On April 10, teams of high school and college students from across the state gathered at The University of Tulsa for the fifth Oklahoma High-Performance Computing Competition. This competition challenged teams to solve one of two problems using a defined set of hardware resources.
“The Tandy School of Computer Science is proud and honored to have created and to host the Oklahoma High-Performance Computing Competition,” said Associate Professor Peter J. Hawrylak. “Every year, this competition provides superb educational and learning opportunities for students from across Oklahoma to apply skills learned in the classroom to meaningful problems.”
Hawrylak noted that the growing cyber economy requires graduates skilled in high-performance computing and parallel programming and that this competition helps Oklahoma remain competitive within the United States and worldwide. The competition also provides participants with excellent experiences in teamwork and team building as well as networking opportunities with high-performance computing professionals from Oklahoma.
“I am deeply thankful for the support that the Tandy School of Computer Science has provided over the last six years for the competition and for their support of the high-performance computing minor,” said Hawrylak. “This minor program gives undergraduate students in the College of Engineering and Natural Sciences an avenue to obtain skills in high-performance computing.”
Using a Raspberry Pi cluster and specified networking hardware, teams had complete freedom to deploy the resources to produce a computing cluster as well as full control over the software stack, libraries and programming languages they used. This yielded a variety of solutions and solution approaches.
Teams entered at the high school level in Track 2, the Traveling Salesperson Problem; the undergraduate level in Track 1, the determinant problem; and the undergraduate level in Track 2, the Traveling Salesperson Problem. Track 1 required finding the determinant of a matrix in the shortest amount of time. This is a core operation in many of the applications currently run on supercomputers. Track 2 entailed solving the Traveling Salesperson Problem for a list of cities in a specified amount of time. To succeed, teams had to find the best circuit through all the cities (solution) they could within the time limit.
The Panther Super Computing team from Lookeba-Sickles High School took first place in the Track 2 category at the high school level. The Bulldog Supercomputing Team from Southwestern Oklahoma State University took first place in the Track 2 category at the undergraduate level.
Team James from TU took first place in the Track 1 category at the undergraduate level. The team comprised computer science senior James Hale and John Hale, the Tandy Professor of Bioinformatics and Computational Biology and chair of the Tandy School of Computer Science.
“Through competing the high-performing computer competition, I was able to meet new people and learn parallel computing techniques through their presentations,” said James. “This also presented an opportunity for me to work on a high-performing computer problem; test its performance via various metrics (e.g., speedup); and discuss my work with an audience composed of domain experts, beginners and everyone in between. I would encourage anyone interested in high-performance computing to get involved next year.”
Solving problems through collaboration
Brock Associate Professor of Computer Science Mauricio Papa remarked on the wide range of skills that those working in high-performance computing need to develop: “Solving problems in high-performance computing requires special skills. It is not just about knowing the procedure (or algorithm) to solve a problem. It is also about knowing how to program, think in parallel, multitask, communicate and coordinate in a way that efficiently solves the problem at hand using all the resources a computing cluster has to offer.”
John Hale noted how the competition provides an opportunity for participants to develop the teamwork, collaboration and soft skills beyond just technical knowledge that employers highly value: “The wonderful thing about this competition is the emphasis on collaboration and learning. It is competitive, true, but teams learn from each other and, as result, benefit that much more from shared struggles and experiences.”
Are you interested in a career in programming? If so, consider TU’s minor in high-performance computing. Learn more.