SCIENCE

This summer, over 1,000 graduate students and researchers registered for courses offered by the Virtual School of Computational Science and Engineering. The courses were designed for those who want to learn new techniques for applying high-performance computing systems to their work. A team led by the University of Michigan and the National Center for Supercomputing Applications at Illinois coordinated the VSCSE's summer program -- which included significant contributions by staff among 21 geographically dispersed campuses across the country.

This year's program included three, week-long Summer Schools:

• Petascale Programming Environments and Tools
• Big Data for Science
• Proven Algorithmic Techniques for Many-core Processors

Participants attended the Summer Schools at one of ten state-of-the-art, distributed high-definition (HD) classrooms located at academic and research institutions across the country. Technical sessions were presented by leading researchers in computational science and engineering, with hands-on lab sessions that utilized cutting edge, HPC systems provided by TeraGrid resource providers.

Petascale Programming Environments and Tools introduced over 200 participants to the challenges, opportunities, techniques, and resources for scaling computational science codes to perform on petascale computing systems. A follow-up to 2009's Scaling to Petascale Summer School, this course also encouraged participants to bring their own code and apply newly learned techniques during lab time. The course, offered at ten locations across the United States, was taught by experts in their fields from NCSA, the National Institute for Computational Science (NICS), Oak Ridge National Laboratory (ORNL), the Texas Advanced Computing Center (TACC), the Center for Computation & Technology (CCT/LSU), and the Pittsburgh Supercomputing Center (PSC).

Considered the "fourth paradigm," data-intensive computing is poised to transform scientific research. In Big Data for Science, over 200 participants learned how to navigate a computing landscape that is becoming increasingly data-centric and how to employ new tools based on new models of computation. They also practiced using data-intensive computing languages in hands-on programming labs. The course was organized by Geoffrey C. Fox and Judy Qiu at the Pervasive Technology Institute, Indiana University. There were sessions led by 20 experts in the field. It was also offered at ten locations throughout the United States.

The third Summer School offered more than 200 participants an opportunity to gain hands-on experience with Proven Algorithmic Techniques for Many-core Processors. A follow-up to 2009's Many-core Processors Summer School, the course used the CUDA programming model to employ graphics processors for scientific computing. The instructors were Wen-mei Hwu, principle investigator of the Illinois CUDA Center of Excellence and professor of electrical and computer engineering at Illinois, and David Kirk, NVIDIA fellow. The course was offered at ten locations across the United States, with the University of Illinois at Urbana-Champaign serving as the host.

The VSCSE program also offered three online short courses on MPI, OpenMP, and CUDA that were each offered over the span of two weeks and were designed to help participants meet course prerequisites. Over 100 individuals enrolled in each of the online prep courses.