Purdue University

openmpi

Link to section 'Description' of 'openmpi' Description

An open source Message Passing Interface implementation.

Link to section 'Versions' of 'openmpi' Versions

• Bell: 3.1.4, 3.1.6, 4.0.5, 4.1.3, 4.1.5
• Scholar: 2.1.6, 3.1.6, 4.1.5
• Gilbreth: 3.1.6-gpu-cuda10, 3.1.6-gpu-cuda11, 4.1.5-gpu-cuda11, 4.1.5-gpu-cuda12
• Negishi: 4.1.4
• Anvil: 3.1.6, 4.0.6, 4.1.6
• Gautschi: 4.1.6, 5.0.5

Link to section 'Module' of 'openmpi' Module

You can load the modules by:

module load openmpi

Link to section 'Compile MPI Code' of 'openmpi' Compile MPI Code

$ mpif77 program.f -o program

$ mpif90 program.f90 -o program

$ mpif90 program.f95 -o program

$ mpicc program.c -o program

$ mpiCC program.C -o program

Link to section 'Run MPI Executables' of 'openmpi' Run MPI Executables

Create a job submission file:

#!/bin/bash

#SBATCH  --nodes=2
#SBATCH  --ntasks-per-node=128
#SBATCH  --time=00:01:00
#SBATCH  -A XXXX

srun -n 256 ./mpi_hello

SLURM can run an MPI program with the srun command. The number of processes is requested with the -n option. If you do not specify the -n option, it will default to the total number of processor cores you request from SLURM.

To run MPI executables, users can also use mpirun or mpiexec from openmpi. Note that mpiexec and mpirun are synonymous in openmpi.

mpirun -n number-of-processes [options] executable