Customer Solution Case Study
NASCAR Team Turns to High-Performance Computing to Sharpen Competitive Edge
Overview
Country or Region: United States
Industry: Media & Entertainment—Automotive
Customer Profile
Chip Ganassi Racing is a competitive automobile racing team that competes in the NASCAR Nextel Cup, NASCAR Nationwide, Indy Racing League, Indy Pro, and Rolex Grand-AM series. The team has 425employees and is based in Concord, North Carolina.
Business Situation
Run times for vehicle simulations took upto 24 hours on an engineer’s laptop, which prevented Chip Ganassi Racing fromgetting the most out of its simulation software.
Solution
The team adopted a high-performance computing solution based on Windows® Compute Cluster Server 2003, which is helping the team be more competitive by enabling it to run multiple simulations before each race.
Benefits
n Racing simulations are 38 times faster
n Broad application support extends rangeof uses
n Easy IT management and administration / “With simulation times reduced from 24 hours to about 30 minutes, we can run multiple simulations for each race and better tune the simulations for each car, track, and set of track conditions.”
Mark Paxton, Research and Development Engineering Manager, NASCAR Team, Chip Ganassi Racing
Although Chip Ganassi Racing has used simulation software for several years, the software’s usefulness was diminished by the massive computing power required. A single simulation run could take up to 24 hours, which restricted the team to only one run per car before each race and limited the number of variables that could be analyzed. Chip Ganassi Racing now performs simulations on Windows® Compute Cluster Server 2003 running on Dell servers, which has reduced simulation times from 24 hours to about 30 minutes. The team now can run multiple simulations for each race and better tune the simulations for each car, track, and set of track conditions, enabling it to be more competitive at the track. Chip Ganassi Racing is already finding additional uses for the cluster, which is easily managed and supported in the same way as its other Windows-based systems.
Situation
Since its founding as a one-car Indy Car team 18 years ago, Chip Ganassi Racing has grown into a highly competitive racing team that competes in the NASCAR Nextel Cup, NASCAR Nationwide, Indy Racing League, Indy Pro, and Rolex Grand-AM series. Chip Ganassi Racing has won more than 80 races and seven championships, and continues to add to its winning record—most recently with former F1 star and NASCAR rookie Juan Pablo Montoya winning the Toyota/Save Mart 350 at Infineon Raceway for his first NASCAR Nextel Cup.
Throughout Chip Ganassi’s ownership of the team, he has continued to innovate, constantly searching for ways to make his cars go faster. Ganassi is recognized as an early adopter of driver and pit crew development programs, and was one of the first team owners to create an advanced research and development group. He also built in-depth partnerships with many of his technology industry sponsors to benefit from their expertise.
One such application of technology is the use of computer simulations to optimize performance at the track. By modeling track characteristics, expected race-day conditions (such as temperature), and how performance under those scenarios is affected by wheel camber, tire pressure, spring rates, and other variables that affect suspension geometry, teams can come up with optimal starting configurations that can help them win races.
Although Chip Ganassi Racing has used such simulations since 2000, in the past, the software’s usefulness has been limited by the large amount of computing power required. “We used to run the simulations on an engineer’s laptop,” says Mark Paxton, Research and Development Engineering Manager for the NASCAR Team at Chip Ganassi Racing. “It required up to 24 hours for a single simulation, which essentially limited us to a single pass per car per race. Even then, we were limited to sweeps of limited granularity for only 12 variables, even though there are more than 50 settings on a car that we can adjust. The process was too slow to run sweeps with a finer granularity, examine more variables, or rerun a simulation on-the-fly if track conditions changed or the team requested it—all potential capabilities that would have helped us to compete.”
Solution
Chip Ganassi Racing worked with Microsoft and Stackpole Engineering Services, the vendor of its simulation software, to modify the software to run on Windows® Compute Cluster Server 2003. The team now runs simulations on a five-node high-performance computing (HPC) cluster, with run times for a simulation pass reduced to less than an hour.
Project inception dates back to the 2006 racing season when a Microsoft executive was talking to a member of the team at a NASCAR race. When the conversation turned to the team’s use of simulation, the Microsoft executive suggested that the team try his company’s new HPC offering. “Two years ago, this project wasn’t even on my mind,” says Paxton, “in that I thought HPC was only for big companies that can afford multimillion-dollar supercomputers. After we realized that HPC was accessible and affordable, we saw the potential competitive advantage it would bring and started pushing hard. My only concerns were that the solution be easy to use, easy to manage, and that we could easily modify the application—and Windows Compute Cluster Server 2003 enabled us to meet all three requirements. We never really considered Linux, which we don’t use at all.”
