[Paraview] [Xdmf] Xdmf/HDF5 access during a long run?

[Eric E. Monson]

Right now we're generating one big HDF5 file, and then one to  
three .xmf files for the Xdmf grids (time series), which reference  
that HDF5 file.

-Eric

On Sep 9, 2008, at 1:10 PM, Berk Geveci wrote:

> Are you generating one big HDF5 file or one per time step or one per
> time step per process?
>
> On Tue, Sep 9, 2008 at 12:22 PM, Eric E. Monson  
> <emonson at cs.duke.edu> wrote:
>> Hello,
>>
>> So far we have just used Xdmf/HDF5 and ParaView to view data from
>> small runs on a single machine, where we generate a time series as  
>> the
>> simulation progresses by appending snapshots of the data to the same
>> HDF5 file. As we work towards larger simulations (on a cluster) which
>> will run much longer, my collaborators are wondering whether we will
>> be able to view any of the data before the whole run is over? This
>> could be a snapshot of the current state, or the ability to animate
>> the data up to the most recent time step, and it wouldn't have to be
>> an auto-update, as in "Animating Live Data" (although they would like
>> this ability, too :).
>>
>> Do any of you have experience checking on your Xdmf/HFD5 data while
>> it's being generated? And, if so, are there any special tricks to it?
>> (I know I had trouble reading an HDF5 file one time that I hadn't
>> closed properly, so I wouldn't be surprised if there are issues
>> regarding whether the file is closed between writes, but I don't know
>> if there are other issues, too.)
>>
>> Thanks for the help,
>> -Eric
>>
>> ------------------------------------------------------
>> Eric E Monson
>> Duke Visualization Technology Group
>>
>>
>>
>> _______________________________________________
>> Xdmf mailing list
>> Xdmf at lists.kitware.com
>> http://www.kitware.com/cgi-bin/mailman/listinfo/xdmf
>>