|Volume 7, Issue 1 -
Jack DongarraDistinguished Professor, Computer Science, University of Tennessee;
Distinguished Scientist, Mathematical Sciences Section, Oak Ridge National Laboratory
Jack Dongarra holds a joint appointment with the University of Tennessee (UT) and Oak Ridge National Laboratory (ORNL) under the UT/ORNL Science Alliance Program. He specializes in numerical algorithms in linear algebra, parallel computing, use of advanced computers, programming methodology, and tools for parallel computers.
Current research also involves the development, testing, and documentation of high-quality mathematical software. His group, the Innovative Computer Laboratory, functions as the CRPC's University of Tennessee site and is part of the university's Computer Science Department. This group is involved in several projects mentioned in past newsletters, including BLACS (Basic Linear Algebra Communication Subprograms), BLAS, PVM (Parallel Virtual Machine), MPI (Message Passing Interface), LAPACK and ScaLAPACK, and templates for the solution of linear systems.
In discussing his work, Dongarra commented, "We all know that highly parallel computers promise orders of magnitude higher speed in many important scientific and engineering computations, but there is still too little software available to realize this promise. My current research addresses this problem in three ways. First, we design and produce numerical linear algebra algorithms and software for high- performance computers. Our team has released both LAPACK, which is the fastest available software on high-performance workstations, vector machines, and shared-memory machines, as well as ScaLAPACK, which is for distributed-memory machines. Second, we are developing enabling technologies in the form of software tools to help in using parallel computing. These tools, such as PVM and MPI, are being used by thousands of scientists and engineers daily. Third, we are developing mechanisms such as Netlib and the NHSE that will provide access to the enabling technologies that have been developed for the HPCC effort."
Regarding NHSE, Dongarra's experience in establishing the Netlib software repository has been extremely valuable in the development of the NHSE system. (see here for a related article.) He is also one of the organizers for the Parkbench parallel benchmark working group and has produced a table of the sites that have the 500 fastest supercomputers worldwide.
Dongarra is currently a member of the CRPC's Executive Committee. Among his other accomplishments, he is co-editor-in-chief of the International Journal of Supercomputer Applications, Netlib, and SIAM Series on Software, Environments and Tools for Scientific Computing. He is an editor for SIAM Review, IEEE Parallel and Distributed Computing, Parallel Computing, Journal of Distributed and Parallel Computing, Journal of Supercomputing, International Journal of High-Speed Computing, Parallel Processing Letters, Journal of Numerical Linear Algebra with Applications, Numerical Linear Algebra with Applications, Numerical Algorithms, and Applied Numerical Mathematics. Dongarra received his Ph.D. from the University of New Mexico in 1980.
For additional information see: http://www.netlib.org/utk/people/JackDongarra/.
Table of Contents