[dune-pdelab] question about slow routines in pdelab!!!

[Nagaiah Chamakuri]

Dear all,

Last week, I have written in my previous mail(to dune mailing list) that
ALUGRID is slower than UG grid.
http://lists.dune-project.org/pipermail/dune/2011-October/009320.html
To check this, I was doing some simple test using UG and ALU grids. And
realized
found that some FE routines are taking long time for ALUGRId comparing with
UG grid.
Specifically, "*const typename Traits::FiniteElementType &fe =
pfem->find(*it);*"
For me it is hard to understand  that why does "*FE type* (pfem->find)"
takes longer for ALUGRID.
I would expect that UG and ALU should take same (or comparable) time.

Second request: Does anybody have obtained good parallel efficiency for
example
problem in dpelab-howto using UG or ALU grids ( example :
nonoverlappingsinglephaseflow problem
in dune-pdelab-howto/src/convection-diffusion ).

Any comments/information would be highly appreciated.

Best Regards
Nagaiah
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.dune-project.org/pipermail/dune-pdelab/attachments/20111027/a988adec/attachment.htm>