OpenMMPol is an open-source library to interface quantum chemical software with atomistic polarizable embedding. With OpenMMPol any quantum mechanical method that is able to provide the electrostatic potential, field, and field gradient for a given electronic density could be coupled to polarizable embedding with AMOEBA (and other forcefield). Through simple interface functions, it allows to compute the QM/AMOEBA contribution to the energy and to the selected Hamiltonian.
OpenMMPol also implements all the non-electrostatic terms of the MM force field (Van der Waals and bonded interactions). This allows the host code to compute the full potential for the embedded system.
OpenMMPol is written in Fortran and is distributed with interfaces to C, Fortran and Python3.
OpenMMPol is written and maintained by the MoLECoLab (Modeling Light & Environment in Complex Systems) research group at the University of Pisa (molecolab.dcci.unipi.it).
Documentation is generated with FORD and is available at project documentation page
OpenMMPol is free software, you can use it under the terms of the LGPL-3.0 License.
OpenMMPol is mantained as a git repository. For reproducibilty reasons we strongly encourage to build the library from a cloned repository. OpenMMPol is versioned using semantic versioning, a tag in the git repository is added for each version. At configure time a version identifier is hard-coded into the library and headers file. If the repo is not in a tagged commit, a suffix specifieng the number of commits from last tag, the last commit identifier and the state of the repository will be added to the version identifier.
A version identifier reads as <MAJOR>.<MINOR>.<PATCH>.r<N>.1234abc.dirty for a dirty repo at commit 1234abc that is N commits beyond the <MAJOR>.<MINOR>.<PATCH> tag.
In non-git release the version identifier can be hard-coded in include/version.h using the preprocessor variable _OMMP_VERSION:
#define _OMMP_VERSION "0.1.2"
Please cite the following papers if you use the library:
Polarizable embedding QM/MM: the future gold standard for complex (bio)systems? M. Bondanza, M. Nottoli, L. Cupellini, F. Lipparini and B. Mennucci Phys. Chem. Chem. Phys. 22, 14433-14448 (2020) doi:10.1039/D0CP02119A
A QM/MM Approach Using the AMOEBA Polarizable Embedding: From Ground State Energies to Electronic Excitations J. Chem. Theory Comput. 12, 3654-3661 (2016) D. Loco, É. Polack, S. Caprasecca, L. Lagardère, F. Lipparini, J.-P. Piquemal and B. Mennucci doi:10.1021/acs.jctc.6b00385
To build and install the package, you just need a compiler that supports standard Fortran2008ts (currently the
library is tested with gnu 7.5.0, intel 2021.7.0 and nvidia 10.2.89), cmake >= 3.20,
lapack libraries and optionally hdf5 library.
To install the requirements on OpenSuse Leap just use the following command:
zypper in gcc gcc-c++ gcc-fortran make cmake python lapack-devel liblapack3 hdf5 hdf5-devel zlib-devel
To build the package, just use cmake in the standard way; from the root directory of the git repository:
$ mkdir build
$ cd build
$ cmake ..
At configuration time you can control several build options:
-DCMAKE_BUILD_TYPE=<RELEASE|DEBUG> # to control the compiler flags (debug generate a much slower code with a more strict control)
-CMAKE_INSTALL_PREFIX=<PATH_TO_INSTALL> # eg. / /usr /usr/local /home/user/.local etc.
-DTESTLANG=<C|F03> # Language for the test programs used in make test / ctest
-DCMAKE_C_COMPILER -DCMAKE_CXX_COMPILER -DCMAKE_Fortran_COMPILER # Compilers to be used; eg. to use intel/nvidia compilers instead of defaults
Once configured you can build the library with
$ make
And install it with:
$ make install
To compile the python module you need the pybind11 package and numpy; to install those requirements on OpenSuse Leap use:
zypper in python-pybind11-common-devel python3-numpy python3-pybind11 python3-pybind11-devel
To compile and install the python packge, from the root of repository, just use:
pip install .