NVIDIA shows off a water-cooled supercomputer as powerful as 150 MacBook Pros, for self-driving cars.
Yesterday, January 4th 2016, NVIDIA has unveiled its new DRIVE PX 2, currently the fastest, most powerful supercomputer for self-driving cars. The new water-cooled computer boasts the ability to process 24 trillion deep learning operations per second, which is the equivalent of the accumulated processing power of 150 MacBook Pro laptops.
This is the second iteration of the DRIVE supercomputer, whose first incarnation was unveiled last summer, and already adopted by more than 50 automakers and tier 1 suppliers, including BMW, Daimler, Audi, and Ford.
DRIVE PX 2 is capable of processing input from 12 cameras, lidar, radar and ultrasonic sensors. The resulting data is then used to detect surrounding objects, and calculate the optimal traveling speed, as well as braking distance, towards a full range of scenarii:
“...DRIVE PX 2's deep learning capabilities enable it to quickly learn how to address the challenges of everyday driving, such as unexpected road debris, erratic drivers and construction zones. Deep learning also addresses numerous problem areas where traditional computer vision techniques are insufficient -- such as poor weather conditions like rain, snow and fog, and difficult lighting conditions like sunrise, sunset and extreme darkness...” reads NVIDIA’s official press release, published yesterday.
What this means for drivers is a safer approach by autonomous vehicles, in respect to dealing with unforeseen circumstances created by the human element.
Connected autonomous cars
DIGITS is another facet of the technology involved in the creation of DRIVE PX 2, which involves the development of a neural network that is available everywhere, on PCs, supercomputers, and cloud services like Amazon, or Facebook Big Sur. This means that updated neural network data is capable of being pushed to autonomous vehicles, to progressively improve the dataset available to self-driving vehicles, and accelerate their learning process.
"...NVIDIA's GPU is central to advances in deep learning and supercomputing...” NVIDIA CEO Jen-Hsun Huang, ”...We are leveraging these to create the brain of future autonomous vehicles that will be continuously alert, and eventually achieve superhuman levels of situational awareness. Autonomous cars will bring increased safety, new convenient mobility services and even beautiful urban designs -- providing a powerful force for a better future...."
NVIDIA DriveWorks
Developers interested in experimenting with DRIVE PX 2 hardware programming, will have access to NVIDIA DriveWorks’s complete toolkit, which includes drivers, modules and libraries for accelerated computing and testing of the hardware on self-driving vehicles.
Object detection, segmentation and classification, as well and map localization and path planning are among some of the featured modules included in the toolkit, and available to NVIDIA deep learning developers and testers.