Hi, Am 23.11.2011 um 09:07 schrieb mahbube rustaee:
> Excuse me Mr. Reuti for your time and I 'm appreciate with your kindness. > > I compiled intel mpi openmpi-1.4.2 by --with-sge and it works via CLI > correctly (openmpi integrate with sge). > I modified PE mpifillamd such: > > pe_name mpifillamd > slots 9999 > user_lists NONE > xuser_lists NONE > start_proc_args /bin/true > stop_proc_args /bin/true > allocation_rule $fill_up > control_slaves FALSE please set: control_slaves TRUE Otherwise Open MPI tries to run under SGE as it detects its presence, but on the other hand SGE won't accept the tightly integrated tasks. > job_is_first_task FALSE > urgency_slots min > accounting_summary FALSE > > I compiled my program with new open mpi. > and my script is: > #!/bin/bash > #$ -S /bin/bash > #$ -N Det2 > #$ -cwd > #$ -j y > #$ -pe mpifillamd 100 > . $HOME/.intelbash > . ~/openmpi_intel_1.4.2.sh > which mpirun > echo $LD_LIBRARY_PATH > mpirun -n $NSLOTS mpi-integ-sge-intel.comp > > Output is: > > /home/mrustaee/PF/openmpi-1.4.2/intel/bin/mpirun > /home/mrustaee/PF/openmpi-1.4.2/intel/lib:/opt/intel/Compiler/11.1/069/lib/intel64:/opt/intel/Compiler/11.1/069/ipp/em64t/sharedlib:/opt/intel/Compiler/11.1/069/mkl/lib/em64t:/opt/intel/Compiler/11.1/069/tbb/intel64/cc4.1.0_libc2.4_kernel2.6.16.21/lib:/opt/intel/Compiler/11.1/069/lib/intel64:/opt/intel/Compiler/11.1/069/ipp/em64t/sharedlib:/opt/intel/Compiler/11.1/069/mkl/lib/em64t:/opt/intel/Compiler/11.1/069/tbb/intel64/cc4.1.0_libc2.4_kernel2.6.16.21/lib > error: executing task of job 1227 failed: execution daemon on host > "amd-7-4.local" didn't accept task > -------------------------------------------------------------------------- > A daemon (pid 31144) died unexpectedly with status 1 while attempting > to launch so we are aborting. > > There may be more information reported by the environment (see above). > > This may be because the daemon was unable to find all the needed shared > libraries on the remote node. You may set your LD_LIBRARY_PATH to have the > location of the shared libraries on the remote nodes and this will > automatically be forwarded to the remote nodes. > -------------------------------------------------------------------------- > -------------------------------------------------------------------------- > mpirun noticed that the job aborted, but has no info as to the process > that caused that situation. > -------------------------------------------------------------------------- > error: executing task of job 1227 failed: execution daemon on host > "amd-7-3.local" didn't accept task > mpirun: clean termination accomplished > > ----------------------------------------- > LIBRARY_PATH do not confilit other. intelbash shell sets intel library path > and openmpi_intel_1.4.2.sh sets open mpi library path! > > -------------------------------------------------------------------------------------- > when I qsub a script without -pe option and I run my job by hostfile such: > #!/bin/bash > #$ -S /bin/bash > #$ -N Det2 > #$ -cwd > #$ -j y > . $HOME/.intelbash > . ~/openmpi_intel_1.4.2.sh > which mpirun > echo $LD_LIBRARY_PATH > mpirun -n 300 --hostfile machines mpi-integ-sge-intel.comp In this case I fear it will pick any machine in the cluster, but not necessarily the ones selected by SGE. Just use `mpirun`, even the -n option can be left out IIRC but it won't hurt. http://www.open-mpi.org/faq/?category=running#run-n1ge-or-sge -- Reuti > everything is ok!. machines is a list of hosts that qsub couldn't run this > program on it. > > what are happening !? > I can't catch that error! > > Thx so much > _______________________________________________ > users mailing list > [email protected] > https://gridengine.org/mailman/listinfo/users _______________________________________________ users mailing list [email protected] https://gridengine.org/mailman/listinfo/users
