21 Dec 2016

Fixed a bug with the new neighboring refactor where it was sometimes not finding a build option when using the pair hybrid/overlay command with a manybody potential only assigned to a subset of atom types and pairwise potentials assigned to other types.

17 Dec 2016

Added a length option to the python command to allow strings longer than 64 characters to be returned by the Python function to LAMMPS.

For the USER-DPD package, this update also changed the way the internal energy of DPD particles is partitioned in the package's fix eos/cv, fix eos/table, fix eos/table/rx commands. This will change the output of simulations which use those commands, relative to previous versions.

This patch also includes a few small bug fixes for the 13Dec patch for the neighbor list refactoring in several of the accelerator packages that use it.

13 Dec 2016

Redesigned and restructured the various classes related to neighbor list building. Simulation results and performance should be unchanged, but internally it is now easier for developers to add new neighboring methods and various packages to add new styles for building lists.

There is also more output to the screen when a simulation is setup, with respect to how many and what kinds of neighbor lists LAMMPS is using internally.

Stan Moore (Sandia) fixed a bug with the kspace_style ewald/disp command for 2d slab geometries where the slab correction was not being computed on every timestep.

30 Nov 2016

New compute global/atom command which can assign a value from a global vector or array to each atom, creating a per-atom vector or array as an output. This can be useful for calculating per-atom properties based on what chunk an atom is in, since chunk properties are global values. E.g. computing the displacement of each atom from a chunk's center-of-mass.

Bug fix for dumping atoms in descending sorted order, via the dump_modify sort command.

22 Nov 2016

New temper/grem and fix grem commands have been added by David Stelter (Boston U), Edyta Malolepsza (Broad Institute), and Tom Keyes (Boston U). These allow multi-replica simulations using the generalized replica exchange method (gREM) originally developed by Kim, Keyes, Straub, J Chem Phys, 132, 224107 (2010). It uses non-Boltzmann ensembles to sample over first order phase transitions. This is done by defining replicas with an enthalpy-dependent effective temperature.

An update to the fix ipi command was provided by Michele Ceriotti (EPFL).

A new version v1/40 of the moltemplate tool was provide by Andrew Jewett.

17 Nov 2016

Several new pair styles have been added:

9 Nov 2016

Ariel Lozano has contributed two pair styles that are variants of a Born potential with damped shifted force Coulombic terms, the second for core/shell models.

5 Nov 2016

Axel Kohlmeyer and Aidan Thompson optimized several of the 3-body potentials with logic for the pairwise vs 3-body terms and their use of the neighbor list, resulting in significant speed-ups in some cases. This includes the pair sw, pair tersoff, and pair vashishta commands and their variants.

Stan Moore added a Kokkos version of the fix reaxc/species command.

27 Oct 2016

Added some functions to the library interface for LAMMPS. In particular a create_atoms() function that allows a driver program to pass in a set of atoms (similar to if they had been read from a data file), e.g. to replace the existing atoms and run a new simulation.

20 Oct 2016

Lars Pastewka (Karlsruhe Institute of Technology, KIT) has created NetCDF dump styles via new dump netcdf and dump netcdf/mpiio commands which are in a new USER-NC-DUMP package. Settings for the local NetCDF library are in lib/netcdf (and could be reused for other styles using NetCDF). The settings use the automatic configuration via the nc-config tool included in a typical netcdf install.

Andrew Jewett added mixing options to pair_style gauss.

BACKWARD COMPATIBILITY: The pair_style table command was looking for a keyword FP, when it should be FPRIME, as the doc page indicates. This is now changed require the full FPRIME in the tablulation file.

18 Oct 2016

Added a fix halt command which allows for checking a condition every N steps to halt a running simulation. The condition can compare a variable evaluation to a value, so is quite general. The "halt" can terminate LAMMPS or exit the current run.

Thanks to Carsten Svaneborg (SDU) and Axel Kohlmeyer (Temple U) for contributing ideas and code to this discussion/implemntation.

13 Oct 2016

Dan Bolintineanu (Sandia) made the new fix wall/gran/region command more robust when restarting with complex or nested union/intersect regions.

