Fast and Robust Iterative Solvers for Large Linear Systems

For the past several years, Roland Freund, Prof. Gene Golub of Stanford and Noel Nachtigal of RIACS (Research Institute for Advanced Computer Science) have collaborated on theory and software for iteratively solving large systems of linear equations (specifically, LAL, "look ahead Lanczos", QMR "quasi-minimum risidual", the coupled variant, and the subroutines thereof). Freund recently joined AT&T Bell Laboratories in Department 11274, and we would like to continue this fruitful collaboration, so we have timely access to this basic technology. They have submitted a joint National Science Foundation proposal, "Robust Iterative Solvers for Partial Differential Equations and for Other Linear Systems Arising in Large-Scale Applications." AT&T Bell laboratories is the only non-University partner to this proposal.