Thinking inside the box
Find out how students tackled the ultimate DIY challenge by building a “baby supercomputer” with big potential for their artificial intelligence research.
As SMU’s powerful new NVIDIA DGX SuperPOD supercomputer research system launched on campus, students assembled their own “baby supercomputer.” Small but mighty, it’s capable of running and training artificial intelligence (AI) and machine learning (ML) models with the potential to make an array of modern conveniences even better at what they do.
“Our student team already has access to a really powerful supercomputer on campus, but having this miniature version gives them a chance to administer their own supercomputer, which is a novel experience,” explains physicist Eric Godat ’18. He’s the team lead for SMU’s Office of Information Technology Research and Data Sciences Services and director of its Student Technology Assistant in Residence (STAR) Program.
Godat describes the pint-sized computer as a “sandbox” for testing students’ machine learning models. And the plan is to take it on the road for workshops around campus and beyond to teach how to build models and do AI research.
Multiple NVIDIA Jetson modules – similar to the hardware used in the Nintendo Switch hand-held game console – give the Hyer Performance Cluster, the project’s official name, its strength.
Senior computer science major Conner Ozenne ’23 took on the DIY challenge as a STAR research project.
“It was a great opportunity to apply what I’ve learned in computer science classes,” he says. “It also helped further my understanding of some foundational principles like networking and parallel computing.”
He had to think inside the box – the 14-by-14-by-16-inch transparent acrylic housing, to be exact – to fit all the components together as a cluster supercomputer.
“Initially, a lot of it was just calculations: How can we make this fit in the box?” Ozenne says.
He used the precision laser cutter in the Deason Innovation Gym campus makerspace to fabricate the airtight container.
“All of the sides were jagged puzzle pieces. If the cut was off by a millimeter, they wouldn’t fit together,” he says.
Ozenne also had to handle everything from budget management to supply-chain hiccups like finding the correct parts online. Resources didn’t always post accurate specifications, so there was a lot of returning and reordering.
“Another issue we had to figure out was how to power all these little computers at once without starting a fire, which is always a concern,” Ozenne says.
Good news: It all came together fire free.
Godat, who received his Ph.D. in physics from SMU, leads about six students in the STAR program each year. The “baby” now hums and blinks atop a desk in his space on the first floor of Ford Hall for Research and Innovation. The STAR team is currently benchmarking it – measuring performance and capabilities.
The small-scale computer’s outsized potential comes in sifting through loads of data and finding patterns at the same time – essential operations for training AI and ML models. Virtual assistants like Siri, GPS apps like Google Maps and those annoying CAPTCHA tests that determine you are not a robot are among the millions of examples of AI in daily life.
Simply put: The “baby” will help students discover ways to make machines, apps and more even smarter and more accurate.
NVIDIA DGX SuperPOD fuels a transformational high-performance computing ecosystem for SMU and North Texas. This partnership increases SMU’s current supercomputer memory tenfold and sets the stage for artificial intelligence (AI) and machine learning to run 25 times faster than current levels.
Get ready for jobs that don’t yet exist. The nearly unlimited combinations of majors and minors at SMU will put you on the path to success. Because of the flexibility of the SMU Common Curriculum, you can graduate with multiple competencies that employers value..