Aidan fixed a bug with fix gcmc when used with the fix shake command.

12 Oct 2016

Jim Larentzos (ARL) added a new fix dpd/energy command to the USER-DPD pacakge, which can be used for energy-conserving DPD time integration for standard energy dissipative DPD models, without use of the Shardlow method used by other commands in the USER-DPD package. See the new command's doc page for details.

11 Oct 2016

Added a LAST option to the dump_modify thresh command, which allows dumping of atoms with a specific property that has changed (or stayed the same) in a logical sense (<, >, ==, !=, etc), since the last time a dump was performed. As examples, this can be used to only dump particles that have crossed a periodic boundary or entered/exited a region.

Also added more complete info printed to the screen and logfile about which fix info is re-used by your script when reading a restart file. As explained on the read_restart doc page, fixes may store either global or peratom info in the restart file which can be used to reset the state of a fix of the same style and ID that you define in your input script after reading a restart file.

6 Oct 2016

Dan Bolintineanu (Sandia) has developed a new fix wall/gran/region command which allows regions to be used as boundaries on granular particles. Since regions can translate or rotate, or even change in size over time, this enables more complex boundary/particle interactions than the fix wall/gran command. Union and intersect style regions can be used to model complex boundaries. Dan also extended some of the regions defined by LAMMPS to allow for an open flag, that removes one or more of their faces, so they be more easily used as containers for granular particles.

There is a new examples/granregion directory with example scripts. Movies of those script simulations were added to the Movies page.

Fixed a bug with the pair_style rebo potential, where one (important) parameter needs to be different than the corresponding AIREBO parameter. Thanks to Antonino Favata (Sapienza University of Rome) for calling our attention to this.

Updated many of the example log files in the examples dir.

5 Oct 2016

Fixed some issues and adjusted the way some of the weight options for the balance and fix balance commands operate, especially for the application of the scale factor used by some of the weight styles. Clarified the details in the respective doc pages.

Added a logical XOR option (exclusive or) to variable formulas, the if command, and the dump_modify thresh command.

30 Sep 2016

Update to Colvars library used by USER-COLVARS package. Forgot to include two CMAP CHARMM potetnial files in the 26Sep16 patch.

29 Sep 2016

Removed the restriction for the prd command that atom sorting be turned off via the atom_modify sort command.

28 Sep 2016

Removed the restriction for the neb command for performing nudged elastic band calculations, that each replica needs to run on a single processor. Each replica can now run on any number of processors, as specified by the -partition command-line switch.

27 Sep 2016

Added weight options to the balance and fix balance commands for static/dynamic load-balancing. These include options for weighting by particle group, neighbor count, CPU time, or a custom atom-style variable that defines per-atom weights.

Weighting can be useful in these kinds of scenarios:

Thanks to Iain Bethune (EPCC) and Axel Kohlmeyer (Temple U) for suggesting and extending these ideas in the two commands.

26 Sep 2016

The following folks have contributed a CHARMM force-field fix for CMAP interactions that we are now releasing as a fix cmap command:

CMAP interactions are 5-body crossterms that are defined for two overlapping 4-body dihedrals. The tools/ch2lmp/charmm2lammps.pl Perl script was augmented to output these terms into LAMMPS data files when PDB files are processed with the CHARMM force field. Two CMAP parameter files for CHARMM 22 and 36 were also added to the potentials directory.

This effort also identified a deficiency in the pairwise CHARMM force fields currently implemented in LAMMPS, with respect to force-switching contributions, which will be remedied and released soon.

21 Sep 2016

Added options in the doc dir to fetch the doc pages from the LAMMPS web site, as well as locally build the old HTML version. You can also (already) locally build the new version of the HTML pages, so the doc pages are being removed from the public SVN/Git repositories, replaced by these new various options.

20 Sep 2016

Rodrigo Freitas (UC Berkeley) has provided an upgraded version of his fix ti/spring command in the USER-MISC package. The doc page for the command cites a 2016 paper from his group describing how to use the command to perform nonequilibrium thermodynamic integration. There is also an updated example script in examples/USER/misc/ti. The updated command can also be used with EAM potentials, which now have a hook to allow them to be scaled when used with thermodynamic integration.

