Wisconsin Applied Computing Center harnesses supercomputing for engineering and much more
The Wisconsin Applied Computing Center only launched in late 2012, but it’s already helping researchers across disciplines harness the power of supercomputers.
“More and more people are relying on computers to model and simulate physics of interest, rather than running experimental tests,” says Dan Negrut, an associate professor of mechanical engineering and WACC co-founder. “The mission of WACC is to learn how computers can run virtual experiments that are impossible, extremely expensive, or extremely time-consuming to perform in reality.”
And, as computers take on a larger role in scientific research, it’s useful to bring together people who apply computing in a variety of ways. WACC serves three main purposes: giving researchers access to a computer cluster that boasts 56 NVIDIA graphics processing units and 1,200 CPU cores; promoting the science and engineering that Wisconsin researchers are doing with computing; and creating an interdisciplinary discussion to help researchers—in engineering fields and beyond—make better use of computers.
“The point of WACC is that now there’s a core group of faculty that represents all the major themes in computational engineering,” says Bernard A. and Frances M. Weideman Professor Vadim Shapiro. “All the themes are amplified together, and we can talk about large computational problems.”
WACC’s supercomputer itself ranks among the most powerful on the UW-Madison campus. Negrut and his colleagues built it with the help of a $120,045 grant from the Army Research Office, and used hardware donated by the companies NVIDIA and Advanced Micro Devices.
Negrut and Shapiro make up WACC’s core faculty, along with Mechanical Engineering Assistant Professor Mario Trujillo, Associate Professor Krishnan Suresh and Professor Darryl Thelen, Computer Science Assistant Professor Eftychios Sifakis, and Associate Scientist Radu Serban. Though the group currently is heavy on ME faculty, Negrut points out that it already represents a wide range of research interests that can harness supercomputers, from computational geometry to fluid dynamics. In the ongoing fall 2013 semester, Negrut adds, engineering faculty are using WACC’s computer to teach 40 students about advanced computing.
In fact, WACC has opened up its supercomputer to dozens of collaborators from such fields as chemistry, biochemistry, economics and mathematics. Shapiro will use some WACC resources as he works on an $800,000 National Science Foundation project to develop new computational models for materials and manufacturing, and Thelen and partners have received a $518,066 National Institutes of Health grant to model the effects of ACL reconstruction surgery on patients’ knees. Negrut and Sifakis are collaborating on an Army project to improve the ability to use computer simulation to gauge the mobility of ground vehicles operating on deformable terrain.
WACC’s founders hope to have an impact beyond engineering by focusing on the “applied” part. While researchers have more and more computing power at their disposal, the real long-term task is figuring out how best to use it.
“This thing is in its infancy,” Negrut says. “There’s a lot of hardware that is very powerful, but you rarely have the right software to leverage the hardware. You need to be able to model the physics and write software that mimics that. It’s a tedious process.”
In May 2013, WACC hosted the Wisconsin Forum on Advanced Computing, inviting researchers from academia and industry to give brief presentations on how their research uses computer simulations and other supercomputer applications. Speakers included representatives from 3M, the United States Army, NVIDIA, Caterpillar, Xerox, Boeing, and GlaxoSmithKline.
In summer 2013, Negrut also launched a WACC “summer school” program, recruiting six graduate students to spend four weeks doing research at UW-Madison. One came from Rensselear Polytechnic Institute in upstate New York, while the rest traveled to Madison from Poland, Italy, Spain and the Czech Republic.
Their work included inquiries into granular flow, computer graphics, and finite element analysis, and each student had a specific project and gave a presentation at the end of the program. “The experience can be exclusively about how to use the hardware,” says Negrut. “But I was equally fine with them just taking their existing research and understanding how it can benefit from advanced computing.”
Negrut, who says many European countries don’t yet offer researchers a great deal of supercomputer resources, wants to make sure the students can continue using what they’ve learned once they return to their home universities and to their own graduate research endeavors.
This year, all six visiting students had engineering backgrounds, but Negrut hopes to draw students from other fields, including computer science. WACC also has experimented with hosting one-week outreach programs for high-school students.
As the center nears its one-year anniversary, Negrut is keeping its focus broad, which he thinks that will make WACC stronger in the long run.
“The hope is that as more colleagues understand the potential of computer simulation and modeling, they’ll decide to take advantage of WACC’s accumulated knowledge,” he says.