“Lonestar 4″ Supercomputer Deployed for Science Community

April 4, 2011

AUSTIN, Texas — The Texas Advanced Computing Center (TACC) at The University of Texas at Austin today will dedicate Lonestar 4, its newest supercomputer, as part of the advanced computing arsenal for the state of Texas and the nation.

The National Science Foundation (NSF), The University of Texas at Austin, The University of Texas System, the UT Institute for Computational Engineering and Sciences, Texas A&M University, Texas Tech University and multiple technology partners committed $12 million to deploy this new supercomputer. It replaces the previous Lonestar system that was one of the most productive platforms of the NSF TeraGrid network for almost four years.

"Austin and Texas have long been leaders in technology R&D, information technology, and in particular, supercomputing," said William Powers Jr., president of The University of Texas at Austin. “We're getting accustomed to being at the front of the supercomputing pack, as it should be. Lonestar 4 solidifies this position of leadership."

TACC Director Jay Boisseau said, "Supercomputers like Lonestar 4 enable researchers to make breakthrough discoveries that advance our knowledge in science and engineering, and often produce transformational impacts for society as well. We're proud to make Lonestar 4 available to researchers in Texas and across the country, and we will continue to increase its capabilities to facilitate new research, education and discovery."

Researchers are conducting research on the newly deployed system across all scientific disciplines. Lonestar 4 is being used to model several phenomena in solid earth geophysics, including seismic wave propagation, mantle convection and the dynamics of polar ice sheets.

"More accurate models of the dynamics of these geophysical processes can lead to a better understanding of seismic hazard and future sea level rise," said Omar Ghattas, the Jackson Chair in Computational Geosciences in the departments of Geological Sciences and Mechanical Engineering and in the Institute for Computational Engineering and Sciences (ICES) at The University of Texas at Austin. 

Clint Dawson, also in ICES, and a professor of aerospace engineering and engineering mechanics, and head of the Computational Hydraulics Group, is conducting hurricane storm surge studies in 2-D and 3-D.  

"With Lonestar 4, we're able to do 3-D runs now in much less time, and we can compare results between 2-D and 3-D to see how much 3-D effects matter," Dawson said. "It's important for two reasons — evacuation as storms approach land and for studying new hurricane protection systems."

TACC partnered with Dell, Intel, Mellanox Technologies and Data Direct Networks to deploy a system designed for optimal performance across a wide spectrum of scientific applications. Lonestar 4 is the third largest system on the NSF TeraGrid and ranks among the most powerful academic supercomputers in the world. The system's capabilities include:

  • 302 teraflops peak performance
  • 44.3 terabytes total memory
  • 1.2 petabytes raw disk

The system's primary computing power comes from 1,888 Dell M610 PowerEdge blade servers, each with two six-core Intel Xeon 5600 "Westmere" processors. DataDirect Networks provides the high-speed disk storage and a Mellanox 40Gb/s InfiniBand network integrates the components to enable very high-performance computing on a wide range of applications. Lonestar 4 will provide almost 200 million processor core hours per year to the national scientific community.

"Geophysical simulations are characterized by a number of computational challenges, including a wide range of length and time scales, highly heterogeneous media, a need for dynamically adaptive resolution and assimilating sparse observational data into the simulations," Ghattas said. "All of these significantly stress the hardware system. Lonestar 4's much greater memory bandwidth, faster CPU clock speed, and faster interconnect relative to other TeraGrid systems combine to promise substantially faster turn-around time for our simulations."

For more information, contact: Faith Singer-Villalobos, Texas Advanced Computing Center, 512 232 5771.