High-Performance Computing | News

Texas A&M System Building Up HPC and Supercomputing in IBM Deal

• By Dian Schaffhauser
• 01/29/14

The Texas A&M University System is engaging with IBM in a wide-ranging agreement that expands the use of IBM technologies by the institution and could station IBM people on campus to work with university researchers. The partnership is intended, according to a prepared statement, to "align skills, assets and resources to pursue fundamental research, applied development, educational reach and sustainable commercial activities."

The Texas A&M Engineering Experiment Station (TEES), which conducts research for practical application, will be heavily involved in the new initiative. "This is a unique opportunity to meet the needs of engineering, geosciences and agriculture and life sciences researchers to expand in areas not feasible before with small-scale HPC systems," said Katherine Banks, director of TEES and vice chancellor of the College of Engineering.

"With the combined research capabilities of both institutions and ready access to state-of-the-art computing technology, we feel this collaboration could produce significant scientific insights leading to industry-changing solutions and material economic impact," added IBM's Vice President of High Performance Analytics and Cognitive Markets, William LaFontaine.

Among the projects that will benefit from the new infrastructure are research in food growth, disease tracking, energy resource management and new materials development.

On the technology side, IBM will provide a number of new hardware systems, including:

• Blue Gene/Q, a supercomputer designed for use in analytics, complex modeling, and simulation with 2,048 16-core compute nodes in two racks (scalable, according to the company to 512 racks), with the power to run a material sciences problem that previously took weeks to solve in a "fraction of an hour";
• PowerLinux 7R2 servers, connected by 10 gigabit Ethernet into a system specifically for Linux workloads;
• A NeXtScale system with 900 IBM System x compute notes, some set up as a university high performance computing cloud and others for general-purpose computing needs for the geosciences and open source analytics initiatives;
• A General Parallel File System with five System x GPFS Storage Servers to provide 5 petabytes of shared storage. The same storage system will include FlashSystem 820 with 10 terabytes of flash storage.

The hardware will be managed with IBM Platform Computing software, including Symphony for parallel computing and LSF for high performance computing workloads.

The agreement comes during a time when the engineering college at Texas A&M is positioning itself for major growth, pushing to draw 25,000 students — more than double its current count — by the year 2025 and expanding its research and technology faculty. Although details are vague, IBM will be working with the A&M System to assess new computing technologies that could be used to advance the research mission.