Nvidia has come out swinging this year at CES 2016. While all may have expected the announcement of a new desktop or mobile graphics architecture and all that, no one expected this upgrade. Nvidia CEO Jen-Hsun unveiled Nvidia’s newest pet project, a supercomputer for self-driving cars that is being called the Drive PX2, the sequel to last year’s model the Drive PX.
The processing power in this thing is insane. Nvidia claims it has a processing power equivalent to that of 150 MacBook pros all on a tiny little computer that you could hold in both of your hands. The Drive PX 2 has 12 CPU cores and is capable of 8 teraflops of processing power.
Perhaps the most interesting feature about is that it’s so powerful that it requires liquid cooling so that it doesn’t melt all over the insides of your car.
“All of this within the size of a school lunchbox,” Huang boasts of his creation.
Nvidia claims that to regulate all of the sensors and functionality of a smart car; it needs superior processing power to buy autonomous. That prevents it from requiring pesky cloud connectivity or other dependencies.
Nvidia seeks to use this technology to enable cars to learn from each other and become truly smart. The idea is that all cars will connect back to a mainframe, and each Nvidia-powered supercar will learn from the experiences of the others, thus becoming, even more, intelligent. This feature is being referred to as the “Driven” by Huang and Nvidia.
It’s been interesting to see Nvidia evolve outside of the gaming and performance computing markets to something that’s practical to anyone. And it’s, even more, awesome that they’re able to take what Google started and possibly even improve on it. Let’s just hope that the smart cars don’t get too smart and take over the world.