×

Numerical linear algebra for high-performance computers. (English) Zbl 0914.65014

Software - Environments - Tools. 7. Philadelphia, PA: SIAM, Society for Industrial and Applied Mathematics. xviii, 342 p. (1998).
This book presents techniques for solving matrix problems on parallel and vector computers. It begins with an overview of computer architecture as of 1998 and the issues involved in minimizing overhead and evaluating performance. Then it presents techniques for direct solution of dense matrix problems and least squares problems, algorithms for direct solution of sparse linear systems, preconditioned Krylov subspace methods for iterative solution, and Lanczos, Arnoldi, Davidson, and Jacobi-Davidson methods for eigenvalue and generalized eigenvalue problems. Available software packages (e.g., BLAS, LAPACK, ScaLAPACK, JDQR, ARPACK) are surveyed.
Reviewer: W.Gander (Zürich)

MSC:

65Fxx Numerical linear algebra
65-01 Introductory exposition (textbooks, tutorial papers, etc.) pertaining to numerical analysis
65Y10 Numerical algorithms for specific classes of architectures
65Y05 Parallel numerical computation
65Y15 Packaged methods for numerical algorithms
PDFBibTeX XMLCite
Full Text: DOI