| |
|
Science Forum Index » Math - Numerical Analysis Forum » problem about FGMRES method in SPARSKIT
Page 1 of 1
|
| Author |
Message |
| Sam |
 Posted: Fri Mar 23, 2007 10:16 am |
|
|
|
Guest
|
Hi, there,
Did somebody here use SPARSKIT software package? I want to compare
some iterative solvers for the linear equations from my area. So I
download the SPARSKIT and use 'runrc'(this subroutine is in itaux.f)
as a template.
In my test, I find that the performence of FGMRES is better than that
of GMRES. I know FGMRES method can allow preconditioner to change at
every step. But in my test only the solver is changed, and the others,
such as preconditioner, parameters of solver, are all unchanged.
So under this condition, can I think that when I use FGMRES method in
my test, preconditioner didn't change because I only offer one
preconditioner to FGMRES. I am not sure whether this understanding is
right or not. Can you do me a favor? Thanks. |
|
|
| Back to top |
|
| |
|
Page 1 of 1
All times are GMT - 5 Hours
The time now is Tue Dec 02, 2008 2:03 am
|
|
