NVIDIA unveils Tegra K1 192-core mobile processor; ADAS applications on the way to autonomous driving
At CES in Las Vegas, NVIDIA unveiled the Tegra K1 mobile processor, a 192-core chip featuring the same NVIDIA Kepler architecture that powers the NVIDIA GeForce GTX 780 Ti GPU. Among its targeted applications—notably mobile gaming—are a variety of automotive applications.
Tegra K1 will drive camera-based, advanced driver assistance systems (ADAS) such as pedestrian detection, blind-spot monitoring, lane-departure warning and street sign recognition, and can also monitor driver alertness via a dashboard-mounted camera.
To process the steady deluge of sensor and camera data required by a self-driving car, NVIDIA is bringing highly energy-efficient supercomputer technology inside the vehicle. Tegra K1 solves this by providing 10 times the computing power of previous mobile processors without consuming additional energy.—Taner Ozcelik, vice president and general manager of the Automotive business at NVIDIA
The Tegra K1 is also the first mobile processor to support CUDA—the parallel computing platform at the heart of modern computer vision. In its automotive-grade form, the chip is hardened to withstand a wider range of temperatures and withstand harsher operating conditions.
In addition to computer vision, the Tegra K1 processor excels at creating razor-sharp, photo-real 3D graphics. Using NVIDIA’s Material Definition Language—which simulates how light reflects and refracts off of actual materials—digital instrument clusters and infotainment systems can be customized with a wide range of virtual materials, from copper, titanium and brushed aluminum, to stitched leather and carbon fiber.
Libraries of photo-realistic rendered material provide a nearly limitless range of customizable digital cockpit configurations. These allow drivers to configure their gauges, controls and dials to suit their unique tastes.
Audi and NVIDIA have a long, deep partnership, in which we’ve utilized three generations of Tegra to bring industry-leading capabilities to the instrument cluster, infotainment and rear seat entertainment systems. Tegra K1 opens a new chapter for Audi to deliver revolutionary supercomputing advances to the car, paving the way to piloted driving experiences.—Ricky Hudi, chief executive engineer of electrics/electronics at Audi AG
ADAS solutions currently in the market are based mainly on proprietary processors. NVIDIA Tegra K1 moves beyond this by providing an open, scalable platform. In addition, NVIDIA designed the Tegra K1 processor to be fully programmable, so it can be enhanced via over-the-air software updates that support new functionalities as they become available from automakers.
Tegra K1 will be available to automotive OEMs and Tier 1 suppliers as a visual computing module (VCM), a design first introduced two years ago. The Tegra K1 VCM delivers a full computer system for the vehicle, capable of running various operating systems including QNX, Android, Linux or Windows.
The Tegra VCM allows in-vehicle systems to be easily upgraded as newer hardware becomes available, closing the gap between the rapid pace of innovation in the consumer electronics space and the longer development lifecycles typically experienced in the automotive industry.
Today there are more than 4.5 million cars on the road powered by NVIDIA processors, including the newest models from Audi, BMW, Tesla Motors and Volkswagen.
For other applications, the Tegra K1 is offered in two pin-to-pin compatible versions. The first version uses a 32-bit quad-core, 4-Plus-1 ARM Cortex A15 CPU. The second version uses a custom, NVIDIA-designed 64-bit dual Super Core CPU. This CPU (codenamed Denver) delivers very high single-thread and multi-thread performance. It is based on the ARMv8 architecture, which brings the energy-efficient heritage of ARM processor technology to 64-bit computing.
The 32-bit version is expected in devices in the first half of 2014, while the 64-bit version is expected in devices in the second half of the year.
The Kepler GPU at the heart of Tegra K1 is 1.5 times more efficient than other mobile GPUs, according to NVIDIA. This results in faster performance in the same power envelope and a better experience for gaming and GPU-accelerated applications.