[Paraview] ParaView not building "pvserver"

Mon Sep 30 13:05:02 EDT 2013

Hi Bill,

I'd suggest looking at the settings at
http://open.cdash.org/viewNotes.php?buildid=3043527. Burlen set up that
dashboard for building on Hopper at NERSC and I often check his settings
when I try to do other Cray builds. It may not be exactly what you need but
it should be fairly close.

Regards,
Andy

On Mon, Sep 30, 2013 at 12:59 PM, Burlen Loring <bloring at lbl.gov> wrote:

>  Hi Bill,
>
> It sounds like you've not configured your build correctly.
>
>
> do I need to have PARAVIEW_USE_MPI
>
> yes. it's Cray what's the point of running a serial build there?
>
>
> I'm using is MPICH 3.0.4, and I've compiled
> it myself
>
> don't do that, :-). Better use highly optimized Cray MPT on Cray systems.
>
> You need to set MPI_C_INCLUDE_PATH/MPI_CXX_INCLUDE_PATH and
> MPI_C_LIBRARIES/MPI_CXX_LIBRARIES.
>
> Make sure your using PrgEnv-gnu then you can dump the libraries and paths
> that are used with "CC --verbose". Also CRAY_LD_LIBRARY_PATH can be helpful
> for getting the right system paths.
>
>
> the
> Plugin's don't compile.
>
> If you're trying to static link, that's a known issue and was just
> recently fixed. IMO It's better to have them off when static linking
> anyway, cause they'll *all* be automatically loaded on the server side and
> you'll need to do the same in the client(or it will crash) which is a bit
> of a pain, and confusing for new users.
>
>
> On 09/29/2013 08:30 PM, William Sherman wrote:
>
> Hello everyone,
>
> So I've been working on getting ParaView compiled on a new
> Cray XE/XK7 system, and I seem to have made some progress,
> but not enough to have ParaView running.
>
> So my main problem is that when I do get the make process
> to complete, it doesn't produce the "pvserver" executable.
> Here's what it does produce:
>     % ls bin
>     H5detect               vtkmkg3states-pv4.0
> vtkWrapHierarchy-pv4.0
>     H5make_libsettings     vtkParseJava-pv4.0         vtkWrapJava-pv4.0
>     protoc                 vtkParseOGLExt-pv4.0
> vtkWrapPythonInit-pv4.0
>     smTestDriver           vtkProcessShader-pv4.0     vtkWrapPython-pv4.0
>     vtkEncodeString-pv4.0  vtkpython                  vtkWrapTclInit-pv4.0
>     vtkHashSource-pv4.0    vtkWrapClientServer-pv4.0  vtkWrapTcl-pv4.0
>
> So there's no "paraview" itself because I turned off the
> QT-GUI flag.  As this is an HPC machine, I expect to use
> it in server mode.
>
>
> There is another problem, which I'm not sure whether it's
> related or not, and that is that I can't get it to compile
> when the PARAVIEW_USE_MPI flag is on.
>
> So I guess my first question is: do I need to have PARAVIEW_USE_MPI
> on in order to get "pvserver" built?  In theory, I can choose
> to run with only one "pvserver" process, in which case it doesn't
> need MPI, but in general of course one would run more than one.
> But I think in other circumstances I've compiled "pvserver" without
> MPI, so I'm not sure whether this is the problem.  (But I'm not
> sure what's going on with MPI.)
>
>
> I'll attach the "CMakeCache.txt" file since that may provide a
> clue to someone who knows what they're looking for -- I didn't
> notice anything that would indicate why no "pvserver".
>
>
> As for the MPI issue, I continue to get a message of this type:
>
> ---------
> [ 19%] Building CXX object
> VTK/Parallel/MPI/CMakeFiles/vtkParallelMPI.dir/vtkMPIController.cxx.o
> Linking CXX shared library ../../../lib/
> libvtkParallelMPI-pv4.0.so/usr/bin/ld:
> /N/soft/cle4/paraview/MPI/install-3.0.4/lib/libmpich.a(setbotf.o):
> relocation R_X86_64_32 against `mpifcmb2_' can not be used when making a
> shared object; recompile with
> -fPIC/N/soft/cle4/paraview/MPI/install-3.0.4/lib/libmpich.a: could not read
> symbols: Bad value
> collect2: error: ld returned 1 exit status
> make[2]: *** [lib/libvtkParallelMPI-pv4.0.so.1] Error 1make[1]: ***
> [VTK/Parallel/MPI/CMakeFiles/vtkParallelMPI.dir/all] Error 2
> make: *** [all] Error 2
> ----------
>
> So the MPI version that I'm using is MPICH 3.0.4, and I've compiled
> it myself, and I set environment variables to cause the -fPIC compile
> flag to be used, but that doesn't seem to help.  Not sure what's
> going on there, or what command to run to verify that the -fPIC
> flag took.
>
>
> I'm using this version of gcc/g++ for both paraview and MPICH:
>     g++ (GCC) 4.8.1 20130531 (Cray Inc.)
>
>
> There is one other thing I'm mention in case it's related -- the
> Plugin's don't compile.  I have to disable all the plugins in order
> to get through the compilation process.
>
>
> Thoughts?
>
>     Thank you in advance,
>     Bill
>
> --
> Bill Sherman
> Sr. Technology Advisor
> Advanced Visualization Lab
> Pervasive Technology Inst
> Indiana University
> shermanw at indiana.edu
>
>
> _______________________________________________
> 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
>
>
>
> _______________________________________________
> 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
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.paraview.org/pipermail/paraview/attachments/20130930/2d4cac9e/attachment-0001.htm>