[petsc-users] Strange behavior of MatLUFactorNumeric()

[Hong Zhang]

Jinquan:

> ** **
>
> I have a strange observation on using MatLUFactorNumeric() for dense
> matrices at different order N.  Here is the situation I have:
>

Superlu and mumps are sparse direct solvers, not for dense
matrices.

> **1.       **I use ./src/mat/tests/ex137.c as an example to direct PETSc
> in
>
    petsc-3.3/src/mat/examples/tests/ex137.c is a
"Tests MatCreateMPISBAIJWithArrays()."
Which version of petsc do you use?

> selecting superLU-dist and mumps.  The calling sequence is ****
>
> MatGetOrdering(A,…)****
>
> MatGetFactor(A,…)****
>
> MatLUFactorSymbolic(F, A,…)****
>
> *MatLUFactorNumeric*(F, A,…)****
>
> MatSolve(F,b,x)****
>
> **2.       **I have three dense matrices A at three different dimensions:
> N=75, 2028 and 21180.  ****
>
> **3.       **The calling sequence works for N=75 and 2028.  But when
> N=21180, the program hanged up when calling *MatLUFactorNumeric*(…).
> Seemed to be a segmentation fault with the following error message:
>
If your matrices are sparse, i.e., using petsc default aij format,
then superlu/mumps should work for  N=21180.
Start with small number of processors, e.g., 4, 16 etc.
If superlu_dist hangs, does mumps also hang?

** **
>
> Does anybody have similar experience on that?
>
Segmentation fault occurs for various reasons.
Give a try to mumps first.

Hong
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20120815/f8c7a778/attachment.html>