The first Duke Research Computing Symposium took place on 8 January, and it was a great success — actually exceeding expectations with heavy turnout, fine talks, and a busy poster competition.

Actually, I will confess that I had difficulty setting my expectations. I knew from many, many conversations that happenings in research computing at Duke were on people’s minds, and I knew of solid and growing support from Duke’s administration and leadership for research computing on campus. But I also fretted about the time of year; we began planning for the event shortly before Thanksgiving break when I knew from experience that the normal inertia of Fall Semester turned attentions to wrapping up classes, completing finals, and heading off for the holiday break.

Who in the word would spend time away from vacation to pull together a poster? Who would even bother to register for an event held right at the onset of the Spring Semester?

Who? Well, many people, it turned out. Over 160 people attended the session. (I figured 100 would be just nifty.) We had 33 posters in the competition, 13 more than my modest target. And we had fourteen displays from groups on campus with service that were relevant to researchers with a computational bent. We filled up the spacious “Great Hall” of the Mary Duke Biddle Trent Semans Center for Health Education on Duke’s West Campus.

I was particularly impressed by the energy of the crowd and its relative diversity in terms of disciplines and types of research. The Duke School of Medicine — which in the past was less well represented in University offerings in research computing — counted about one-third of the total attendance. There were substantive (often quite technical and sometimes animated) exchanges among people conducting computationally intensive research. People who had no other opportunity to see each other’s work … well, they saw each other’s work.

The poster session, of course, had Amazing Prizes, but despite their amazing qualities the real delight was in the exchange — something that all who attended witnessed. University leaders took the role of judging, and the judging process was designed to be more-or-less invisible and non-stressing — seamless with the normal scholarly interactions of doing an event. The judges were Prof. Larry Carin (ECE, Vice Provost for Research), Prof. Craig Henriquez (BME), Dr. Julian Lombardi (CS, OIT), and Prof. Thomas Nechyba (Economics, SSRI). They had the tough job of identifying three of the posters for the sake of awarding the prizes — no mean task given the talents that we saw in the session.

First prize (four-year use of a beefy node on the Duke Compute Cluster)

Christa Kelleher, Brian McGlynn, Thorsten Wagener. Refining process representation in high-resolution models of headwater catchments using internal catchment diagnostics.

Second/Third (virtual machine or data storage)

Thomas Joseph Cahill, Alex Thomsen, Georgis Skiniotis, and Robert Lefkowitz. Functional consequences of different conformational states of [Beta]2-Adrenergic Receptor — [Beta]-Arrestin complexes.

and

Curtis Lee, John Dolbow, Hans Johansen. The narrow-band gradient augmented level set method and applications.

“Lightning talks” by researchers made up much of the proceedings (agenda is here), and these were accessible and enjoyable (sometimes a difficult thing to manage among a diverse group of scholars).

And next year? We’re doing it again — next time with greater lead time and well practiced execution, we think. I’d like to see the range of scholarly disciplines that are involved widen significantly, especially among the social sciences and the humanities. And perhaps structuring the offerings to attend to more specific interests would be good. Do you have an idea? If so, let me know. I’ve got a few months to noodle and plot and plan.

Mark R. DeLong (mark.delong@duke.edu)