HPC/Applications/lammps/Package OMP

From CNM Wiki
< HPC‎ | Applications‎ | lammps
Revision as of 23:17, November 5, 2012 by Stern (talk | contribs) (→‎Package OMP)
Jump to navigation Jump to search

Package OMP

LAMMPS modules since 2012 are compiled with yes-user-omp, permitting multi-threaded runs of selected pair styles, and in particular MPI/OpenMP hybrid parallel runs. To set up such runs, see HPC/Submitting and Managing Jobs/Advanced node selection.

Be careful how to allocate CPU cores on compute nodes.

  • The number of MPI tasks running on a node is determined by options to mpirun.
  • The number of threads that each MPI task runs with is usually determined by the environment variable OMP_NUM_THREADS, which is 1 by default on Carbon.
  • The LAMMPS package omp command has a mandatory argument Nthreads to either override OMP_NUM_THREADS, or use it when set to *.

Usage

  1. Use the command package omp Nthreads mode near the beginning of your LAMMPS control script.
  2. Do one of the following:
  3. In the job file or qsub command line, reserve entire nodes.
  4. Call the lmp_openmpi (regular) binary.

Input file example

Examples:

package omp * force/neigh
package omp 4 force

Job file example

  • Single node:
#PBS -l nodes=1:ppn=8
…
lmp_openmpi -suffix omp -in infile
  • Multiple nodes:
# See HPC/Submitting and Managing Jobs/Advanced node selection#OpenMP/MPI hybrid for #PBS directives.
…
mpirun … \
       lmp_openmpi -suffix omp -in infile