CES 2015: Nvidia targets Tegra X1 mobile supercomputer chip at future cars
Tegra X1 doubles the processing power of its predecessor but consumes the same amount of energy
LAS VEGAS: Nvidia has revealed its Tegra X1 mobile processor, claiming to deliver supercomputer power to fuel the graphics and artificial intelligence capabilities in future cars.
Revealed at CES in Las Vegas on Sunday, the Tegra X1 sports the same 256-core Maxwell graphics processing architecture found in powerful PC graphics cards, but with an 8-core, 64-bit processor.
Nvidia claims that the Tegra X1 is the world's first mobile supercomputer chip and the most "super energy efficient architecture in the world".
Nvidia co-founder and chief executive Jsen Huang said that the power envelope of the Tegra X1 is around 10W, the same as its predecessor the Tegra K1, but offers "double the performance".
Huang explained that the Tegra X1 will be targeted at the automotive industry to provide better graphical interfaces and enable cars to "see and understand what is happening around" them through better processing of data gathered by sensors.
"So what are we going to do with all that horsepower? Sure, we don't need it for phones," said Huang.
"We believe future cars will be the most advanced computers in the world, with rich displays all round and more computing horsepower than any other systems."
Huang explained that side mirrors could become smart mirrors, complemented by rear seat entertainment and curved sides that could become displays.
"We can imagine curved displays within cars to expand rapidly. The combined resolution inside your car will grow like nothing you've ever seen. We'll become accustomed to rich displays," said Huang.
To enable this vision, Nvidia has developed Drive CX, a platform powered by Maxwell which includes software called Drive Studio to "take computer graphics to whole new levels".
Nvidia said that designers won't have to understand the coding and tech involved in Drive CX, as they can work with Design Studio and translate their vision for the future into something they can easily deploy.
Huang also unveiled Drive PX that uses a dual Tegra X1 chipset to combine the data gathered from a car's sensors and cameras with "deep neural learning" capacities.
The hardware platform effectively allows a car to classify various objects and obstacles on the road ahead and change behaviour to suit the situation, for example noting the difference between a civilian car and a police vehicle.
With Drive PX, Nvidia looks to be providing the foundations of a platform that bridges the gap between in-car system automation and fully autonomous vehicles.
The video below showcases the Tegra X1.
Nvidia is not the only company exploring the relationship between technology and cars at CES 2015.
Hyundai revealed that its BlueLink app, which allows a car to be located, locked and unlocked from a smartphone, will be available for smartwatches running Google's Android Wear.
Google has been developing its own driverless car over the past 12 months, choosing complete car autonomy over selected system automation.