[Dune] GMSH reader fails in reading the same file for AlbertaGrid<2, 3> but not for ALUSimplexGrid<2, 3>

[Sacconi, Andrea]

Hi Martin,

yes, I have ALBERTA 2.0.1 installed on my laptop.
Because I thought it was the last version ... but now I see that there is something new that can be downloaded!

I'll try and see what happens with my .msh files.

Cheers,

Andrea
__________________________________________________________

Andrea Sacconi
PhD student, Applied Mathematics
AMMP Section, Department of Mathematics, Imperial College London,
London SW7 2AZ, UK
a.sacconi11 at imperial.ac.uk

________________________________________
From: Martin Nolte [nolte at mathematik.uni-freiburg.de]
Sent: 20 July 2012 19:01
To: Sacconi, Andrea
Subject: Re: [Dune] GMSH reader fails in reading the same file for AlbertaGrid<2, 3> but not for ALUSimplexGrid<2, 3>

Hi Andrea,

not checking exactly, I would guess that you are using ALBERTA 2.0 (because
version 3.0 does not require the elements to be "oriented"). This internal
requirement by the ALBERTA library (version 2.0). In the special case of a 2d in
3d grid, it is possible to generate an "orientation" using the vector product
(this is, e.g., done internally in ALUGrid), but it does not work for a 4d
world. Since the restriction is gone in ALBERTA 3.0, I did not bother to
implement it, though.

I hope this helps,

Martin

On 07/20/2012 07:36 PM, Sacconi, Andrea wrote:
> Hi everyone,
>
> I'm experiencing a problem, again, with the GMSH reader.
>
> You will find attached two .msh files.
>
> I managed to import a .msh file for a boundaryless surface (representing a
> sphere, an ellipsoid, ...) into ALUSimplexGrid<2, 3>, eliminating from the .msh
> all the elements that are not triangles (otherwise the segments would be
> interpreted as boundary objects, but here there is no boundary at all). Thanks
> to Andreas Dedner for the useful suggestion.
>
> Anyway, if I try to use the same file with AlbertaGrid<2, 3>, this error message
> appears on my screen:
>
> Loading macro grid: ../data/ellipsoid.msh
> Reading 2d Gmsh grid...
> version 2.2 Gmsh file detected
> file contains 92 nodes
> file contains 178 elements
> number of real vertices = 91
> number of boundary elements = 0
> number of elements = 178
> Error: GridError [setOrientation:macrodata.cc:259]: Surface grid cannot be oriented.
>
> It seems that the same file gives problems with the orientation of the elements,
> but only in the AlbertaGrid case.
> Here my questions:
>
> 1) are there any specific requirements about the orientation I should pay
> attention to? Check also the (long, sorry!) post scriptum.
> 2) why doesn't ALUGrid detect the problem?
>
> Cheers,
>
> Andrea
> ________________________
>
> PS, if I try with another file, the problem doesn't appear. For example:
>
> Loading macro grid: ../data/mesh_GMSH_coarse_new.msh
> Reading 2d Gmsh grid...
> version 2.2 Gmsh file detected
> file contains 19 nodes
> file contains 32 elements
> number of real vertices = 18
> number of boundary elements = 0
> number of elements = 32
> macro_test_2d: WARNING in ./../2d/macro_2d.c, line 451
> There is a cycle beginning in macro element 14.
> Correcting refinement edges....
> reorder_2d: Current elem: 0, best_edges: 0 3 3
> Current elem: 1, best_edges: 2 3 3
> Current elem: 2, best_edges: 1 3 3
> Current elem: 3, best_edges: 2 3 3
> Current elem: 4, best_edges: 0 3 3
> Current elem: 5, best_edges: 2 3 3
> Current elem: 6, best_edges: 0 1 2
> Looking at best_edges[6][0]...
> Found compatibly divisible neighbour 7!
> Current elem: 8, best_edges: 0 3 3
> Current elem: 9, best_edges: 1 3 3
> Current elem: 10, best_edges: 1 3 3
> Current elem: 11, best_edges: 2 3 3
> Current elem: 12, best_edges: 0 3 3
> Current elem: 13, best_edges: 1 3 3
> Current elem: 14, best_edges: 1 3 3
> Current elem: 15, best_edges: 2 3 3
> Current elem: 16, best_edges: 0 3 3
> Current elem: 17, best_edges: 0 1 2
> Looking at best_edges[17][0]...
> Found compatibly divisible neighbour 19!
> Current elem: 18, best_edges: 1 3 3
> Current elem: 20, best_edges: 0 3 3
> Current elem: 21, best_edges: 1 3 3
> Current elem: 22, best_edges: 1 3 3
> Current elem: 23, best_edges: 2 3 3
> Current elem: 24, best_edges: 0 3 3
> Current elem: 25, best_edges: 2 3 3
> Current elem: 26, best_edges: 2 3 3
> Current elem: 27, best_edges: 1 3 3
> Current elem: 28, best_edges: 0 3 3
> Current elem: 29, best_edges: 1 3 3
> Current elem: 30, best_edges: 1 3 3
> Current elem: 31, best_edges: 2 3 3
> check_mesh: checking mesh 'DUNE AlbertaGrid'
> checking done; no error detected
> Grid file read correctly.
>
> __________________________________________________________
>
> Andrea Sacconi
> PhD student, Applied Mathematics
> AMMP Section, Department of Mathematics, Imperial College London,
> London SW7 2AZ, UK
> a.sacconi11 at imperial.ac.uk
>
>
> _______________________________________________
> Dune mailing list
> Dune at dune-project.org
> http://lists.dune-project.org/mailman/listinfo/dune

--
Dr. Martin Nolte <nolte at mathematik.uni-freiburg.de>

Universität Freiburg                                   phone: +49-761-203-5630
Abteilung für angewandte Mathematik                    fax:   +49-761-203-5632
Hermann-Herder-Straße 10
79104 Freiburg, Germany