# vtkMarkBoundaryFilter

**URL:** <https://discourse.vtk.org/t/vtkmarkboundaryfilter/8335>\
**Category:** Development\
**Tags:** proposal\
**Created:** [April 20, 2022, 11:17am UTC](https://discourse.vtk.org/t/vtkmarkboundaryfilter/8335 "2022-04-20T11:17:12Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![andreasbuykx](https://discourse.vtk.org/user_avatar/discourse.vtk.org/andreasbuykx/32/8137_2.png) [@andreasbuykx](https://discourse.vtk.org/u/andreasbuykx)\
**Post date:** [April 20, 2022, 11:17am UTC](https://discourse.vtk.org/t/vtkmarkboundaryfilter/8335/1 "2022-04-20T11:17:12Z")

</div>

In my FEA application I need to compute the (still unstructured, i.e. possibly non-linear) surface of unstructured grids and this often is time consuming. I was looking for whether vtkUnstructuredGridGeometryFilter was SMP enabled but unfortunately it isn’t. However I also discovered that the fairly new vtkMarkBoundaryFilter.

Reading through the code I found out that 1 dimensional cells in a vtkUnstructuredGrid are treated in a different way than in vtkGeometryFilter and the likes: in vtkMarkBoundaryFilter the lines are only marked if their end-points are not connected to other points, whereas in vtkGeometryFilter etc the lines are copied unconditionally. Note that 0 dimensional cells _are_ marked unconditionally.

It would be nice if the vtkMarkBoundaryFilter could support the geometry-filter behavior as well, e.g. by optionally defining a cell dimension below which all cells are included. Also it would be cool if the vtkUnstructuredGridGeometryFilter could be parallellized…