[petsc-dev] hdf5 self-describing output?

[Jed Brown]

On Mon, Oct 17, 2011 at 03:48, Richard Katz <richard.katz at earth.ox.ac.uk>wrote:

> I think that PetscViewerBinaryMatlabXXX() is a good model for this hdf5
> Viewer.  Each binary output file is accompanied by a light-weight descriptor
> file.
>
>
> http://www.mcs.anl.gov/petsc/petsc-2/snapshots/petsc-current/docs/manualpages/DM/PetscViewerBinaryMatlabOpen.html
>
> Incidentally, I guess the reason that PetscViewerBinaryMatlabXXX() exists
> outside of the VecView() and PetscBagView() context is that they must be
> aware of DMDAs.  It seems a bit strange that there is no generic
> "VecDMDAView()"
>

It's spelled VecView() and any viewers that need you to call something else
should be fixed.


>
> > I have a dream of eventually having model code (via the DM interface)
> callable by the vis plugin so that we could store only state variables (e.g.
> density, momentum, energy) and compute derived quantities (e.g. velocity,
> pressure, temperature, lift, drag, fluxes, stresses) in a consistent way. As
> it is, you have to write loads of redundant information into files if you
> want to visualize it. That's lame.
>
> Does this violate PETSc policy of being science-neutral?  Or do you mean
> this is all implemented in application-code?  Seems to me that the user
> should be responsible for physics-based post-processing of the output
> (though maybe I have misunderstood you).
>

I only mean to provide hooks so that users can provide this code. It's too
much work these days to get decent visualization if you have to roll
everything yourself.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20111017/f64ff648/attachment.html>