Scholar cluster now features GPUs, increased storage

August 14, 2019

Purdue students using Scholar, a cluster supercomputer ITaP makes available for classroom use, can now use GPUs to study subjects such as machine learning, cryogenic electron microscopy and data science.

Nvidia sponsored Purdue’s most recent student cluster competition team. After the competition was over the company donated the Tesla V100 GPUs the team used in its machine to Scholar.

“This generous donation will open up support for an entirely new subset of our community,” says Preston Smith, ITaP’s director of research services and support.

In the spring of 2019, Scholar supported 79 classes and over 3,000 students.

Wen Jiang, a professor of biological sciences, is using the new GPU-enabled Scholar to teach undergraduate and graduate classes that involve cryo-EM image processing and 3-D reconstruction of the structures of macromolecular complexes and viruses.

In the past, Jiang has taught his classes using Scholar’s traditional CPUs, but as the image-processing software he uses increasingly becomes optimized for GPUs, having access to GPU nodes is crucial to address complex problems and increase the speed of computation.

“The addition of GPUs to the Scholar cluster will allow my students to keep up with the fast-progressing cryo-EM field,” says Jiang.

In addition, Scholar now has increased storage capabilities, thanks to support from The Data Mine, a Purdue learning community that introduces students in a variety of different fields to concepts in data science. The Data Mine is part of Purdue’s Integrative Data Science Initiative to help support and coordinate research, education and broader in engagement in data science, with a goal of bringing data science to all.

Scholar is available to any faculty member at Purdue who is interested in teaching subjects that could benefit from access to high-performance computing. To enroll a course, an interested instructor can use ITaP’s automated add-a-class page for Scholar, simply by providing a CRN number.

To learn more about Scholar, contact Smith, psmith@purdue.edu or 49-49729.

Originally posted: August 14, 2019  12:48pm