** Next:** Summary
**Up:** Top
** Previous:** Algorithm Performance

To demonstrate the performance of the algorithm two sets of examples
are presented.^{3} In
the first set, triangulations of elementary surfaces
have been carried out. The relation between the number of elements and
generation time is given separately
for uniform meshes (Fig. 4) and graded meshes
(Fig. 5).
Mesh size gradation was prescribed at a discrete
location and along a curve on the surface with the same order of
magnitude. Better performance in the case of
``square'' and ``circle'' examples originates from the fact that both
surfaces have been recognized to be planar. It is also distinct that
discretization of surfaces with curvature in both directions is more demanding.
In the second set of
examples, more realistic models have been discretized. Sequences of
uniform meshes of SGI logo (Fig. 7) and junction of two pipes
(Fig. 8), and graded mesh of chair and
teapot (Fig. 9) have been generated. The gradation in
latter two examples have been done on basis of surface curvature.
The relation between the number of elements and time
consumed for various mesh densities is plotted in Fig. 6
and approaches to a linear one. The rate of the performance is again
depending on the model complexity.

#### Footnotes

- ... presented.
^{3}
- The example meshes have been generated on
SGI workstation with 150 MHz IP22 processor and 32 MBytes RAM.

** Next:** Summary
**Up:** Top
** Previous:** Algorithm Performance
*Daniel Rypl *

2005-12-03