Solving some of the world’s most complex scientific and business problems isn’t a task for the faint of heart. These days it requires massive amounts of brainpower—and computing power.

So when leaders at the Peter Kiewit Institute (PKI), an information technology and engineering learning center that’s part of the University of Nebraska, Omaha (UNO), decided to build one of the world’s largest computing clusters in 2006, it was about more than dollars—it simply made good sense for the school. The center, they theorized, would receive worldwide attention while providing greater learning opportunities for students.

You can safely say the theory has proven true. The supercomputer, ranked in the top 50 globally and the largest existing Dell/AMD cluster, is changing the face of education, research and business. It creates new opportunities for businesses, researchers and students to tackle complex problems that exceed the capabilities of a single computer. Gallup and Microsoft® are among the entities contracting to use the 1,151-node facility, which sits in the 2,000-square-foot glass-enclosed Holland Computing Center (HCC) at UNO’s South Campus. “We’re entering a new era in research and creating opportunities that weren’t imaginable only a few years ago,” states Jim Skirvin, president of the Holland Computing Center.

The computing cluster combines leading-edge thinking with state-of-the-art technology. Funding for the project—nearly $20 million in cash and in-kind contributions—came from an array of donors, many of whom live in Omaha and/or attended the university. Much of the technical prowess and gear came from AMD, Dell, Cisco Systems, Force10 Networks and American Power Conversion. The cluster uses Dell PowerEdge SC1435 servers with Dual- and Quad-Core AMD Opteron^™ processors. Says Greg Branch, business development manager at AMD, “HCC has established itself as an undisputed leader in high-end computing.”

Although the supercomputer is already helping customers manage huge data sets and address real-world computing challenges, it is designed to open the door to greater opportunities in the future. It’s likely to be used to solve physics problems, sift through climatological data, handle sophisticated business analytics and conduct molecular modeling and complex simulations required by scientists and government agencies.

Power Play

The number of supercomputing clusters is growing at a rate of approximately 20 percent a year, according to IT research firm IDC, and institutions such as the University of Nebraska are realizing the benefits. Scientists, researchers, businesses, government agencies and others are constantly looking for more-powerful tools to crunch numbers and sift through mountains of data. Whether the task is spotting global-warming trends, understanding consumer behavior or managing complex medical research data, there’s almost no limit to the compute power that’s needed to drive research, conduct sophisticated modeling and create what-if scenarios.

HCC switched on its supercomputer in 2007. By establishing itself as a Top 50 computing center, school administrators reasoned, the university’s reputation and standing would rise among applicants, alumni, researchers and the business community. “Not a lot of people outside of the local region have heard of the University of Nebraska, Omaha, and we hope to change that,” Skirvin says.

After raising the requisite funds, including $3.7 million from retired advertising executive and philanthropist Richard Holland, PKI set out to design a system that would send ripples through the computing world. In November 2006, the institute’s leaders sorted through various proposals and selected Dell PowerEdge SC1435 servers powered by Dual-Core AMD Opteron processors. The solution fit the school’s budget, but, more importantly, Dell and AMD made it clear that they were committed to providing the technical expertise and resources needed to succeed.

Originally, HCC planned to install 800 nodes. However, school officials decided that they wanted to crack the supercomputer Top 50 list and another 350 nodes would put them somewhere in the ballpark—while attracting worldwide attention. Richard Holland, who gained naming rights with his multimillion-dollar donation to the institute, dubbed the supercomputer FireFly. In November 2007, PKI grabbed the No. 43 spot on the Supercomputing Top 500 list.

In addition to gaining the distinction of being the largest Dell/AMD cluster in the world, the facility is currently the largest and fastest Microsoft-based supercomputer in the world. PKI is a test site for Microsoft Windows^® High Performance Computing (HPC) Server 2008 and is also handling extensive scalability testing for Microsoft while conducting significant Linux^® testing to better understand how to use the operating system within a supercomputing cluster. What’s more, students from the University of Nebraska-Lincoln’s College of Engineering and the University of Nebraska at Omaha’s College of Information Science and Technology increasingly rely on the supercomputer for an array of
research and academic tasks, including scientific analysis, engineering simulations and learning how to use sophisticated IT resources effectively.

In the months ahead, Skirvin expects to attract insurance companies, pharmaceutical firms, engineering organizations and government agencies—all of which will benefit from FireFly’s enormous processing power and pay the center to use computing resources that otherwise aren’t available to them. HCC is now marketing the center to the business community as well as to research institutes and government agencies.

AMD and HCC Drive Results

HCC had no shortage of vendors and options available when it solicited bids for equipment. However, the institute opted for AMD technology for a simple reason: “AMD platforms allowed us to upgrade from dual-core processor technology to quad-core technology seamlessly,” Skirvin explains. “We don’t have to pull out a server and replace it. We simply slide in a Quad-Core
AMD Opteron processor to replace the Dual-Core AMD Opteron processor that we bought the servers with, and then we update the BIOS and
are ready to put it to work.” These processor upgrades are expected to enable HCC to push the computing capacity from 21.5 teraflops to beyond 60 teraflops and likely boost the center into the top 20 worldwide, Skirvin notes.

But there are other advantages as well. “One of the buzzwords in industry is performance per watt,” Branch points out. “The AMD Opteron processors offer outstanding performance with low power consumption. They can help reduce space needs, equipment costs and energy expenses. Ultimately, they provide low total cost of ownership.” Finally, AMD’s commitment to providing ongoing support helped cement the relationship. “We needed to know that we had a commitment from the companies we worked with, and AMD has met the challenge,” Skirvin says.

Delivering on the promise of the technology and building a world-class computing environment was no simple task. But the way Skirvin sees it, the opportunities are enormous. “We have built a system that allows scientists, business leaders and others to take computing to the next level,” he explains. “Organizations that tap into the power of supercomputing are likely to realize a competitive advantage. They’re more efficient and better positioned for success.”

For a look inside the supercomputer, click here