This plug-in provides a graphical user interface (GUI) for setting up complex simulations using LAMMPS. It uses a sub-flowchart that provides steps such as constant pressure and temperature (NPT) dynamics which give access to the functionality in LAMMPS in a more consistent and understandable way than the inscrutable fixes that LAMMPS uses.

These sub-flowcharts mirror the main flowchart in form and function and can use the same variables such as temperature and pressure that are accessible anywhere in the flowcharts. This allows "programming" a LAMMPS workflow in the same familiar way that SEAMM uses to represent the overall workflow.

Free software: BSD license
Documentation: https://molssi-seamm.github.io/lammps_step/index.html
Code: https://github.com/molssi-seamm/lammps_step

Features

Use of any forcefield supported by the forcefield plug-in:
- PCFF
- OpenKIM: EAM, MEAM, LJ, ReaxFF
Molecular statics: minimization
Molecular dynamics: NVE, NVT, and NPT with any of the approaches supported in LAMMPS
Automatic statistical analysis of averages from MD
- Detection of equilibration
- Mean and standard error of the mean for the sampling after equilibration
- Autocorrelation function and time
- Statistical inefficiency
- Plotting of results in the Dashboard
Using property values to drive MD. Rather than running MD for a length of time, automatically run long enough to determine a set of properties within given error bars.

Acknowledgements

This package was created with Cookiecutter_ and the molssi-seamm/cookiecutter-seamm-plugin_ project template.

.. _Cookiecutter: https://github.com/audreyr/cookiecutter .. _molssi-seamm/cookiecutter-seamm-plugin: https://github.com/molssi-seamm/cookiecutter-seamm-plugin

Developed by the Molecular Sciences Software Institute (MolSSI_), which receives funding from the National Science Foundation_ under award ACI-1547580

.. _MolSSI: https://www.molssi.org .. _National Science Foundation: https://www.nsf.gov

======= History

2026.3.26 -- Bugfix: MDI stress not calculated and error catching * MDI requires an option in fix mdi/qm for it to request the virial; it is not automatic for e.g. NPT calculations. The code now adds the keywords for periodic calculations but not molecular ones. * Added code to catch errors in the analysis of the GPU performance in cases where data is missing or cannot be interpreted correctly, so that calculation continues without error. * Fixed an issue returning the stdout and stderr from the LAMMPS/MDI calculation. LAMMPS logs to stdout and MDI to stderr to keep the streams separate.

2026.3.24 -- Bugfix: Issues calculating the stress and its sign. * Added an option to the fix mdi/qm line to ensure that MDI calculates the stress. * Fixed the sign issue in the stress between LAMMPS and MDI.

2026.3.23 -- Improved the performance output and results for MD

2026.3.22.1 -- Fixed issue running LAMMPS using a conda environment * Fixed a problem running LAMMPS directly in a conda environment, due to the construction of the command line having an error.

2026.3.22 -- Added support for MDI and GPUs, mainly for MLFFs * Added monitoring of GPU utilization and memory utilization * Added more output about the performance of the run to the job output. * Added the ability to run MACE MLFFs using MDI. This avoids needing a GPU version of LAMMPS that supports ML-IAP, making it much easier to run these MLFFs.

2026.2.25 -- Added support for ML-IAP forcefields. * Handle forcefields with the extension .pt and containing 'mliap' in their name using ML-IAP.

2026.1.28 -- Bugfix: Fixed error in analysis of non-periodic molecules * Recent additions to the analysis introduced a bug for non-periodic systems, where in some circumstances the code crashed due to interpreting the cell info in LAMMPS dump files.

2026.1.8 -- Improved handling of atom position and velocity trajectories. * Added the time and units for the time to the atomic poisition and velocity trajectories used in e.g. diffusivity to ensure the correct times are used.

2025.12.9 -- Enhancements: Running on GPUs and output from minimization * Ensured that at least 1 GPU is used when running using PyTorch potentials. This allows running from the commmand-line or locally in addition to using SLURM, which already worked. * Added the cell parameters, volume, and density, plus the change in each, whne running minimizations. These are now results that can be stores in tables and saved in variables.

2025.11.22 -- Bugfixes: volume units and converting pressure units. * Fixed the incorrect units for volume in the metadata for LAMMPS. * Fixed a crash caused by LAMMPS storing large pressures as integers, not floating point numbers.

2025.11.19.1 -- Bugfix: Properties not handled in MD runs * Due to an inadvertant error in the previous release, properties were not placed in local variables, tables or as properties on the configuration if requested.

