complex sparse matrix
Hello, I expect to solve a large complex linear system, which is sparse and symetric but nonhermitian, what kind of numerical method work well for such a problem? I know that iterative method exist that deal with nonsymetric matrices (BiCG or GMRES), but are they really efficient? Which other method may I use? Thanks, Stephane

Re: complex sparse matrix
QMR perhaps. Usually BICGSTAB works well but the convergence depends a lot of the preconditioners. Algebraic multigrid (AMG) is an option although MG based on coarse discretizations (if you matric results from discretizing some type of conservation equations) would is a much better choice. Netlib has some routines that you can try out.

All times are GMT 4. The time now is 18:28. 