[mpich-discuss] installation problem on AIX
Eric Sun
eric.sunyy at gmail.com
Thu Mar 8 10:13:04 CST 2012
I think I didn't express clearly. I didn't try to use MPD. I think I used
the default launcher Hydra. But, I got a hang when using mpiexec to start
my MPI code. Can you give me some suggestion on how to check the reason of
the hang?
Thank you very much!
On Thu, Mar 8, 2012 at 10:49 AM, Dave Goodell <goodell at mcs.anl.gov> wrote:
> MPD is deprecated, use hydra instead:
> http://wiki.mcs.anl.gov/mpich2/index.php/Frequently_Asked_Questions#Q:_I_don.27t_like_.3CWHATEVER.3E_about_mpd.2C_or_I.27m_having_a_problem_with_mpdboot.2C_can_you_fix_it.3F
>
> The ".POSIX" warnings are only issues in the old build system. The newer
> build system (1.5a2 and up) shouldn't have this problem. I would ignore
> those messages unless they clearly are causing a problem for you.
>
> -Dave
>
> On Mar 8, 2012, at 9:11 AM CST, Eric Sun wrote:
>
> > Hi Good Morning!
> >
> > I did "make testing". It seems similar problem happened. I had a hang.
> Below is all the screen output from "make testing" before the hang. I'm
> guessing that it's probably a communication problem. Like in the previous
> version of MPICH2, I need mpdboot to set up a parallel environment first.
> But, I have no idea how to check this. I found I don't have mpdboot common
> in my installation.
> >
> > "Makefile", line 468: warning: .POSIX directive is not first non-comment
> line.
> > (cd test && make testing)
> > "Makefile", line 67: warning: .POSIX directive is not first non-comment
> line.
> > (NOXMLCLOSE=YES && export NOXMLCLOSE && cd mpi && make testing)
> > "Makefile", line 111: warning: .POSIX directive is not first non-comment
> line.
> > ./runtests -srcdir=. -tests=testlist
> -mpiexec=/gpfs/small/STED/shared/mpich2/pwr5/bin/mpiexec
> -xmlfile=summary.xml
> > Looking in ./testlist
> > "Makefile", line 78: warning: .POSIX directive is not first non-comment
> line.
> > Processing directory attr
> > Looking in ./attr/testlist
> >
> >
> > On Thu, Mar 8, 2012 at 8:48 AM, Rajeev Thakur <thakur at mcs.anl.gov>
> wrote:
> > This example may have some stdin/stdout issue. Try some other F90
> program. Or run "make testing" in the top-level mpich2 directory, which
> will run the whole test suite (may take an hour).
> >
> > Rajeev
> >
> >
> > On Mar 7, 2012, at 10:40 PM, Eric Sun wrote:
> >
> > > Anthony,
> > >
> > > Changing all compilers to the "_r" version makes the difference. Now I
> can compile my Pi.f90 with no problem! Thank you very much!
> > >
> > > Now the problem is that when I run the code using "mpiexec -f hostfile
> -n 4 Pi.x", there is nothing output. It looks like a hang there. I have
> attached all c.txt, m.txt, mi.txt and config.log files in case you may need
> them for diagnosis.
> > >
> > > Thanks again!
> > >
> > >
> > > On Wed, Mar 7, 2012 at 10:45 AM, Anthony Chan <chan at mcs.anl.gov>
> wrote:
> > >
> > > Send us your configure and make output files as stated in README.
> > > Did you try build mpich2 with _r version of compilers ?
> > >
> > > A.Chan
> > >
> > > ----- Original Message -----
> > > > I configure and make MPICH2 using:
> > > > [bash] export CC=xlc
> > > > [bash] export CXX=xlC
> > > > [bash] export F77=xlf
> > > > [bash] export FC=xlf90
> > > > [bash] ./configure --prefix=/gpfs/small/shared/mpich2/pwr5 --with-
> > > > device=ch3:nemesis
> > > > [bash] make
> > > > [bash] make install
> > > >
> > > > Before I compile my Pi.f90, I did:
> > > > [bash] export PATH=/gpfs/small/shared/mpich2/pwr5/bin:$PATH
> > > >
> > > > Then, to compile my code:
> > > > [bash] mpif90 Pi.f90
> > > >
> > > > Thanks for your help!
> > > >
> > > > On Tue, Mar 6, 2012 at 11:54 PM, Rajeev Thakur <thakur at mcs.anl.gov>
> > > > wrote:
> > > >
> > > > > What exact options did you pass to configure, and how exactly are
> > > > > you
> > > > > compiling your program?
> > > > >
> > > > > On Mar 6, 2012, at 10:41 PM, Eric Sun wrote:
> > > > >
> > > > > > Just tried. It didn't work ... The same error.
> > > > > >
> > > > > > On Tue, Mar 6, 2012 at 11:10 PM, Rajeev Thakur
> > > > > > <thakur at mcs.anl.gov>
> > > > > wrote:
> > > > > > Try giving the full path to mpif90.
> > > > > >
> > > > > > On Mar 6, 2012, at 9:51 PM, Eric Sun wrote:
> > > > > >
> > > > > > > Thanks for your quick reply!
> > > > > > >
> > > > > > > Yes. But, I also tried "mpif77" and got the same error.
> > > > > > >
> > > > > > >
> > > > > > > On Tue, Mar 6, 2012 at 10:33 PM, Dave Goodell
> > > > > > > <goodell at mcs.anl.gov>
> > > > > wrote:
> > > > > > > Are you building your application with "mpif90"?
