Sorry guys,
I found another problem with apbs-mpi.
It's because apbs-mpi.info is replacing /sw/bin/apbs instead of putting
another binary called, e.g., /sw/bin/apbs_mpi then when running apbs (with
mpi) the output dx file will be appended with -PE0, no matter what one
defines in the apbs.in input file, e.g.:
output.dx
with apbs (with mpi) will became:
output-PE0.dx
and if one uses many cores then there will be output-PE1.dx and so on.
This is not a problem at first for those who use apbs at command line but
apbs plugin for pymol and vmd will be disrupted.
Eventually, I will suggest at pymol and vmd maling list the suggestion of
enabling their plugin to be able to use multi core (for abps mpi enable).
But because of the very peculiar way how apbs with mpi is developed (it
needs to be addressed at the apbs.in input file as well and not only by
mpirun -c #cpus command as many programmes like mdrun_mpi from gromacs) this
is likely not simple.
Cheers,
Alan
On Mon, Aug 17, 2009 at 07:57, Alan <[email protected]> wrote:
> I think it's fixed at least with openmpi since apbs-mpi is working as
> expected AFAIK.
> Cheers,
> Alan
>
>
> On Mon, Aug 17, 2009 at 03:57, William G. Scott <
> [email protected]> wrote:
>
>> Hi Alan:
>>
>> Sorry, I was away in the most inaccessible part of Utah I could find. I
>> think Jack Howarth fixed it in my absence, but if not, please let me know.
>> Sorry for this continuing nightmare.
>>
>> Bill
>>
>>
>>
>> On Aug 14, 2009, at 12:35 AM, Alan wrote:
>>
>> Starting to answer myself, I did this test with a custom compiled version
>>> of
>>> apbs:
>>> ~/Programmes/apbs-1.1.0-source% ./configure --enable-fast
>>> --with-openmpi=/usr
>>> make all
>>> cd examples/actin-dimer
>>> /usr/bin/mpirun -c 8 ../../bin/apbs apbs-mol-parallel.in
>>>
>>> WORKED!
>>>
>>> Than I realise that "--with-openmpi=" flag is missing in
>>> apbs-mpi-openmpi,
>>> although it's using mpicc and mpif77, but it seems not to be enough,
>>> specially because mpi is not compatible with python as I read from the
>>> "configure" script. And looking better at the configure script, this
>>> mpif77/python combination only pass because '--with-openmpi' is not
>>> declared
>>> and to test to avoid this incompatibility is not checked. So, I am sorry
>>> Bill Scott, but "apbs-mpi" is not OK yet.
>>>
>>> HOWEVER, if I put "./configure --enable-fast --with-openmpi=/sw" I got
>>> this
>>> error:
>>>
>>> [snip]
>>> checking for --with-openmpi option... /sw
>>> checking for /sw/include/mpi.h... yes
>>> checking for /sw/lib/libmpi.a... no
>>> checking for /sw/lib/libmpi.so... no
>>> checking for /sw/lib/libmpi.dylib... no
>>> configure: error: Error...Could not find OpenMPI libraries in /sw/lib
>>>
>>> Well, libmpi.dylib is in /sw/lib/openmpi and frankly after reading
>>> ./configure --help all combinations I tried and I didn't get it to work
>>> with
>>> Fink openmpi libs.
>>>
>>> Any help here would be very appreciated.
>>>
>>> Cheers,
>>> Alan
>>>
>>> On Thu, Aug 13, 2009 at 17:22, Alan <[email protected]> wrote:
>>>
>>> Hi There,
>>>> So I installed apbs-mpi-openmpi and it seems to work fine but when
>>>> trying
>>>> this link to run in multi-core machine:
>>>>
>>>>
>>>> http://apbs.wustl.edu/MediaWiki/index.php/How_do_I_run_a_calculation_that%27s_too_big_for_my_computer%3F
>>>>
>>>> I got this message and nothing done.
>>>>
>>>> [snip]
>>>> Reading PQR-format atom data from complex.pqr.
>>>> 11754 atoms
>>>> Centered at (2.518e+00, -2.465e+00, 1.674e+01)
>>>> Net charge -2.40e+01 e
>>>> NOsh_setupCalcMGPARA: Oops! You're trying to perform an 'mg-para'
>>>> (parallel) calculation
>>>> NOsh_setupCalcMGPARA: with a version of APBS that wasn't compiled with
>>>> MPI!
>>>> NOsh_setupCalcMGPARA: Perhaps you meant to use the 'async' flag?
>>>> NOsh_setupCalcMGPARA: Bailing out!
>>>> [snip]
>>>>
>>>> Checking my apbs and libmpi is there:
>>>>
>>>> ldd `which apbs`
>>>>>
>>>> /sw/bin/apbs:
>>>> /System/Library/Frameworks/vecLib.framework/Versions/A/vecLib
>>>> (compatibility version 1.0.0, current version 242.0.0)
>>>> /usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version
>>>> 111.1.4)
>>>>
>>>> /System/Library/Frameworks/Accelerate.framework/Versions/A/Frameworks/vecLib.framework/Versions/A/libBLAS.dylib
>>>> (compatibility version 1.0.0, current version 218.0.0)
>>>> /sw/lib/openmpi/libmpi_f77.0.dylib (compatibility version 1.0.0, current
>>>> version 1.0.0)
>>>> /sw/lib/openmpi/libmpi.0.dylib (compatibility version 1.0.0, current
>>>> version 1.0.0)
>>>> /sw/lib/openmpi/libopen-rte.0.dylib (compatibility version 1.0.0,
>>>> current
>>>> version 1.0.0)
>>>> /sw/lib/openmpi/libopen-pal.0.dylib (compatibility version 1.0.0,
>>>> current
>>>> version 1.0.0)
>>>> /usr/lib/libutil.dylib (compatibility version 1.0.0, current version
>>>> 1.0.0)
>>>> /sw/lib/gcc4.4/lib/libgfortran.3.dylib (compatibility version 4.0.0,
>>>> current version 4.0.0)
>>>> /sw/lib/gcc4.4/lib/libgcc_s.1.dylib (compatibility version 1.0.0,
>>>> current
>>>> version 1.0.0)
>>>>
>>>> om-mpirun -c 2 apbs ... doesn't work either.
>>>>
>>>> Any ideas?
>>>>
>>>> Cheers,
>>>> Alan
>>>> --
>>>> Alan Wilter S. da Silva, D.Sc. - CCPN Research Associate
>>>> Department of Biochemistry, University of Cambridge.
>>>> 80 Tennis Court Road, Cambridge CB2 1GA, UK.
>>>>
>>>>> http://www.bio.cam.ac.uk/~awd28<<
>>>>>>
>>>>>
>>>>
>>>
>>>
>>> --
>>> Alan Wilter S. da Silva, D.Sc. - CCPN Research Associate
>>> Department of Biochemistry, University of Cambridge.
>>> 80 Tennis Court Road, Cambridge CB2 1GA, UK.
>>>
>>>> http://www.bio.cam.ac.uk/~awd28<<
>>>>>
>>>>
>>
>
>
> --
> Alan Wilter S. da Silva, D.Sc. - CCPN Research Associate
> Department of Biochemistry, University of Cambridge.
> 80 Tennis Court Road, Cambridge CB2 1GA, UK.
> >>http://www.bio.cam.ac.uk/~awd28<<
>
------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Fink-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/fink-users
