[Paraview] Question about Paraview and parallel processing

[Berk Geveci]

Hi Adam,

The legacy VTK format was never upgraded to support domain
decomposition. The VisIt folks added an extra layer to do that. In
ParaView, you need to use another format. I would recommend the new
VTK format (in your case pvts). See
http://www.vtk.org/VTK/img/file-formats.pdf for more information. I
don't know what your workflow is. If you want to convert existing VTK
files, it is very easy to do that with some Python scripting. Let us
know if you need help.

If you absolutely must stick to the legacy format, I actually lied.
There are ways of reading partitioned VTK files into ParaView but it
is not as trivial as the solution above.

Best,
-berk

On Wed, Jan 2, 2013 at 3:45 PM, Adam Larios <alarios at math.tamu.edu> wrote:
> To whom it may concern,
>
> I have some data which takes on different values in space, which normally
> would be easy to view in Paraview.  The problem is that the space (which is
> a big cube) is decomposed into different subdomains since the data was
> processed in parallel, and each subdomain is saved in a different file.
> Currently, paraview interprets each subdomain as a different time-step.  Is
> it possible to get paraview to load all of the parallelized data and view it
> as a single time step?
>
> Of course, I could parse all of these files into one big file, but I would
> like to avoid this extra step if necessary, since the files are updated
> regularly, and an extra post-processing step would slow things down
> significantly.
>
> My files are in legacy VTK format.  They look like this, for example (the
> real runs are much larger than this):
>
>     #==== myData_00.vtk ====#
>     vtk myData
>     ASCII
>     DATASET STRUCTURED_GRID
>     DIMENSIONS      12       12       12
>     POINTS    1728 float
>     0.9519978E-01  0.9519978E-01  0.9519978E-01
>     0.2855993E+00  0.9519978E-01  0.9519978E-01
>     0.4759989E+00  0.9519978E-01  0.9519978E-01
>     ......
>
>     #==== myData_01.vtk ====#
>     vtk myData
>     ASCII
>     DATASET STRUCTURED_GRID
>     DIMENSIONS      12       12       12
>     POINTS    1728 float
>     0.9519978E-01  0.9519978E-01  0.2189595E+01
>     0.2855993E+00  0.9519978E-01  0.2189595E+01
>     0.4759989E+00  0.9519978E-01  0.2189595E+01
>     ......
>
> and so on up to myData_07.vtk.  Visit can stitch all of these domains
> together with a simple file that tells what all the other files are:
>
>     #==== myData.visit ====#
>     !NBLOCKS      8
>     myData_00.vtk
>     myData_01.vtk
>     myData_02.vtk
>     myData_03.vtk
>     myData_04.vtk
>     myData_05.vtk
>     myData_06.vtk
>     myData_07.vtk
>
> Is there anything I can do to get Paraview to interpret these as a single
> data set?  Any help would be much appreciated!  (Also, my apologies in
> advance if this is not the right place to post this kind of thing.)
>
> Best regards,
> Adam
> _______________________________________________
> 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