Gmx energy LJ (SR) Calculation

User discussions
1

GROMACS version: 2022.3
GROMACS modification: No

Hello,

I am looking at simple MD simulations of homogeneous liquid Argon to understand GROMACS’s parameters before moving onto more complex systems.

Currently, I am having an issue replicating the gmx energy LJ (SR) calculation. My total LJ calculation for all unique pairs should overestimate GROMACS’s, then by excluding pairs outside the cut-off and subtracting the potential shift be brought to GROMACS’s LJ (SR). However, my total LJ calculation underestimates GROMACS’s immediately.

Could someone help me understand why based on the below discussion?

Method

  • Using gmx trjconv to generate .gro files of the coordinates of n particles every ns of the 11 ns simulation;
  • Reading the coordinates into Python and calculating the total LJ between all unique pairs at each ns frame

Results
Figure 1

  • The red points are my calculated total LJ for all unique pairs;
  • The blue data is the LJ (SR);
  • The green line is the LJ (SR) average + the Dispersion Correction;
  • The orange data is the total Potential

Thoughts

  • I compared the positive and negative contributions in my total LJ. The positive contributions are only about 50 kJ/mol, so I don’t think I am somehow overly including positive contributions
  • Pairs outside the cut-off have negative energy, so excluding them would make my values less negative
  • The shift, V(r_cut-off), is negative, so subtracting it from contributions would make my values less negative

Non-Default Parameters

integrator = md
dt = 0.001
nsteps = 11000000

cutoff-scheme = verlet
nstlist = 10
pbc = xyz

vdw-modifier = Potential-shift
rvdw = 1.0
DispCorr = EnerPres

tcoupl = Nose-Hoover
nsttcouple = 5
tc-grps = system
tau-t = 0.1
ref-t = 117.33

Thanks!
Tom