Self-learning with Volta
NVIDIA’s Volta GPU architecture to power DRIVE PX Xavier A.I. car supercomputer.
With NVIDIA’s next-generation Volta GPU architecture now out of the bag, the company announced DRIVE PX Xavier, its new A.I. automotive supercomputer that is designed for use in self-driving cars.
“This is the greatest SoC endeavor I have ever known, and we have been building chips for a very long time,” Jen-Hsun Huang, CEO of NVIDIA, said to the attendees at the GPU Technology Conference (GTC) Europe this year.
Xavier is the latest SoC to be based on the new Volta GPU architecture. It features a custom 8-core CPU architecture, and a new computer vision accelerator capable of delivering up to 20 TOPS (trillion operations per second) of performance, all while consuming only 20W of power. Much like its desktop GPU counterparts, Xavier packs seven billion transistors, and is manufactured on a 16nm FinFET die lithography. Volta itself was announced during the company’s inaugural GTC 2017 in San Jose, California earlier last month, and is the most advanced GPU that NVIDIA has built to-date.
Creating an A.I. for a self-driving car can be tedious, and the SoC that powers the brain of an automobile’s A.I. supercomputer requires an incredible amount of computeintense processes. As the realm of self-driving vehicles become increasingly more mission-critical, faster and smarter A.I. supercomputers are necessary in order for a vehicle to learn how to navigate its way in and around a designated perimeter.