After spending much of April kicking the tires of his university’s new supercomputer, Manish Parashar and his students spent last week test driving the new tool.
As eager as he is to start using the computer, the Rutgers University professor is just as interested in handing over the keys. The school hopes the computer will open up a new range of partnerships with industry, and boost the university’s reputation and its graduates’ marketability.
The computer — an IBM Blue Gene/P supercomputer known as Excalibur — is the first step in the establishment of the university’s new Rutgers Discovery Informatics Institute, which Parashar will direct. It was purchased as part of a partnership with IBM, and is roughly 10 times more powerful than any other computer on the school’s campus.
Parashar said the computer has a wide range of potential applications. Its computing power has the ability to complete billions of computations at a much faster pace than smaller computers, making it ideal for researchers who need to test millions of scenarios.
Parashar said Rutgers is keen to lend that power to corporate partners, both by handing over the reins and by designing research at a more comprehensive level. To start, Rutgers is targeting the life sciences and financial industries for partnerships, and is creating advisory panels in both sectors to help plan and promote partnership opportunities, and find out what the industries need. The first panel, the life sciences group, meets next month.
Michael J. Pazzani, vice president for research and economic development at Rutgers, said those industries were chosen in part because they’re located nearby, and in part because they each have a need for such computing power.
“In pharma, there are things like gene discovery or molecular modeling that require a lot of computing power,” he said. “In the financial services industry, it might be doing Monte Carlo simulations — looking at your portfolio and assuming the price of oil doubles, what effect will that have on your portfolio?”
The school already is working with a number of companies, including Bristol-Myers Squibb and Xerox Corp. Nathan Gnanasambandam, a senior research scientist at Xerox, in Rochester, N.Y., said having access to computers like the ones at Rutgers is increasingly vital in an industry where quick information is a must.
“Xerox has large banks as clients, and they actually use this information for their day-to-day information,” he said.
Gnanasambandam said the industry is shifting due to the massive amount of data that’s now available, but that data is only valuable if banks can analyze it in time to make relevant decisions.
Pazzani said an even larger shift, called “open innovation,” is also taking place, as traditional barriers to academic collaboration with industry fade as companies outsource more of their R&D.
That’s key, he said, as “no matter how large your company is, there are probably people outside who are smart and who have more specific skills than people within the company.”
Pazzani said universities can be particularly helpful for riskier early-stage research — companies could outsource the initial work, then bring it within the company if that research proves successful.
Pazzani said such collaboration hasn’t always worked, in part because universities were seen as bureaucratic and in part because schools hadn’t developed the right partnerships to work with industry.
“One of the reasons universities have a bad reputation is they first tended to apply the pharma model to all industries,” he said.
The “pharma model” is built around the fact that drug companies spend billions of dollars and several years on their products, so it’s critically important that their research is unique and protected. But Pazzani said other types of partnerships can make more sense outside of pharma.
For instance, “in the tech world, they might be willing to pool the intellectual property,” Pazzani said. “Five companies pay something to join the center … and then all of the companies can use the patents that come out of it. Whereas that would never work in the pharma world.”
Parashar said Excalibur is meant to be the first step in an ongoing process to boost the high-performance computing capabilities of the school. He said Rutgers is hoping to upgrade to a newer model, perhaps within the next year to 18 months. Other computers with different configurations and strengths also could be added in the future.
“This machine was more to get people aware of it, get people using it and get our experience,” he said, “but then we want to grow it into a system that’s significantly larger, both from a data perspective as well as a computing perspective.”
E-mail to: email@example.com
On Twitter: @JaredKaltwasser