Guy Helmer
Ames Laboratory Research

I have long been interested in high performance computing, so the opportunity to do research in the Scalable Computing Laboratory at the Ames Laboratory has been an incredible opportunity. I originally started doing research under Dr. John Gustafson's direction in 1997, but he has since moved on to other things and I have continued my research in performance measurement tools.

NetPIPE

The main focus of my work is NetPIPE, the Network Protocol Independent Performance Evaluator. I have been working on extending NetPIPE to evaluate the communications performance of entire systems, such as clusters interconnected with commodity network components. The SCL was an early adopter of Gigabit Ethernet and we used NetPIPE to characterize the performance advantages of Gigabit Ethernet while supplying useful information to hardware developers to help further improve performance.

ABC

 
[ABC Computer at Howe Hall, Nov 2 1999]
Holly Helmer (right) and I discuss the ABC with visitors after the posthumous presentation of the Iowa Award to John Vincent Atanasoff on November 2, 1999
Between 1994 and 1997, the Ames Laboratory built a working replica of the Atanasoff Berry Computer, the first electronic digital computer. In August and September of 1997, I was privileged to assist with the final debugging and preparation of the ABC replica before its unveiling and demonstration to the world in Washington, DC at the National Press Club on October 7, 1997.

Charles Shorb, a fellow graduate student, had nearly completed debugging the machine before leaving to begin a full-time job in Oregon. However, a significant amount of tedious wiring repair needed to be completed, and I was picked for the job. Fortunately, my careful, dilligent work paid off and the ABC worked for the crowd in Washington, DC. Charles returned to the ISU campus for two weeks in May, 1998 and we video taped the operation the ABC for posterity.

Cluster Cookbook

Technology transfer is an important issue in research environments, and Ames Laboratory has a rich body of knowledge dealing with clusters. I started writing the Cluster Cookbook to guide others in the selection of components and construction of a cluster for parallel computation.

Other Projects

My other activities at Ames Lab have included:

Publications



Last Updated November 12, 1999