Texas A&M Researcher Receives $1.3 Million to Make Supercomputing Easier
COLLEGE STATION -- Fifteen years ago, undergrad Wolfgang Bangerth immersed himself in creating software that would give researchers working on complex equations containing billions of variables additional ease and efficiency.
Now a mathematics professor at Texas A&M University, Bangerth boasts a robust software program, deal.II, used by hundreds of researchers around the world studying a variety of topics, from plant root growth and glacier mechanics to heart muscle fiber simulation and even the impact of air pollution on Roman statues.
Bangerth's globally beneficial efforts recently were enhanced with a $1.3 million grant from the National Science Foundation (NSF) to expand the software‘s uses and make it more adaptable for and accessible to researchers.
"For us, this is validation that what we are doing is right," Bangerth said. "We're helping others. We want to provide the infrastructure for other researchers in as broad a way as possible. We want to give researchers the building blocks they need to do their research. There's a need for this. There's not a lot of software for people doing research on projects that have billions of variables."
Researchers working on such complex problems use specialized computing clusters -- essentially tens of thousands or more computers stacked and grouped together to pool processing power. Some of the large supercomputing operations use more than a million processor cores, usually in warehouse-like rooms equipped with a powerful cooling system. For example, a recent, partly NSF-funded addition at the University of Texas at Austin, Stampede, has 102,400 processor cores.
The specialized research done on these systems has world-changing potential. Bangerth, for instance, is working with a center at the University of California, Davis to study a hypothesis about the cause of earthquakes, an area that so far has baffled scientists. The computing power is needed to accurately model the turbulent convections of heat ebbing and flowing from the Earth's bowels to its surface -- a task Bangerth says requires at least a thousand computers to provide a simulation with adequate resolution.
Although computing power is important, Bangerth notes it's the software that helps researchers harness the true potential of this hardware, which is accessible to all but of maximum benefit to a handful of research groups with the specialized knowledge, ability and resources to write the necessary high-level code from scratch.
Enter Bangerth, his team and deal II.
Given his expertise in computational mathematics, Bangerth is uniquely positioned to write such software that could be accessible to far more researchers. He's interested in programming, and, although by the strictest of definitions is a mathematician, he is at heart an engineer wanting to solve real-world problems. Because he also is well versed in many subjects, including physics and chemistry, Bangerth offers an interdisciplinary understanding of the myriad computing needs of sundry researchers.
"I think we made a name for ourselves as a group of people who provide high-quality software and have the expertise," Bangerth said. "When we wrote that proposal to NSF, that's what they recognized: Here are people who could provide an essential service to the community by creating these building blocks, and they have the expertise to do it."
Texas A&M's internationally renowned core expertise in computational mathematics represents the fruition of a future rooted in the past -- specifically, the late Dr. Richard E. Ewing, a former dean of the College of Science and prolific visionary who was one of the first mathematicians to embrace what was then an uncharted interdisciplinary area. As a convert to academia from the corporate world, Ewing recognized the need for mathematicians capable of keeping one foot in their field while maintaining the flexibility to put the other into applied areas, from modeling to computing, that could benefit both industry and society.
Bangerth said a couple dozen researchers across the world currently help him with refining the software. The trio that leads the effort includes Bangerth, Texas A&M visiting assistant professor of mathematics Timo Heister and former Texas A&M mathematics professor Guido Kanschat, who was Bangerth's adviser when the undergraduate devised deal.II. On this recent NSF grant, he'll be assisted by Yassin Hassan from the Texas A&M Department of Nuclear Engineering and Michael Heroux from St. John's University.
The software is known as deal.II because it builds on an earlier program, DEAL, that Kanschat helped create. The team that devised it dissolved in the late 1990s, and that's when Kanschat and Bangerth started spending several hours a week discussing deal.II. Kanschat says Bangerth had the vision to make the software library more broadly accessible.
"We turned out to be a perfect team," said Kanschat, now a mathematics professor at Heidelberg University, where Bangerth earned his doctorate in 2002. "I brought in the older experience and mathematical knowledge, and he the vision for dissemination and application and knowledge about new concepts in software design. It's fair to say that deal.II not only made a scientific impact, but also changed the lives of many graduate students and postdocs, since it allowed them to focus on interesting mathematics instead of reinventing the wheel."
The software is open source, meaning it's available for free to anyone.
"It's a way for us to share what we do as academic researchers," Bangerth said. "My personal view is, I'm paid by the taxpayer, and whatever I do should be out there in the open for everybody. That's my goal: creating knowledge for the taxpayer. That's why we do it open source."
For his work on deal.II, Bangerth and his team were the recipients of the 2007 Wilkinson Prize for Numerical Software, an award given every four years for an outstanding piece of numerical software. The honor, named after numerical analyst James Hardy Wilkinson, is given by Argonne National Laboratory, the National Physical Laboratory, and Numerical Algorithm Group.
Bangerth has been a member of the Texas A&M Department of Mathematics faculty since 2005. Prior to that, he spent three years as a postdoctoral research fellow at the University of Texas at Austin with joint appointments at the Institute for Computational Sciences and Engineering and the Institute for Geophysics.
To learn more about Bangerth's research and deal.II, visit http://www.math.tamu.edu/~bangerth/.
For more information about Texas A&M Mathematics, visit http://www.math.tamu.edu/.
# # # # # # # # # #
About 12 Impacts of the 12th Man: 12 Impacts of the 12th Man is an ongoing series throughout the year highlighting the significant contributions of Texas A&M University students, faculty, staff and former students on their community, state, nation and world. To learn more about the series and see additional impacts, visit http://12thman.tamu.edu/.
About Research at Texas A&M University: As one of the world's leading research institutions, Texas A&M is in the vanguard in making significant contributions to the storehouse of knowledge, including that of science and technology. Research conducted at Texas A&M represents an annual investment of more than $700 million. That research creates new knowledge that provides basic, fundamental and applied contributions resulting in many cases in economic benefits to the state, nation and world. To learn more, visit http://vpr.tamu.edu.
Contact: Vimal Patel, (979) 845-7246 or email@example.com or Dr. Wolfgang Bangerth, firstname.lastname@example.org