Nvidia officials came to the 2015 Consumer Electronics Show and showed off technology that they hope will drive the company’s technology deeper into the automotive world.
At his keynote session in Las Vegas Jan. 4, Nvidia CEO Jen-Hsun Huang unveiled the Tegra X1, a mobile chip that he said essentially puts the power of a supercomputer into the pockets of its users. Based on the Maxwell architecture—the successor the Kepler architecture—the Tegra X1 includes 256 cores, including eight 64-bit ARM-based “Denver” CPU cores.
The new chip comes a year after Huang introduced the Tegra K1, which was aimed not only at mobile devices but also embedded systems and automobiles, hinting at how important company officials see the automotive space as a growth market.
The Tegra X1 offers more than a teraflop of performance, offering more compute power than the world’s fastest supercomputer—the IBM-based ASCI Red system—from 15 years ago.
“Slated to arrive in products during the first half of the year, Tegra X1 provides more power than a supercomputer the size of a suburban family home from 15 years ago,” Brian Caufield, editor of Nvidia’s blog, wrote in a post.
Caufield noted that the Maxwell architecture has been used over the past year in some of the top gaming graphics cards, and Nvidia officials noted that the chip can be used in mobile devices. However, the bulk of Huang’s address focused on what the company’s technology will do for automotive systems in the future.
The car industry is fast becoming a target for tech vendors, given the rapidly growing number of systems that are being put into automobiles and the push toward self-driving vehicles. Nvidia wants to be a major player in the industry, according to the CEO.
“Your future cars will be the most advanced computer in the world,” Huang said, according to Caufield. “There will be more computing horsepower inside a car than anything you own today.”
Toward that end, Huang also introduced other Nvidia products aimed at the car industry. Nvidia’s Drive PX is a technology powered by two Tegra X1 chips that can take input from 12 high-resolution cameras, crunch the data and use the information to help drivers as they steer. Eventually, the technology will be used in self-driving cars, according to Nvidia.
Another automotive platform is Drive CX for in-cabin infotainment systems that will process 17 million pixels on a display, 10 times the number of pixels in current high-end vehicles—or about two 4K displays at 60 frames a second.
Supporting all this is Nvidia’s Drive Studio software, which developers can leverage to build 3D cockpits and integrate such features as navigation and infotainment with technologies like Surround Vision (which gives the driver a real-time, 360-degree view of the car).
Huang demonstrated the capabilities of Drive PX, using a smartphone to maneuver a virtual car through a digital garage. By hitting a button, the car was called back.
“When you’re done with dinner, you say ‘Come back to me’ … and it becomes an auto-valet,” he said. “The car meanders back out and gets back to the driver.”