dolfin team mailing list archive

[Anders Logg <logg@xxxxxxxxx>]

On Fri, Feb 15, 2008 at 03:49:34PM +0000, Garth N. Wells wrote:
> 
> 
> Anders Logg wrote:
> > On Fri, Feb 15, 2008 at 10:28:44AM -0500, Jake Ostien wrote:
> >> Anders Logg wrote:
> >>> When I run the DG demo, it looks like the assembly over interior
> >>> facets is very slow.
> >>>
> >>> I suspect this is because we are not computing the sparsity pattern
> >>> correctly before assembly (so we won't get the correct number of
> >>> nonzeros per row).
> >>>
> >>> Does anyone else see this. Ideas for fixes? Kristian?
> >>>   
> >> I have noticed that there is a significant speedup when I omit my 
> >> interior facet terms, also.
> >>
> >> Jake
> > 
> > We had this problem initially when we implemented DG / interior facet
> > terms but we fixed it at some point. It shouldn't be slow.
> > 
> > So at some point we broke it.
> > 
> 
> It's back up to speed now.
> 
> Garth

Yes, here too!

-- 
Anders