[Trilinos-Users] crash using Amesos_superludist
Eric Bavier
bavier at cray.com
Fri Feb 15 23:55:41 MST 2013
On 02/15/2013 04:08 PM, Robert Crockett wrote:
> Hello,
>
> I am using superLU_dist through Amesos, and am seeing a crash in
> parallel. The same problem run using superLU in serial completes
> without issue.
>
> The error is at the NumericalFactorization step - it exits with an
> error code of -1, indicating an issue with the operator. Attached is a
> stack trace from a single processor run, still using superLU_dist.
The NumericFactorization step in Amesos_Superludist exits with -1 either
if 1) The Epetra_Operator is not an Epetra_RowMatrix, or 2) if the shape
of the matrix is deemed inappropriate, which is determined by checking
whether the number of points in the operator domain map is the same as
the number of points in the range map.
> One specific question I have is whether superLU_dist requires the
> global row numbers to be contiguous. I ask because I do not see this
> crash in problems where that is the case.
I do believe that superLU_dist requires the row numbers to be
contiguous, though if the matrix is redistributed before the
factorization, the matrix values will be imported into a matrix with
contiguous, linear maps. But this isn't really an issue since it seems
Amesos is exiting before even reaching this step.
Have you tried the SuperLU_Dist interface in Amesos2?
`~Eric
More information about the Trilinos-Users
mailing list