Volkswagen Accelerates Aerodynamics Concept Design with NVIDIA V100 GPUs on AWS

Designing a fuel-efficient and stylish vehicle is a massive undertaking that requires teams from many parts of a manufacturer to work in unison.

To accelerate the process, Volkswagen, Altair, and Amazon recently showed a proof of concept that can more rapidly test the aerodynamics of vehicle designs in simulation.

“As vehicle manufacturers seek to increase the aerodynamic performance of vehicles, they create aerodynamic concept designs and run transient aerodynamic simulations to understand the impact of a particular design on performance,” the team stated in a case study highlighting the work. “

As a first step, Volkswagen turned to Altair ultraFluidX, a computational fluid dynamics solver that leverages the power of NVIDIA GPUs, to accelerate simulation time from days to hours.

“Forward-thinking car manufacturers such as Volkswagen AG have turned to technology to address the inherent conflict between the resources and time required to complete a transient vehicle aerodynamics simulation versus the response time that’s allowable from a design and styling perspective in the early stages of a vehicle development process.”

Using the Amazon Web Services cloud EC2 p3.16xlarge instances, which are comprised of the latest generation NVIDIA Tensor Core V100 GPUs, the team ran 10 parallel 8x NVIDIA V100 GPU servers to perform over 100 simulation runs for two design of experiments (DoE) cycles.

Each cycle needed to be completed within five business days, so speed and reliability were critical.

Diagram: Rear lift coefficient Clr vs. drag coefficient Cd for the Jetta PoC  


“The robustness of the solution was proven when one node experienced a system issue during the DoE, but then restarted itself and resubmitted the job automatically, and then ran and finished at 40 percent reduced speed,” the team stated. “The remaining 9 instances continued their operation properly and compensated for the slow node. So, even with an occurring system issue, the study was able to finish within four days without user interaction–-meeting the criteria outlined for successful PoC execution,” the team explained.  

For the automotive industry, having the ability to simulate the aerodynamics of a vehicle on NVIDIA GPUs is game-changing, the team stated.

According to Altair, using ultraFluidX on GPUs instead of a CPU-based CFD solver, Volkswagen could save up to 70 percent in existing hardware costs. When scaled to the 100+ simulations performed in this work, this can add up to a 5-digit USD amount.

“We were able to run 200 car shape variants in a time frame that would normally correspond to only a few runs with our current operational tools,” says Dr. Henry Bensler, head of CAE Methods at Volkswagen Group Research.

The work was recently published on Amazon’s Web Services Partner Network blog and an NVIDIA blog.

Read more>