BACKWARD COMPATIBILITY: This patch removes the fix ti/rs command, which has been deprecated in light of the enhanced fix ti/spring command.

15 Sep 2016

Fixed a bug with the comm_style tiled command where atoms that should be be deleted when crossing non-periodic boundaries, were sometimes not.

Thanks to the user Qifang on the LAMMPS mail list for flagging the issue.

13 Sep 2016

Anders Hafreager (U Oslo) added a tabulated version of the Vashishta potential as pair_style vashishta/table. It can run 3-5x faster with little loss in accuracy if sufficiently long tables are used.

Efrem Braun (UC Berkeley) contributed some code for added options to the compute bond/local commmand, which include partitioning the kinetic energy of the atomss into translational (of the center-of-mass), vibrational, and rotational energy. This can be useful for monitoring how energy is partitioned bewteen different degress of freedom.

7 Sep 2016

Fixed an issue with the compute omega/chunk command and variable group omega function when operating on chunks/groups of just 2 atoms, or linear molecules (e.g. for chunk = molecule). The conversion of angular momentum to omega needs to be handled differently in those cases to account for the singular inertia matrix.

Thanks to Efrem Braun (UC Berkeley) for pointing out this issue.

6 Sep 2016

Stan Moore (Sandia) has added support within the KOKKOS package for long-range electrostatics via the kspace_style pppm/kk command. He also fixed some issue with the recently added Kokkos version of pair_style reax/c.

27 Aug 2016

Steve Strong and Joel Eaves (U Colorado) added a fix flow/gauss command which uses Gaussian Dynamics (GD) to model fulid flow through pores, pipes, and channels.

23 Aug 2016

A small bug fix for indexing multiple computes in the fix ave/histo command

20 Aug 2016

Added a pbc option to the dump_modify command to force output atoms to be inside the simulation box, if desired. Without the option, atoms can be slightly outside the box, as explained on the dump doc page.

13 Aug 2016

Various small changes.

5 Aug 2016

Aidan Thompson (Sandia) has added a fix controller command, which can be used in a general manner to guide a simulation to a desired target.

It uses a control loop feedback mechanism known as a proportional-integral-derivative (PID) controller. The basic idea is to define a "process variable" which is a quantity that can be monitored during a running simulation. A desired target value is chosen for the process variable. A "control variable" is also defined which is an adjustable attribute of the running simulation which, which the process variable will respond to. The PID controller continuously adjusts the control variable based on the difference between the process variable and the target.

The examples/pid directory contains a script that used to fix controller to implement a simple thermostat. Many other control mechanisms are possible.

3 Aug 2016

Added options to the thermo_modify format and dump_modify format commands, and made their syntax consistent with each other.

Each now takes a line option to specify the format of an entire line of output, int and float options to specify the format of each integer or floating point value output, an M string option to change the format of a single value in column M of the output, or a none option to restore the default format for all output values.

BACKWARD COMPATIBILITY: The dump_modify format command previously only allowed the format for an entire line to be specified. Using that option now requires the "line" setting be used after "format", to distinguish it from the other options that were added.

1 Aug 2016

Enhanced the syntax for many commands to allow multiple selected values from a compute or fix vector, or multiple columns from a compute or fix array, to be specified with a single argument, using the same wildcard asterisk syntax that the pair coeff command uses. E.g. specifiying c_ID[*] as a single argument gets expanded into multiple arguments c_ID[1], c_ID[2], ..., using these wildcard rules:

where N is the length of the vector or number of columns in the array.

These commands now have the enhanced syntax:

Also removed a restriction on the header line printed before thermodynamic data is output during a run. The per-column fields were limited to 8 characters each which could mean long IDs for computes, fixes, or variable names were truncated. Now the full column name, as specified in the thermo_style custom command, is output. This includes the "c_", "f_", or "v_" prefix for compute, fix, variable quantities, so there will be no ambiguity if a compute and fix have the same ID.

