NCSA awarded grant to examine effective practices in industrial HPC

01.29.16 -

by Susan Szuch

Being a principal investigator for a $300,000 National Science Foundation (NSF) award to evaluate and share practices in partnerships between high performance computing (HPC) and industry was not on Merle Giles' radar before he started working with NCSA.

"My background is business—it's not technology and it’s not science. My previous lives were not in something related to HPC," Giles says. "But it's the most exciting thing I've done in my life, because it’s very impactful."

The project comes on the heels of the release of the National Science Computing Initiative (NSCI). President Obama issued the NSCI at the end of July to ensure the United States continues to be a leader in the HPC field.

Giles' project fulfills the initiative's need for more effective practices with industry, and will "help inform the community as to what those practices look like."

The project is meant to evaluate current HPC-centered partnership programs, determine best practices in partnerships and to create a report that can serve as a reference for other HPC partnerships. To do this, Giles has to gather data from various HPC centers.

"NSF has a high interest in understanding how industry engages with HPC centers. So our funding was to tell the story and gather data about these most effective and least effective—maybe both ends of the spectrum—to gather stories around effective practices with industry’s engagement with HPC centers," Giles says. "We have our own stories here but we will go around the globe."

In addition to examining how to improve engagement, the project will also examine the economic impact of industry.

"Economic impact has become important so that we can do more than just the science, more than just the research. How can we take that research and go another step forward, and make an impact economically so that economic impact goes through companies?" Giles says.

In the past, companies would be opposed to sharing information with the public because they wanted to keep their information secret from competitors. The book Giles edited with director of Dutch HPC Anwar Osseyran was an early example of information sharing within industrial HPC.

"We've begun to open (information) in what we might say is a pre-competitive way, and share some things that if others would say, 'Ah, that looks like me, and I could do that, too' now that becomes valuable," Giles says.

At the SC conference in 2014, Giles showed others what would be an early draft of the proposal, though it got pushed to the back burner until the release of the NSCI.

"(When the NSCI began), we got into some deeper conversations about how there's an urgency to go with this, so let's do this," Giles says. "So it was very good timing, we were actually spun up and ready to rock and roll, and we twisted it a little bit, and got it to where NSF said, 'Let's go. Let's make it happen.'"