[Paraview] [ParaView] Unofficial ParaView 4.1.0 SDK for Windows 32 and 64-bit

Fri Aug 8 14:20:59 EDT 2014

> The ParaView dev team switched from Open-MPI 1.6.2 backwards to 1.4.4, 
> precisely because the support was dropped.
I'm not sure about this, Kitware folks could clarify. If ParaView 
developers are picking an MPI implementation, then it should probably be 
MS MPI. It's the only option going forward, as other porjects have 
dropped support some time ago. Using the aging versions of OpenMPI/MPICH 
are going to be problematic as windows platform compilers etc evolves.

> Problem is that the source code for the current MS-MPI builds does not 
> seem to be publicly available,
Is that really a problem?


On 08/08/2014 09:59 AM, Bruno Santos wrote:
> As far as I can understand/deduce:
>
>  1. The ParaView dev team switched from Open-MPI 1.6.2 backwards to
>     1.4.4, precisely because the support was dropped. I guess the
>     1.4.4 version provided a more stable'ish working version than 1.6.2.
>  2. MS-MPI originated from MPICH2 source code for Windows. And MPICH2
>     is no longer supported on Windows, because MS-MPI is the go-to
>     build of MPICH2 for Windows.
>  3. Problem is that the source code for the current MS-MPI builds does
>     not seem to be publicly available, which might be why ParaView
>     isn't linking to it in the first place. Although I vaguely
>     remember a ParaView version that was linked to a stub MPI,
>     possibly because of that...
>
> On 08-08-2014 17:21, Burlen Loring wrote:
>> OpenMPI dropped support for windows some time ago. Would it not make 
>> sense to use MS MPI?
>>
>> https://www.open-mpi.org/software/ompi/v1.6/ms-windows.php
>> http://msdn.microsoft.com/en-us/library/bb524831(v=vs.85).aspx 
>> <http://msdn.microsoft.com/en-us/library/bb524831%28v=vs.85%29.aspx>
>>
>> On 08/08/2014 07:30 AM, Robert Maynard wrote:
>>> This is really cool! We should try to integrate the open mpi changes
>>> as those seem really reasonable.
>>>
>>>
>>> As far as building Qt, this is required for two reasons. The first is
>>> that the Qt 4.8 series only has 32bit official binaries for windows,
>>> and we would need 32 and 64bit. If we moved to Qt 5.X it would be
>>> possible to use the official binaries, as long as moved to VS
>>> 2010/2012.
>>>
>>> Now the reason that we currently build with VS 2008 is because of hard
>>> requirements in the python build system. I know that people are
>>> investigating a new build system for python that supports VS 2010 /
>>> VS2012.  So if ParaView is ever able to use the new python build
>>> system, and officially moves over to Qt 5.X we should transition to
>>> use the official Qt binaries for windows.
>>>
>>> On Wed, Aug 6, 2014 at 8:59 AM, Bruno Santos
>>> <bruno.santos at bluecape.com.pt>  wrote:
>>>> Greetings to all!
>>>>
>>>> We at blueCAPE have made available an unofficial SDK of ParaView 4.1.0 for
>>>> Windows 32 and 64-bit, based on the official SuperBuild and respective build
>>>> options. It's available here:
>>>> https://code.google.com/p/unofficial-paraview-dev-install/wiki/Downloads#ParaView_4.1.0
>>>> Instructions on how these packages were built is provided here:
>>>> https://code.google.com/p/unofficial-paraview-dev-install/wiki/Notes410
>>>>
>>>> According to our tests, it's possible to use this SDK to build plug-ins and
>>>> any other associated applications, which will work with the respective
>>>> official ParaView 4.1.0 builds for Windows. A proof of concept we've used is
>>>> the vtkPOFFReader plug-in which was successfully built with these SDK
>>>> packages:
>>>> https://code.google.com/p/unofficial-paraview-dev-install/wiki/vtkPOFFReaderBuilds
>>>>
>>>> Note: We roughly used the same development environment as the original
>>>> ParaView 4.1.0 binaries for Windows:
>>>> http://www.paraview.org/Wiki/ParaView_Binaries  - although we've used MSVC
>>>> 2008 C++ Express and did not have a respective FORTRAN compiler (which
>>>> requires the Professional version either way), which is why we needed to do
>>>> some adjustments to build Open-MPI with success.
>>>>
>>>> Furthermore, here's a bit of a back story and the current status of this
>>>> endeavour:
>>>>
>>>> Back in January 2012 we at blueCAPE began to try to bring back the
>>>> development installers of ParaView for Windows, which was last officially
>>>> available for ParaView 3.8.1.
>>>> The overall features available in the development installers we created were
>>>> still missing important files, because we were still basing ourselves on the
>>>> infrastructure that was still remnant from ParaView 3.8.1.
>>>> Nonetheless, we made available our findings and builds for ParaView 3.12.0,
>>>> 3.14.0, 3.14.1 and 3.98.0, where the last release we managed to build was
>>>> back in January 2013.
>>>> ParaView fortunately kept evolving since then. We missed the versions 3.98.1
>>>> and 4.0.1, due to time constraints on our side and a few failed builds.
>>>> As far as we can see, ParaView 4.1.0 already has a pretty good SDK structure
>>>> which evolved overtime. The official SuperBuild repository also already
>>>> provides many of the necessary features for making the SDK releases, at
>>>> least for the Unix-like environments.
>>>> This Summer 2014 we at blueCAPE finally managed to get around back to trying
>>>> and adjust the SuperBuild settings to build and deliver a ParaView 4.1.0 SDK
>>>> for Windows and believe we have successfully achieved it. The changes have
>>>> been shared and referenced in the following wiki page at our project:
>>>> https://code.google.com/p/unofficial-paraview-dev-install/wiki/Notes410
>>>> In addition to this, after some reviewing and building tests with the latest
>>>> commits on the master branch from the SuperBuild repository, we have adapted
>>>> most of the changes we've developed for the SDK build above and have
>>>> submitted two commits to Kitware Gerrit, namely:
>>>>
>>>> http://review.source.kitware.com/#/t/4487/
>>>> http://review.source.kitware.com/#/t/4488/
>>>>
>>>> We haven't managed to create the NSIS and ZIP packages through CMake/CPack,
>>>> but our main concern has mostly been on having a working SDK.
>>>>
>>>> Hopefully this can help bring sooner the ParaView SDK in future versions in
>>>> a more official way... so that we won't have to look into generating the
>>>> NSIS/ZIP packages ourselves, nor build Qt 4.8.6 ;)
>>>>
>>>> If anyone has any questions, feel free to ask!
>>>>
>>>> Best regards,
>>>> Bruno Santos
>>>>
>>>> PS: I haven't sent this email to the ParaView Developers mailing list simply
>>>> because previous emails I've sent on this topic have been to this main
>>>> mailing list.
>>>>
>>>>
>>>> _______________________________________________
>>>> Powered bywww.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://public.kitware.com/mailman/listinfo/paraview
>>>>
>>> _______________________________________________
>>> Powered bywww.kitware.com
>>>
>>> Visit other Kitware open-source projects athttp://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://public.kitware.com/mailman/listinfo/paraview
>>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/paraview/attachments/20140808/1a155b4d/attachment-0001.html>