[Trilinos-Users] Trilinos + threaded BLAS/LAPACK?

[Heroux, Michael A]

Nico,

What are the specifics of your situation?  Do you have one MPI rank per
core, or per node?  In other words, are there cores available for threaded
BLAS to use?  If so, then you should see performance improvement from
threaded BLAS and LAPACK, if the corresponding non-MPI use would be improved
by threading.

Mike


On 2/24/11 1:48 PM, "Nico Schlömer" <nico.schloemer at ua.ac.be> wrote:

> Hi all,
> 
> a quick one here:
> Is there any data on the performance of Trilinos (read: {T,E}petra) when
> compiled with threaded BLAS/LAPACK as opposed to serial? Would that
> interfere with MPI at all?
> 
> Cheers,
> Nico
> 
> _______________________________________________
> Trilinos-Users mailing list
> Trilinos-Users at software.sandia.gov
> http://software.sandia.gov/mailman/listinfo/trilinos-users