dolfin team mailing list archive

Thread
Date

[Question #155458]: control over mpirun

To: dolfin@xxxxxxxxxxxxxxxxxxx
From: Bento <question155458@xxxxxxxxxxxxxxxxxxxxx>
Date: Mon, 02 May 2011 13:44:18 -0000
Reply-to: question155458@xxxxxxxxxxxxxxxxxxxxx
Sender: bounces@xxxxxxxxxxxxx

New question #155458 on DOLFIN:
https://answers.launchpad.net/dolfin/+question/155458

Hi,

is it possible to get a better control over the parallel capabilities of Dolfin? I have the following in mind:

1) One of the meshes I am working with has some internal boundaries that would be canonical choices for splitting the mesh. But when I start the process with mpirun, the mesh is just split arbitrarily. Can I control the areas that are distributed to the different CPU kernels for example with a meshfunction?

2) I have a set of several time-dependent PDEs that live on different meshes. The assembling and solving processes in all the time steps could be done in parallel. Is it possible to do this?

Thank you!

-- 
You received this question notification because you are a member of
DOLFIN Team, which is an answer contact for DOLFIN.