[petsc-users] AMG for PETSc

[Travis Austin]

Hi Mark,

AMG for complex systems is an active research area.  Neither Hypre or ML support directly solving the complex systems.  Both at this point would
require recasting the system as a real system but, at least for Hypre, building PETSc with HYPRE with complex turned on causes problems as you
cannot build Hypre with complex types.  We have been exploring hooking a real build of HYPRE into a complex build of PETSC but our progress lately
has stalled due to other commitments.  In any case, I'd suggest the reference comparing real equivalent versus complex.

MacLachlan and Oosterlee. Algebraic multigrid solvers for complex-valued matrices. SIAM Journal on Scientific Computing (2008) vol. 30 (3) pp. 1548-1571

Travis

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Travis Austin, Ph.D.
Tech-X Corporation
5621 Arapahoe Ave, Suite A
Boulder, CO 80303
austin at txcorp.com
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^


On Mar 23, 2011, at 5:48 AM, Nun ion wrote:

> Hello all, 
> 
> I am interested in using AMG for PETSc, where my matrices are complex.  I looked at the documentation for Hypre and I'm not sure that it supports complex arithmetic.  I also looked at ML from trilinos, however the only work around would be to recast my problem as a real system (doubling the MATVEcs).  Any suggestions?
> 
> Thanks
> 
> Mark

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20110323/385ca3a6/attachment.htm>