[petsc-users] ILUTP

[Torquil Macdonald Sørensen]

On 30/10/13 19:01, Xiaoye S. Li wrote:
> The sequential SuperLU has ILUTP implementation, not in parallel
> versions. PETSc already supports the option of using SuperLU. 
>
> In SuperLU distribution:
>   EXAMPLE/zitersol.c : an example to use GMRES with ILUTP
> preconditioner (returned from driver SRC/zgsisx.c)
>   SRC/zgsitrf.c : the actual ILUTP factorization routine
>
> Sherry Li
>

Thanks! I installed the sequential SuperLU and managed to get it working
with PETSc. I also managed to get convergence with the ASM
preconditioner, but ILUTP is much faster, and I can use it when I'm
doing uniprocessor runs. It is convenient that I don't have to keep a
separate program version for those cases. Now I can do PETSc-based
solving in all cases, and relegate GetFEM/GMM to only finite element
matrix assembly.

It was a bit of a hassle to get a working install of both PETSc and
GetFEM due to SuperLU-related linking problems. The built-in SuperLU in
GetFEM caused problems when linking a program to both GetFEM and PETSc.
But in the end I managed to conjure up a SuperLU shared library, and get
both GetFEM and PETSc to use that.

Best regards
Torquil Sørensen