[Gmsh] Node ordering for quadratic elements

[Christophe Geuzaine]

Hi Kris - I found the problem: there is indeed a nasty little bug for 
higher order elements with nodes associated with faces, which occurs 
only when we need to change the ordering of the nodes to guarantee a 
positive Jacobian (this explains why some meshes were OK, while others 
were not).

We will release a patched version as soon as possible (this will also 
fix another annoying bug with high order elements in 2.1.0, which leads 
to crashes with quad faces).

Many thanks for submitting the bug report!

Christophe


(The reason it manifests itself only in one of your meshes is that the)

Kris Van den Abeele wrote:
> Hi Christophe,
> 
> I verified the node ordering in the meshes I attached in my previous
> email. The mesh for the sine bump (sineBumpHexaCurved3D.msh) has the
> ordering you described. The mesh for the cilinder (halfcylinder_3D.msh)
> has not. The nodes associated to the second and fourth face (1256 and
> 2367) are swapped, as well as the nodes associated to the third and
> fifth face (1458 and 3478).
> 
> Best regards,
> Kris
> 
> 
> On Sun, 2008-02-24 at 18:56 +0100, Christophe Geuzaine wrote:
>> Kris Van den Abeele wrote:
>>> Hi,
>>>
>>> My name is Kris Van den Abeele, I am a PhD student at VUB (research
>>> group of Prof Lacor). I am working on the development high-order
>>> Spectral Volume and Spectral Difference methods, which I implement in
>>> the VKI COOLFluiD code. I've been using Gmsh (Version 2.0.8, on Linux)
>>> as a mesh generator for a long time now and I am very happy with it.
>>>
>>> Lately, I've been creating meshes consisting of quadratic (27 nodes)
>>> hexahedral elements, and I've encountered a problem. On the website, I
>>> have not found the order in which the nodes of such a quadratic
>>> hexahedron are written to the mesh file. I have two different meshes,
>>> and it seems that for two different meshes, the order in which the nodes
>>> are outputted differs (more specifically for the nodes that lie in the
>>> center of a face).
>> Hi Kris - The second order nodes are given in the following order: 
>> first, the nodes associated with edges are given in the order of the 
>> edges; then, the nodes associated with faces are given in the order of 
>> the faces. The ordering of the edges and the faces is given here:
>>
>> http://geuz.org/gmsh/doc/texinfo/gmsh_10.html#SEC71
>>
>> Can you check if your meshes are consistent with the docs? If not, it's 
>> a bug and we will fix it.
>>
>> Take care,
>> Christophe
>>
>>
>>> In attachment, I have included the .geo and .msh (Version 1.0) files
>>> corresponding to the two meshes.
>>>
>>> Could you have a look at this, or maybe send me the order in which the
>>> nodes are outputted?
>>>
>>> Thanks in advance!
>>> Best regards,
>>> Kris Van den Abeele
>>>
>>>
>>>
>>>
>>> ------------------------------------------------------------------------
>>>
>>> _______________________________________________
>>> gmsh mailing list
>>> gmsh at geuz.org
>>> http://www.geuz.org/mailman/listinfo/gmsh
>>


-- 
Prof. Christophe Geuzaine
University of Liege, Electrical Engineering and Computer Science
http://www.montefiore.ulg.ac.be/~geuzaine