Tesla Inc shares rose 13% on the news of the Hertz order, pushing its market value past $ 1,000 billion. Musk’s net worth of $ 288.6 billion now exceeds the market value of Exxon Mobil Corp or Nike Inc.