Blue Waters Symposium a success | National Center for Supercomputing Applications at the University of Illinois
Blue Waters Symposium a success
05.28.14 - Permalink
by Nicole Gaynor
The symposium, held May 13-15 in Champaign, Ill., gathered many of the country’s leading supercomputer users to share what they have learned using Blue Waters and discuss the future of supercomputing.
On May 13, 2014, Blue Waters supercomputer users and many of the NCSA staff who support their work converged in Champaign, Ill., for the second annual Blue Waters Symposium. The ensuing three days were filled with what many of them would later refer to as a wonderful variety of science talks and opportunities for networking and collaboration.
Efficient discovery through supercomputing
Argonne National Laboratory's Charlie Catlett gave the first plenary talk on the impact of supercomputing on studies of the natural and built environment of urban areas. Catlett commented that “success creates opportunity.” In the context of Blue Waters, the success that is the supercomputer created opportunities to answer questions that were previously out of reach, which lead to opportunities to improve medicine, engineering, and a host of other fields.
Indeed, the science talks ranged from high-energy physics to molecular dynamics to climate science and even political science.
Slide after slide, Blue Waters Professor Klaus Schulten, who is also a professor of biophysics at the University of Illinois and researcher at the Beckman Institute, gave examples of discoveries that Blue Waters has enabled: HIV infiltrates human cells and the knowledge to formulate a vaccine for Rabbit Hemorrhagic Disease, among others.
Tom Jordan, director of the Southern California Earthquake Center and professor of Earth sciences at the University of Southern California who is studying earthquakes by modeling the physical waves from the earthquake, noted that the GPU+CPU architecture on Blue Waters reduced his wall time by a factor of four compared to CPUs alone. Brian Thomas, professor of mechanical science and engineering at the University of Illinois, also remarked that his code ran 20 to 30 times faster on Blue Waters thanks to GPUs.
Blue Waters enables more efficient science, summarized Paul Woodward, professor of astronomy at the University of Minnesota. Researchers can run simulations quickly and then have more time to draw meaning from the results while someone else runs their simulations. NCSA Director Ed Seidel added that big computing and big data will revolutionize science, whether physical or social, by making possible the formerly impossible. Many problems are too complex to solve without such resources.
A few talks touched on social sciences that initially seem incongruous with supercomputing. For example, Shaowen Wang, director of the new CyberGIS Center at the University of Illinois, is leading an exploration into minimizing bias in voting districts. Later in the same session, Keith Bisset, of the Network Dynamics and Simulation Science Laboratory at Virginia Tech, said he simulated one scenario of disease propagation for the entire U.S. population for four months in just 12 seconds using 352,000 cores. He estimated that the world population would take 6-10 minutes per scenario, though he emphasized that a realistic assessment of disease threat would require many such runs.
Shared evenings, common goals
The most popular speaker of the symposium was Irene Qualters, the National Science Foundation Division of Advanced Cyberinfrastructure director. She spoke Thursday morning about the future of supercomputing at NSF and encouraged users to work with the NSF to ensure that the future of supercomputing met their needs. The symposium and NCSA’s Private Sector Program annual meeting met for dinner Tuesday at Allerton Park and Wednesday at Memorial Stadium, combining the most advanced computational science teams in the country, according to Seidel.
Seidel remarked after Wednesday’s dinner that he heard a common need from PSP and Blue Waters users: an all-around system that not only can run simulations, but also analyze and visualize data. One of the challenges in the era of Big Data is transferring data between specialized systems.
Science talks throughout the symposium bespoke the advances that Blue Waters enabled. Additionally, researchers envisaged what they could achieve with the next generation of supercomputers, looking forward to the future of large-scale computing. Three working groups discussed Blue Waters 2.0, many-core computing, and data @scale and will release reports in the coming weeks on their findings.
NCSA plans to continue these conversations and connections next year at the 2015 Blue Waters Symposium and lead large-scale computing through the petascale and beyond.
Blue Waters users consist of 120 teams and 575 researchers from 56 universities across the United States, in addition to international collaborators, said Seidel. More than 180 people attended the symposium.
The working group reports, extended abstracts from the science teams, and several other articles that look back on Blue Waters’ first full year are being compiled and will be available in a single publication this summer. Look for Proceedings of the 2014 Blue Waters Symposium on the Blue Waters portal or in print from NCSA.
Blue Waters is supported by the National Science Foundation through awards ACI-0725070 and ACI-1238993.