Pressing the button pops up the configuration layer (figure 8.21) for the Dirichlet minimizers, it contains some rulers and checkboxes which allow to change what is displayed and printed during the minimizations.
Figure 8.21: Config-Layer for the Dirichlet-Minimizers
The ruler determines the number of iterations that should be applied, if its value is n and a minimization method is applied to the result is . If the checkbox is selected then for each step of the CG method the step number and the residuum are printed to the shell GRAPE was started in. The checkbox and the ruler determine which steps of the CG method should be displayed -- every display intervalth step if the checkbox is on.
The configuration layer can of course be closed with the button, but it is a good idea to keep it open during the minimization since most of the options can still be changed while the minimization is running. If for example displaying each step slows down the minimization too much it can be disabled completely by switching off the checkbox or the number of steps displayed can be reduced with the ruler.
The checkbox allows to enable a special feature: if it is on the surfaces of the minimization are stored as frames of a Surface instance, the minimization process then can be played back as a video (see section 4.3.5.2). This only works for the methods on a Surface instance, the minimization is always applied to the last frame and the result is added as a new frame at the end of the sequence.
When the minimization is started all menus and layers but the configuration layer are disabled and a layer containing information about the minimization process (figure 8.22) is opened.
Figure 8.22: Info Layer for the Dirichlet-Minimizers
It shows the frame and loop, the object (there can be two objects at each frame, i.e. TimeStep, the pre_object and the post_object) that is currently minimized and the step and residuum of the CG method. The minimization process can be suspended temporarily with the checkbox or aborted completely with the button.
Copyright © by the Sonderforschungsbereich 256 at the Institut für Angewandte Mathematik, Universität Bonn.