scientific article; zbMATH DE number 176387
[For the entire collection see Zbl 0753.00017.] An elimination algorithm is proposed for the \(LU\) decomposition of upper Hessenberg matrices on a shared memory multiprocessor where the number of processors is not related to the problem dimension. As an illustration for application of the \(LU\) solver serves a five-point difference approximation of an elliptic partial differential equation in a rectangle where two tridiagonal systems arise in one ADI iteration. A generalization to solving systems with a banded Hessenberg matrix is presented. What concerns the complexity, the direct solver yields best estimations for solving tridiagonal systems known hitherto.
This page was built for publication:
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4035536)