BACKWARD COMPATIBILITY: The fix ave/time and fix ave/histo commands previously allowed an entire vector or array of input values (from a compute or fix) to be specified using just the ID of the compute or fix. You now have to use the wildcard syntax above to have the same effect.

BACKWARD COMPATIBILITY: This patch can change the format of one line in the screen or logfile output, namely the header line that describes the columns of thermodynamic output. If your output includes quantities from a compute or fix or variable, then that line will have a slightly different syntax, as described above. If you have a post-processing parser that reads that line or uses it to identify columns of data, you may to need to adjust your parse semantics.

30 Jul 2016

Maintenance patch with a few small tweaks to enable this version to be tagged as stable.

28 Jul 2016

Dave Heine (Corning) figured out we were using an incorrect equation for the pumping term in the pair lubricate/poly and pair lubricateU/poly commands. This patch replaces it with a different equation from the same Kim and Karilla textbook: Microhydrodynamics: Principles and Selected Applications. Mineola, NY: Dover Publications, 2005.

27 Jul 2016

A new dipole/dlm option was added to the fix nve/sphere and fix nvt/sphere commands (see this link for option details) by Iain Bethune (EPCC). This allows the direction of the dipole moment for the sphere to be integrated by the more accurate Dullweber-Leimkuhler-McLachlan (DLM) integration method.

Aidan Thompson (Sandia) added an axes option to the compute centro/atom command to allow 3 directional symmetry vectors to also be output to give information about an atom's local orientation relative to its neighbors.

22 Jul 2016

Added an optional cutoff keyword to the compute pair/local and compute property/local command, which allows the cutoff for interactions between pairs of particles to be included based on a per-atom-type cutoff or the radii of the 2 particles. The latter is typically more useful for granular systems and can reduce the size of the output list of interactions dramatically for polydisperse systems.

Some other updates to the Python wrapper from Richard Berger (Temple U) are also included.

15 Jul 2016

Various and sundry changes. Axel Kohlmeyer (Temple U) added a Gaussian distribution option to the fix deposit command. Andrew Jewett added a dihderal_style spherical command. Updates to the USER-COLVARS and USER-QMMM and USER-DPD packages.

13 Jul 2016

For granular models, the fix wall/gran command was not using the correct rigid-body mass for spherical particles (in rigid bodies) bouncing off walls, as it is for sphere/sphere interactions. Now it is. Also changed the syntax so that the user specifies what style of particle/wall interaction is desired. This allows use of different models for sphere/sphere and sphere/wall interactions.

BACKWARD COMPATIBILITY: An extra required argument was added to the fix wall/gran command, as explained above.

12 Jul 2016

Added some internal support for a new fix cmap command we will releasing adding soon. This includes an added fix option for the compute pe and compute pe/atom commands, to control whether global or per-atom energies from fixes are added.

BACKWARD COMPATIBILITY: The compute_modify thermo command has been removed. The new fix options described above effectively replace it.

9 Jul 2016

Some small upgrades to the USER-DPD and USER-OMP packages. Also some low-level infrastructure added for neighbor lists and pair hybrid to support some new ideas we are working on.

1 Jul 2016

This patch includes some upgrades and bug fixes to the GPU, USER-INTEL, and USER-DPD packages, from Trung Nguyen, Mike Brown, and Axel Kohlmeyer respectively.

28 Jun 2016

Chris Stone (Computational Science & Engineering, LLC) made some optimizations to the reactive chemistry DPD options in the USER-DPD package. It includes support for dense and sparse stoichiometric matrices as well as an additional kinetics solver.

Mike Brown (Intel) has added an option in the package command to use LRT mode (Long Range Thread) with the USER-INTEL package. This is similar to the run_style verlet/split command except it generates an additional thread for some of the kspace_style pppm operations instead of using a separate MPI process. The idea is to use this with hyperthreading on Intel processors or any CPUs supporting SMT so that if PPPM routines are waiting on MPI or memory, you can context switch quickly to threads doing other calculations.

27 Jun 2016

Ray Shan (Materials Design) and Stan Moore (Sandia) have created a Kokkos version of the pair_style reax/c command. This means the ReaxFF can now be run using OpenMP or on GPUs or Intel Phis.

