[Paraview] Minimal setup for coprocessing in a Fortran solver

Tue Sep 18 23:58:01 EDT 2012

I'm happy it's working for you. I've been slow to respond as some of the
symbols from the FortranAdaptor library aren't getting exported properly
for newer GCC compilers.

There is some prototype code to get data from the supercomputer and display
it on a ParaView client as well as code to modify create and modify
existing coprocessing pipelines but for now the coprocessing tools should
be viewed as a batch system for extracting data or screenshots.

Andy

On Tue, Sep 18, 2012 at 8:46 PM, Renato Elias <rnelias at gmail.com> wrote:

> lol, It just worked \o/
>
> I thought coprocessor would open ParaView to make the visualization
> interactively but it produced the vtkPolydata files and stored them in the
> directory where I called TestPVCoProcessing program. I guess it's working
> as expected :0)
>
> Well, thanks again!
>
> Renato.
>
>
> On Tue, Sep 18, 2012 at 9:23 PM, Renato Elias <rnelias at gmail.com> wrote:
>
>> Hi Andy, or anyone that could help me...
>>
>> I've found some errors and fixed. It now compiles, links and runs without
>> erros or warnings but I'm not quite sure how to connect ParaView with the
>> simulation code. Do I have to open ParaView and leave it waiting for
>> coprocessor calls or the coprocessor adaptor will do it automatically?
>>
>> Regards
>>
>> Renato
>>
>>
>> On Tue, Sep 18, 2012 at 11:17 AM, Renato Elias <rnelias at gmail.com> wrote:
>>
>>> Hi Andy,
>>>
>>> thanks for helping. I was doing my homework and building a very simple
>>> Fortran90 example aside PV distribution. This example follows attached.
>>>
>>> As you suggested, I've built this example using the ParaView source code
>>> structure. The user just needs to uncompress the zip file in the
>>> CoProcessing/Adaptors/FortranAdaptors folder that it'll create an
>>> EdgeCFDAdaptor and insert an BUILD_EDGECFD_ADAPTOR entry to the CMake setup
>>> for ParaView building. The adaptor itself compiles without any problem. I
>>> also included a Test folder with a small Fortran program that should mimic
>>> a transient simulation calling ParaView coprocessing. Now, I'm a little bit
>>> held trying to understand the next steps. I've created a simple pipeline
>>> and exported the coprocessing.py Python script to be used by the adaptor
>>> but I'm not quite sure about how I'd make things work together. Do I need
>>> to start ParaView first? Does the coprocessor open PV automatically?
>>>
>>> Regards
>>>
>>> Renato.
>>>
>>> obs.: It should be great have more Fortran90 examples in the ParaView
>>> distribution since there's still a bunch of scientific codes running in
>>> Fortran.
>>>
>>>
>>> On Fri, Sep 14, 2012 at 3:04 PM, Andy Bauer <andy.bauer at kitware.com>wrote:
>>>
>>>> Hi,
>>>>
>>>> The quickest way to get things working together may be to do your
>>>> adaptor in ParaView, similar to the Phasta example. After that, what I
>>>> would do is just add in an executable in that same directory that depends
>>>> on your adaptor and see the verbose output from linking that executable and
>>>> just add that in to your makefile. I assume the same strategy would work
>>>> for a VS project but I don't know that very well so I can't be sure.
>>>> Obviously this isn't very portable but should at least get you going on a
>>>> specific machine. CMakeifying your simulation code would make things better
>>>> long term and there's been a lot of improved Fortran support as of late. If
>>>> you're linking with C or C++ code it can even take care of the Fortran to C
>>>> name mangling as well now.
>>>>
>>>> Andy
>>>>
>>>> On Fri, Sep 14, 2012 at 1:37 PM, Renato Elias <rnelias at gmail.com>wrote:
>>>>
>>>>> Hi there, long time not playing with ParaView and it has evolved so
>>>>> much...
>>>>>
>>>>> ok, just starting to play with coprocessing in our solver and I have
>>>>> very first questions before starting. Our solver is entirely written in
>>>>> Fortran90. Some time ago I've tried to use CMake as its building system but
>>>>> I got stucked with Fortran modules compilation and linking, thus, I've just
>>>>> given up. Our build system using plain makefiles and VS projects is so
>>>>> simple that I'm not encouraged in changing everything to CMake in a short
>>>>> time.
>>>>>
>>>>> After taking a look in the Phasta example, I started to adapt it to
>>>>> our solver but I'm not quite sure about what I'll have to do to link such
>>>>> adaptor against our code. Is it possible to compile just the adaptor using
>>>>> cmake and link the produced object file to our code or I'll have to make
>>>>> CMake take care of the entire building system?! To make it short, what's
>>>>> the minimalist setup to make compile coprocess support in a fortran code?
>>>>>
>>>>> -- An adaptor to bridge solver calls to ParaView (ok, I'm sure I'll
>>>>> need this)
>>>>> -- A bunch of VTK libraries (which?! Am I forced to use CMake?! I hope
>>>>> not...)
>>>>> -- Python (it seems that it's necessary only to create the pipeline
>>>>> scripts)
>>>>>
>>>>> any guideline would be appreciated.
>>>>>
>>>>> thanks
>>>>>
>>>>> --
>>>>> Renato N. Elias
>>>>> =============================================
>>>>> Professor, at Department of Civil Engineering
>>>>> COPPE/Federal University of Rio de Janeiro
>>>>> PO Box 68506, Rio de Janeiro, RJ 21945-970, Brazil
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> 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
>>>>>
>>>>>
>>>>
>>>
>>>
>>> --
>>> Renato N. Elias
>>> =============================================
>>> Professor, at Department of Civil Engineering
>>> COPPE/Federal University of Rio de Janeiro
>>> PO Box 68506, Rio de Janeiro, RJ 21945-970, Brazil
>>>
>>>
>>
>>
>> --
>> Renato N. Elias
>> =============================================
>> Professor, at Department of Civil Engineering
>> COPPE/Federal University of Rio de Janeiro
>> PO Box 68506, Rio de Janeiro, RJ 21945-970, Brazil
>>
>>
>
>
> --
> Renato N. Elias
> =============================================
> Professor, at Department of Civil Engineering
> COPPE/Federal University of Rio de Janeiro
> PO Box 68506, Rio de Janeiro, RJ 21945-970, Brazil
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.paraview.org/pipermail/paraview/attachments/20120918/8aa56d00/attachment-0001.htm>