Are there are tuning parameters than I can use to reduce the amount of memory used by OpenMPI?  I would very much like to use OpenMPI instead of MVAPICH, but I’m on a cluster where memory usage is the most important consideration. Here are three results which capture the problem:

 

With the “leave_pinned” behavior turned on, I get good performance (19.528, lower is better)

 

mpirun --prefix /usr/mpi/intel/openmpi-1.2.8 --machinefile /var/spool/torque/aux/7972.fwnaeglingio -np 28 --mca btl ^tcp  --mca mpi_leave_pinned 1 --mca mpool_base_use_mem_hooks 1 -x LD_LIBRARY_PATH -x MPI_ENVIRONMENT=1 /tmp/7972.fwnaeglingio/falconv4_ibm_openmpi -cycles 100 -ri restart.0 -ro /tmp/7972.fwnaeglingio/restart.0

Compute rate (processor-microseconds/cell/cycle):   19.528

Total memory usage:    38155.3477 MB (38.1553 GB)

 

 

Turning off the leave_pinned behavior, I get considerably slower performance (28.788), but the memory usage is unchanged (still 38 GB)

 

mpirun --prefix /usr/mpi/intel/openmpi-1.2.8 --machinefile /var/spool/torque/aux/7972.fwnaeglingio -np 28 -x LD_LIBRARY_PATH -x MPI_ENVIRONMENT=1 /tmp/7972.fwnaeglingio/falconv4_ibm_openmpi -cycles 100 -ri restart.0 -ro /tmp/7972.fwnaeglingio/restart.0

Compute rate (processor-microseconds/cell/cycle):   28.788

Total memory usage:    38335.7656 MB (38.3358 GB)

 

 

Using MVAPICH, the performance is in the middle (23.6), but the memory usage is reduced by 5 to 6 GB out of 38 GB, a significant decrease to me.

 

/usr/mpi/intel/mvapich-1.1.0/bin/mpirun_rsh -ssh -np 28 -hostfile /var/spool/torque/aux/7972.fwnaeglingio LD_LIBRARY_PATH="/usr/mpi/intel/mvapich-1.1.0/lib/shared:/usr/mpi/intel/openmpi-1.2.8/lib64:/appserv/intel/fce/10.1.008/lib:/appserv/intel/cce/10.1.008/lib" MPI_ENVIRONMENT=1 /tmp/7972.fwnaeglingio/falconv4_ibm_mvapich -cycles 100 -ri restart.0 -ro /tmp/7972.fwnaeglingio/restart.0

Compute rate (processor-microseconds/cell/cycle):   23.608

Total memory usage:    32753.0586 MB (32.7531 GB)

 

 

I didn’t see anything in the FAQ that discusses memory usage other than the impact of the “leave_pinned” option, which apparently does not affect the memory usage in my case.  But I figure there must be a justification why OpenMPI would use 6 GB more than MVAPICH on the same case.

 

Thanks for any insights.  Also attached is the output of ompi_info –a.