Nvidia Aims To Ride Pegasus to Self-Driving Car Glory

Well, Nvidia is still making those graphics chips, but it has also found out a way to branch out in new and growing business areas. One of those is the market for autonomous-driving cars, and at the GPU Tech Conference in Munich on Tuesday, Nvidia left no doubt about the levels it thinks it can reach with self-driving cars.
The highlight was a new chipset that Nvidia calls Pegasus, named after the winged horse of Greek mythology. Nvidia said Pegasus is capable of supporting “Level 5 autonomy.” And now you are probably asking yourself just what “Level 5 autonomy” is.
The Holy Grail of the self-driving car is, well, a car that can really drive itself. It’s 2017, after all. Shouldn’t we have cars that can drive themselves by now? Nvidia says Pegasus can make true autonomous-driving cars a reality, and that’s what Level 5 represents.
That means a car without a driver. Or, without some of the traditional accessories that drivers use, such as steering wheels, gas or braking pedals, or mirrors. To do all of this, you need something that can run a lot of operations at once to make sure the car stays safely on the road. And Pegasus is capable of handling 320 trillion operations every second.
Nvidia said the first applications for Pegasus will be what it calls “robotaxis”: robotic taxis that will pick you up when you call for a ride. Even though fleets of such vehicles are still years away from taking over our city streets, Nvidia said Pegasus will be made available for automakers starting in the second half of 2018.
Investors liked the idea of what Pegasus might do for Nvidia. They sent the company’s shares to an all-time high on Tuesday of $192.95. Nvidia shares ended the day up by almost 2 percent, at $188.93.
As he showed off Pegasus, Nvidia Chief Executive Jensen Huang said in no uncertain terms how the technology will impact driving in the coming years.
“In the old world, the more powerful your engine, the smoother your ride will be,” Huang said. “In the future, the more computational performance you have, the smoother your ride will be.”
With 320 trillion computations going on every second, a car running with Pegasus under the hood should be able to make that ride as smooth as Huang expects it to be.
© 2017 San Jose Mercury News

syndicated under contract with NewsEdge/Acquire Media. All rights reserved.
Image credit: Product shot by NVIDIA.


爱加密与安百科技达成战略合作 共筑信息产业应用安全+
网络安全法普法宣传 004《网络安全法》的突出亮点
土库曼斯坦总统送普京软萌小狗 画风有点萌