Red Bull Racing revs up high-performance IT
The Formula 1 racing team is using new grid management software to improve the throughput of its high performance computing environment for car design and analysis.
Red Bull Racing, the Formula 1 race team, has revealed it is using new grid management software to improve the management and use of its high performance computing (HPC) infrastructure.
The team has been has been using a 1,024-core HPC cluster running AMD processors on an IBM server platform for some months now and needed to find a more automated way of running its computational fluid dynamics (CFD) models on the platform.
Steve Nevey, business development director of Red Bull Technology, which manages the technology partnerships of both Red Bull Racing and sister team, Scuderia Toro Rosso F1, told IT PRO that having already used two similar 250-core clusters installed two years go, its latest HPC platform allowed the team to run larger CFD models, faster.
"The problem emerged when the guys would set a big job to run overnight," he said. "Being able to run bigger jobs faster meant they would have to hang around to kick off another job. We needed to automate this with software to handle the scheduling of CFD jobs."
The team took on load sharing facility (LSF) software from specialist provider Platform Computing to manage job scheduling more effectively. As a result, Nevey said it has managed to improve the throughput of CFD jobs on it HPC platform. "We've been able to optimise the AMD/IBM cluster implemented last year to run CFD jobs by 20 per cent," he said.
He explained how maximising CFD analysis throughput is key to successful car design because it allows Red Bull Racing's designers and engineers to streamline their development of key characteristics that simultaneously maximise downforce and reduce drag the cars create on the track, as well as shorten production time.
Nevey said another benefit of the Platform LSF software is that Red Bull Technology is now able to use hardware more efficiently by selecting the best hardware to run simulation processes at each development stage. Maximizing throughput also reduces downtime in support of the development, from building the CFD mesh to post processing.
"We've been able to demonstrate better return on investment," he said. "By optimising the cluster, we are also building fewer 60-per-cent models and optimising the time it allows us to spend in wind tunnel."
Nevey added that the team is now looking to run more structural and vehicle dynamics modelling on the new cluster because of the LSF software's added management capabilities. And it will also look to use it to add the spare processing power of its computer aided design (CAD) PC workstations when not in use to grow its virtual pool of compute power in future.
The ultimate law enforcement agency guide to going mobile
Best practices for implementing a mobile device programFree download
The business value of Red Hat OpenShift
Platform cost savings, ROI, and the challenges and opportunities of Red Hat OpenShiftFree download
Managing security and risk across the IT supply chain: A practical approach
Best practices for IT supply chain securityFree download
Digital remote monitoring and dispatch services’ impact on edge computing and data centres
Seven trends redefining remote monitoring and field service dispatch service requirementsFree download