[vtkusers] parallel mesh reading and writing

[Karl Merkley]

On Jan 25, 2007, at 3:38 AM, Jean Favre wrote:

> On 25, Jan 2007 11:27 AM, John Biddiscombe <biddisco at cscs.ch> wrote:
>
>
>> If the geometry is fixed for subsequent time steps, then the scalars
>> can
>> be written out as groups in time order. If the geometry is changing,
>> new
>> groups would need to hold geom and scalars per time step. some
>> convention would need to be agreed on on how to lay it out.
>
> Someone has already gone through this effort. Check out the following
> link which might be helpful
>
> http://hdf.ncsa.uiuc.edu/HDF5/papers/prototypes/mesh/
>

I have followed through on that link previously.  They document one  
possible arrangement but it appears that the effort has not been  
pursued since the prototype proposal in 2002.  The mailing-list is  
dead and one of the principals indicated in a private email that he  
had done nothing more.  If someone knows more than this I would still  
be interested.

I have also investigated a couple of other possibilities.

XMDF (not to be confused with XDMF!) is an open source project with  
maintenance funded by the US Army Corp of Engineers.  It is primarily  
used for ground water modeling although it appears to be useful as  
general mesh storage format. I'm not yet sure how complete the data  
model is for an arbitrary FE model (boundary conditions, material  
properties, etc.)

Another option may be CGNS (CFD General Notation System).   This is  
promoted by NASA and according to their website they are now a "AIAA  
recommended practice."   It appears they are in transition to an HDF5  
based model but they are still supporting an alternate system during  
the full transition.  Once I stumbled across CGNS I actually found a  
thread on this list talking about it.  I will follow through up on  
that old thread.

     Karl