[Paraview] Test Data for Particle Tracer

[Berk Geveci]

I think we should create a new data "repository" for large data. I
don't think repository should be under version control since scm
software do not usually deal well with such files. The tests then
should check if there a CMake variable pointing to such files and run
the tests if yes. This way we can run tests with files bigger than
tiny (which is what we currently allow in the cvs). What do you think?
Having said that, is there no way of generating a smaller artificial
dataset for testing the particle tracker. I am more concerned with how
long it will take to run such a test. Testing time is precious since
we are running a lot of dashboards.

-berk

On Sat, Aug 9, 2008 at 4:34 AM, John Biddiscombe <biddisco at cscs.ch> wrote:
> I have a small dataset which I use for testing the particle tracer. It's
> about 1GB (60 timesteps). Now that I'm ready to commit the latest bunch of
> improvements to the particle tracer code, I'd like also to add a test or two
> for it.
>
> 1GB seems a tadge on the large side for the standard vtkData cvs repository.
> I've got another that I'll strip out all the scalar fields from and reduce
> the number of time steps (50 currently). I should be able to get it down it
> to a few hundred MB.
>
> Question : Usually we try to keep datasets down to a few MB for vtk tests.
> What is the absolute maximum size that I can get away with for a dynamic
> mesh, multi-block, vector field dataset for testing. Of course we could
> limit it to (say) 5 timesteps, but it'd be nice to have a big decent dataset
> for testing. Many of the bugs etc don't crop up until particle cross
> boundaries etc and a good long run is nice.
>
> Kitware guys - Guidelines?
>
> thanks
>
> JB
>
>
> --
> John Biddiscombe,                            email:biddisco @ cscs.ch
> http://www.cscs.ch/about/BJohn.php
> CSCS, Swiss National Supercomputing Centre  | Tel:  +41 (91) 610.82.07
> Via Cantonale, 6928 Manno, Switzerland      | Fax:  +41 (91) 610.82.82
>
>
> _______________________________________________
> ParaView mailing list
> ParaView at paraview.org
> http://www.paraview.org/mailman/listinfo/paraview
>