After a few phone calls, a Microsoft Senior Technical Evangelist paid a visit to Stackpole Engineering Services, where he helped developers modify the company’s software to run on the cluster. An engineer from Dell worked with the project team to understand how the cluster would be used, and then provided a hardware design and sizing recommendations for the cluster, suggesting that Chip Ganassi Racing use quad-core Xeon processors to maximize processing power. “Dellmade the process of sizing the HPC cluster and specifying the hardware simple and easy,” says Paxton. And it didn’t take long to get our existing simulation software running on Windows Compute Cluster Server 2003. End to end, the entire project only took a few weeks.”
March 2007 marked the first successful simulation on the team’s new HPC cluster, which consists of five Dell 1950 server computers, each configured with two quad-core Xeon processors and eight gigabytes of RAM. During that initial test run, a simulation that used to take 24 hours was completed in 37 minutes and 30 seconds—or 38 times faster.
“Today we use the cluster for all five of our NASCAR Teams,” says Paxton. “Each week, the teams provide their starting setup for each car to a central simulation support group, which uses the cluster to provide the teams with a printout that specifies each car’s sensitivity to setup adjustments.”
Benefits
By taking advantage of high-performance computing, Chip Ganassi Racing can run simulations faster and more often, enabling the team to more efficiently optimize race car performance and ultimately be more competitive at the track. Thanks to the broad application support for Windows Compute Cluster Server 2003, the team is already finding new uses for the cluster that will further increase its ability to compete. And, the team’s IT staff is finding that Windows Compute Cluster Server 2003 is easy to manage and support.
“We’re all about making cars go faster, not building fancy computer systems,” says Paxton. “With Windows Compute Cluster Server 2003, we were able to easily put a solution in place that helps our teams better prepare for race day. I can’t say that we’ve won a race solely because of those new capabilities, but they definitely have helped. There are all kinds of factors aside from suspension geometry that decide who wins a race, and we now have more time to focus on those other things.”
Racing Simulations Are 38 Times Faster
Chip Ganassi Racing is taking advantage of the faster simulation times provided by its HPC cluster to get more out of its simulation software, in turn improving the team’s ability to compete. The ability to run simulations virtually on demand, take into account a greater number of variables, and increase the granularity of parameter sweeps are all ways that the team is turning the capabilities provided by Windows Compute Cluster Server into a competitive advantage.
“With simulation times reduced from 24 hours to about 30 minutes, we can run multiple simulations for each race and better tune the simulations for each car, track, and set of track conditions,” says Paxton. “Faster simulation times also leave our car teams with more time to apply the results and rerun simulations if issues arise at the track or expected race-day conditions change.”
Broad Application Support Extends Range of Uses
Thanks to the broad application support provided for Windows Compute Cluster Server 2003, Chip Ganassi Racing is already finding additional uses for its HPC cluster. For example, one of the team’s research and development engineers is working on a new, proprietary simulation program that will deliver additional capabilities above and beyond what the company can do today.
“We’re taking advantage of the Microsoft HPC porting lab and the ease of development on Windows to create our own state-of-the-art simulation software,” says Eric Strauss, a Chip Ganassi Research and Development Engineer based at the team’s Indianapolis location, who is developing the new application. “It will provide more detailed simulation and analysis capabilities, which we’ll be able to make easily accessible to track engineers.”
Easy IT Management and Administration
Just as Paxton’s team—and the race car engineers it supports—is benefiting from the solution’s performance and ease of use, IT personnel are thankful for its ease of management and support. “The whole idea of an HPC cluster was kind of mysterious at first, leaving me wondering how we would manage and support it,” says Michael Carbone, IT Manager at Chip Ganassi Racing. “However, the process is no different than for any other of our 40 or so Windows-based servers. We control access to it with Active Directory® [directory service], apply software updates with Windows Server® Update Services, and back up the cluster in the same way that we do for all other systems. From an IT perspective, our new HPC cluster hasn’t created any more work or caused any problems.”
Carbone, himself a race fan, also has a personal perspective on the improved simulation speeds the team’s new HPC cluster is delivering. “It used to take 24 hours to simulate traveling 100 miles on a track, which is the equivalent of going about 4 miles per hour, or as fast as I can walk,” he says. “Today, the same simulation runs at about 200 miles per hour. Now that’s NASCAR speed!”
Microsoft Server Product Portfolio
For more information about the Microsoft® server product portfolio, go to:
www.microsoft.com/servers