LU Decomposition on multiple processors

[Lucas Campos]

Dear all,

I am using the PETSc interface to solve my system matrix, like so

 SolverControl solver_control;

 PETScWrappers::SolverPreOnly solver(solver_control, mpi_communicator);

 PETScWrappers::PreconditionLU preconditioner(system_matrix);

 solver.solve(system_matrix, distributed_dU, system_rhs, preconditioner);

where all vectors and matrices are either PETScWrappers::MPI::Vector PETScWrappers::MPI::SparseMatrix.

Today I was checking the documentation and was struck by this sentence[1]:

A class that implements the interface to use the PETSc LU preconditioner. The LU decomposition is only implemented for single processor machines. It should provide a convenient interface to another direct solver.

Is that correct? If so, what is the recommendation for direct solvers in parallel?

Bests,

Lucas

 

[1]: https://dealii.org/8.4.1/doxygen/deal.II/classPETScWrappers_1_1PreconditionLU.html

[Jean-Paul Pelteret]

Dear Lucas,

I don’t have much familiarity with our PETScWrappers, but after a quick look through the documentation I guess that the most convenient interface to a parallel direct solver would be that of PETScWrappers::SparseDirectMUMPS.

Best,

Jean-Paul

--
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see https://groups.google.com/d/forum/dealii?hl=en
---
You received this message because you are subscribed to the Google Groups "deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dealii+un...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Lucas Campos]

Dear Jean-Paul,

Thanks for the indication! I just modified the program to use MUMPS, and there was no difference,

neither in the speed nor in the solution. Right now I am interpreting this as a positive thing, as it 

means the old version was correct.

However I still do not understand what that line in the documentation means. Maybe it is a leftover?

Thanks a lot,

Lucas

To unsubscribe from this group and stop receiving emails from it, send an email to dealii+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see https://groups.google.com/d/forum/dealii?hl=en
---

You received this message because you are subscribed to a topic in the Google Groups "deal.II User Group" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/dealii/ktJUFbymYeM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to dealii+unsubscribe@googlegroups.com.

[Timo Heister]

> However I still do not understand what that line in the documentation means.
> Maybe it is a leftover?

Yes it probably is. Thanks for pointing that out. I just submitted a
fix here: https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_dealii_dealii_pull_5453&d=DwIBaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=4k7iKXbjGC8LfYxVJJXiaYVu6FRWmEjX38S7JmlS9Vw&m=stXjSgb1AyhPUy5isNdAOPkur6U5zQZ5EJcnwR63nM8&s=QIkQSB_k2y6BJY8ykryOERkGsWxOKmaV_eawFWCfbC4&e=


Best,
Timo