Mismatch Improper Energy between LAMMPS and CP2K

[Trevor Gayer]

Hi CP2K users,

I am attempting to move a system from LAMMPS to CP2K so that I can perform QM/MM calculations. Before I implement the QM portion I wanted to first make sure that the MM energies match when I make the switch. I noticed that the Improper energy is significantly different between the two. 

LAMMPS is explicit that the first atom listed in the improper is the central atom. The CP2K documentation does not specify which atom is the central atom. Also, what is the difference between an improper torsion and an out-of-plane bend in CP2K? They have the same functional forms, and both use the angle between two planes, but they require atom kinds listed in different orders and sometimes compute different values.

My workflow is as follows:

1. Use VMD to get pdb and psf files from LAMMPS data files

2. Python script to modify pdb and psf files to change atom kinds and residue names (I don't change any of the improper information generated by VMD)

3. Run CP2K with pdb and psf files

Below are comparisons between the two software for a single molecule extracted from my larger system.

LAMMPS:

E_bond = 11.851723

E_angle = 23.60059

E_dihed = 11.19677

E_impro = 122.73291

CP2K:

E_bond = 11.8517

E_angle = 23.6006

E_torsion = 11.1968

E_imptors = 134.0361

E_opbend = 126.9725

Attached is a zip file containing my LAMMPS data, input, and output files as well as my CP2K pdb, psf, input, and output files. All files are for a single molecule extracted from a larger system that I have.

Thank you,

Trevor