The Effective Fragment Potential (EFP) method allows one to describe large molecular systems by replacing chemically inert part of a system by a set of Effective Fragments while performing regular ab initio calculation on the chemically active part [1-8]. The LIBEFP library is a full implementation of the EFP method. It allows users to easily incorporate EFP support into their favourite quantum chemistry package. LIBEFP is used by many major quantum chemistry packages, such as Q-Chem, PSI4, NWChem, GAMESS, Molcas, and others.

Detailed description of methods and algorithms can be found in two LIBEFP papers:

• Kaliman and Slipchenko, JCC 2013
• Kaliman and Slipchenko, JCC 2015

The latest release can be downloaded here.

To build LIBEFP from source you need the following:

• C compiler (with C99 standard and OpenMP support)

• POSIX complaint make utility (BSD make or GNU make will work) or CMake

• BLAS/LAPACK libraries (required when linking with LIBEFP)

If you are going to compile EFPMD program (required for tests):

• Fortran 77 compiler

First, copy the configuration file which suits you from config directory to the top source code directory. Rename the file to config.inc and edit it according to your needs. All available options are explained in comments. Defaults usually work well, but you may need to change the MYLIBS variable to link with additional libraries required by your setup. You may also need to add additional include directories to MYCFLAGS (using -I flag) or library search path to MYLDFLAGS (using -L flag).

To compile issue:

make

If you only need the library you can use:

make libefp

To run the test suite (optional) issue:

make check
make checkomp    # to test OpenMP parallel code
make checkmpi    # to test MPI parallel code

Finally, to install everything issue:

make install

For CMake instructions, see README-cmake.md.

The description of public LIBEFP API is available here. Also see this file for step-by-step instructions. Fortran bindings to LIBEFP are available in interface/efp.f90.

The EFPMD program is a molecular simulation package based on LIBEFP. It supports EFP-only molecular simulations such as geometry optimization and molecular dynamics. EFPMD is a part of the LIBEFP distribution. See this file for more information.

Parallel scaling of LIBEFP is shown below. Benchmarks were done on the SDSC Gordon supercomputer using EFPMD.

LIBEFP comes with a library of ready-to-use fragments. If you decide to generate custom fragment parameters follow the instructions below.

LIBEFP uses EFP potential files in format generated by GAMESS quantum chemistry package (see http://www.msg.ameslab.gov/gamess/). A version of GAMESS from August 11, 2011 is the currently a recommended and tested version. A set of pre-generated library fragments are available in the fraglib directory. If you want to generate parameters for custom fragments you should create GAMESS makefp job input similar to the fraglib/makefp.inp file. Using this input file as a template you can create EFP parameters for arbitrary fragment types.

After you created .efp file using GAMESS you should rename the fragment by replacing $FRAGNAME with your name of choice (e.g. rename $FRAGNAME to $MYH2O).

For a complete description of EFP data file format consult FRAGNAME section in GAMESS manual (see http://www.msg.ameslab.gov/gamess/).

• The main design principle for the LIBEFP library is Keep It Simple. All code should be easy to read and to understand. It should be easy to integrate the library into programs written in different programming languages. So the language of choice is C and no fancy Object-Oriented hierarchies.

• Be consistent in coding style when adding new code. Consistency is more important than particular coding style. Use descriptive names for variables and functions. The bigger the scope of the symbol the longer its name should be. Look at the sources and maintain similar style for new code.

• As with most quantum chemistry methods, EFP can require large amounts of memory. The guideline for developers here is simple: ALWAYS check for memory allocation errors in your code and return EFP_RESULT_NO_MEMORY on error.

• The code is reentrant which means that it is safe to use two different efp objects from two different threads. NEVER use mutable global state as it will break this. Store all mutable data in the efp object.

• Use -Wall -Wextra -Werror flags to make sure that compilation produces no warnings. Use make check to make sure that the code passes test cases.

1. Fragmentation Methods: A Route to Accurate Calculations on Large Systems. M.S.Gordon, D.G.Fedorov, S.R.Pruitt, L.V.Slipchenko. Chem. Rev. 112, 632-672 (2012).

2. Effective fragment method for modeling intermolecular hydrogen bonding effects on quantum mechanical calculations. J.H.Jensen, P.N.Day, M.S.Gordon, H.Basch, D.Cohen, D.R.Garmer, M.Krauss, W.J.Stevens in "Modeling the Hydrogen Bond" (D.A. Smith, ed.) ACS Symposium Series 569, 1994, pp 139-151.

3. An effective fragment method for modeling solvent effects in quantum mechanical calculations. P.N.Day, J.H.Jensen, M.S.Gordon, S.P.Webb, W.J.Stevens, M.Krauss, D.Garmer, H.Basch, D.Cohen. J.Chem.Phys. 105, 1968-1986 (1996).

4. Solvation of the Menshutkin Reaction: A Rigorous test of the Effective Fragment Model. S.P.Webb, M.S.Gordon. J.Phys.Chem.A 103, 1265-73 (1999).

5. The Effective Fragment Potential Method: a QM-based MM approach to modeling environmental effects in chemistry. M.S.Gordon, M.A.Freitag, P.Bandyopadhyay, J.H.Jensen, V.Kairys, W.J.Stevens. J.Phys.Chem.A 105, 293-307 (2001).

6. The Effective Fragment Potential: a general method for predicting intermolecular interactions. M.S.Gordon, L.V.Slipchenko, H.Li, J.H.Jensen. Annual Reports in Computational Chemistry, Volume 3, pp 177-193 (2007).

7. Water-benzene interactions: An effective fragment potential and correlated quantum chemistry study. L.V.Slipchenko, M.S.Gordon. J.Phys.Chem.A 113, 2092-2102 (2009).

8. Damping functions in the effective fragment potential method. L.V.Slipchenko, M.S.Gordon. Mol.Phys. 107, 999-1016 (2009).