18 Jun 2016

Axel Kohlmeyer (Temple U) added a timeout option to the timer command to trigger LAMMPS to exit when a time limit is reached. This can be useful when running in batch queues. Axel also implemented a respa option for the fix_modify command which allows user control over which level of the rRESPA timestep hierarchy certain fixes are invoked at. A variety of fixes now support that option; see their doc pages for details.

Various other small bug fixes and code clean-ups are also included in this patch.

17 Jun 2016

Removed the newton off requirement for pairwise interactions for the granular pair styles with history dependence. The newton flag can now be either on or off.

Added a compute rigid/local command to allow access to the per-rigid-body attributes of the fix rigid/small command and its variants. Note that these attributes were already available for the fix rigid command and its variants, via the fix itself.

Tegar Wicaksono (UBC) added a fix orient/bcc command, as a complement to the fix orient/fcc command.

14 Jun 2016

Added a fix spring/chunk command, to allow multiple chunks of atoms (e.g. molecules) to each be tethered by their center-of-mass to initial positions.

Mike Brown (Intel) sent an update for the USER-INTEL package. In addition to various performance improvements and small bug fixes, it adds bond_style fene/intel command.

8 Jun 2016

Peter Wirnsberger (University of Cambridge) has written a new fix ehex command which implements an enhnaced heat exchange algorithm which should have less of an energy drift than the HEX algorithm implemented by the fix heat command.

Examples of its use are illustrated in the examples/KAPPA and examples/HEAT directories.

7 Jun 2016

Jim Larentzos (ARL) and Tim Mattox (Engility Corp) and their collaborators at ARL and Engility have extended the USER-DPD package with reactive models, so that reactive DPD simulations can be performed.

This adds several new commands:

as well as the example directory examples/USER/dpd/dpdrx-shardlow.

6 Jun 2016

Maintentance patch with minor code clean-ups and small bug fixes.

31 May 2016

Deprecated the fix ave/spatial command. Its functionality (and more) is avaialable thru the compute chunk/atom and fix ave/chunk commands.

Also deprecated the USER-CUDA package. There is support for GPUs in the GPU and KOKKOS packages, which are more actively developed and supported.

Also a variety of small updates that have been mentioned on the mail list in the last few weeks, which Axel has been diligently patching.

BACKWARD COMPATIBILITY: The fix ave/spatial command has been removed. If you use it, a message about the new commands to use instead will be output. The USER-CUDA package has also been removed.

14 May 2016

Maintenance patch, with some small changes and clean-up.

12 May 2016

Jim Larentzos (ARL) and Tim Mattox (Engility Corp) provided an update to their USER-DPD package, which includes new documentation for a pair multi/lucy command and new examples (see examples/USER/dpd). A pair dpd/conservative command was deprecated.

11 May 2016

Stefan Paquay (Eindhoven University of Technology, The Netherlands) has added a USER-MANIFOLD package which allows particles to move while constrained to a manifold (2d surface). The package implements several fixes:

and a manifold class, that allows different surfaces tob be defined or added by users. See the src/USER-MANIFOLD/README file to get started, and some animations of biological simulations on the Movies page) of the web site.

10 May 2016

Restructured the doc dir to contain "src" with all the files used to produce the doc and web pages, "html" with the HTML version of the doc pages, and "utils" with various tools and formatting from Richard Berger (Temple U) for creating them.

3 May 2016

Upgrade to the Kokkos library included in LAMMPS.

30 Apr 2016

Axel Kohlmeyer (Temple U) added an rmass option to the fix property/atom command to enable easy adding of per-atom mass as a property to an atom style that does not have it.

28 Apr 2016

Some updates and a bug fix to the fix colvars command. Also a bug fix for the molecule command, when it is used to pour or creating clumped rigid bodies from a molecule template file. It was computing the moments of inertial incorrectly (if there were not specified in the template file) if the body was composed of particles with per-type mass. Per-atom mass was correct, which is the usual case for granular rigid bodies.

Also moved the dump xtc command from the XTC package to the MISC package, so the extra package could be removed.

