NVIDIA has created what it’s calling the world’s 22nd fastest supercomputer. Dubbed the DGX SuperPOD, the supercomputer provides artificial intelligence (AI) infrastructure that meets the demands of the company’s autonomous-vehicle deployment program. The system was built with 96 NVIDIA DGX-2H supercomputers and Mellanox interconnect technology. Delivering 9.4 petaflops of processing capability, it has the power to train the large number of deep neural networks required for safe self-driving vehicles.
According to NVIDIA, AI training of self-driving cars is the ultimate compute-intensive challenge. A single data-collection vehicle generates 1 Terabyte of data per hour. Years of driving over an entire fleet quickly boosts that Petabytes of data. That data is used to train algorithms on the rules of the road and to find potential failures in the neural networks operating in the vehicle, which are then re-trained in a continuous loop.
Powered by 1,536 NVIDIA V100 Tensor Core GPUs interconnected with NVIDIA NVSwitch and Mellanox network fabric, the DGX SuperPOD can tackle data with peerless performance for a supercomputer its size. The system works around the clock, optimizing autonomous driving software and retraining neural networks at a fast turnaround time than. For example, the DGX SuperPOD hardware and software platform takes less than two minutes to train ResNet-50. When this AI model came out in 2015, it took 25 days to train on the then state-of-the-art system, a single NVIDIA K80 GPU. The DGX SuperPOD delivers results that are 18,000x faster. For more info, visit the NVIDIA self-driving cars page . Embedded Technologies Expo & Conference
The inaugural event will take place June 25-27 in San Jose, CA!
Embedded Technologies Expo & Conference (ETC), in the largest embedded and IoT market in North America, is the ONLY event focused on what is most important to designers and implementers – education and training. Attendees will experience over 100 hours of unparalleled education and training covering embedded systems, IoT, connectivity, edge computing, AI, machine learning, and more. Co-located with Sensors Expo & Conference, attendees will have the opportunity to see hundreds of leading exhibitors and network with thousands of industry peers and innovators.