page Hello World(s)! Hello-ing all over the cluster.

---

6. Hello World(s)!

Objective: Implement the basic “hello world” program using C and Python

There are many ways to tackle a problem. Even a basic “hello world” program that runs across multiple compute nodes can have more than one way to solve it. Two examples will be shown.

Implementing Hello World with C

resources: mpicc, sbatch

Log in as the user user instead of admin this time. Under /shared, create a new directory (you can name it something like hello_mpi_c or hello_world_c). Change to this directory and create a new file, hello_mpi.c:

#include <stdio.h>
#include <mpi.h>

int main(int argc, char** argv) {
  int node;
  MPI_Init(&argc, &argv);
  MPI_Comm_rank(MPI_COMM_WORLD, &node);

  // this is the stuff that's parallelized
  printf("Hello World from compute node: %d!\n", node);

  MPI_Finalize();
}

Now, compile the program.

mpicc hello_mpi.c

This should create a new file called a.out. This is your program. Finally, create a shell script used to submit this program to Slurm (name it sub_mpi.sh):

#!/bin/bash

# Define submission allocation
#SBATCH --nodes=4
#SBATCH --ntasks-per-node=4

cd $SLURM_SUBMIT_DIR

# print the hostname of the submission node
echo "submitted from $(hostname)"

# run the program
mpirun a.out

Finally, submit a new job to the cluster.

sbatch sub_mpi.sh

This should submit a job, and it should return “Hello World” 16 times with a different number after the phrase.

Implementing Hello World with Python (if you installed it)

Resources: mpirun

Create a new directory under /shared like you did before, but use a different name to indicate that it’s a python application. Create a new file hello_mpi.py:

#!/usr/bin/env python

from mpi4py import MPI
import sys

size = MPI.COMM_WORLD.Get_size()
rank = MPI.COMM_WORLD.Get_rank()
name = MPI.Get_processor_name()

sys.stdout.write(
  "Hello World! I am process %d of %d on %s.\n"
  % (rank, size, name))

Copy the sub_mpi.sh you made for the C program into this directory, and edit the last line:

mpirun python hello_mpi.py

Submit the job to the cluster:

sbatch sub_mpi.sh

As before, you should see “Hello World” return 16 times.

MPI Examples

MPI examples can be found on the LLNL HPC Tutorials Page. Experiment and run a few of these examples.