Nvidia Corp. (Santa Clara, Calif.) has announced its most advanced processor yet, the Tegra X1, and is emphasizing the mobile processor's suitability to take Nvidia away from gaming and into solving problems, such as automated driving, in the real world.
Nvidia did speak up about suitability of its previous processor, the 28nm Tegra K1, for automotive applications but that was mostly as a graphics engine for detailed cockpit displays. With the X1 Nvidia is aiming the GPU cores on-board as a teraflops capable processor that can do "deep learning" and computer vision applications.
The company describes "deep learning" as the ability to differentiate various types of vehicles and even the ability to differentiate between a a parked vehicle and one that is about to pull out into traffic. It is not clear whether this is based on an ability to run artificial neural network (ANN) software or not.
And alongside the chip announcement Nvidia has announced two automotive computers, the Drive PX and the Drive CX that are based on Tegra X1 or optionally in the case of Drive CX on the Tegra K1.
256-core GPU on-chip
The Tegra X1 is a "big-little" octacore 64bit processor with four Cortex-A57 cores with 2Mbytes of level 2 cache and four Cortex-A52 cores with 512Kbytes of level 2 cache. It also has a 256 core graphics processing unit compliant with Nvidia's 'Maxwell' GPU architecture. The chip is being manufactured for Nvidia by Taiwan Semiconductor Manufacturing Co. Ltd. using its 20nm CMOS manufacturing process and should become available in the first half of 2015.
The 256-core Tegra X1 provides twice the performance of its predecessor, the Tegra K1, which is based on the previous-generation Kepler architecture, although some of the performance improvement will have come from the migration from 28nm CMOS to 20nm CMOS. The chip consumes less than 10 watts, Nvidia said.
Tegra X1 supports the following graphics standards: Unreal Engine 4, DirectX 12, OpenGL 4.5, CUDA, OpenGL ES 3.1 and the Android Extension Pack.
More information can be found at the Tegra X1 website
Computers that drive cars
The Drive PX computer is based on the Tegra X1 and is intended as an auto-pilot development platform Nvidia said. One of the applications the Drive PX and its Tegra X1 has been used for is Auto-Valet, a self-parking application.
"Audi and Nvidia share a common belief that machine learning is a powerful enhancement to our zFAS piloted driving technology. Thus, Audi sees Drive PX as a crucial tool for further research and development," said Ricky Hudi, executive vice president of electronics development at Audi AG, in a statement issued by Nvidia.
The Drive CX is described as cockpit computer that offers complete hardware and software platform for computer vision for navigation and the graphics for infotainment, digital instrument clusters and driver monitoring. It supports 360 degree all-around vehicle vison and a digital rear-view mirror. Available with either the Tegra X1 or Tegra K1 processors, the Drive CX can process and send information to16.8 million pixels on multiple displays – more than 10 times that of current model cars.
"We see a future of autonomous cars, robots and drones that see and learn, with seeming intelligence that is hard to imagine," said Jen-Hsun Huang, CEO of Nvidia in a statement. "To achieve this dream, enormous advances in visual and parallel computing are required. The Tegra X1 mobile super chip, with its one teraflops of processing power, is a giant step into this revolution." Huang added: "With vast arrays of cameras and displays, cars of the future will see and increasingly understand their surroundings. Whether finding their way back to you from a parking spot or using situational awareness to keep out of harm's way, future cars will do many amazing, seemingly intelligent things."
Both the Drive PX and Drive CX computers come with software from Nvidia or third-party suppliers and will be available in the second quarter of 2015.
Related links and articles: