In the framework of my protect I started test simulations using LAMMPS with reax/c as an interaction potential. Looking into output files I realize that the timestep of output file from reax/c differs from the one which set as an input - 0.25 (output) vs 0.1 (input) fs (I use real units). What is the correlation between LAMMPS and reax input data. It looks like 0.25 fs is a minimum value for timestep ? Can be that true?
Here I put data: