Skip to content

Commit

Permalink
Setting summit env for OOM errors
Browse files Browse the repository at this point in the history
OLCF recommends setting "OMPI_MCA_io" env variable to "romio314"
to prevent out of memory issues with the code. In our testing we
have found that setting this env gets rid of OOM errors with
certain E3SM simulations (ECP simulation runs).

However also note that setting this environment variable reduces
the performance of parallel HDF5 (when using NetCDF4P PIO iotype
to write data).

See Issue #2856
  • Loading branch information
jayeshkrishna committed Apr 22, 2019
1 parent c6709af commit b57d606
Showing 1 changed file with 2 additions and 0 deletions.
2 changes: 2 additions & 0 deletions cime/config/e3sm/machines/config_machines.xml
Original file line number Diff line number Diff line change
Expand Up @@ -3078,6 +3078,8 @@
</environment_variables>
<environment_variables mpilib="!mpi-serial">
<env name="HDF5_PATH">$ENV{OLCF_HDF5_ROOT}</env>
<!-- OMPI_MCA_io env is required due to OOM errors -->
<env name="OMPI_MCA_io">romio314</env>
<env name="PNETCDF_PATH">$ENV{OLCF_PARALLEL_NETCDF_ROOT}</env>
</environment_variables>
</machine>
Expand Down

0 comments on commit b57d606

Please sign in to comment.