[petsc-users] command line option for superlu_dist

[Barry Smith]

On Jan 5, 2010, at 1:00 PM, hxie at umn.edu wrote:

>
>
>> 1) The solution norm you report is NaN, which indicates memory  
>> corruption in
> your code
>
> I guess the reason may be a division by zero. The diagonals of the  
> matrix have some zeros.

    SuperLU_Dist should take this into account, but ...

    Send us the matrix as I requested via the ftp site.

>
>
>> 2) The number of allocated nonzeros is really suspicious. There are  
>> only 2
> nonzeros per row?
>
> At least 5.

   Don't worry about this. This is an error in the way we print  
information about the "factored" matrix.


>
> Hui
>