Dear LAMMPS users and developers,
In molecular dynamics (MD) simulations, the time integration and particle tracking steps are computationally expensive and become increasingly costly as system size and simulation time grow.
In our current workflow, we sometimes realize after a simulation has finished that we forgot to compute a specific quantity—for example, the center-of-mass position of a selected group of atoms. In such cases, we currently have no choice but to re-run the entire simulation from scratch just to obtain that single missing value.
This seems highly inefficient, and I would like to ask whether LAMMPS offers a more practical solution.
Specifically, is there a way to save the full trajectory (or selected per-atom data) during the simulation so that, afterward, we can compute additional properties—such as the center of mass—without having to redo the time integration? For instance, can we use a dump file or a restart file to reconstruct atomic positions and then perform calculations efficiently?
If such an approach exists, what is the recommended strategy in terms of file format or post-processing tools ?
Thank you in advance for your suggestions.