REPACCS CPU

Work on REPACSS is managed by the Slurm workload manager. After logging in, you land on a login node, which is meant only for preparing work like editing files, compiling code, and staging data, not for running it. All computation must be submitted to the compute nodes as a Slurm job. Jobs run either as batch scripts (submitted with sbatch) or as interactive sessions (started with the interactive wrapper, which REPACSS recommends over calling salloc directly). CPU work runs in the zen4 partition, which is also the cluster's default, so a job that doesn't name a partition is sent there.

Slurm schedules by priority rather than strict first-come-first-served: jobs gain priority the longer they wait (aging), and a backfill mechanism lets smaller or shorter jobs fill gaps ahead of larger ones, so tightly-scoped jobs often start sooner. The zen4 partition handles the full range of CPU workloads such as serial, multithreaded (OpenMP), distributed (MPI), and hybrid MPI+OpenMP, along with job arrays and dependencies for managing many related runs. One REPACSS-specific detail worth knowing: each Slurm "CPU" is a hyperthread, so for threaded code you'll generally want --cpu-bind=cores to pin threads to physical cores.

Jobs are charged against your ACCESS allocation at 1 ACCESS credit per CPU core-hour, so one full 256-core node used for one hour draws 256 credits. Walltime limits aren't stated in the user guide, so confirm those with sinfo/scontrol or the REPACSS team.

The following is an example on how to create and compile a test program and the basics of a job script:

1. Create a file named mpi_program.c with the following content:

#include <stdio.h>
#include <mpi.h>
int main(int argc, char **argv) {
   int rank, size;
   char processor_name[MPI_MAX_PROCESSOR_NAME];
   int name_len;
   MPI_Init(&argc, &argv);
   MPI_Comm_rank(MPI_COMM_WORLD, &rank);
   MPI_Comm_size(MPI_COMM_WORLD, &size);
   MPI_Get_processor_name(processor_name, &name_len);
   printf("Hello from processor %s, rank %d out of %d processors\n",
          processor_name, rank, size);
   MPI_Finalize();
   return 0;
}

2. Load the required modules and compile the program:

module load gcc/15.2.0
module load mpich/4.3.2
mpicc mpi_program.c -o mpi_program

3. Create a file named mpi_job.sh with the following contents:

#!/bin/bash
#SBATCH --job-name=mpi_job
#SBATCH --output=mpi_job.out
#SBATCH --error=mpi_job.err
#SBATCH --partition=zen4
#SBATCH --time=01:00:00
#SBATCH --nodes=2
#SBATCH --ntasks=8
#SBATCH --cpus-per-task=1
#SBATCH --mem-per-cpu=2G

# Load modules
module load gcc/14.2.0
module load mpich/4.1.2

# Run MPI program using srun (for OpenMPI/MPICH)
srun -n 8 ./mpi_program

4. Make the script executable and submit it using sbatch:

sbatch mpi_job.sh

For more details, please visit Credits for REPACSS Resources page, REPACSS Jobs Basics, and the Job Examples page.