27 Apr 2016

A new write_coeff command from Axel Kohlmeyer (Temple U) to allow writing out the coefficient settings for a style, similar to what the write_data command does for a data file.

22 Apr 2016

Another patch with a variety of small changes and updates to a few files and packages, many of which have been discussed on the mail list.

Also two new commands: compute dipole/chunk for calculating the dipole moments of a chunk of atoms, and pair morse/soft for the USER-FEP package. The former from Axel Kohlmeyer (Temple U) and the latter from Stefan Paquay (TU Eindhoven).

15 Apr 2016

From Axel Kohlmeyer (Temple U), update to the COLVARS library and USER-COLVARS package. Some performance enhancements to pair_style meam for use with the GNU gfortran compiler. A few other small bug fixes from the mail list.

7 Apr 2016

Patch with a variety of small changes and updates to a few files and packages, many of which have been discussed on the mail list.

Also a new bond_write command from Axel Kohlmeyer (Temple U).

28 Mar 2016

Added compute_bond (revised), compute_angle, compute_dihedral, compute_improper commands for use with the hybrid version of the corresponding bond, angle, dihedral, improper styles. The new or revised commands allow the energy contribution of each of the hybrid sub-styles to be tallied and output.

Note that the compute_bond command is not new, but its syntax has changed, to allow all the bond sub-style contributions to be referenced by a single compute.

BACKWARD COMPATIBILITY: This changes the syntax of the relatively new compute_bond command. The output this compute creates is now a global vector instead of a global scalar.

23 Mar 2016

Maintenance patch that cleans-up and fixes various small issue. It includes some missing new files from recent patches that were inadvertently not checked into the repository.

22 Mar 2016

Mike Brown (Intel) sent some performance enhancements to the USER-INTEL package.

21 Mar 2016

Thomas O'Connoer from Mark Robbins group (JHU) has provided a pair_style airebo/morse variant which replaces the LJ interactions with Morse, for better high pressure behavior.

There is also an upgrade to the KOKKOS package for acclerated styles.

Axel Kohlmeyer (Temple U) added a "compute pe/mol/tally command variant to the growing options for pairwise tallying.

This patch also includes a variety of other small updates and small bug fixes, mostly handled by Axel.

14 Mar 2016

Axel Kohlmeyer (Temple U) added a compute_bond command to query the contributions of each bond style when using bond_style hybrid.

Stefan Paquay (TU Eindhoven) contributed an OMP-enabled version of the pair_style morse/smooth/linear command.

Fixed a bug in the smoothed Coulombic portion of the pair_style lj/charmm/coul/charmm command.

Thanks to Sridhar Kumar Kannam for pointing out the issue with CHARMM smoothing term.

10 Mar 2016

Axel Kohlmeyer (Temple U) created a faster way to create the dependency lists when building LAMMPS from source code, which doesn't create a long list of *.d files when you issue a "make". This makes small changes to all the Makefile.machine files in src/MAKE. Old style Makefile.machine files will still work.

Stefan Paquay created a nice tutorial on how to use GitHub to sumbmit new features or bug fixes to LAMMPS for perusal by the developers. It's now included in the doc pages, linked from this section.

1 Mar 2016

Added vector-style variables to the variable command. They produce an N-length vector of global values, with the same style formulas used by equal-style and atom-style variables. They can reference global vectors produced by computes, fixes, and other vector-style variables. They can also be now be used as input by these commands, either as vectors or by subscripting them to refernce a scalar value from a global vector:

Stefan Paquay provided a new pair_style morse/smooth/linear command which adds a smoothed cutoff to the Morse potetnial.

25 Feb 2016

Aidan Thompson (Sandia) has enabled the fix gcmc command to work with triclinic simulation cells.

Fixed a bug with the USER-CUDA package not compiling, due to including some not-ready-to-release code.

22 Feb 2016

Carsten Svaneborg (SDU) suggested adding "zero" styles for pair, bond, angle, etc interactions and sent some code for them. In the pairwise case, this allows a cutoff to be defined for neighbor list builds, but no pairwise forces are computed. This is so that other commands that require neighbor lists can operate as usual. For the bond, angle, etc cases, this allows the compute bond/local, etc commands to tally geometry statistics on the list of bonds, angles, etc defined in a data file, without computing any interactions.

