[Gmsh] optimization hangs

[Christophe Geuzaine]

Giuseppe Borzì wrote:
> I think I have found a bug in gmsh, version 1.57.0 and 1.58.0. Open the 
> file wg1.geo (a ridged waveguide), mesh it and then click on the 
> "Optimize quality" button, it starts what seems to be a neverending 
> optimization. At least on my AMD Athlon(tm) XP 1800+ I had to stop it 
> when I saw the following line in top
>  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
> 15074 gborzi    25   0  127m 107m  25m R 95.8 10.6   7030:41 gmsh
> Hope this helps.
> 

It seems to work fine on my Athlon XP 1700+ (running Fedora Core 2): the 
optimization completes in 150 seconds and the final mesh contains around 
130000 elements (the worst having a quality measure of 0.396). I used 
the following command:

gmsh wg1.geo -3 -optimize

The result is similar on my powerbook (running MacOS X 10.3)...

Christophe



> Regards
> 
> 
> ------------------------------------------------------------------------
> 
> lc = 1.0e-3;
> Point(1) = {0.0,9.017e-3,9.524e-3,lc};
> Point(2) = {0.0,10.033e-3,9.524e-3,lc};
> Line(1) = {1,2};
> Extrude Line {1,{0.0,0.0,-7.619e-3}};
> Extrude Line {3, {5.08e-3,0.0,0.0}};
> Extrude Line {4, {5.08e-3,0.0,0.0}};
> Extrude Line {2, {5.08e-3,0.0,0.0}};
> Line(18) = {5,7};
> Line Loop(19) = {6,14,-10,-18};
> Plane Surface(20) = {19};
> Point(9) = {-46.0e-3,0.0,0.0,lc};
> Point(10) = {-46.0e-3,19.05e-3,0.0,lc};
> Line(21) = {9,10};
> Extrude Line {21, {0.0,0.0,9.524e-3}};
> Extrude Line {23, {97.08e-3,0.0,0.0}};
> Extrude Line {21, {97.08e-3,0.0,0.0}};
> Extrude Line {24, {97.08e-3,0.0,0.0}};
> Line(38) = {14,18};
> Line Loop(39) = {30,34,-38,-26};
> Plane Surface(40) = {39};
> Line Loop(41) = {36,-38,-28,22};
> Line Loop(42) = {7,18,-11,-1};
> Plane Surface(43) = {41,42};
> Surface Loop(44) = {43,-37,25,-33,-40,29,-5,13,-17,-20,-9};
> Volume(45) = {44};
> Physical Surface(1) = {37,33,43,29,9,13,5,20,17};
> Physical Surface(2) = {25};
> Physical Surface(3) = {40};
> Physical Volume(1) = {45};
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> gmsh mailing list
> gmsh at geuz.org
> http://www.geuz.org/mailman/listinfo/gmsh


-- 
Christophe Geuzaine
Applied and Computational Mathematics, Caltech
geuzaine at acm.caltech.edu - http://geuz.org