Electric car giant Tesla has renewed its supercomputer called Dojo. A 28 percent GPU boost has been given to the neural network unit used to process data from the tools. With this update, 1600 new graphics cards have been added to the Dojo.
Tesla’s supercomputer Dojo is the seventh in the world
Tesla applied some hardware changes to his machine, which he used as a supercomputer. The electric car giant said that the GeForce RTX 30 series has increased the number of Nvidia A100 GPUs using the same architecture to 7360.
According to the statement given by Tesla Engineering Manager Tim Zaman, Tesla Dojo, which has 7360 Nvidia A100 graphics cards, ranks seventh in the world in GPU ranking. However, the company also said that the 28 percent growth will continue.
The neural network unit, called the Dojo by Elon Musk, stores and analyzes data from Tesla drivers. These data are thought to be especially important for the development of Tesla’s autonomous driving technology called FSD.
We don’t yet know Tesla’s prospects for the Dojo supercomputer. But in an earlier statement, company CEO Elon Musk said, “Tesla is developing a neural network training computer called Dojo to handle really huge amounts of video data.”
While the Dojo currently relies on the Nvidia A100 GPU, Tesla has a goal of developing its own chip. Tesla stated that the neural network will be powered by the D1 Dojo chip and this chip will have a speed of 362 TFLOPs. The current performance of the machine is uncertain.
What do you think about this subject? Don’t forget to share your views with us in the comments!
3960
