LAS VEGAS — Nvidia introduced its next-generation mobile processor, Tegra K1, at the International Consumer Electronics Show (CES) in Las Vegas. The 192-core super chip, built on Kepler architecture, brings “the heart of GeForce and the soul of Tesla to mobile computing,” said Nvidia CEO Jen-Hsun Huang at the company's press event on Sunday.
The Tegra K1 comes in two versions: a 32-bit, quad core 4-Plus-1 ARM Cortex A15 CPU, slated to go into production in the first half of 2014, and a 64-bit version built on dual Denver CPUs to be available in the latter half of 2014. The 64-bit Tegra is designed for “high performance single threaded and very high performance multithreaded processes,” Huang said.
At the most basic level, Tegra (and a combination of other Nvidia products such as Shield, which allows for ultra-high quality game streaming) enables console and computer games to be played on a mobile platform without quality degradation. Tegra K1 is the first mobile processor to deliver the same graphics features as consoles such as the Xbox and PlayStation, but with faster performance.
“We can take absolutely anything that runs on PC or high-end console and run it on Tegra… I didn’t think that we’d be at this level on mobile for another three to four years,” Epic Games Founder Tim Sweeney said.
The processor is a major step forward in game development, allowing for a new level of graphics display that solves the problem of frame stutter and lag. Huang showed Epic Games’ Unreal Engine 4 running on a Tegra K1 mobile processor, highlighting capabilities for photo-real displays alongside global illumination and physics simulations using multiple render targets.
“Tegra K1 makes possible and turns your PC into an amazing game console,” Huang said. “It self-detects and reconfigures games so you can get best game rate.”
Nvidia introduced its next-generation mobile processor, Tegra K1, at International CES. Nvidia CEO Jen-Hsun Huang said the 192 core super chip, built on Kepler architecture, brings “the heart of GeForce and the soul of Tesla to mobile computing.”
Huang demonstrated several real-time generated graphics, including a face with subtle shadows and changes in skin coloration based on lighting. Other demonstrations highlighted texture and pupil adaptation. These aimed to simulate the variations in detail we see in different light.
An automotive Tegra platform is already sampling and “makes it possible to run the types of applications that dedicated (automated driver assistance systems) boxes do today,” Huang said. Pedestrian detection, blind-spot monitoring, lane departure warning, and other signals would be made possible through optical flow, histogram, and feature detection on Tegra’s programmable platform.
“We will see a rapid rate of algorithm advancement,” Huang said. “This brings to bear our last decade of work in computer vision and GPU computing.”
The Tegra K1 comes in two versions: a 32-bit, quad core 4-Plus-1 ARM Cortex A15 CPU, slated to go into production in the first half of 2014, and a 64-bit version built on dual Denver CPUs. The 64-bit Tegra is designed to be “high performance single threaded and very high performance multi-threaded processes,” Huang said.
"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," said Taner Ozcelik, vice president and general manager of the automotive business at NVIDIA. "Tegra K1 solves this by providing 10 times the computing power of previous mobile processors without consuming additional energy."
Additionally, the Tegra K1 Visual Computing Module brings real time rendering to car design. Dubbed “Project Mercury,” Nvidia’s CES demonstration included customizable digital dashboards and photo-realistic rendered materials.
“We see no reason why industrial design should end at physical construction of the car,” Huang said.
— Jessica Lipsky, Associate Editor, EE Times