Advay Shindikar, SDSC Communications, contributed to this story

At Illinois Wesleyan University, Associate Professor of Computer Science Mark Liffiton is taking his deep-learning course to new heights. By utilizing the power of a U.S. National Science Foundation ACCESS allocation on Jetstream2 at Indiana University, Liffiton has created a course to equip his students with the power to tackle real-world problems.

“I designed the course around Jupyter Notebooks and the PyTorch framework, and we initially used Google Colab to access compute and GPU resources we don’t have available on campus,” Liffiton said.  “While the resources provided by Colab are generous, the usage limits proved insufficient even for some of the course assignments, and they would have severely limited what the students could achieve in their final projects.  I transitioned to using an ACCESS allocation on Jetstream2 to let my students properly explore and develop with powerful systems and without fear of being cut off.”

The students in Liffiton’s class were able to use Jetstream2 to meet their hefty GPU requirements. He said that because of this ACCESS allocation on that supercomputer, his students were able to engage much more deeply with deep learning.

“They have the time they need to solve problems and really work with the skills and concepts in the course, and they are able to work with far more complex models and larger datasets than they otherwise could, opening up lots of options for their projects,” Liffiton said. 

Our ACCESS allocation is the key to letting us move beyond the toy models and problems we would otherwise be stuck with. We are very grateful to the National Science Foundation for funding this program that allows for broader, more equitable access to supercomputers.

–Mark Liffiton, associate professor of computer science, Illinois Wesleyan University

Project Details

Resource Provider Institution(s): Indiana U (Jetstream2)
Affiliations: Illinois Wesleyan University
Funding Agency: NSF
Grant or Allocation Number(s): CIS230388

The science story featured here was enabled by the U.S. National Science Foundation’s ACCESS program, which is supported by National Science Foundation grants #2138259, #2138286, #2138307, #2137603, and #2138296.