[Gmsh] File formats

dolfyn info at dolfyn.net
Tue Apr 7 12:21:26 CEST 2009


Hi all,

I'm puzzling about the file formats. At the moment the *.geo input 
creates a *.msh file; this file is converted into the format for 
the CFD solver which produces a result which is written into a 
*.msh file (other formats can be selected as well; but the main 
idea I'm after here is the feedback from the solver to gmsh). 
However, the original *.geo file, and the first *.msh file, contain 
the 'physical surface' information, this information is lost in the 
process (point and lines are useless to the solver; the 'pysical 
surfaces' are the boundary conditions). In the documentation a 
legacy format *.pos is mentioned; as the original data/file is 
still there it is not necessary to replicate this in the result 
file so this could have been a good idea. But it's a 'legacy 
format'... 

In the same line: the idea that one can read in a background 
result/mesh is very appealing. In the texts it is done for 2D 
meshes; does this also work for 3D meshes? Tet or...? And how about 
extruding a 2D mesh; the result is in 3D; create a cut plane and 
use this 2D plane?

Finally, the results like velocity vectors/components, pressure, 
temperature, effective viscosity etc etc are written either 
as '$ElementData' or as '$NodeData$. When one starts gmsh with this 
dataset all 'views' are shown (and stored); that consumes memory, 
takes time. Is there another way? One a view is 'removed' it cannot 
be 'recreated'. Is that correct?

Thanks in advance!
Henk