Fixed a bookkeeping bug for variables that reference thermodynamic energy quantities, where sometimes the energy values were not re-computed for the current timestep.

Thanks to Peter Wirnsberger for pointing out the bookkeeping bug and providing an example script.

17 Feb 2016

Richard Berger (JKU) and Daniel Queteschiner (DCS Computing) added a dump custom/vtk command, which can output snapshots in the format used by the popular VTK visualization toolkit or other visualization tools that use it, e.g. ParaView. This command is in a new USER-VTK package because it requires you link LAMMPS with the VTK library. See lib/vtk/README for more details.

Axel Kohlmeyer added a ndx2group command to the USER-COLVARS package to complement the group2ndx command.

16 Feb 2016

New stable release with updated error messages, benchmark log files, example log files. This version will also be posted to SourceForge for 2016.

15 Feb 2016

Mike Brown (Intel) made an update to the USER-INTEL package, which also added some new styles:

Trung Nguyen added some integration fixes and a temperature compute for body-style particles in LAMMPS:

3 Feb 2016

James Larentzos (ARL), Timothy Mattox (Engility Corp), and John Brennan (ARL) have added a USER-DPD package. It implemets a dissipative particle dynamics model with constant temperature, pressure, energy, and enthalpy options. It also uses a Shardlow splitting algorithm to efficiently integrate the equations of motion.

See the src/USER-DPD/README file to get started. Example input scripts are in the examples/USER/dpd dir.

29 Jan 2016

Fixed a bug in processing dihedrals from multiple data files.

Also changed the style name of some USER compute commands to be more LAMMPS compatible and ditto for their options in the set command, converting underscores to "/", namely:

BACKWARD COMPATABILITY: This changes the style syntax of the compute meso and compute smd and set commands for the meso/smd properties, as listed above.

Thanks for Jake Gissinger (U Colorado) for posting a scripts that triggered the problem.

26 Jan 2016

Mike Brown (Intel) provided some new bond styles for the USER-INTEL package:

25 Jan 2016

The previous patch added a method to the interface for pair styles. Inadvertanly, some existing pair styles used a method with the same name for other purposes, so that method needed to be renamed in those pair styles.

22 Jan 2016

Aidan Thompson (Sandia) added an option to the compute voronoi/atom command to enable it to produce a list of geometric neighbors (Vornoi faces) for each atom. This is local data, which can be output by the dump local command.

15 Jan 2016

Various small bug fixes that have recently been flagged on the mail list. Removed the USER-OMP variants of the pair line/lj and pair tri/lj commands, since they have loop dependencies which prevent threading.

14 Jan 2016

Added options to the dump image command to enable atom_style line, tri, and body particles to be drawn with lines, triangles, or body components as defined by the corresponding body style.

Added code to the rotate options of the fix move and displace_atoms commands, so that update the orientiation and rotational motion of particles defined as atom_style ellipsoid, line, tri, and body particles.

Modified the pair_style line/lj command to allow more control over the size of line-segment sub-particles and their Lennard-Jones interactions with each other and with point particles.

BACKWARD COMPATABILITY: This changes the syntax of the pair_coeff command for pair_style line/lj, adding extra arguments.

13 Jan 2016

Added some new options to the set command, namely a theta/random keyword and the ability to set the quaternions of any particle that store one, including body particles.

Added some new keywords to the fix store/state command to sync it with new options in the dump custom command, namely xsu, ysu, zsu, mu, and diameter.

Added a pbc option to the compute chunk/atom command to allow the bin/sphere and bin/cylinder chunk styles to compute the bin an atom belongs to and account for periodic boundaries.

Added support to the compute msd/chunk command to work with restart files, by storing the original center-of-mass positions of the chunks in the restart file.

11 Jan 2016

Nicolas Salles (LAAS-CNRS, France) has fixed some issues with the second moment tight binding QEq (SMTBQ) potentials previously released by their group, which had to do with running them in parallel.