Jobscript

ORCA does not distribute the source code. The executable installed at Hipergator is built with OpenMPI 4.1.1 (the compiler is unspecified and assumed to be GNU/Linux x86_64). ORCA can run in parallel. However, there several important things that users need to know. ORCA calls mpirun internally, and most computer centers including Hipergator prefer mpirun (or srun with slurm controller) invoked externally. As the following example shows we are not using srun, instead directly call ORCA binary "orca" with the full path (essential!!). ORCA documentation does not give examples of how to use it with multiple nodes, but you may visit the ORCA forum to see related issues and solutions.

The input file must contain the parallel configuration section.

%pal nprocs n 
end

*where n is the number of processors requested. The number of processors must match with the total number of tasks requested in the slurm configuration.

Orca on a single node, with multiple cores:

!/bin/bash
#SBATCH --job-name=parallel_job      # Job name
#SBATCH --mail-type=END,FAIL         # Mail events (NONE, BEGIN, END, FAIL, ALL)
#SBATCH --mail-user=usename@ufl.edu  # Where to send mail
#SBATCH --nodes=1                    # Run all tsks on a single node
#SBATCH --ntasks=2                   # The total number of tasks  
#SBATCH --mem-per-cpu=500mb          # Memory per processor
#SBATCH --time=00:05:00              # Time limit hrs:min:sec
#SBATCH --output=parallel_%j.log     # Standard output and error log
pwd; hostname; date

echo "Running orca test calculation on a with four CPU cores"
echo "Date              = $(date)"
echo "Hostname          = $(hostname -s)"

echo "Working Directory = $(pwd)"
echo ""
echo "Number of Nodes Allocated      = $SLURM_JOB_NUM_NODES"
echo "Number of Tasks Allocated      = $SLURM_NTASKS"
echo "Number of Cores/Task Allocated = $SLURM_CPUS_PER_TASK"
echo ""

module load gcc/14.2.0 openmpi/5.0.7 orca/6.0.1

which mpirun; echo $PATH; echo $LD_LIBRARY_PATH
export ORCA_DIR=/apps/gcc/14.2.0/openmpi/5.0.7/orca/6.0.1
export OMPI_MCA_coll_hcoll_enable=0
export SRUN_CPUS_PER_TASK=1
$ORCA_DIR/orca job.inp  > job.ou
date

Orca on multiple nodes, with multiple cores:

!/bin/bash
#SBATCH --job-name=parallel_job      # Job name
#SBATCH --mail-type=END,FAIL         # Mail events (NONE, BEGIN, END, FAIL, ALL)
#SBATCH --mail-user=usename@ufl.edu  # Where to send mail
#SBATCH --nodes=2                    # The number of nodes              
#SBATCH --ntasks-per-node=8          # Maximum number of tasks on each node
#SBATCH --cpus-per-task=1            # number of CPUs per task. 
#SBATCH --mem-per-cpu=500mb          # Memory per processor
#SBATCH --time=00:05:00              # Time limit hrs:min:sec
#SBATCH --output=parallel_%j.log     # Standard output and error log
pwd; hostname; date

echo "Running orca test calculation on a with four CPU cores"
echo "Date              = $(date)"
echo "Hostname          = $(hostname -s)"

echo "Working Directory = $(pwd)"
echo ""
echo "Number of Nodes Allocated      = $SLURM_JOB_NUM_NODES"
echo "Number of Tasks Allocated      = $SLURM_NTASKS"
echo "Number of Cores/Task Allocated = $SLURM_CPUS_PER_TASK"
echo ""

module load gcc/14.2.0 openmpi/5.0.7 orca/6.0.1

which mpirun; echo $PATH; echo $LD_LIBRARY_PATH
export SRUN_CPUS_PER_TASK=1
export ORCA_DIR=/apps/gcc/14.2.0/openmpi/5.0.7/orca/6.0.1
export OMPI_MCA_coll_hcoll_enable=0
$ORCA_DIR/orca job.inp  > job.ou
date

Disclaimer: The above slurm configurations are hypothetical. The user must customize it based on the size of the calculation, available resources etc.