HPC@ISU: A New Direction for High-Performance Computing
If you wanted to buy a supercomputer in 1985, you might have been very interested in the Cray 2, a liquid-cooled behemoth capable of 1.9 billion floating-point operations per second and containing more memory than all previous Cray systems combined. It was the very definition of a "supercomputer".
But that's where the problem comes in. Definitions are (more or less) constant, but the things they represent are not. If we lock our definition to 1985, chances are good that you either have, or know someone who has, a supercomputer; in the form of an Apple iPad 2. It's as fast as the Cray was, has more memory, a much better display, and boasts a footprint smaller than one of the Cray technical manuals.
And while the yesterday's supercomputer is today's Angry Birds game machine, current high-performance systems are literally a million times faster, and new and innovative designs on the horizon point to even greater capability. The future of supercomputing is also the future of advancing knowledge in the fields of science and engineering.
So what does that mean to Iowa State? We are, by classification and direction, a university with a strong commitment to research and the scientific method. And taking a leadership role in fields like bio-engineering, materials science, and information technologies requires computing power; a lot of it. It will require vision, collaboration, and dedication to develop, harness, and direct this vital resource.
Overseeing this task will be a group named "HPC@ISU". High Performance Computing At Iowa State University is a new steering committee formed by Vice President for Research and Economic Development Sharron Quisenberry and Jim Davis, Chief Information Officer. The Chair of the committee is Arun Somani, Distinguished Professor in the Department of Electrical and Computer Engineering and the members represent a broad cross-section of technology leaders from across the university. In addition to the Chair, they are: Wolfgang Kliemann, Mark Gordon, Joe Colletti, Jim Reecy, Rodney Fox, Srinivas Aluru, Glenn Luecke, Bruce Harmon, and ex-officio members Chitra Rajan, Jim Davis, and Lynette Sherer.
Currently, the committee is moving forward with two main initiatives. The first is a collaborative effort with researchers at Iowa State to develop a proposal to submit to the National Science Foundation. This request for a Major Research Instrumentation grant would help fund the establishment of a new central computing facility open to researchers in all disciplines.
The second area of work involves developing a community cluster model for high performance computing. The goal is to develop a resource that will meet the demand for high performance research computing while freeing researchers from the need to maintain their own systems.
Although they are only three to five years old, Iowa State's existing supercomputers are beginning to show their age, leapfrogged by technical advances and changes. HPC@ISU will be working to retire or repurpose those systems as they lay the foundation for sustainable and expandable high performance computing at Iowa State.