软件包: libtrilinos-shylu-dev (13.2.0-1ubuntu1)

solving sparse linear systems - development files

ShyLU is a package for solving sparse linear systems. It can be used either as a preconditioner, or as a solver. Currently, it is recommended as an Ifpack preconditioner. ShyLU uses a hybrid direct/iterative approach based on Schur complements. The goal is to provide robustness similar to sparse direct solvers, but memory usage more similar to preconditioned iterative solvers. ShyLU was designed as a node-level solver. It can use both MPI and threads in several ways. ShyLU was designed (a) to solve difficult but medium-size problems, and (b) to be used as a subdomain solver or smoother for very large problems within an iterative scheme. It is a purely algebraic method so can be used as a black-box solver. ShyLU has been particularly successful in electrical circuit simulation (Xyce).

This package provides headers.