High-Performance Computer Aids Researchers Across Campus
Ten AU researchers have been quietly celebrating the purchase of a new high-performance computer (HPC), which will be fully operational by October 2011. Now they would like to let the secret out.
The refrigerator-size HPC consists of a mass of cables and wires that connect “blades” of CPUs—stripped-down computers optimized to maximize performance and minimize use of physical space and energy.
“Each blade can do eight computations at a time,” says economics professor Mary Hansen. “We will get at least 15, but perhaps more, blades yielding at minimum 120 computations at once.”
Translation: the HPC will mean a 60-fold increase in efficiency for university researchers. Tasks that typically would take weeks on a standard computer will now take hours, boosting the rate of research and overall productivity.
Funded by a $260,745 grant from the National Science Foundation, the supercomputer supports an AU strategic goal, says Hansen, the principle investigator.
“This computer allows researchers across many disciplines to do research that we couldn’t do before.” She explains, “AU’s recently adopted strategic plan calls for faculty to do research that has high impact, but AU could neither attract nor retain truly talented computational researchers without a system like this. Particularly in terms of policy-related computational research, that is hugely important.”
From science and technology to the humanities, says Hansen, the HPC “increases productivity and quality in every field that’s touched by computational work.”
Economics professor James Bono, for example, is researching how to allocate a limited number of airplane landing sites when airports are disrupted by weather or other circumstances. How do you maximize benefits to consumers and minimize losses to airlines? “His research will have broad impact because it will create a means by which researchers and policy makers can design better functioning, continually improving regulated systems, such as energy grids and air traffic control,” says Hansen.
Some of the most interesting projects involve data visualization. SETH dean Sarah Irvine Belson will create a tool for analyzing the actions of students with behavioral disabilities and rendering them in three-dimensional animation, which will require intensive computation. But she won’t have to wait long for the computer to “draw” images; they will just pop up. Animation will allow teachers and students to jointly observe behavior patterns and devise more effective teaching methods.
Chemistry professor Kathryn Muratore’s research could potentially lead to the design of new enzymes useful in drug development and in water cleanup. Her computational method combines information from sequence databases, substrate specificity data, and molecular structures from the RCSB Protein Data Bank. Muratore will use the HPC system for programming and testing code to automate 3D visualization of the covarying residues within the protein and simulate the protein’s stability.
Public health is another area of research to benefit from the HPC. Betty Malloy, professor of mathematics and statistics, works on epidemiological models of the effects of exposure to carcinogens. With the new computer system, she will be better able to test models that calculate the odds of someone becoming ill after exposure.
Hansen and her colleague, economist Kara Reynolds, share the long-term goal of informing public debate on policies that impact children, the poor, and the unemployed. Through their structural modeling project, they will be the first to measure the extent to which eligibility for federal support influences services provided to all children in foster care and their health outcomes.
“AU is making investments in the type of infra-structure that pushes all schools forward, including all of the academic clusters within the College,” Hansen says. Everyone will benefit from this machine, even researchers outside AU. “You can’t get this type of award without having broad impact and contributing to the public good.”
The next steps are installation of the HPC and training faculty and student researchers how to use the system. The real joy, says Hansen, has been to see people come together around a research goal and talk to each other across disciplines.
“Professor Muratore in chemistry and Professor Irvine Belson in education, for example. Both want to use the computer for data visualization. If it weren’t for their work on this project, they would not have met each other. They’ll now be able to solve their problems much more quickly as part of a community of researchers.”
“What’s most exciting is that this project will bring many very talented researchers together to solve problems,” says Hansen. “That will have much more impact than any of us working alone.”
Seminars on use of the HPC system will be offered in spring 2011 and continue through the 2011–2012 academic year. For information, email firstname.lastname@example.org.