First, use a simple threading approach.
Next step could be to port the filter to GPU. But we should wait for the LME/Erlangen filters to be ported. They should already have a GPU version of this.
First, use a simple threading approach.
Next step could be to port the filter to GPU. But we should wait for the LME/Erlangen filters to be ported. They should already have a GPU version of this.
Simple threading does not improve the performance (even increased runtime), due to the vtk data structure, which has to be connected after the generation.
I figured out that a small refactoring can improve the performance slightly and implemented the delaunay triangulation.
[59683d]: Merge branch 'bug-11021-DelaunayTriangulation'
2012-02-23 11:06:36 Thomas Kilgus [10fcf4]
Removed includes.
2012-02-23 11:05:31 Thomas Kilgus [9aeef6]
Implemented delaunay triangulation. Improved performance of the filter slightly.
[f3e5e4]: Merge branch 'bug-11021-DelaunayTriangulation'
2012-02-23 13:18:11 Thomas Kilgus [18e548]
COMP: initialize variables correctly.
[81d6b8]: Merge branch 'bug-11021-RevertToOldTriangulation'
2012-03-01 15:06:08 Thomas Kilgus [fb869e]
Revert changes in DistToSurfaceFilter