Timo,

> One option is to write into a temporary vector without ghost elements
> and skipping values that are not locally_owned. Before writing out the
> solution, one assigns the temporary vector into (the ghosted)
> joint_solution. I have no time to do work on that over the next days,
> though.

Ah, good point indeed. I think the way we initialized the vector so far is 
somewhat fishy anyway -- there is a good reason I believe why we disallowed 
writing into ghosted vectors ;-)

I can do the change one of these days, it shouldn't be too hard, but I wanted 
to ask if anyone sees a reason why what different processors write into these 
vectors might not be unique? So far, every processor would simply write out 
what it put into the vector, but this may no longer be true under the scheme 
you suggest...

Best
 W.

-------------------------------------------------------------------------
Wolfgang Bangerth                email:            [email protected]
                                 www: http://www.math.tamu.edu/~bangerth/

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

Reply via email to