[Gmsh] Triangles in .msh file

[mianzhi]

Hello,
Hopefully my answer would be helpful.
1. All Tri.s are on the boundary.
2. The 2nd "tag" of each Tri., the "geometric entity", tells you which 
boundary surface the Tri. belongs to.
(find "number-of-tags" on this page 
http://geuz.org/gmsh/doc/texinfo/gmsh.html )
3. You may see the "geometric entity index" of all boundary surfaces in 
gui of gmsh via enabling "Tools/Options/Geometry/Surface numbers".

Mianzhi

On 2011年05月01日 13:52, Gmsh User wrote:
> Hello everyone,
> I had one quick question regarding the output in .msh file. I see that .msh file
> contains nodal numbers and nodal coordinates of different types of elements. For
> instance, if you mesh a cube with tetrahedrons, you will obtain nodal numbers
> and coordinates for following types of elements: points, lines, triangles and
> tetrahedrons.
> Here is my question: Are the output triangles in .msh file ONLY surface
> triangles of the mesh, or are those ALL possible triangles (including the
> interior triangles that form the faces of an interior tetrahedron) in the mesh?
> The reason I am asking this question is because I also need surface triangles
> along with tetrahedrons in my calculation to implement boundary conditions, and
> it is really helpful if I am getting all the surface triangles(only surface
> triangles) along with tetrahedrons right off the bat in .msh file.
> But what really confuses me is the fact that surface triangles (assuming
> triangles in .msh files are indeed the surface triangles only) obtained from
> Gmsh are NOT the same as surface triangles obtained by running a MATLAB program
> called surftri.m, which is included in distmesh package (developed by Per-Olof
> Persson and Gilbert Strang), on tetrahedron meshes obtained from Gmsh. surftri.m
> uses very simple algorithm (faces of tetrahedrons that do not repeat are surface
> triangles) to determine surface triangles from a mesh of tetrahedrons.
>
>
> Does anyone of you who share a similar experience but have managed to fix this
> problem be willing to share what exactly is going on with this problem? I'd
> appreciate it.
> Thanks in advance!
>
>
>
> _______________________________________________
> gmsh mailing list
> gmsh at geuz.org
> http://www.geuz.org/mailman/listinfo/gmsh
>
>