On 12/29/25 09:58, Daniel Arndt wrote:
On 12/21/25 11:21, Daniel Arndt wrote:
Running that program in a profiler shows that almost all time is spent
in dealii::internal::p4est::functions<dim>::connectivity_join_faces.
That's just an alias for a p4est function, right? So the time is actually
spent in p4est?
Yes, exactly, that's what my profiler was showing. I didn't dive
deeper into p4est to understand why that is or if there is a better
interface to call.
Ah, that's a bummer. Someone with more time might have to dig into p4est,
and/or the interface between p4est and deal.II to see how this could be
improved :-(
p4est used to have some algorithms that were written with the assumption that
the coarse mesh is relatively small, until we came along with meshes that had
a few 100,000 cells and that blew up quadratic algorithms. The p4est folks
eventually fixed the cases we pointed out, but it is quite possible that the
function here is one that has simply never been tried with large coarse
meshes. You may want to open an issue with the p4est folks explaining the
situation and seeing whether they have an easy solution.
Best
W.
--
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see
https://groups.google.com/d/forum/dealii?hl=en
---
You received this message because you are subscribed to the Google Groups "deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion visit
https://groups.google.com/d/msgid/dealii/45ac1ba8-49ab-41a3-8121-35b4395e772c%40colostate.edu.
