[CIG-SHORT] mpi missing error when trying to install nemesis

[Brad Aagaard]

Niz-

What was PETSC_ARCH set to when you configured PETSc? The value used at 
that time must match the one used when configuring PyLith.

When you configured PyLith PETSC_ARCH was set to "gcc-4.4.3".

Brad


On 10/05/2010 06:00 PM, Niz wrote:
> Brad
>
> I adapted the instructions detailed in Linux_Ubuntu_64.
> I reached the Last Step, to install the Pylithdev where I  ran:
>           cd $HOME/pylith/scratch/build/$TOOLS_FORMAT/pylith-dev
>          ${HOME}/pylith/src/pylith-dev/configure  --config-cache
> --enable-swig --enable-testing --enable-cubit --prefix=${TOOLS_DIR}
> CPPFLAGS="-I${TOOLS_DIR}/include "  LDFLAGS="-L${TOOLS_DIR}/lib "
> CC=mpicc CXX=mpiCC  CFLAGS="-O2 -fPIC" CXXFLAGS="-O2 -fPIC"
>
> so he replies:
> ...
> checking /home/tora/pylith/tools/gcc-4.4.3/bin/nemesis sysconfig... ok
> checking for PETSc dir... /home/tora/pylith/src/petsc-dev
> checking for PETSc arch... gcc-4.4.3
> checking for PETSc config... no
> configure: error: PETSc config dir not found; check PETSC_ARCH
>
> What did I miss? [I joined the config.log]
>
> Thanks
> Niz.
> -----------------------------------------
> make check PETSc was fine:
> tora at dragon:~/pylith/src/petsc-dev$ make
> PETSC_DIR=/home/tora/pylith/src/petsc-dev PETSC_ARCH=linux_gcc-4.4.3
> test
> Running test examples to verify correct installation
> .C/C++ example src/snes/examples/tutorials/ex19 run successfully with
> 1 MPI process
> C/C++ example src/snes/examples/tutorials/ex19 run successfully with 2
> MPI processes
> Fortran example src/snes/examples/tutorials/ex5f run successfully with
> 1 MPI process
> Completed test examples
>
> PS:I compiled spatial-data with mpicc/mpiCC, Python import
> spatialdata/FIAT/pyre was ok. Nemesis import mpi was ok too. FIAT is
> the 0.3.5 (I uninstalled the 0.9), and uninstalled all mpich I could
> find. And changed PYTHON_VERSION to 2.6 (instead of 2.6.5)  as you
> just suggested to Birendra.
> ----------------------------
> 2010/10/5, Brad Aagaard<baagaard at usgs.gov>:
>> Birendra-
>>
>> You should not rename the lib/python2.6 folder. You should change
>> PYTHONPATH instead. The python2.6 folder was created automatically by
>> "make install".
>>
>> Brad
>>
>>
>> On 10/05/2010 11:52 AM, Birendra jha wrote:
>>> Brad,
>>>
>>> I think I found the error. It is "python2.6" in site-packages folder vs
>>> "python2.6.5" in the PYTHONPATH. My python version is 2.6.5 which is what
>>> I have in PYTHONPATH but the folder name in $TOOLS_DIR/lib is python2.6.
>>>
>>> When I rename this folder to python2.6.5 (should I rename PYTHONPATH
>>> instead??), I can import spatialdata from within python and run make for
>>> pylith.
>>>
>>> I will keep you posted if I finish the make install and make check
>>> successfully.
>>>
>>> Thanks
>>> Birendra
>>>
>>>
>>>
>>> --- On Tue, 10/5/10, Brad Aagaard<baagaard at usgs.gov>   wrote:
>>>
>>>> From: Brad Aagaard<baagaard at usgs.gov>
>>>> Subject: Re: [CIG-SHORT] mpi missing error when trying to install nemesis
>>>> To: "Birendra jha"<bjha7333 at yahoo.com>
>>>> Date: Tuesday, October 5, 2010, 11:03 PM
>>>> Birendra-
>>>>
>>>> The config.log shows that the spatialdata Python files are
>>>> not found. This means either the PYTHONPATH is wrong or the
>>>> files are not being installed into the install directory.
>>>>
>>>> What is the output of
>>>>
>>>> python
>>>> import spatialdata
>>>>
>>>>
>>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> CIG-SHORT mailing list
>>> CIG-SHORT at geodynamics.org
>>> http://geodynamics.org/cgi-bin/mailman/listinfo/cig-short
>>>
>>
>> _______________________________________________
>> CIG-SHORT mailing list
>> CIG-SHORT at geodynamics.org
>> http://geodynamics.org/cgi-bin/mailman/listinfo/cig-short
>>
>
>