[Gmsh] GMSH questions

Christophe Geuzaine cgeuzaine at ulg.ac.be
Thu Jul 19 10:20:17 CEST 2007


idesbald van den bosch wrote:
> Well, can you try both GMSH 2.0.8 and 1.65 on the attached file 
> containing only flat surfaces?
> 
> For me:
> 1) GMSH 1.65, took 2 min 20 sec., 624 MB max mem
> 2) GMSH 2.0.8, took 16 min using MeshAdapt+Delauney, 147 MB max mem

With 2.0.8 and "-algo del2d" it takes 33 seconds and 147Mb on my laptop 
(mesh contains 1.5 million triangles). Beware that "del2d" is 
experimental and that the generated elements are oriented randomly.


> 
> so the speed is degraded by a factor of 8 (bad), but the memory is 
> divided by 4 (good).
> 
> On the other hand, is the daily release available? where?

2.0.8 is pretty much in sync with the nightly release right now.


> 
> Ides
> 
> On 7/18/07, *Jean-Fran çois Remacle* <remacle at gce.ucl.ac.be 
> <mailto:remacle at gce.ucl.ac.be>> wrote:
> 
>     idesbald van den bosch a écrit :
>      > Hello dear GMSH users,
>      >
>      > I am a GMSH user since quite a while now, and I have a question
>      > concerning the 2.0.x series. It seems that v2.0.X is much more
>      > RAM-memory efficient than its v1.65 counterpart. Unfortunately it is
>      > also much slower (~3 times from my limited experience).
>      >
>     The 2D mesh generator is slower, but I think it is more robust and
>     enables to mesh general curved surfaces (NURBS ...).
>     In the daily release, I have put back a classical isotropic delaunay
>     that is much faster even than the v1.65 triangulator. Yet,
>     it only works for planar surfaces for now, and for meshing the sphere
>     (look at the attachement). When I'll find some time, I'll
>     generalize this algo for general curved surfaces.
> 
>     Normally, the 3D algorithms are much faster and more robust that the
>     ones of previous releases.
>      > 1) does anyone have the same observation?
>      > 2) is the slowness a necessary trade-off for memory efficiency?
>      >
>     No, it's just the algorithm.
> 
>      > Thanks
>      >
>      > Ides van den Bosch
>      > Bilkent Computational Electromagnetics Lab
>      > Bilkent, Ankara
>      > Turkey
>      >
>     ------------------------------------------------------------------------
> 
>      >
>      > _______________________________________________
>      > gmsh mailing list
>      > gmsh at geuz.org <mailto:gmsh at geuz.org>
>      > http://www.geuz.org/mailman/listinfo/gmsh
>     <http://www.geuz.org/mailman/listinfo/gmsh>
>      >
> 
> 
>     --
>     --------------
>     Prof. Jean-François Remacle
>     Department of Civil Engineering and
>     Center for Systems Engineering and Applied Mechanics
>     Université catholique de Louvain, Belgium
>     Tel : +32 10 472082
>     Fax : +32 10 472210
> 
> 
> 
> 
> 
> 
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> 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