NVIDIA Parker SoC Aims To Be The Brains Under The Hood Of Autonomous Cars

nvidia drive px2 3
It seems as though computers have taken over more and more aspects of our driving experience in recent years. Many of the advanced features are being introduced to improve safety; meaning that you can find many mainstream and luxury vehicles that offer adaptive cruise control, lane keep assist, cross-traffic alert, emergency braking, and even auto parking functionality.

But the holy grail is of course fully autonomous driving. Google is perhaps the leader in this arena with its fleet of self-driving Toyota and Lexus hybrids. Tesla is the closest to providing this functionality to regular consumers with Autopilot, and now NVIDIA is looking to provide the hardware to spread fully autonomous driving to the masses.

drive px 2

NVIDIA announced its Drive PX 2 autonomous car processing engine back at CES 2016, and deemed it the “world’s first in-car artificial intelligence supercomputer.” Now NVIDIA is lifting the cloak on the chips that are powering this powerful platform. At the heart of Drive PX 2 are two Parker processors and two Pascal-based GPUs. As you might expect, Parker is part of NVIDIA’s Tegra family of mobile SoCs, and is based on the company’s next generation Denver processor architecture (Denver 2.0). Two 64-bit Denver 2.0 cores are paired with a quartet of 64-bit ARM Cortex A57 processors.

22 Parker diagram 1

Parker can deliver up to 1.5 teraflops of performance for deep-learning “self-driving AI cockpit systems” according to NVIDIA. Likewise, Parker offers anywhere from 50 to 100 percent better multi-core performance compared to mainstream mobile SoCs. Topping things off are two, 256-core Pascal GPUs, which have more than enough power to handle the complex deep learning algorithms that have to be processed to make autonomous driving possible. In addition, the GPUs are also used to power HD displays used for infotainment and instrument clusters.

parker specifications two

“Working in concert with Pascal-based supercomputers in the cloud, Parker-based self-driving cars can be continually updated with newer algorithms and information to improve self-driving accuracy and safety,” writes NVIDIA. “In fact, DRIVE PX 2 delivers an unprecedented 24 trillion deep learning operations per second to run the most complex deep learning-based inference algorithms. Such systems deliver the supercomputer level of performance that self-driving cars need to safely navigate through all kinds of driving environments.”

According to NVIDIA, 80 automakers, tier 1 suppliers and research centers that have backed Drive PX 2 for autonomous vehicle integration. The furthest along is perhaps Volvo, which will use Drive PX 2 for its Drive Me autonomous XC90 pilot-program in 2017. However, if you want to take a ride in an autonomous Volvo XC90 a bit sooner, Uber is offering free rides in Pittsburgh, Pennsylvania later this month.