Phone: (434) 982-2207
Fax: (434) 982-2214
Email: jmo@cs.virginia.edu
Office: Olsson Hall, UVa
ames Ortega received a BS in Mathematics from the University of New Mexico and a PhD in
Mathematics from Stanford
University in 1962. He was at the University of Maryland from 1964-73,
becoming a professor of Computer Science and Applied Mathematics in
1969. From 1973-77, he was the founding director of ICASE at NASA's Langley Research Center. In
1977, he became head of the Mathematics Department at North Carolina State and then joined the
University of Virginia in 1979 as
Charles Henderson professor and chairman of Applied Mathematics and
Computer Science. He served concurrently as Associate Dean of
Engineering and Applied Science in 1980-82, and was chairman of Applied
Mathematics in 1984-89. From 1993-96 he served as chair of Computer
Science, and then he retired. He has directed 19 PhD theses (in Applied
Mathematics). He is author or coauthor of nine books and over 40
papers.
rtega's research involves the development of
scientific computing algorithms for parallel and/or vector
high-performance computers. In this area, typical model problems arise
from structural mechanics or fluid dynamics and involve solving very
large (several hundred thousand or more) systems of linear or nonlinear
equations. The primary algorithms that have been investigated are based
on the conjugate gradient method with preconditioning. This usually
requires solving very large sparse triangular systems of equations which
is a major challenge for parallel computing. One approach is to reorder
the equations by a multicoloring strategy. This often provides ideal
parallelism but at the cost of decreasing the rate of convergence of the
conjugate gradient method. One of Professor Ortega's research projects
was to find different reorderings which give adequate parallelism
without the bad side effect of degraded convergence. Current machines
used were the Intel Paragon and the IBM SP.