[Gmsh] Gmsh and postprocessing

Martin Vymazal martin.vymazal at vki.ac.be
Sat Dec 8 19:22:45 CET 2007

Dear Christophe,

 I'm actually using one of your 'nightly builds'. But the documentation
does not say anything about the major enhancements of gmsh postprocessor
you talk about, so I supposed the functionality was very similar to that
of gmsh version 2.0 (stable).
 Of course I can display the data I want to see using gmsh as it is at the
moment, but I didn't like the necessity of having such an overhead
(either defining 9 values/per node and not using all of them, or storing
nodal coordinates in multiple separate files - one file per output
variable). And it is quite unusual to keep mesh geometry and output
information in different files.
 I've tried Paraview several times before, but it is really hard to use it
with no documentation. I agree that the VTK toolkit seems to be a good
choice. One thing that I miss in Paraview and that gmsh has is postscript
 Concerning other file formats - have you heard of CGNS ? The cgns project
originated as a joint effort of several major industrial companies,
including e.g. Boeing. Its goal was to create a general file format that
could be used by everyone. Cgns is oriented towards CFD problems, but I
think it might be suitable for your purposes also. It is a library of
routines written in C (and in Fortran, alternatively), that store mesh
and simulation output in binary files. You can also define boundary
conditions, your own datastructures etc. CGNS is free and very well
documented. If you're interested, visit www.cgns.org.
 I guess this sounds a bit like an advertisement, but cgns may be an
interesting alternative. It is also supported by third-party software
producers as for example Tecplot.
 I wish you good luck with your work on gmsh postprocessor. It is very
nice of you and your colleagues to share the results of your work with