Nvidia has provided a supercomputer for accelerating the development of autonomous driving.
Nvidia is also working on the development of autonomous driving and provides computing capacity for some industry stakeholders. Specifically for over 225 companies, including Audi, Mercedes Benz. Toyota. TuSimple. Volvo or ZF.
This computing capacity is used for example Artificial intelligence in the image processing, To push the development of this, Nvidia has now brought the supercomputer DGX “SuperPOD” into play.
This computer is ranked 22nd in the ranking of the world’s fastest computing units and includes an AI infrastructure. It has 96 individual Nvidia DGX-2H computers and a Mellanox connection technology. The computing power is 9.4-petaflops, which is mainly used for the training of Artificial Intelligence. The individual DGX-2H units each require only 12 kW of power.
This one meets the requirements in the field of Autonomous driving Reasonably, so Nvidia, because a single data collection vehicle generates a terabyte of data per hour. These datasets are a challenge for computing power, but not so much for the Nvidia supercomputer. This is needed to prepare the artificial intelligence for their tasks. In this way, the autonomous driving should succeed better, as would be a People at the wheel put.
So the SuperPOD should optimize autonomous driving. This succeeds faster than before, in an endless loop. This will enable artificial intelligence to train better and ultimately develop driving skills.
Also the fastest board computerwhich is currently available, comes from Nvidia. It was built and used in cooperation with the supplier ZF Friedrichshafen. The development of supercomputers, such as the now-introduced DGX SuperPOD, helped advance the design for large artificial intelligences. Now one hopes for progress in autonomous driving.
source (English)
About David Fluhr
I’ve been writing about Autonomous & Connected Driving since 2011 and also talk about it on other sites like the Smart Mobility Hub. I studied social sciences at the HU Berlin and since 2012 I am a freelance journalist. Contact: mail@autonomes-fahren.de