2025.11.19 -- Enhancements to properties and storing trajectories in the database

This release enhances the LAMMPS step with property calculations and trajectory storage
capabilities. The main focus is adding formation energy calculations (DfE0) using
tabulated element energies, storing stress as a 6-component Voigt vector, and
enabling trajectory import into the database as system configurations.

* Added formation energy (DfE0) and atomization energy calculations using tabulated
  reference data for atomic energies
* Implemented trajectory save functionality to store MD trajectories as
  configurations in the database 
* Reorganized stress tensor output to follow Voigt ordering (Sxx, Syy, Szz, Syz,
  Sxz, Sxy) and added stress vector property

2025.11.9 -- Bugfix: Error with units in trajectories when using OpenKIM

 This release adds support for volume units across all LAMMPS unit systems to fix
 a bug in trajectory analysis when using OpenKIM potentials that use "metal" units.
 The changes enable proper unit conversion for volume quantities during trajectory
 analysis. Key changes are

 * Added "volume" definitions to all seven LAMMPS unit systems
 * Added volume dimensionality mappings in the unit conversion dictionaries
 * Implemented unit conversion logic in trajectory analysis to convert from LAMMPS
   units to SEAMM units

2025.11.6 -- Bugfix: Error in writing extxyz trajectory files. * Fixed the condition check from "Pxx" to "Sxx" to properly detect stress data * Corrected the case of the stress component variable from 'sxy' to 'Sxy' for consistency

2025.11.5 -- Bugfix: Fixed errors with trajectories using OpenKIM potentials

Fixed an error writing the dumpfile for trajectories for e.g. MEAM potentials where there are no charges, and the Open KIM atom type does not support charges on the atoms.
Fixed an error with the element labels for OpenKIM potentials.

2025.9.24 -- Bugfix: Error with cell angles for sheared cells

Fixed an error with sheared cells that actualy had 90 degree angles, which were incorrectly made to be 0 degress :-)

2025.9.21.1 -- Added cell angle parameters to properties

The cell lengths, but not angles, were available as properties. This update adds the angles.

2025.9.21 -- Bugfix: error in NPT dynamics

Fixed an error that slipped through and caused the code to crash.

2025.9.20 -- Completed handling of stress in NPT dynamics

Added the ability to define the offdiagonal stresses XY, XZ, and YZ, as well as to fix any of the directions. This is now handled in the same way as the Minimization step handles the stress.

2025.9.2 -- Bugfix: error in the stress parameters in the GUI

There was an error in the setup of the stress parameters, which caused a crash. This is now fixed.

2025.8.27 -- Added cell optimization to minimization step

Added the ability to optimize the cell during minimization, giving control over applied pressure or stress as well as fixing lattice directions.
Allowing "=" as the first character of a parameter to signal that it is a variable or expression that needs to be evaluated before use.

2025.8.21 -- Added stress and improved graph handling

Added stress for periodic systems to the state trajectory, and added to the properties and results so that it can be saved and used elsewhere.
Improved the formatting of graphs and added options for writing the graphs to a variety of formats such as HTML, PDF, and JPG.

2025.8.2 -- Initial pass at trajectories.

Added option to save the geometry & energies for a trajectory, and converting to the extended xy format that e.g. ASE uses, and the MACE ML code uses as input.
Bugfix: an error slipped into the tabulated angle term used for PF6- when formatting the forcefield parameters and tracking the provenance. This is fixed.

2025.7.23 -- Added a complete description of number of cores used.

Added a detailed description of how the codes decided on the number of cores to use, so that users can understand and change the limits in seamm.ini to their desired values.
Changed the default number of atoms per core (natoms-per-core) to 100.

2025.6.25 -- Bugfix: ReaxFF memory allocation

Apparently the memory handling for reafxx in LAMMPS is kludgy, and there is an empirical fudge-factor that helps solve errors with e.g. numbers of bonds. This version increases the factor to 1.6 until we get more experience.

2025.6.23 -- Bugfix: handling of GPUs and KOKKOS arguments

Fixed issue detecting GPUs caused by differences in SLURM enviroment variables
Fixed inconsistencies of dash vs underscore in configuration arguments cmd-args and gpu-cmd-args, which are used for e.g. Kokkos.

2025.6.21 -- Added support for MACE PyTorch potentials.

2025.5.27 -- Bugfix: error in formatting forcefield parameters

The cleanup of the format of 'structure.dat' accidentally converted some floating point numbers into integers, with loss of precision. This is now fixed.

2025.5.26 -- Added more metadata to the structure.dat file.

