Were it not for the flurry of blinking lights, "El Gato" wouldn't appear much more exciting than a row of three refrigerator-sized racks sitting in a warehouse. But underneath its black metal case lies a powerful universe. Three things are required to access it: a user account, a password and a big idea – like, what does it take for a supernova to blow up? Or, what is the cosmos made of?
University of Arizona assistant professor of astronomy Brant Robertson is propelling the UA’s computational facilities to the forefront of contemporary computer technology. Robertson and collaborators obtained a $1.3 million grant from the National Science Foundation to help fund the implementation of a new generation of supercomputers on campus. Support from the Office of the Senior Vice President for Research provided vital funding to complete the project.
The Extremely LarGe Advanced TechnOlogy system, known as "El Gato," was ranked 336 among the world’s top 500 fastest supercomputers and placed seventh in the world for "green" – energy efficient – supercomputing systems on the Green500 list.
El Gato was created specifically for the UA, and its processing power is more than 13 times greater than the previous generation of high performance computing systems in the UA's Research Data Center. The key to El Gato's exceptional performance lies in the fact that in addition to CPUs – central processing units – it uses GPUs, which stands for graphics processing units. Commonly found in smartphones, game consoles and personal computers, El Gato's unique GPU-CPU hybrid system is designed for its calculation power rather than its ability to display graphics.
"For the price, this computer is very, very fast, and it's very green," Robertson said. "The graphic processing units enable you to speed up your calculations up to 300 times faster compared with central processing units."
According to Robertson, El Gato makes the UA stand out among peer research institutions.
"The top 100 or so of these machines are located either at national research facilities, like NSF, NASA, the Department of Energy, or similar agencies in other countries, and large corporations," Robertson said. "For example, oil companies use them to analyze seismic data to locate underground oil reservoirs, while financial institutions employ supercomputers to figure out how fast stock markets are changing, to gain an advantage in trades."
El Gato's hybrid processor architecture is a huge asset to researchers whose work involves advanced computations where multiple calculations must run simultaneously. The increased processing power of the new computer system allows researchers to run more complex computations at a faster rate and receive more detailed results.
Such simulations are needed to study many complex phenomena, from the evolution of organisms to the formation of black holes.
"El Gato allows us to perform bigger calculations, to look at finer details, and to include more features in our models," said Robertson, who studies how galaxies evolve over time and how dark matter may be distributed in the cosmos. "All of this enables us to do research we couldn't' do otherwise."
While the scientific focal points of El Gato include answering questions in theoretical astrophysics and computer science, any faculty member or student performing research at the UA can use El Gato through the "windfall" usage policy at the UA Research Data Center, which allows them to run their programs when reserved time is going unused. El Gato account holders may additionally request reserved time for special projects that make use of the system’s unique architecture.
"Anyone who is eligible for a high-performance computing account can use the system," he said. "We have reserved 30 percent of the total usage time for the UA researchers and their collaborators. Most of the time it is used by people not associated with the grant, which is exactly what we wanted."
El Gato came about through a collaboration of astronomers, computer scientists and engineers.
"We combined our technical expertise to make sure the computer fits in with the existing technology at the UA," Robertson said. "It works extremely well, and I really do hope that El Gato is helpful to the University and we can encourage people to make full use of the system.”
Robertson said he hopes that El Gato will serve as a springboard for researchers to develop expertise in writing software code for a GPU/CPU architecture, which is different from what a conventional, CPU-based computer understands.
"We want everyone at the UA to train with it, so they can apply for time on the larger national-level machines, which is good for their research," Robertson said.
To that end, a symposium will be held in September, during which experts from El Gato's supplier companies – IBM, Nvidia and Intel – will visit the UA to teach people how to use it most effectively.
The El Gato project is led by Robertson and co-principal investigators professors of astronomy Feryal Ozel and Dimitrios Psaltis; Paul Cohen, professor and director of the School of Information: Technology, Science, and Arts; and Derek Masseth, senior director at University Information Technology Services. Additional personnel from the College of Science, SISTA, and UITS contributed significantly to the project. El Gato was installed in December 2013 in the UA Research Data Center, a facility available to all UA researchers, supported by the Office of the CIO and the Office of the SVPR and managed and operated by UITS.