As high performance computing technology continues to evolve, so, too, does its role in various industries. In particular, the automotive manufacturing sector has been relatively quick to adopt new HPC solutions as they become available, and to make greater use of these resources, as Greg Clifford, manufacturing segment manager with Cray, recently highlighted for Automotive IT News.
An evolving technology
Clifford explained that HPC tools have become particularly important for automotive manufacturers in the context of crash simulation. The expert pointed out that while the industry has been relying on HPC systems for advanced simulations for quite a while, advances in HPC technology have provided a major boost to these efforts.
"Just five years ago a simulation based on MPI parallelism would use only a few dozen cores," Clifford wrote. "Today, parallel simulations using 256 compute cores are routine and there is a growing demand for scaling to thousands of cores per simulation."
As Clifford noted, this increase in computing capabilities enables automotive manufacturers to create high-fidelity crash simulations that incorporate as many as 10,000 different elements.
This provides tremendous value for the automotive manufacturers.
"Organizations would commonly create and prototype a design and then test it in physical ways to figure out the nuances of how the new design functions – but the costs of this trial-and-error process can be huge," the writer explained. "High performance computing systems enable the use of predictive assessments, eliminating uncertainty when developing new vehicles and creating manufacturing strategies."
The growing significance of HPC solutions in the automotive industry is not likely to taper off in the near future. Instead, Clifford predicted that the focus of these solutions will shift toward predictive simulation. Rather than analyzing products or features after development, predictive software will enable manufacturers to accurately determine the performance of these offerings in the early design stages.
As HPCwire pointed out, these predictive capabilities will become increasingly important for companies in this sector as they continue to rely on new materials, such as steel aluminum alloys and honeycomb materials. These are complex, relatively untested components, so the ability to predict their behavior early is key to efficient utilization.
In order for automotive manufacturers to continue to see these benefits, though, they need to adapt their HPC strategies and continue to pursue the most useful HPC resources.
Clifford argued that the improved fidelity and ensemble analysis capabilities needed to maximize the value of crash and other simulations demands consistently improving computing power. Manufacturers that do not regularly upgrade their HPC resources will struggle to maximize their benefits when pursuing these strategies.
Additionally, organizations need to invest in the right complimentary tools for their HPC systems. Advanced debugging solutions that handle multiple threads, processes, and cores are a key example. These tools are essential for ensuring that a given HPC solution does not suffer from faulty code, memory issues, or other defects. Without such tools, effective, efficient debugging becomes impossible, potentially undermining the value of the HPC tools themselves.