Hi everyone, As the author of pygmsh [1] I sometimes get user complaints about how slow mesh generation is. The way pygmsh works is that it generates a geo-file in memory, writes that out, has gmsh run over it to generate a msh-file, then read in and parse that file to generate the nodes and cells in memory. I'm wondering if this could be improved.
What comes to mind straight away is passing the geo-file as a string to gmsh, and to retrieve the msh-file as a string from gmsh. This way, the slow I/O on the hard drive is avoided. Is that at all possible with gmsh? Even better would be to retrieve the mesh data directly from gmsh, but I guess for this it'd need tighter integration with Python which just isn't there. Cheers, Nico [1] https://github.com/nschloe/pygmsh
_______________________________________________ gmsh mailing list [email protected] http://onelab.info/mailman/listinfo/gmsh
