The steady advancement of computational and storage technologies available in modern supercomputing is changing how colleges and universities train future scientists, engineers and other professionals. It is also leading to a shift in how they enable faculty members and commercial partners to do the kind of ground-breaking research that has come to be expected in academia. At Cray, we want to both anticipate and embrace advances, and we recently hosted a higher-education roundtable discussion at the XSEDE Conference to engage in some serious talks about how supercomputing technologies need to change in light of new needs in academia.
The discussion was well attended with representatives from colleges, universities and supercomputing labs sponsored by academic institutions. The day got going with discussions from both Intel and Cray about key future decisions pertaining to HPC systems in academia. Jeff Brooks talked about Cray XC30™ supercomputers, while John Lee discussed the CS300™ cluster supercomputers. With these foundational presentations in the morning, the afternoon portion of the meeting progressed to some real-world examples of Cray machines being used in higher education research institutions.
Michael Norman, director of the San Diego Supercomputer Center, talked about how the Gordon supercomputer at the facility, a Cray system, is being used to meet faculty needs and enable research in a variety of scientific fields. Byoung-Do Kim, deputy director of HPC and advanced research computing at Virginia Tech then spoke on the university’s efforts to make supercomputing more accessible to the masses. Rounding out the group, Craig Stewart, executive director from Indiana University, provided an informative peak into how the contributions offered by supercomputers in academia have fueled the institutions major investments in its Cray Big Red II system.
These system- and use-case talks were followed by a glimpse into the technology trends that Cray is following which was provided by Giri Chukkapalli from Cray’s CTO Office.
The panel discussion, which I moderated, was an engaging finish of the event. The panelists: Jay Boisseau (TACC), John Towns (NCSA), Thomas Hauser (University of Colorado), Barry Schneider (NSF), and Greg Peterson (UT/NICS) discussed the transformative nature of big data and data-intensive computing in academic research centers. This led to talks about the complexity heterogeneous computing concepts and the unique challenges of small institutions entering the HPC game. We also spent a long time talking about the growing need for newer programming languages, something that Cray is working to address through its Chapel initiative.
All told, the roundtable gave listeners valuable information on how they can take advantage of supercomputers and allowed us to get direct feedback from academic institutions about what they really need and how we can deliver it more effectively.
The entire day was a worthwhile experience for everybody involved. From Cray’s perspective, we were thrilled by the opportunity to educate attendees on our entire portfolio of solutions, including devices specifically designed for the kind of problems higher education research organizations are facing. We also enjoyed hosting a forum for real peer-to-peer discussions about HPC implementation in the sector, an opportunity that is a rare, and highly valued.
Talking about broad, overarching industry trends was also important. Having well-informed participants and guests let us get into some meaningful conversations about where academic research institutions are heading, how supercomputing is progressing and what we can do to align our strategies accordingly.
At Cray, we value the opportunity build a foundation for innovation in the sectors we support. As a result, we are incredibly thankful for all who attended this event and brought their unique perspectives to the table. In particular, I would like to thank the speakers who graciously gave us their time and energy to provide the foundational material to make the event a success.