Added the original parameters, when available, plus the version of each parameter to the 'structure.dat' file input to LAMMPS to aid in understanding which forcefield parameters were used and where they came from.
Cleaned up the format of the 'structure.dat' file to make it more readable.

2025.5.7 -- Bugfix: error if not saving configuration in minimization

The code crashed in the minimization step if the structure was discarded.

2025.4.9 -- Adding support for ReaxFF forcefields.

Added support for ReaxFF forcefields.
Added control for how charges are handled, and options for the various charge equilibration methods in LAMMPS.
For ReaxFF, include the approximate enthalpy of formation, if available, and added it as a property.
Improved output for most substeps and included the forcefield used in the output.

2025.3.17.1 -- Bugfix: Incorrect handling of Dreiding hydrogen bonds

Fixed an issue with the Dreiding forcefield where having no hydrogen bonds in the system caused a crash.

2025.3.17 -- Bugfix: LAMMPS installer did not have correct conda environment

2025.3.16 -- Added Dreiding forcefield and support for structure handling

Added support for the Dreiding forcefield.
Added standard support for structure handling
Added the RMSD between initial and final structues in minimization, and expanded the possible minimizers and results from minimization.

2025.2.7 -- Installation issue

Encountered problems with the OpenMPI version of LAMMPS and with the installation of OpenKIM, so fixed the CONDA environment file to correct.
Changed logging at the INFO level to DEBUG to reduce the output when using geomeTRIC.

2024.8.22 -- Bugfix: error in number of cores.

Fixed an error if options for number of cores were strings, not numbers.

2024.7.31 -- Bugfix and improvements for parallel runs

There was an issue saving the pressure and volume as properties due to a mismatch in the names used. This has been corrected.
Improved how the code determines the number of processors to use for parallel runs, giving salience to the number of atoms in the system, but limiting to the LAMMPS and global limits on numbers of cores as well as the hardware available.

2024.7.25 -- Bugfix and improvements

Bugfix: Fixed issue with the initial seamm.ini file, created if it is missing from the installation.
Added the ability to set the number of points in the trajectories rather than the sampling rate.
Added diagnositic information and timings to available results.

2024.7.21.1 -- Minor internal change for GUI

Switched to new functionality in the SEAMM widgets to simplify the layout of the trajectories panel.

2024.7.21 -- Improved handling of trajectories and results

Improved control over the trajectories of positions, velocities, etc. to allow the user to give the number of points in the trajectory rather than the time interval of samples
Added volume of the cell to properties, and the cell lengths, density, and volume for NVE and NVT, where those parameters don't vary but are nonetheless useful in subsequent analysis.

2024.6.28.1 -- Internal release to fix issue making Docker image.

2024.6.28 -- Added energy and forces to properties

Added ability to get the energy and forces from single point calculations to supprt e.g. energy scans.
Fixed issue with assigning atoms types if they have not been assigned but are None
Updated for change in the order of units in the new version of pint
Improved analsys based on the output file.

2024.3.22 -- Corrected issue with e.g. heat flux calculations

Corrected an issue running LAMMPS via Python, introduced in the new scheme for executing. It ignored parallelism.

2024.3.21 -- Switched to new installation scheme

Fully support ~/SEAMM/lammps.ini
Updated to new installer
Support for Conda and Docker installation.

2024.1.18 -- Restructured to support running in containers.

2023.11.7 -- Bugfix: properties that are constant

A property, such as the total energy, can be a constant over an MD run due to precision of the trajectory. This caused errors because the autocorrelation function is not defined. These cases are now detected and the ACF not calculated for them.

2023.11.6 -- Bugfix in thermal conductivity

Due to change in input file name.

2023.9.6 -- Corrected issues with final coordinates; added velocities

There was a problem with getting the final coordinates from a dump file.
Added saving and reusing velocities so now a second LAMMPS step will by default use the velocities from the previous step, which is what you would expect.

2023.8.31 -- Bugfix: not reading structure correctly after dynamics

2023.8.27 -- Added support for tabulated angle potentials.

Support for the CL&P-OPLSAA potential for octahedral PF6-

2023.8.21 -- Bugfix: x-axes length on graphics incorrect

The size of the x-axes of the trajectory graphs were wrong, often much too large, compressing the actual data near the beginning of the graph.
Fixed an issue with systems with no non-bonds.

