[BUG] pair_style ylz needs explicit symmetrization of the interaction cutoff
maitane-muba opened this issue · 0 comments
Summary
MD trajectories generated using pair style ylz with single particle type and two particle types behave differently for identical seeds, initializations and pair_style coefficients. Pair style seems to incorrectly handle more than one particle type.
LAMMPS Version and Platform
LAMMPS version: 2 Aug 2023 - Update 2
Operating system: macOS Sonoma 14.3.1
Expected Behavior
A simulation with a single type of particle, e.g., type 1, should be identical to a simulation with two types of particles, e.g., type 1 and type 2, if the coefficients for interactions between types 1-1, 1-2 and 2-2 are set to be the same.
Actual Behavior
Trajectories do not agree.
For two simulation set ups with identical seeds, initializations and pair style coefficients, one with N = 10 particles of type 1, and another one with N = 5 particles of type 1 and N = 5 particles of type 2, computation of the forces shows discrepancy in force calculation even for time = 0.
Steps to Reproduce
Two systems are simulated:
- in.Test_SingleType simulates a box with N = 10 ellipsoids of type 1 interacting via ylz potential (corresponding initialization file = initial_conditions_SingleType.dat)
- in.Test_TwoTypes simulates a box with N = 5 ellipsoids of type 1 and N = 5 ellipsoids of type 2 (corresponding initialization file = initial_conditions_TwoTypes.dat)
Comparison of the force output in the respective Trajectory.dump files shows discrepancy.
Further Information, Files, and Links
Attached: files needed to reproduce bug.