We have tested with PGI 6.0 on our Opteron cluster (which it looks like
you're running on), and it works for us. So I'm hoping that this is a
local configuration issue that we can help you figure out easily.
Your laminfo output shows that you should have several RPIs available
to run (e.g., tcp). The "hello" program sample included in LAM/MPI
uses MPI_INIT (i.e., it doesn't request MPI_THREAD_MULTIPLE), so it
shouldn't be an issue there (unless you have the environment variable
LAM_MPI_THREAD_LEVEL set...?).
Also double check that you're compiling and running with the same
installation of LAM (the one from /usr/local/bin).
On Jun 1, 2005, at 10:37 AM, Yongsheng Zhao wrote:
>
> Sorry, forget to attach laminfo:
>
> yzhao_at_master:~> laminfo
> LAM/MPI: 7.1.1
> Prefix: /usr/local
> Architecture: x86_64-unknown-linux-gnu
> Configured by: root
> Configured on: Tue May 31 15:30:40 EDT 2005
> Configure host: master
> Memory manager: ptmalloc2
> C bindings: yes
> C++ bindings: yes
> Fortran bindings: yes
> C compiler: pgcc
> C++ compiler: pgCC
> Fortran compiler: pgf95
> Fortran symbols: underscore
> C profiling: yes
> C++ profiling: yes
> Fortran profiling: yes
> C++ exceptions: yes
> Thread support: yes
> ROMIO support: yes
> IMPI support: no
> Debug support: no
> Purify clean: no
> SSI boot: globus (API v1.1, Module v0.6)
> SSI boot: rsh (API v1.1, Module v1.1)
> SSI boot: slurm (API v1.1, Module v1.0)
> SSI coll: lam_basic (API v1.1, Module v7.1)
> SSI coll: shmem (API v1.1, Module v1.0)
> SSI coll: smp (API v1.1, Module v1.2)
> SSI rpi: crtcp (API v1.1, Module v1.1)
> SSI rpi: lamd (API v1.0, Module v7.1)
> SSI rpi: sysv (API v1.0, Module v7.1)
> SSI rpi: tcp (API v1.0, Module v7.1)
> SSI rpi: usysv (API v1.0, Module v7.1)
> SSI cr: self (API v1.0, Module v1.0)
>
>
> Yongsheng Zhao
>
> -----Original Message-----
> From: lam-bounces_at_[hidden] on behalf of Yongsheng Zhao
> Sent: Wed 6/1/2005 10:25 AM
> To: lam_at_[hidden]
> Cc:
> Subject: LAM: Lam MPI with PGI run time problem
> Hello, there:
>
> We installed lam/mpi with pgi compilers. Installation is good. But we
> got
> problem when we run a program. Here are the detail:
>
> We installed it following the instruction as:
> export PGI=/scracth/pgi
> export PATH=/scracth/pgi/linux86_64/6.0/bin:$PATH
> export cc=pgcc
> export CC=pgCC
> export FC=pgf95
> export F77=pgf95
> ./configure --disable-tv-queue --with-exceptions
> make
> make install
>
> Everything OK here
>
> Then we tested it with the hello example
> master:/scratch/lam-7.1.1/examples/hello> mpicc -o hello hello.c
> It's OK
> yzhao_at_master:/scratch/lam-7.1.1/examples/hello> lamboot -v lamhosts
>
> LAM 7.1.1/MPI 2 C++/ROMIO - Indiana University
>
> n-1<10024> ssi:boot:base:linear: booting n0 (master)
> n-1<10024> ssi:boot:base:linear: finished
>
> It's OK too.
>
> Then we got problem to run it.
> yzhao_at_master:/scratch/lam-7.1.1/examples/hello> mpirun C hello
> -----------------------------------------------------------------------
> ------
> No SSI RPI modules said that they were available to run.
>
> This should only happen if you request an MPI thread level that is too
> high (for example, none of the SSI RPI modules that are included with
> LAM/MPI currently support MPI_THREAD_MULTIPLE). If this happened and
> you did not call MPI_INIT_THREAD with MPI_THREAD_MULTIPLE, something
> catastrophic happened during SSI module selection such that no RPIs
> were able to run.
> -----------------------------------------------------------------------
> ------
> -----------------------------------------------------------------------
> ------
> It seems that [at least] one of the processes that was started with
> mpirun did not invoke MPI_INIT before quitting (it is possible that
> more than one process did not invoke MPI_INIT -- mpirun was only
> notified of the first one, which was on node n0).
>
> mpirun can *only* be used with MPI programs (i.e., programs that
> invoke MPI_INIT and MPI_FINALIZE). You can use the "lamexec" program
> to run non-MPI programs over the lambooted nodes.
>
>
> Anybody know what happened and how to solve it? Help will be highly
> appreciated.
>
> Regards
> Yongsheng Zhao
>
>
>
> <winmail.dat>_______________________________________________
> This list is archived at http://www.lam-mpi.org/MailArchives/lam/
--
{+} Jeff Squyres
{+} jsquyres_at_[hidden]
{+} http://www.lam-mpi.org/
|
