Q&A with the Winner of the Parallel Universe Computing Challenge 2013

Catching Up with Georg Hager, Captain of Last Year’s Winning Team

By Mike Bernhardt, Community Evangelist for Intel’s Technical Computing Group

As we ramp up our planning for the second annual Intel Parallel Universe Computing Challenge (PUCC) at SC14, we caught up with Georg Hager, team captain for last year’s champion Gaussian Elimination Squad and senior research scientist at Germany’s Erlangen Regional Computing Center. Georg took a few moments from celebrating Germany’s World Cup championship to answer some questions about the PUCC.

Q: Why do you think this event was so popular at SC13?

A: People are interested in performance, that's why they attend SC (mostly). But there is such little competition outside the academic peer review that an idea such as the PUCC was bound to attract large crowds. Let's be honest, most "fun" events at conferences are more or less embarrassing, but the PUCC was different. It was an all-new flavor of fun that was previously unseen at SC. And it wasn't even a part of the official schedule.

Q: Why is the Gaussian Elimination Squad excited about defending their title in the 2014 Parallel Universe Computing Challenge?

A: The PUCC is about showing knowledge and experience in the field of HPC. This is exactly what we are trying to build in the German institutions that were part of the team at SC13, and so we are eagerly waiting for our next chance to show that we have done well on that.

Q: Apart from winning the competition last year, what stands out as a highlight from the 2013 competition?

A: In my personal opinion, one highlight (and a little bit of a surprise) was the large attention the event got on the show floor. There were paper sessions and other events in parallel, but still many people found the PUCC interesting enough to sacrifice a little of their precious conference time and watch the teams fight. And I should add that the German team had a lot of supporters in the audience (including a cheerleader, with pompoms and all 😉 ).

Q: SC14 is using the tagline, “HPC Matters.” Please tell us why you think HPC matters.

A: HPC has become a crucial tool in research and engineering. Using parallel computers and high-performance programming techniques we can tackle computational problems that are impossible to do by hand or on a desktop computer, simulate experiments that are too dangerous or overly expensive to carry out in real life, and even "play god" by messing with the parameters of the universe 😉 All this contributes to the goal of science exploration and discovery to know more.

And knowing more is always better than knowing less. Decades ago we had human computers, slide rules, mechanical and electromechanical computing devices, pocket calculators. Today we have massively parallel computers. It is like building a house with a shovel and a wheelbarrow, and suddenly someone gives you an excavator. Sure it's expensive, but you can do so much more.

Q: How will your team prepare for the 2014 competition?

A: I don't know yet. If the competition is similar to last year's, a couple of drinks and some pizza up front will do 😉

Q: What advice would you give a team considering participation in this year’s competition?

A: It's like Germany vs. Brazil all over again: Just give up. You will be eliminated.

Well it certainly looks like the Gaussian Elimination Squad has thrown down the gauntlet. If you think you can put together a team to take on these tenacious Germans, get more information here or let us know your interest by completing a PUCC information form.

Georg Hager holds a PhD in computational physics from the University of Greifswald. He has been working with high performance systems since 1995, and is now a senior research scientist in the HPC group at Erlangen Regional Computing Center (RRZE). His daily work encompasses all aspects of HPC user support and training, assessment of novel system and processor architectures, and supervision of student projects and theses. Recent research includes architecture-specific optimization for current microprocessors, performance modeling on processor and system levels, and the efficient use of hybrid parallel systems. His textbook “Introduction to High Performance Computing for Scientists and Engineers” is recommended reading for many HPC-related courses and lectures worldwide. A full list of publications, talks, and other things he is interested in can be found in his blog.

Erlangen Regional Computing Center (RRZE) is the IT service provider for the Friedrich-Alexander University of Erlangen-Nuremberg (FAU) in Bavaria, Germany. With its 35,000 students and about 250 professors, FAU is the second largest university in Bavaria. Apart from standard IT services, RRZE procures and runs HPC systems for the use of local researchers. The HPC Services department at RRZE conducts research, user support, and HPC teaching, and has developed special expertise in performance engineering for scientific computing.