On Wed, Mar 6, 2019 at 11:39 AM Floriane Gidel [RPG] <mmfg@leeds.ac.uk> wrote:

Dear all,

I consider a big mesh which is made of several subdomains (basically a grid, cf attached). My issue is that I do not want to define the subdomains from the big mesh but the contrary: merge all small meshes to obtain the big mesh (because I want all subdomains to be defined on the exact same mesh).

I have one question: Do you care about exactly what is in the small mesh? If not, you could just regularly refine a big mesh to get this.

My first question is therefore: is there a way to merge several small meshes to one big mesh ?

If not, I would like to at least be able to open the solutions on Paraview from one unique file, to avoid opening and setting hundreds of files when I have hundreds of subdomains.

I though that creating a pvtu file from all my vtu files should work, since solving my equations on each small mesh is similar to parallel programming on the big mesh. Can you advise me on how to create this pvtu file out of my vtu files (I attach an example of my vtu file on one small mesh)?

If Firedrake has a way to do this, great. If not, I have a proposal for an easy, but somewhat restrictive way to do it. You could have every process create a serial Plex using his specification for the small mesh. Then you just need to make an SF that matches the boundaries. Everyone has the same mesh, so you know the remote numbering. Its a little kludgy, but this does not strike me as a general thing (I think refinement is the more general solution). Thanks, Matt

Thank you very much,

Floriane

_______________________________________________ firedrake mailing list firedrake@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/firedrake

-- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener https://www.cse.buffalo.edu/~knepley/ <http://www.cse.buffalo.edu/~knepley/>