Research at national labs by computing faculty members, students contributes to climate change accuracy, supercomputer modeling

Two Grand Valley undergraduate students and one faculty member worked over the summer at U.S. Department of Energy national laboratories with research teams that are operating supercomputers, while another faculty member solidified a partnership for future DOE collaborations.

Their work contributed to long-term projects aimed at improving the accuracy of climate change predictions and testing a new language for a high-performance computer. Students Marc Tunnell and Elise Dettling gained valuable experience learning from top scientists and now have a boost on their resumes for graduate school and beyond.

Dettling and Christian Trefftz, professor of computing, spent the summer at Oak Ridge National Laboratory in Tennessee. The Sustainable Horizons Institute, which promotes diversity within STEM, supported their proposal for the 10-week internship.

At Oak Ridge, Dettling and Trefftz worked in William Godoy's lab, performing experiments on new computer languages that will eventually be used for the DOE's Exascale Computing Project, a supercomputer with implications for advancing precision medicine, regional climate, additive manufacturing, the conversion of plants to biofuels and more.

From left are William Godoy, Christian Trefftz and Elise Dettling at the Oak Ridge National Laboratory in Tennessee.
From left are William Godoy, Christian Trefftz and Elise Dettling at the Oak Ridge National Laboratory in Tennessee. Trefftz and Dettling worked in Godoy's lab, performing experiments on new computer languages.
Image credit - courtesy photo

Dettling, a senior mathematics major, said the first two weeks at Oak Ridge were spent in a classroom learning the concepts related to high-performance computing and getting acquainted with other interns in Godoy's lab who came from research universities.

"It was a beautiful introduction to high-performance computing," Dettling said. "I was then able to jump right in. My Grand Valley education gave me the foundation and the skills that led to my success at Oak Ridge."

Dettling is considering an advanced computer science degree but the idea of switching from math to computer science had given her pause. Not any more. "I was nervous about applying to graduate schools but now I have connections and a few offers from different universities. The experience at Oak Ridge made it possible for me to get into graduate school for computer science and boosted my confidence," she said.

Tunnell spent the summer at the Los Alamos National Laboratory in New Mexico. He worked in Darren Engwirda's lab on developing mesh optimization software for the Perlmutter supercomputer that will increase the accuracy of predicting large-scale weather events.

As a double major in computer science and mathematics, Tunnell said his applied math skills were tested but Engwirda's style allowed, and even encouraged, failure.

"It was really an environment where you were allowed to fail," Tunnell said. "There were lots of dead ends on projects but that was to be expected. We were encouraged to try things." The team built a model program that worked well by the end of the summer, he added.

Tunnell, who expects to graduate in December 2023, said he has an offer to return to Los Alamos next summer. He was encouraged to apply by computing faculty members Erin Carrier and Nate Bowman.

portrait of Michelle Dowling, in a gray jacket
Michelle Dowling, assistant professor of computing, worked remotely over the summer with the Pacific Northwest National Laboratory to solidify details of a partnership between the lab and Grand Valley.
Image credit - Kendra Stanley-Mills

Michelle Dowling, assistant professor of computing, worked remotely over the summer with the Pacific Northwest National Laboratory under its visiting faculty program. Dowling initiated a new research collaboration between PNNL and Grand Valley that will begin in the Winter semester and involve students.

The project centers on user studies to determine the effectiveness of PNNL's new system in helping data analysts understand their data. Dowling said the PNNL experience will lend real-world situations in two of her computer science courses.

“Questions analysts have about their data must be clearly and precisely defined or else you risk misunderstanding the question and not helping them properly," she said. "Clearly and precisely defining a question requires very different thinking than what most students in the class are used to, so having these experiences to draw from will be immensely helpful in motivating these topics in the classroom.”

Trefftz was also grateful for his experience at Oak Ridge and said the group, including Dettling, will present their findings at a conference in January.

"It's amazing that these national labs are carrying out the research that is being done at the country's top universities," Trefftz said. "This really expands the opportunities for our faculty researchers."

Subscribe

Sign up and receive the latest Grand Valley headlines delivered to your email inbox each morning.