The present version is pretty inefficient for two reasons:
Each face and line is present multiple times in SubCellData.
We have to sort the arrays in SubCellData to use them.
Instead, it is much faster to simply copy boundary and manifold ids after creating the new Triangulation: we still have the native cells, after all. This makes setting up a spherical Triangulation about twice as fast.
The present version is pretty inefficient for two reasons:
Instead, it is much faster to simply copy boundary and manifold ids after creating the new Triangulation: we still have the native cells, after all. This makes setting up a spherical Triangulation about twice as fast.