Facebook Twitter LinkedIn Google Plus RSS

Super-size it: Rutgers says acquisition of a supercomputer will be good for university research, business community

By ,
Peggy Brennan, associate vice president of economic development at Rutgers: “We want to make it affordable for companies to utilize it.”
Peggy Brennan, associate vice president of economic development at Rutgers: “We want to make it affordable for companies to utilize it.” - ()

Rutgers University recently announced its acquisition of a supercomputer that puts the university's calculating ability on the international stage.

But the technology goes beyond giving its scientists new power to crush complex calculations.

According to Peggy Brennan, associate vice president of economic development at Rutgers, the new technology could be a benefit to companies of all sizes in all industries.

“Because the equipment is so sophisticated and expensive, it’s not something that, in particular, a small business could ever afford to purchase,” Brennan said. “And, even for larger companies, it’s an expensive piece of equipment.

“Plus, you need to have a staff that’s able to manage the system.”

The costs are prohibitive. Supercomputers can cost companies as much as $250 million. And that’s before maintenance. Or even energy consumption, which can be millions more annually.

Because of it, Brennan said, the New Jersey Small Business Development Council and Rutgers are currently developing an “Industry User Program” as a financial model for when the technology is made available next month to businesses throughout the state.

The price tag
Peggy Brennan, associate vice president of economic development at Rutgers University, said the price tag of a supercomputer is too much for many businesses.
And it turns out it’s too big even for the Air Force.
Back in 2010, the Air Force needed a supercomputer, but it didn’t have the budget. To solve the problem, it connected 1,716 Sony Playstation 3s. The result was a computational mega-machine known as the “Condor Cluster” that was built for 10 percent of the cost of a regular supercomputer.

“We want to make it affordable for companies to utilize it,” Brennan said.

According to Rutgers computer science professor Manish Parashar, the school hopes to leverage its new capabilities by collaborating with private industry to train the emerging workforce.

“We can train students on how use next-generation technologies and capabilities to better address problems that are relevant to industry today,” Parashar said. “That’s really what we’re hoping to do — build a whole co-op and interactions of research and collaboration around this instrument that we have so that we can build better partnerships moving ahead.”

That’s because the new supercomputer connects Rutgers to a whole new generation of computing technology, according to Parashar.

“The reason we have been so excited about this is exactly that,” he said. “There are two aspects to it: On one side, it allows the physicists, chemists and biologists to do computations and calculations at a scale and a resolution that was otherwise not possible, whether that’s modeling an engine or a chemical process.”

The second prong of the computer’s capabilities? Helping the university’s partners and other businesses make sense of the ever-growing pool of data.

“The way we’ve architected was to add new capabilities for this data-driven world,” Parashar said. “We’ve tried to build a network in the computer that can solve these data-intensive problems.”

According to Parashar, this represents a shift that the role data plays in computer science, one that’s been occurring over the last decade.

What is a supercomputer?
Supercomputers have been around since the 1960s, but, without any personal experience with them, it might be hard to discern exactly what they can do and what the requirements are to operate a computer that superlative.
“One metric people use is how many computations can it do per second,” said Manish Parashar, professor of computer science at Rutgers University. “There’s a list called the TOP500 that takes one computation and sees how fast the different computers can do that computation and use that to rank all the computers in the world.”
The new supercomputer at Rutgers University has found the school ranking among the most advanced supercomputer locations.

Here’s a look at how it ranks around the world:
2nd Among Big Ten universities
8th Among U.S. academic institutions
49th Among global academic institutions
165th Among supercomputers worldwide
Source: TOP500

                                                                                                             

Usage: A sign of the times
With uses that date back more than 50 years, the primary use for supercomputers in each decade can give insight regarding just how far these room-sized computers have come.
1970s: Weather forecasting, aerodynamic research
1980s: Probabilistic analysis and radiation shielding modeling
1990s: Brute-force code breaking
2000s: 3-D nuclear test simulations
2010s: Complex physics, chemistry, astrological model computations and data input and analysis

“What has changed in the past decade or so is that data is not only the output of a computer, but also the input, which is coming from social media,” he said. “So, you have to build a whole new architecture, which is what we’ve been doing here.”

What industries could find applications in this new technology?

Parashar says the applications are vast.

“We work with a lot of different industries already and we’re open to others,” he said. “You can think of the pharmaceutical industry, where there’s a lot of data and trials in the drug design area and personalized medicine.

“There’s a lot of work going on in the engineering and manufacturing industries, where they do a lot of simulation studies.”

Then, of course, there are the marketing and finance applications.

“There’s a lot of work going on in the marketing area where people are doing data mining to find out what consumer sentiments are, how people are reacting to products,” he said. “Finance is full of these applications in the sense of evaluating risk and making predictions.”

Parashar also noted how insurance companies use large swaths of data to determine instances of fraud.

And, for the emerging workforce of Parashar’s current students, that’s where the opportunity lies.

“The larger companies have in-house capabilities that are often not-advertised and are quite comparable,” Parashar said. “And we’re trying to come up with partnerships so private companies can utilize resources like this in private industry.”

Brennan thinks this could be a boom for the state’s businesses.

“All industry is now relying on advanced computation to make their businesses more efficient and effective,” she said. “By having this equipment available to New Jersey industry, I think this is a tremendous opportunity for the state in terms of really building its industry sectors, having economic impacts and really making New Jersey the place to be because of access to this state-of-the-art equipment.”

E-mail to: andrews@njbiz.com
On Twitter: @sheldonandrewj

More From This Industry

Write to the Editorial Department at editorial@njbiz.com

Leave a Comment

test

Please note: All comments will be reviewed and may take up to 24 hours to appear on the site.

Post Comment
View Comment Policy

Comments

close