This test runs an overdamped Langevin dynamics simulation. The test quantity is then the average temperature of the system. The test passes if the average temperature is within a certain tolerance of the target temperature.
The damping constant is determined by scaling the damping time by the mean atomic mass units of the simulation cell. This provides a reasonable damping constant for the simulation cell, as LAMMPS apply the damping constant in fix langevin to scale the momentum term as (m/damping) v. Therefore for any simulation cell the damping will now effectively be scaled as ((m/mean_mass)/damping) v, providing somewhat equal footing.
Other defaults and assumptions are listed below, and can be modified by searching for the Parameter in the devel.py file.
Defaults:
| Parameter | Value | Description |
|---|---|---|
timesteps |
1e-6 to 1e-3 fs | timesteps to tests simulations on |
temps |
1e1 to 1e4 K | target temperature of the simulation |
n_atoms |
250 | The desired number of atoms in simulation cell, it influences the parameter repl, which sets the replicatio of the cell, such that total number of particles are ~250 |
n_steps |
1000 | Number of steps to run the simulation for |
variance_tol |
0.2 | Tolerance for the average temperature fluctuation, 0.2 means, 0.8 * temp < temps < 1.2 * temp, where temp is the requested temperature |
langevin_damp |
10.0 | timesteps to avgerage over for Langevin damping constant |
This test calculates two properties namely, maximum-stable-timestep, and
all-stable-timestep-per-temperature. The maximum-stable-timestep is the absolute largest timestep that can run the simulation stably, irrespective of the temperature. this serves as a scalar value for comparing the stability of the simulation. The all-stable-timestep-per-temperature lists the combinations of all temperatures and timestemps for which the simulaitons ran stably. This serves as a detailed information on the stability of the simulation.