[Trilinos-Users] Problems with crashing with the ML's MatrixFreePreconditioner (UNCLASSIFIED)
Ramsey, James J CIV (US)
james.j.ramsey14.civ at mail.mil
Wed Jul 3 11:52:17 MDT 2013
Classification: UNCLASSIFIED
Caveats: NONE
-----Original Message-----
I've tried two different types of Epetra_CrsGraph with the
MatrixFreePreconditioner: one like that used to construct the
Epetra_CrsMatrix, and one like that used to construct the Epetra_FEVbrMatrix
(with a block size of 3).
-----End Original Message-----
Okay, I've tried a new graph, which seems to work, at least in the sense of
not crashing and of returning the same results that I get when not using the
preconditioner. What I did was to create a graph *without* applying the
boundary conditions that I apply when creating the graph for the
Epetra_FEVbrMatrix.
However, performance isn't particularly better than not using a
preconditioner. I've got far fewer *iterations* (~50 iterations with ML vs.
~300 without), but each iteration takes longer. Currently, I'm using the
Chebyshev smoother and "METIS" as the aggregation type. The other parameters
are left at their defaults.
Classification: UNCLASSIFIED
Caveats: NONE
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5692 bytes
Desc: not available
Url : https://software.sandia.gov/pipermail/trilinos-users/attachments/20130703/adb1fd21/attachment.bin
More information about the Trilinos-Users
mailing list