[Trilinos-Users] AztecOO Convergence

[Ammar T. Al-Sayegh]

----- Original Message ----- 
From: "Mike Heroux" <maherou at sandia.gov>
To: "'Ammar T. Al-Sayegh'" <alsayegh at purdue.edu>; <trilinos-users at software.sandia.gov>
Sent: Tuesday, January 03, 2006 11:58 AM
Subject: RE: [Trilinos-Users] AztecOO Convergence

> Sorry for the delayed reply.  We were all on break the past week.

Hope you all enjoyed your holidays. Glad to see you back!


> After looking at this a bit more, I think you are probably better off, at
> least initially, redistributing your original user matrix so that it has the
> distribution that you want for the preconditioner.  It is still possible to
> leave the vectors with their original distribution by calling the version of
> Epetra_CrsMatrix.FillComplete that takes the domain and range map arguments,
> making sure that the maps you pass in for the domain and range are those of
> your vectors.

So for Ax = b, if I want the preconditioner to be computed using a single
processor, I must redistribute A to one processor only? Wouldn't that
cause AztecOO to use that node only for the computation and will also cause
the fast cache advantage if the matrix is big? How can this be done for that
case? create a single-node map, then set the map of A to it after instantiating
A with the multi-processor map? Guess I'm a bit lost on implementing this
approach. Any code examples?


-ammar