RE: [Libmesh-devel] Triangle Interface

[KIRK, B. (JSC-E. (NASA) <ben...@na...>]

It would be interesting to have something like 

MeshTools::Generation::build_delaunay_triangulization (mesh, begin, end)

where begin and end are node iterators.  (That is all that should be needed,
right?)

You could then do

Mesh quad_mesh(2);
Mesh tri_mesh(2);

MeshTools::Generation::build_delaunay_triangulization(tri_mesh,
quad_mesh.nodes_begin(), quad_mesh.nodes_end());


BTW,  I've downloaded the latest PETSc and will work though that crap
again...  Of course, since everyone uses their build system this is a
non-issue, right? :-) 2.2.x sure was a short-lived release cycle.

-Ben





-----Original Message-----
From: lib...@li...
[mailto:lib...@li...] On Behalf Of John
Peterson
Sent: Monday, May 02, 2005 10:11 AM
To: lib...@li...
Subject: [Libmesh-devel] Triangle Interface


Hi,

There is now an initial interface to the Delaunay
triangulation library known as "Triangle", written
by J.R. Shewchuk (http://www-2.cs.cmu.edu/~quake/triangle.html).

The code can be found in src{include}/mesh/mesh_triangle_support.C{h}.
The current way to use Triangle is through the

MeshTools::Generation::build_delaunay_square(...)

routine.  This routine operates similarly to the existing
build_square(...) routines, except that it generates "unstructured looking"
Delaunay triangulations with TRI3 elements.  Use the test code attached at
the end of this mail to check it out.  I hope it will be useful in other
parts of the library to have fast built-in Delaunay routines available.

Additional enhancements will hopefully provide support for second-order
(TRI6) elements, as well as the ability to generate triangulations of
arbitrary 2D point sets.  AFAIK, triangle deals exclusively with points in
R2, so you can't use it to generate triangulations of surfaces in R3, but I
think tetgen can possibly be used for that.  At some point we should put
tetgen in the contrib/ directory as well, it is much easier to configure and
interface with software if we don't have to guess what version is out there
(PETSc, I'm looking in your direction here ...)

-John


P.S. As already mentioned on the user's list, PETSc 2.3.0 is
     out, and they have changed a bunch of the Vec* and Mat*
     routines that we use.  Unless any PETSc experts want to
     volunteer *cough* Ben *cough* I'll take a swing at d/ling
     and compiling against the new version, which BTW doesn't
     use BOPT!  What is the world coming to?







#include "mesh.h"
#include "mesh_generation.h"

int main(int argc, char** argv)
{
  libMesh::init(argc,argv);
  {
    Mesh mesh(2);

    unsigned int nx=5, ny=5;
    Real xmax =1.0, ymax=1.0;
    if (argc >= 3)
      {
	nx = atoi(argv[1]);
	ny = atoi(argv[2]);
      }

    if (argc >= 5)
      {
	xmax = atof(argv[3]);
	ymax = atof(argv[4]);
      }
    
    MeshTools::Generation::build_delaunay_square(mesh,
						 nx, ny,
						 0., xmax,
						 0., ymax);
    // Print the results to GMV
    mesh.write("triangulated_square.gmv");
    
    
  }
  return libMesh::close();
}



-------------------------------------------------------
This SF.Net email is sponsored by: NEC IT Guy Games.
Get your fingers limbered up and give it your best shot. 4 great events, 4
opportunities to win big! Highest score wins.NEC IT Guy Games. Play to win
an NEC 61 plasma display. Visit http://www.necitguy.com/?r=20
_______________________________________________
Libmesh-devel mailing list
Lib...@li...
https://lists.sourceforge.net/lists/listinfo/libmesh-devel