HPC GROMACS

Treat GROMACS as a staged workflow from preprocessing to analysis.

Start

1. Read references/workflow-manual.md before building a run pipeline.
2. Read references/file-and-parameter-matrix.md when mapping .mdp, topology, structure, and index responsibilities.
3. Read references/ensembles-and-analysis.md when choosing minimization, NVT, NPT, production, and standard analysis commands.
4. Read references/cluster-execution-playbook.md when staging a GROMACS workflow for scheduler-backed cluster execution.
5. Read references/error-recovery.md when grompp, mdrun, or analysis commands fail.

Additional References

Load these on demand:

• references/topology-and-system-building.md for pdb2gmx, box setup, solvation, and ion insertion workflows
• references/restarts-and-output-control.md for checkpointing, append behavior, and long-job continuation
• references/mdp-coupling-and-constraints-matrix.md for thermostat, barostat, constraints, and timestep compatibility
• references/cluster-execution-playbook.md for grompp to mdrun staging, continuation, and cluster launch choices

Reusable Templates

Use assets/templates/ when a concrete starting scaffold is needed, especially:

• em.mdp
• nvt.mdp
• npt.mdp
• md_prod.mdp
• gromacs-mdrun-slurm.sh

Guardrails

• Do not skip grompp validation and jump straight into mdrun.
• Do not mix force fields, water models, and topology assumptions casually.
• Do not copy .mdp cutoffs, timesteps, or coupling settings without checking the intended force field and ensemble.
• Do not treat warnings from grompp as harmless by default.

Outputs

Summarize:

• input structure and topology sources
• .mdp stage and ensemble
• generated run artifact names such as .tpr
• analysis outputs requested or still needed