Skip to content

ORCA

ORCA is a quantum chemistry package with methods ranging from semiempirical and DFT to coupled-cluster and multireference approaches, with a focus on spectroscopic properties.

Loading the module

module load ORCA/6.0.0-gompi-2023b

To see all installed versions across clusters, check the software catalog or run module spider ORCA.

Always use the full path

ORCA must be invoked as $EBROOTORCA/orca (not just orca) so that it can find its parallel helper binaries. The environment variable $EBROOTORCA is set automatically when you load the module.

Expand to see all available modules on each cluster
ORCA/5.0.4-gompi-2022a ORCA/5.0.4-gompi-2023b ORCA/6.0.0-gompi-2022a
ORCA/6.0.0-gompi-2022a-xtb ORCA/6.1.0-gompi-2023b
ORCA/4.2.1-gompi-2019b ORCA/5.0.0-gompi-2021a ORCA/5.0.1-gompi-2021a
ORCA/5.0.2-gompi-2021a ORCA/5.0.4-gompi-2022a ORCA/6.0.0-gompi-2022a
ORCA/4.0.1.2-OpenMPI-2.0.2 ORCA/4.2.1-gompi-2019b ORCA/5.0.0-gompi-2021b
ORCA/5.0.3-gompi-2021a

Batch job examples

Running ORCA on local scratch avoids I/O contention on the shared filesystem. This is the recommended approach for single-node jobs.

orca.slurm
#!/bin/bash
#SBATCH --qos=regular
#SBATCH --job-name=orca
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=24
#SBATCH --cpus-per-task=1
#SBATCH --mem=100G
#SBATCH --time=24:00:00
#SBATCH --output=%x.%j.out
#SBATCH --error=%x.%j.err

module load ORCA/6.0.0-gompi-2023b

SCRATCH_DIR=/lscratch/$USER/$SLURM_JOB_ID
mkdir -p $SCRATCH_DIR

cp $SLURM_SUBMIT_DIR/input.inp $SCRATCH_DIR/
cd $SCRATCH_DIR

$EBROOTORCA/bin/orca input.inp > $SLURM_SUBMIT_DIR/input.out

cp *.gbw *.hess *.xyz $SLURM_SUBMIT_DIR/ 2>/dev/null
rm -rf $SCRATCH_DIR

If you need multi-node parallelism or the local disk is too small, run directly from /scratch:

orca.slurm
#!/bin/bash
#SBATCH --qos=regular
#SBATCH --job-name=orca
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=24
#SBATCH --cpus-per-task=1
#SBATCH --mem=100G
#SBATCH --time=24:00:00
#SBATCH --output=%x.%j.out
#SBATCH --error=%x.%j.err

module load ORCA/6.0.0-gompi-2023b

$EBROOTORCA/orca input.inp > input.out

Parallel keywords in the input file

ORCA's parallelism is controlled from the input file, not the command line. Make sure your .inp file contains a %pal block matching the SLURM request:

%pal
  nprocs 24
end

See also