Hi Georg,
You should post to the forum on www.open-mpi.org . LAM/MPI is a precursor project to OpenMPI and is no longer under development.
Cheers,
Mac
----- Original Message -----
From: lam-bounces_at_[hidden] <lam-bounces_at_[hidden]>
To: lam_at_[hidden] <lam_at_[hidden]>
Sent: Tue Dec 21 03:25:58 2010
Subject: LAM: OpenMPI MPI_ERR_RANK multi-node problem
Hi!
I have installed the Intel Fortran Compiler ifort (now part of the
latest Intel Fortran Composer XE package). I also installed OpenMPI
1.4.3 - my OS is OpenSuse 11.1
Using those two tools I compiled a Fortran-Code to run on a Quad-Core
Intel machine. The code has worked on other machines before. Yet when I
try to run it in parallel mode, I get the following error:
[my-machine:23586] *** An error occurred in MPI_Isend
[my-machine:23586] *** on communicator MPI_COMM_WORLD
[my-machine:23586] *** MPI_ERR_RANK: invalid rank
[my-machine:23586] *** MPI_ERRORS_ARE_FATAL (your MPI job will now abort)
[my-machine:23588] *** An error occurred in MPI_Isend
[my-machine:23588] *** on communicator MPI_COMM_WORLD
[my-machine:23588] *** MPI_ERR_RANK: invalid rank
[my-machine:23588] *** MPI_ERRORS_ARE_FATAL (your MPI job will now abort)
[my-machine:23587] *** An error occurred in MPI_Isend
[my-machine:23587] *** on communicator MPI_COMM_WORLD
[my-machine:23587] *** MPI_ERR_RANK: invalid rank
[my-machine:23587] *** MPI_ERRORS_ARE_FATAL (your MPI job will now abort)
[my-machine:23589] *** An error occurred in MPI_Isend
[my-machine:23589] *** on communicator MPI_COMM_WORLD
[my-machine:23589] *** MPI_ERR_RANK: invalid rank
[my-machine:23589] *** MPI_ERRORS_ARE_FATAL (your MPI job will now abort)
When I just utilize one node, everything works fine.
The commands used to compile the code are:
mpif77 -w -O2 -o my-code.x my-code.f
or
mpif77 -w -O2 -i-dynamic -o my-code.x my-code.f
(I included the i-dynamic option, because I read that I might have a
problem with static libraries, but that does not seem to be the case)
To run the code I use
mpirun -np 4 my-code.x
This works on other machines (with ifort and OpenMPI) but not here. If I
execute
mpirun -np 1 my-code.x
everything works fine.
Any feedback is appreciated.
Thanks a lot!
Georg
_______________________________________________
This list is archived at http://www.lam-mpi.org/MailArchives/lam/
This e-mail and any attachments are for the sole use of the intended recipient(s) and may contain information that is confidential. If you are not the intended recipient(s) and have received this e-mail in error, please immediately notify the sender by return e-mail and delete this e-mail from your computer. Any distribution, disclosure or the taking of any other action by anyone other than the intended recipient(s) is strictly prohibited.
|