2023.6.17 -- Bugfix: more centroid/stress/atom issues

Avoided using centroid/stress/atom for heat flux in standard NVE, NVT, ... dynamics with Class 2 forcefield.
Added option to not use centroid/stress/atom for any forcefield. 2023.6.16 -- Heat flux with PCFF
centroid/stress/atom does not work with Class 2 forcefields, so don't use for PCFF. 2023.5.29 -- Self diffusion and other improvements
Added trajectory panel to support diffusion, viscosity and simple thermal conductivity.
Added support for separate GPU versions of LAMMPS.
Added support for command-line arguments to LAMMPS, mainly used for accelerators.
Added support for using modules.

2023.4.24 -- Support for thermal conductivity

Internal changes to support thermal conductivity with its embedded flowchart.
Added the heat flux substep.
Now delete output and reference files when rerunning, so the output is clean.
Internal changes to support running LAMMPS from a Python driver.
Corrected units of properties returned from LAMMPS when e.g. metal units used.
Added support for Buckingham potentials
Fixed issues with and cleaned up the use of hybrid types for bonds, angles, ....
Fixed issues with the alignment of some of the widgets in the GUI.

2023.4.9 -- Hid the warning from pymbar

Importing pymbar timeseries writes a warning to the terminal about its proper usage. SEAMM already handles the warned case, so the message is simply confusing to users and hence this release hides it.

2023.4.6 -- Better forcefield handling.

Added correct molecule numbers for valence forcefields.
Correctly handle ReaxFF from OpenKim
Updated for some minor changes in OpenKim

2023.2.6 -- Added handling of OPLS-AA forcefield

Added handling of the OPLS-AA forcefield
Moved documentation to new MolSSI theme and diátaxis layout
Cleaned up internale dependencies and workflows for GitHub

2022.10.31 -- Bugfix: properties with commas Properties with commas in their name in data/properties.csv need to have quotes to protect the property name!

2022.10.27 -- Added properties

Added properties to be saved in the database.
Updated calls to pymbar because the names of methods were changed.
Add the missing references for pymbar

2021.2.11 (11 February 2021)

Updated the README file to give a better description.
Updated the short description in setup.py to work with the new installer.
Added keywords for better searchability.

2021.2.4.1 (4 February 2021) Internal patch to fix CI; no changes for users.

2021.2.4 (4 February 2021) Updated for compatibility with the new system classes in MolSystem 2021.2.2 release.

2020.12.4 (4 December 2020) Internal: switching CI from TravisCI to GitHub Actions, and in the process moving documentation from ReadTheDocs to GitHub Pages where it is consolidated with the main SEAMM documentation.

2020.11.2 (2 November 2020) Updated to be compatible with the new command-line argument handling.

2020.10.13 (13 October 2020) Added capability to run MD until a set of user-selected properties are converged to requested accuracy.

2020.9.25 (25 September 2020) Updated to be compatible with the new system classes in MolSystem.

2020.8.2.1 (2 August 2020) Bugfix: Fixed problem with nonbonds and charges just introduced.

2020.8.2 (2 August 2020) Bugfix: Corrected the time units when using metal units with e.g. EAM potentials.

2020.8.1 (1 August 2020) Added support for OpenKIM potentials.

0.9.4 (29 May 2020) Cleaned up the output for the statistical analysis.

0.9.3 (29 May 2020) Fixed issue with settings for bins in LAMMPS for small nonperiodic systems with just a few atoms.

0.9.2 (25 May 2020) Switched to using PYMBAR for detecting covergence to equilibrium for MD runs. This is a more robust solution than the previous approach.

0.9.1 (24 May 2020) Support for rigid water models, such as TIP-3P.

0.9 (15 April 2020) Support for plots in the dashboard of properties from MD. Added option to produce local HTML for the above plots.

0.8.2 (2020-01-25)

No significant changes in functionality.
Incorporating changes to the SEAMM infrastructure, which simplify the code for plug-ins.
Updating the Travis CI to handle incompatible changes in Travis, and to use Conda environments in all steps.

0.7.1 (18 December 2019) Fixed problem with assigning charges to the system.

0.7.0 (17 December 2019) General clean-up of code and output.

0.6 (8 September 2019)

Switched to ConfigArgParse for handling command-line arguments.
Added the locations of LAMMPS executables to a configuration file for easier access.

0.5.2 (31 August 2019) Defined the correct requirements for installation.

0.5.1 (30 August 2019) Bugfix: corrected the name of the LAMMPS executable.

0.5.0 (30 August 2019) Added ability to use serial or parallel versions of LAMMPS based on an environment variable.

0.3.1 (27 August 2019) Added initial, fairly reasonable output.

0.2.1 (29 July 2019) First release on PyPI.

Keywords

FAQs

What is lammps-step?

Is lammps-step well maintained?

Did you know?

Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.

Install