Continental Debuts the Fastest Supercomputer in the Automotive Industry – and It’s Built for AI
The automotive industry finds strong use cases for supercomputers, which can help industrial designers do everything from optimizing engines and aerodynamics to running virtual crash tests. Now, German automotive manufacturer Continental is announcing a new leap forward with the debut of the fastest known supercomputer in the automotive industry.
The new supercomputer (as-yet unnamed) is built from more than 50 networked Nvidia DGX nodes, connected by Nvidia Mellanox InfiniBand and based on the Nvidia DGX SuperPOD reference architecture. The DGX nodes (each of which costs, Continental says, about as much as a luxury sports car) are purpose-built for AI – crucial functionality for Continental, which is aiming to strengthen its deep learning chops in order to run smarter simulations and develop future technologies for applications like self-driving cars. Continental expects, for example, that the investment will enable its engineers to run 14 times more autonomous driving experiments.
“The high-end computer will be used for innovative software disciplines such as deep learning and AI-driven simulations,” explained Christian Schumacher, head of program management systems in Continental’s Advanced Driver Assistance Systems business unit. “With the computing power we have now gained, we can develop the modern systems we need for assisted, automated and autonomous vehicles in a much quicker, more effective and more cost-efficient way. We use these to simulate real-life, physical test drives – and need fewer journeys on the actual road as a result. This significantly reduces the time required for programming, including the training of artificial neural networks.”
While the system is based in a datacenter in Frankfurt, that location was chosen specifically to enhance its accessibility to cloud providers – and, by extension, its accessibility to Continental’s engineers around the world. This tracks with Continental’s general business trajectory, with nearly 40 percent of its 51,000 engineers specializing in software and IT – among them, nearly a thousand experts in AI (a number expected to double by 2022).
“Software is the oxygen of the industry,” said Elmar Degenhart, chairman of Continental’s executive board. “It lays the foundation for entirely new services. Value creation with software is recording double-digit percentage growth each year.”
The power of the system is a point of pride for Continental, which is comparing it favorably to systems on the Top500 list of the world’s most powerful publicly ranked supercomputers (the source of its fastest-in-the-industry claim). Continental also touts the efficiency of the system, with the host datacenter using green energy to power the supercomputer and the system’s GPU-driven design proving comparatively efficient.
Header image: Continental's new supercomputer.