[Paraview] xdmf hdf5 gzip compressed data

Renato Elias rnelias at gmail.com
Wed Feb 18 18:29:56 EST 2009


Hi Johannes,

as you're also working with Fortran, I think I can help you a little bit
more. I wrapped some files (
http://www.nacad.ufrj.br/~rnelias/paraview/xdmf/xdfm-files.zip) and Fortran
routines I use in my program to write the Xdmf as well as the HDF5. It's
also a cavity flow dataset (unstructured grid, 5 time steps partitioned in 4
parallel subdomains). Note that the Xdmf files are written using the
Fortran's FORMAT statement. I know it's a bit uggly but my data does not
vary between runs, so, why would I need to use a library with a bunch of
routines if just a couple of Fortran FORMATs solve my problem? Moreover,
FORMATs are quite portable between systems without the need to compile or
install anything else.

By the way, I loaded your files without any problem...

Just deleted your system path from the HDF5 DataItem reference, for example:

Original:

/home/johannes/workspace/XDMF/xdmf_data/LDCF/LDCF.h5:data/coordinates/x_v

Modified:
          LDCF.h5:data/coordinates/x_v

Voilá, it works ;o)
(thanks God)

Of course, the xdmf file, in this case, must be kept in the same directory
of the HDF5 files it makes references (I have no idea if Xdmf file could
point to HDF5 files in other directories...)

I hope it helps you

Renato N. Elias
==============================================
Researcher at High Performance Computing Center (NACAD)
Federal University of Rio de Janeiro (UFRJ) - Rio de Janeiro, Brazil
http://www.nacad.ufrj.br/~rnelias


On Wed, Feb 18, 2009 at 6:49 PM, Johannes Will
<johannes.will at tu-harburg.de>wrote:

> Hi Renato,
>
> many thanks for your replay, I'd rather have it to do the mistakes myself,
> and be sure PV is able to process compressed files. Got to sell xdmf to my
> manager on Friday!
>
> We use a fortran based code to produce the hdf5 files, that fore we use the
> hdf5 APIs.
>
> I write the xdmf files with python script on my own, since we did not want
> to produce yet another dependency. The files I produce work well, if the
> h5-files are not compressed. I tried to export my case as xdmf from PV3.4
> this produces a new h5 and xdmf file. I tried to compress this new file with
> ' h5repack ', but the resultfile crashes also.
>
> I have attached a h5 file we produced. It is the well known Lid Driven
> Cavity Flow (uncompressed) together with txt-dump and my xdmf file.
>
> Maybe you have an idea were I go wrong.
>
> Kind regards
>
> Johannes
>
>
>
> Quoting Renato Elias <rnelias at gmail.com>:
>
>  Hi Johannes,
>>
>> I use Xdmf/HDF5 in Fortran and my files are compressed. After reading your
>> message I did a test in PV3.4 and the dataset was loaded without any
>> problem. Are you sure that the problem is related to compression? Have you
>> got success in loading uncompressed files? Are you writing your Xdmf files
>> using the library or the files are being made by yourself?
>>
>> Regards
>>
>> Renato.
>>
>> On Tue, Feb 17, 2009 at 6:42 PM, Johannes Will
>> <johannes.will at tu-harburg.de>wrote:
>>
>>  Hallo everybody,
>>>
>>> I am working now for quite a while with xdmf, which I find a very good
>>> tool. I now tried to process compressed (CHUNK, GZIP) h5 data, and
>>> paraview
>>> keeps crashing with a 'segmentation fault'. Has anybody some idea about
>>> that, or yet tried to process compressed data? Any hint is appreciated.
>>>
>>> Many Thanks
>>>
>>> Johannes
>>>
>>> --
>>> Johannes Will
>>>
>>> Pulverhofsweg 5
>>> 22159 Hamburg
>>> Germany
>>>
>>> Tel.: #49-40-644 199 29
>>> Cell: #49-176-23 600 633
>>>
>>> Mail: johannes.will at tuhh.de
>>>
>>>
>>> _______________________________________________
>>> Powered by www.kitware.com
>>>
>>> Visit other Kitware open-source projects at
>>> http://www.kitware.com/opensource/opensource.html
>>>
>>> Please keep messages on-topic and check the ParaView Wiki at:
>>> http://paraview.org/Wiki/ParaView
>>>
>>> Follow this link to subscribe/unsubscribe:
>>> http://www.paraview.org/mailman/listinfo/paraview
>>>
>>>
>>
>
>
> --
> Johannes Will
>
> Pulverhofsweg 5
> 22159 Hamburg
> Germany
>
> Tel.: #49-40-644 199 29
> Cell: #49-176-23 600 633
>
> Mail: johannes.will at tuhh.de
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.paraview.org/pipermail/paraview/attachments/20090218/eac7f98e/attachment.htm>


More information about the ParaView mailing list