Skip to content

CP2K

CP2K is a quantum chemistry and solid state physics software package that can perform atomistic simulations of solid state, liquid, molecular, periodic, material, crystal, and biological systems. CP2K provides a general framework for different modelling methods such as DFT using the mixed Gaussian and plane waves approaches GPW and GAPW. Supported theory levels include DFTB, LDA, GGA, MP2, RPA, semi-empirical methods (AM1, PM3, PM6, RM1, MNDO, …), and classical force fields (AMBER, CHARMM, …). CP2K can do simulations of molecular dynamics, metadynamics, Monte Carlo, Ehrenfest dynamics, vibrational analysis, core level spectroscopy, energy minimisation, and transition state optimisation using NEB or dimer method.

Using CP2K on Cirrus

CP2K is available through the cp2k module. Loading this module provides access to the MPI/OpenMP hybrid cp2k.psmp executable. To see which versions are available: 

$ module avail cp2k
...
   cp2k/2025.2

Optional packages

The centrally installed module versions of CP2K are managed by Spack. To see the list of optional packages built, along with the compiler details, one can use, e.g.: 

$ module load cp2k
$ cp2k.psmp --version
 CP2K version 2025.2
 Source code revision
 cp2kflags: omp libint fftw3 libxc parallel scalapack xsmm
 compiler: GCC version 14.2.1 20240801 (Red Hat 14.2.1-1)
...
showing that CP2K has been built with OpenMP, libint, FFTW3, and so on.

Running parallel CP2K jobs (MPI/OpenMP hybrid)

To run CP2K using MPI and OpenMP, load the cp2k module and use the cp2k.psmp executable.

For larger jobs requiring one or more nodes, an exclusive SLURM submission is appropriate. For example, a job using two nodes (576 cores), with two cores (OpenMP threads) for each MPI process:

Exclusive SLURM job submission script for CP2K 

#!/bin/bash

#SBATCH --export=none
#SBATCH --time=00:20:00

#SBATCH --nodes=2
#SBATCH --exclusive

#SBATCH --ntasks-per-node=188
#SBATCH --cpus-per-task=2

#SBATCH --partition=standard
#SBATCH --qos=standard

#SBATCH --distribution=block:block
#SBATCH --hint=nomultithread

module load cp2k

export OMP_NUM_THREADS=2
export OMP_PLACES=cores

srun cp2k.psmp -i H2O-0512.inp
In this example, SLURM allocates two nodes based on --nodes=2 and --exclusive, and the srun parameters are controlled via--ntasks-per-node=188and--cpus-per-task=2inherited from the environment, together withOMP_NUM_THREADS=2`.

A valid budget may be required, e.g., 

#SBATCH --account=z00`

A relatively small CP2K job requiring less than a full nude should be run in non-exclusive mode. E.g., A job requiring 18 MPI processes each with two threads might be:

Non-exclusive SLURM job submission for CP2K 

#!/bin/bash

#SBATCH --export=none
#SBATCH --time=00:20:00

#SBATCH --partition=standard
#SBATCH --qos=standard

#SBATCH --ntasks=18
#SBATCH --cpus-per-task=2

#SBATCH --distribution=block:block
#SBATCH --hint=nomultithread

module load cp2k

export OMP_NUM_THREADS=2
export OMP_PLACES=cores

srun cp2k.psmp -i H2O-0064.inp
This job will use a total of 36 cores.

See Running Jobs on Cirrus for further general information on SLURM submissions.

Compiling and testing CP2K on Cirrus

Using Spack

If a specific version of CP2K is required, and this is not available as a central module, it is possible to build CP2K using Spack. 

$ module load spack
# spack info cp2k

For further information, see

Builds using the toolchain

The recommendation for developers who wish to build their own version is to use the toolchain build process.

Use a starting point which is PrgEnv-gnu, and use MKL for linear algebra. This involves unloading the the default cray-libsci module and loading MKL, e.g., 

module load PrgEnv-gnu
module load cray-python
module load cray-fftw

module unload cray-libsci
export MKLROOT=/opt/intel/oneapi/mkl/2025.0

source ${MKLROOT}/env/vars.sh
One should then be able to run the toochain build via, e.g., 
./install_cp2k_toolchain.sh --enable-cray --with-mkl=system

Testing

The entire regression test suite can be run using, schematically, 

CP2K_DIR=$(pwd)/cp2k

source ${CP2K_DIR}/tools/toolchain/install/setup

export OMP_NUM_THREADS=2
export OMP_PLACES=cores

${CP2K_DIR}/tests/do_regtest.py \
        --workbasedir=$(pwd) \
        --maxtasks=36 \
        --mpiranks=2 \
        --ompthreads=${OMP_NUM_THREADS} \
        --mpiexec="srun --ntasks=2 --cpus-per-task=${OMP_NUM_THREADS}" \
        ${CP2K_DIR}/exe/local psmp
All tests should pass. One should see, e.g., 
------------------------------- Summary --------------------------------
Number of FAILED  tests 0
Number of WRONG   tests 0
Number of CORRECT tests 4261
Total number of   tests 4261

Summary: correct: 4261 / 4261; 11min
Status: OK