[Trilinos-Users] AztecOO conjugate gradient

[Bart Janssens]

On Fri, May 10, 2013 at 12:13 AM, C.S. Natarajan <csnataraj at gmail.com>wrote:

> My previous e-mail might be a bit misleading..I just mean that 15s doesn't
> sound like too long not that it sounds right.
> Also, maybe a good idea to try out AnalyzeHierarchy method of the
> preconditioner...this way you might get an idea as to which smoother is
> performing well.
>

Thanks, with the 5594877 I meant the global number of rows, forgot the
crucial word there :) The poisson problem here is part of a segregated
solution of the incompressible Navier-Stokes equations, and right now the
Poisson solve is taking about 60% of the time that it takes to advance one
time step, so I'm just wondering if I can expect any real improvement
there. I'll definitely try the  AnalyzeHierarchy method!

Cheers,

Bart
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://software.sandia.gov/pipermail/trilinos-users/attachments/20130510/f41d7113/attachment.html