Retired Carter cluster was nation’s fastest campus supercomputer

  • May 5, 2017
  • Science Highlights

Purdue’s recently retired Carter community cluster will be remembered as a powerful supercomputer that could handle computationally-intensive applications much faster than its predecessors – qualities that made it perfectly suited to Ashlie Martini’s large-scale numerical simulations. Martini used the cluster for her research in tribology, an interdisciplinary branch of engineering dedicated to the study of how moving surfaces interact with each other.

“If we were to run the simulations on a single core, they would be prohibitively time-consuming,” says Martini, a former Purdue faculty member who kept using Carter even after becoming an associate professor of mechanical engineering at the University of California, Merced. “The availability of Carter’s computational resources was fantastic.”

When it was built in April 2012, Carter was the fastest supercomputer solely for use by researchers on a single campus and not part of a national supercomputing center. It debuted as number 54 on the Top500 list of the world’s most powerful supercomputers and remained on the list for several years. It was Purdue’s first supercomputer in the top 100.

Carter remained so powerful for so long in large part thanks to a partnership between Intel, HP and Purdue. The cluster was built with new technology that wasn’t commercially available at the time, which offered a 70 percent improvement in application speed compared to Steele, one of Carter's predecessors.

“This groundbreaking supercomputer shows what is possible when researchers, campus IT staff and corporate partners work together as a close team," Gerry McCartney, Purdue’s chief information officer and vice president for information technology, said in a news release at the time.

Over its five-year lifespan, Carter enabled breakthroughs in many different research areas. It was used by Purdue faculty in fields as diverse as bioinformatics, atmospheric science and agricultural economics. In all, 60 faculty members across 26 departments purchased access to the cluster.

“Carter is like our virtual laboratory, without which we couldn’t conduct our modeling and simulation related research,” says Haifeng Wang, an assistant professor of aeronautics and astronautics, who used Carter to develop, implement, and test simulation models for turbulent reactive flows before they can be applied to real applications in combustion engines. “Carter provided the computational platform for about 10 students in my group to perform their daily research work.”

Even though Carter has been decommissioned, it will live on. Some of its hardware will be sold to Universidad EAFIT in Medellin, Colombia, through a partnership between the two universities. There, it will be used by researchers who are searching for new drugs to treat HIV and predicting seismic activity in order to design buildings that will better withstand earthquakes.

The cluster was named after Dennis Carter, a Purdue alumnus who created the famous “Intel Inside” marketing campaign that led to consumer awareness of Intel microprocessors as a key component of the personal computer. The effort was so successful that, almost 30 years later, personal computers still bear a sticker with the slogan.

After initially making its home in the Mathematical Sciences Building, Carter was moved to a modular data center near the Purdue power plant. Purdue was one of the first universities to move some of its supercomputers to secure, portable, energy-efficient containerized data centers, to make room in MATH for new machines.

Carter was the fifth supercomputer built as part of Purdue’s Community Cluster Program, an innovative shared computing infrastructure that lets faculty access more computational resources than they could afford on their own, without having to maintain the machines. Purdue will build its tenth community cluster in as many years later this year.

“I don’t need to spend my time or my graduate students’ time maintaining the computers,” says Martini, about the benefits of community clustering. “We can focus our efforts where we really want them to be, which is on the research.”

To learn more about Purdue’s community cluster program, contact Preston Smith, ITaP’s director of research services and support, or 49-49729.

Originally posted: May 5, 2017 2:46pm EDT