sieve-dev slow loading of mesh from files

[Matthew Knepley]

On Tue, Oct 28, 2008 at 9:55 PM, Shi Jin <jinzishuai at gmail.com> wrote:
> Hi there,
>
> I am trying to run some large simulations from the code I have
> developed. One problem I realized is that loading the mesh from files
> using
>
> ALE::PCICE::Builder::readMesh()
>
> is very slow. I am not sure whether is related to the fact memories
> are being allocated on the fly for each element or it is about file
> IO. Although I have a fairly large mesh (over 3 million elements), the
> mesh files are not that huge at all (less than 300MB). I am wondering
> if there is any optimization for this. For example, will caching the
> input data at the beginning help?

That method does not deal with ISieves. Thus it does allocation for every
arrow (until Boost multiIndexContainer gets smarter). I thought you had
written output/input for ISieve. Now, if you ,mean you would like to read
in just cells+vertices and get out a fully interpolated mesh without a lot
of allocation, I have no idea how to do this. You might get some benefit
from memory pooling, but it would be a lot of work to put this in (which
is why no Boost things or STL has it).

  Matt

   Matt

> Thank you very much.
>
> --
> Sincerely,
> Shi Jin, Ph.D.
> http://www.ualberta.ca/~sjin1/
>
>



-- 
What most experimenters take for granted before they begin their
experiments is infinitely more interesting than any results to which
their experiments lead.
-- Norbert Wiener