> > > > > > >
> > > > > > > -Dave
> > > > > > >
> > > > > > > On Mar 6, 2012, at 9:15 PM CST, Eric Sun wrote:
> > > > > > >
> > > > > > > > Dear MPICH2 developers and users,
> > > > > > > >
> > > > > > > > I successfully installed mpich2-1.4.1p1 on an IBM Power5 SMP
> > > > > > > > machine
> > > > > running AIX system. The configure line was
> > > > > > > > ./configure --prefix=/install/directory
> > > > > > > > --with-device=ch3:nemesis
> > > > > > > >
> > > > > > > > I had CC=xlc CXX=xlC F77=xlf FC=xlf90
> > > > > > > >
> > > > > > > > "configure" and "make" are quite smooth without error. But
> > > > > > > > when I
> > > > > try to compile a testing MPI code (calculating Pi), I receive the
> > > > > following
> > > > > error message:
> > > > > > > >
> > > > > > > > ** main === End of Compilation 1 ===
> > > > > > > > 1501-510 Compilation successful for file Pi.f90.
> > > > > > > > ld: 0711-317 ERROR: Undefined symbol: .mpi_init
> > > > > > > > ld: 0711-317 ERROR: Undefined symbol: .mpi_comm_rank
> > > > > > > > ld: 0711-317 ERROR: Undefined symbol: .mpi_comm_size
> > > > > > > > ld: 0711-317 ERROR: Undefined symbol: .mpi_type_contiguous
> > > > > > > > ld: 0711-317 ERROR: Undefined symbol: .mpi_type_commit
> > > > > > > > ld: 0711-317 ERROR: Undefined symbol: .mpi_send
> > > > > > > > ld: 0711-317 ERROR: Undefined symbol: .mpi_recv
> > > > > > > > ld: 0711-317 ERROR: Undefined symbol: .mpi_finalize
> > > > > > > >
> > > > > > > > Any suggestion?
> > > > > > > >
> > > > > > > > Thanks!
> > > > > > > > _______________________________________________
> > > > > > > > mpich-discuss mailing list mpich-discuss at mcs.anl.gov
> > > > > > > > To manage subscription options or unsubscribe:
> > > > > > > > https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
> > > > > > >
> > > > > > > _______________________________________________
> > > > > > > mpich-discuss mailing list mpich-discuss at mcs.anl.gov
> > > > > > > To manage subscription options or unsubscribe:
> > > > > > > https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > --
> > > > > > > Yiyang Sun
> > > > > > > Department of Physics, Applied Physics, and Astronomy
> > > > > > > Rensselaer Polytechnic Institute
> > > > > > > Troy, NY 12180, USA
> > > > > > > _______________________________________________
> > > > > > > mpich-discuss mailing list mpich-discuss at mcs.anl.gov
> > > > > > > To manage subscription options or unsubscribe:
> > > > > > > https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
> > > > > >
> > > > > > _______________________________________________
> > > > > > mpich-discuss mailing list mpich-discuss at mcs.anl.gov
> > > > > > To manage subscription options or unsubscribe:
> > > > > > https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
> > > > > >
> > > > > >
> > > > > >
> > > > > > --
> > > > > > Yiyang Sun
> > > > > > Department of Physics, Applied Physics, and Astronomy
> > > > > > Rensselaer Polytechnic Institute
> > > > > > Troy, NY 12180, USA
> > > > > > _______________________________________________
> > > > > > mpich-discuss mailing list mpich-discuss at mcs.anl.gov
> > > > > > To manage subscription options or unsubscribe:
> > > > > > https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
> > > > >
> > > > > _______________________________________________
> > > > > mpich-discuss mailing list mpich-discuss at mcs.anl.gov
> > > > > To manage subscription options or unsubscribe:
> > > > > https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
> > > > >
> > > >
> > > >
> > > >
> > > > --
> > > > Yiyang Sun
> > > > Department of Physics, Applied Physics, and Astronomy
> > > > Rensselaer Polytechnic Institute
> > > > Troy, NY 12180, USA
> > > >
> > > > _______________________________________________
> > > > mpich-discuss mailing list mpich-discuss at mcs.anl.gov
> > > > To manage subscription options or unsubscribe:
> > > > https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
> > > _______________________________________________
> > > mpich-discuss mailing list mpich-discuss at mcs.anl.gov
> > > To manage subscription options or unsubscribe:
> > > https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
> > >
> > >
> > >
> > > --
> > > Yiyang Sun
> > > Department of Physics, Applied Physics, and Astronomy
> > > Rensselaer Polytechnic Institute
> > > Troy, NY 12180, USA
> > > <mpich2_AIX.tar.gz>_______________________________________________
> > > mpich-discuss mailing list mpich-discuss at mcs.anl.gov
> > > To manage subscription options or unsubscribe:
> > > https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
> >
> > _______________________________________________
> > mpich-discuss mailing list mpich-discuss at mcs.anl.gov
> > To manage subscription options or unsubscribe:
> > https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
> >
> >
> >
> > --
> > Yiyang Sun
> > Department of Physics, Applied Physics, and Astronomy
> > Rensselaer Polytechnic Institute
> > Troy, NY 12180, USA
> > _______________________________________________
> > mpich-discuss mailing list mpich-discuss at mcs.anl.gov
> > To manage subscription options or unsubscribe:
> > https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
>
> _______________________________________________
> mpich-discuss mailing list mpich-discuss at mcs.anl.gov
> To manage subscription options or unsubscribe:
> https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
>
--
Yiyang Sun
Department of Physics, Applied Physics, and Astronomy
Rensselaer Polytechnic Institute
Troy, NY 12180, USA
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/mpich-discuss/attachments/20120308/4ab0e9f6/attachment.htm>
More information about the mpich-discuss
mailing list