Investigate GROMACS builtin early stopping capabilities

[steffencruz]

Instead of implementing ES logic externally we prefer to use GROMACS to do this. This should reduce bugginess, instability and complexity.

[steffencruz]

GROMACS, a molecular dynamics package primarily designed for simulating proteins, lipids, and nucleic acids, does not have a direct "early stopping" feature like those found in machine learning frameworks. However, it does support various ways to control and stop simulations based on certain conditions, which can be somewhat similar to early stopping. Here's a breakdown of how you can manage this:

1. Stopping Based on Performance or Stability Criteria:

   • GROMACS doesn't natively support stopping simulations when a certain stability or performance criterion (like convergence in protein folding simulations) is met. Simulations generally run for a predefined number of steps or time duration specified in the input file (.mdp file).
   • Users typically need to manually monitor the simulation outputs and decide when to stop based on their specific criteria, such as the stability of the system energy, RMSD (Root Mean Square Deviation), or other properties indicative of system convergence.

2. Custom Scripts or Extensions:

   • For advanced users, custom scripts can be written to periodically check simulation outputs (like energy, RMSD, etc.) and terminate the simulation if certain criteria are met. This involves using tools that can read and process output files like .log or trajectory files (.xtc, .trr), then using system commands to stop GROMACS.

3. Checkpointing and Resuming:

   • GROMACS supports checkpointing, where it periodically writes out state files (.cpt) that can be used to resume a simulation. This feature can be used in conjunction with custom scripts to stop and later resume simulations based on analysis of intermediate results.

4. Implementation and Default Settings:

   • By default, GROMACS runs until the number of steps defined in the simulation parameters file is reached. There is no built-in early stopping mechanism that automatically terminates the simulation based on molecular stability or convergence criteria.

5. Inspecting Output Files:

   • To inspect simulation progress and determine whether to manually stop it early, you can look at several output files:
     • .log file: Contains detailed logs of the simulation, including performance metrics and warnings.
     • .edr file: Energy file that can be analyzed using gmx energy to extract various thermodynamic properties.
     • Trajectory files (.xtc, .trr): Contain the position (and optionally velocities) of atoms over time. Use tools like gmx traj to analyze trajectories.

6. Modifying the .mdp File:

   • You might want to customize the .mdp file to adjust the output frequency of logs and energy data to monitor the simulation more closely, which can aid in deciding when to apply an early stopping criterion manually.

To conclude, while GROMACS does not support automatic early stopping based on internal criteria like convergence, users can implement custom monitoring and stopping mechanisms using external scripts and checkpointing features. Checking and analyzing the output files regularly during the simulation will be crucial for making decisions about stopping early.