NVIDIA releases supercomputer to support Level 5 drones
Nvidia launched its latest Drive PX Pegasus self-driving computing platform at its annual GPU technology conference in Munich on October 10, according to foreign media reports. The platform can handle 320 trillion operations a second, compared to 30 trillion operations a second on the previous generation platform PX 2.
Nvidia says the new platform is sufficient to drive driverless or support full driverlessness at level 5. Danny Shapiro, president of NVIDIA's automotive division, said in an interview that "after the launch of the first-generation self-driving computing platform, we realized we needed to improve performance even further. And PX Pegasus is a true automotive supercomputer" NVIDIA says 20 of the division's 225 customers are using the PX self-driving computing platform to develop driverless taxis.
In order to improve the safety of driverless cars, these vehicles need to be equipped with cameras, LIDAR, radar and ultrasonic sensors to receive up to 1 terabyte of massive data per second.
Shapiro says the Drive PX Pegasus self-driving computing platform is the size of a license plate, while huge amounts of data are processed instantly on this platform, requiring less energy to familiarize the vehicle with road conditions than before.
More than 25 partners are already using NVIDIA's platform to develop fully self-driving taxis, including startups like Zoox, Optimus Ride and NuTonomy. NuTonomy CEO Carl Egnema said in a statement that the Drive PX Pegasus self-driving computing platform coincides with the requirements of the Level 5 autonomous driving class they want to develop.
Shapiro said the Drive PX Pegasus self-driving computing platform will be available in the second half of 2018, and the existing PX2 platform will be upgraded to the new system at that time.
Recommended>>
1、atomikosJTAXA global transaction
2、TensorFlow from 1 to 2Chapter 1 Vanishing Gradients
3、Science Nonsense Series No 44 on Air
4、SuccinctDataStructure
5、Another CAS AI startup surfaces just won the AIC Challenge in visual perception