Hi Andrew,

> When you say, the partitioning uses user_indices, you are referring to the
> partitioning across processors?

Yes, exactly. It concerns the function that assigns subdomain IDs to cells.

> Will this approach fail? (My code is not complete at present so I can't
> answer my own question just yet).

It will fail if you call any function that uses user indices. A prominent 
example is the already mentioned partitioning algorithm.

> > Someone should implement a decent backup-and-restore functionality for
> > general user data, not indices or pointers.
>
> Agreed. More generally it should allow for projection on refinement as
> well. I would be happy to be involved.

What I had in mind has nothing to do with refinement. It is just a store data, 
do some quick work and restore data approach which will be useful for library 
functions that work on a given fixed mesh and want to use user pointers or 
indices without destroying the previous user data, whatever it may be.

> A silly question perhaps, but what Is the point of a cell user_index?

Whenever you just want to save an integer per cell instead of a lot of data it 
is easier to just assign that integer directly instead of having a pointer to 
it. I think that was the general idea. In the given example it is just set to 
a global contiguous active cell index which deal.II does not offer otherwise. 
(You could also use a dummy DoFHandler with a DG(0) element to achieve the 
same thing.)

Best,
Tobias
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii

Reply via email to