SEMINAR

High Performance Eigensolver for Complex Symmetric Matrices

Department of Comuter Science and Statistics
University of Southern Mississippi
Hattiesburg, Mississippi

ABSTRACT

Complex symmetric matrices often appear in quantum physics in the solution methods of partial differential equations (such as the Schrödinger equation). A new fast and efficient direct eigensolver for this problem has been introduced by Bar-On and Ryaboy. The aim of this presentation is to report on some benchmark tests for computing the full and partial eigenspectrum on a variety of supercomputing machines, i.e. the Cray J-9x, the DEC Alfa 8400, and the SGI Power Challenge 8000 and 10000 machines. It is shown that in all cases, that the new algorithm is much faster than codes available in a standard state of the art eigensolver packages such as LAPACK.

WHERE: TEC 340

WHEN(day): Friday, February 6th, 1998

WHEN(time): 2:00

EVERYBODY IS INVITED