UW expands effort to serve advanced computing needs in research
Miron Livny, a UW–Madison computer science professor, is pictured near an enclosed bank of distributed computing equipment in the Computer Sciences and Statistics Building.
If you’re conducting quantitative research at the University of Wisconsin–Madison, regardless of discipline, chances are there is a “next level” of discovery that could be unlocked through advanced computing.
UW-Madison is part of a six-university consortium supported by the National Science Foundation (NSF) working to eliminate barriers to entry, such as effort, cost and technical know-how, for any scholar looking to tap the potential of grid computing technologies and capabilities.
“Our goal is not only to increase the overall adoption of advanced computing infrastructure on campuses, but to encourage more people from areas traditionally not involved in computing to use these resources,” says Paul Wilson, chair of UW–Madison’s Advanced Computing Infrastructure (ACI) committee and UW–Madison principal investigator of the award.
Paul Wilson
“We’re not just looking at people who are using 100,000 computing hours a day, but people who may just need 10,000 hours a year,” Wilson adds. “By helping them use some of the larger-scale resources we can offer, they can maybe transform the way they do their research.”
The program, announced in March, will fund 12 new facilitator positions (two at each university) to serve as a bridge between researchers and advanced computing resources both on and off campuses, including the Open Science Grid. Other participating campuses in the $5.3 million NSF award include lead institution Clemson University, Harvard University, the University of Utah, the University of Hawaii, and the University of Southern California.
The two positions for UW–Madison will triple the current capacity of campus facilitators. These positions will consult with researchers to identify custom hardware and software solutions to meet research needs and guide the development of new capabilities.
Miron Livny, chief technology officer for the Morgridge Institute for Research (MIR) and the Wisconsin Institute for Discovery (WID), serves as chief scientist for the consortium. Livny’s Center for High-Throughput Computing (CHTC) has been offering this type of facilitation service for more than a decade. In 2013, CHTC supported more than 100 UW–Madison research projects and delivered more than 110 million hours of computing time.
“I always say that if the issue is hardware, the problem is already solved,” Livny says. “Hardware is easy. It is people who are the glue that bring the power of our advanced computing capabilities to the domain scientists.”
The facilitators should work to execute each project in a way that limits the time commitment and cost for researchers.
“For us, the key metric is not how long it took to get the computing done, but how many human interventions were needed to complete the task,” Livny says. “We’re trying to push the number of human interventions to zero.”
Lauren Michael, research computing facilitator for CHTC, says the diversity of UW–Madison research projects has continued to grow. Recent examples include a study on the dynamics of galaxies, from formation to death; a simulation of the optimal parameters for engine efficiency; and an effort to improve cochlear implants by analyzing how the brain processes sound.
The consortium will create a professional network among the facilitators across all six campuses. This peer group will work together and bring different knowledge and expertise to the greater whole.
The Discovery building houses a significant amount of centralized computing resources offered at UW–Madison, and will also serve as home for the facilitators.
Bruce Maas, chief information officer for UW–Madison, sees strategic value in the campus connection to this project, formally known as the Advanced Cyberinfrastructure–Research and Educational Facilitation (ACI-REF) Consortium.
“We are jump-starting a national community of individuals whose role is to support faculty computation needs. The national community is critical to facilitate the frictionless exchange of emerging practices,” Maas says.
“When university faculty and enterprise IT staff, the federal government, and the private sector pool their collective resources, we can accomplish things that none of us could do alone,” adds Maas.
For more information on how you can tap into advanced computing resources, contact Lauren Michael at lmichael@wisc.edu, 608-316-4430.