As an example to demonstrate the effectiveness of the GRS object-oriented
approach together with the KIRKPATRICK point-location for the
interpolation on unstructured grids, the MINIMOS
wrapper
was used to create MINIMOS
input files from a converted TIF file as output by TMA-SUPREM.
This example involved interpolation from various concentration attributes
defined on a triangular grid with 1387 points onto a gradually refined tensor
product grid created by the MINIMOS wrapper. This tensor product grid
started with just 14 points, and finally is refined to 780 points. The
refinement criterion of the tensor product grid is based on the net doping
concentration, and nine steps are needed to reach the final grid. On each
step, the unstructured grid interpolation is used three times and
invoked on all gridpoints of the tensor product grid. Therefore a total of
6594 points is interpolated.
The total execution time of the wrapper is about 48 s, where most of the time
is used by the interpolation. This is about double the time needed for
interpolation from tensor product grids, where point location is considerably
faster. However, an 
behavior is achieved due the KIRKPATRICK point location method, therefore the difference between tensor
product and unstructured grid interpolation is just a constant factor, and
does not depend on the number of grid points, as it would be the case if
unstructured grid point location would require an 
effort.
Fig. 7.13 shows the original boron doping on the triangular grid as converted from the TIF file. After interpolation and gradually refining the grid, the MINIMOS wrapper generates a boron doping on a tensor product grid as shown in Fig. 7.14. Note that the grid lines are dense in areas where it seems unnecessary, which can be accounted to the fact that the net doping is used as a grid refinement criterion, and all output attributes of the wrapper are defined on the same tensor product grid. The final net doping is shown in Fig. 7.15 where the grid lines are dense where the doping values change rapidly.
Figure 7.13: Original boron doping on the triangular grid
Figure 7.14: Interpolated boron doping on the tensor
product grid
Figure 7.15: Final net doping on the tensor product grid