algorithm

[Christophe Geuzaine]

Lin Ji wrote:
> 
> Hi, Christophe,
>    Since I am solving a linear wave equation, would it be faster to solve the
> system by LU factorization than by GMRES (default)? Is GMRES more accurate in
> this case?

LU will be much slower, and accuracy should not be a problem (provided
that the iterative solver converges). You may play with the
preconditioner (increase its size may be benefic if you don't recompute
it at each time step -> cf. ReUseILU flag in one of my preceding
e-mails).

Christophe

-- 
Christophe Geuzaine

Tel: 32 (0) 4 366 37 10    http://geuz.org
Fax: 32 (0) 4 366 29 10    mailto:Christophe.Geuzaine at ulg.ac.be