LIGGGHTS® Version History

Current version in LIGGGHTS®-PUBLIC git repository: LIGGGHTS® 3.8.0 (released 30 November 2017)

LIGGGHTS® 3.8.0: (released by DCS Computing, 30 November 2017)

  • (**) Alexander Podlozhnyuk (DCS Computing) has implemented a major new feature in LIGGGHTS: Superquadric shape particles . There is a new public tutorial to showcase how to use it
  • (**) Rahul Mohanty (P&G and University of Edinburgh) and Tomaz Zorec (University of Ljubljana) added an implementation of the Edinburgh contact model - currently there is no doc yet, this will hopefully be in the next release.
  • (**) Rahul Mohanty (P&G and University of Edinburgh) also added an implementation of the Luding contact model - currently there is no doc yet, this will hopefully be in the next release.
  • (**) Rahul Mohanty (P&G and University of Edinburgh) and Tomaz Zorec (University of Ljubljana) added an implementation of the Thornton/Ning contact model - currently there is no doc yet, this will hopefully be in the next release.
  • Josef Kerbl (DCS) extended the feature (fflag & tflag) to deactivate forces or torques in x,y or z direction for multi-spheres
  • Arno Mayrhofer and Christoph Kloss (DCS) fixed a potential segfault with mesh contact history. Thanks to Moritz Höfert (BASF) for providing a test case
  • Josef Kerbl (DCS) fixed a bug in the argument parsing of compute/coord/atom.
  • Josef Kerbl (DCS) updated fix/cfd/coupling/force to work with the new trackIO model from CFDEMcoupling.
  • Arno Mayrhofer (DCS) and Josef Kerbl (DCS) fixed two issues in parallel reading of binary STL files. Thanks to user lesscharles for pointing out this issue.
  • Arno Mayrhofer (DCS) fixed an input issue to viblin. Thanks to user Weijing for pointing out this issue.
  • Arno Mayrhofer (DCS) fixed an input issue to vibrot. Thanks to user Weijing for pointing out this issue.
  • Arno Mayrhofer (DCS) fixed an issue related to the bounds of Y in CGS. Thanks to user hrvig for pointing out this issue.
  • Christoph Goniva (DCS) added 'additional mass' terms to multisphere integration.
  • Andreas Aigner (DCS) fixed several compiler warnings
  • Arno Mayrhofer & Josef Kerbl (DCS) changed max number of bins from MAX_INT to 8e6
  • Arno Mayrhofer added a auto installation for vtk 8.0.1
  • Arno Mayrhofer added the possibility to compile the libraries with make auto directly (without -f)
  • Arno Mayrhofer updated Makefile.auto to add -rpath linking
  • Arno Mayrhofer added clang compilation capability
  • Josef Kerbl (DCS) made radius field in read_dump mandatory
  • Arno Mayrhofer fixed an issue in the jpg compilation tests. Thanks to user JoG for reporting it.
  • Arno Mayrhofer modified the numeric and inumeric input parsers to be able to handle Windows line endings
  • Arno Mayrhofer added the capability for automatically creating output folders (all dumps and restart writing)
  • Arno Mayrhofer rewrote the python documentation and renamed lammps to liggghts for python. Thanks to user AndresMM for reporting this issue.
  • Enabled Superquadrics by default, NOOPT in USE_DEBUG renamed to FULL

The contibutions marked with (**) have been made as part of the T-MAPPP project. The T-MAPPP – Training in Multiscale Analysis of multi-Phase Particulate Processes (www.t-mappp.eu) is an EU funded Marie Curie Initial Training Network with funding from the European Union’s Seventh Framework Programme for research, technological development and demonstration under grant agreement no ITN607453. See www.t-mappp.eu for more information

LIGGGHTS® 3.7.0: (released by DCS Computing, 07 July 2017)

  • Big improvement for compilation: All contact models can now be run, even though they might not be in the whitelist for optimized runs. In this case they will run unoptimized (without inlining).
  • Christoph Kloss (DCS) fixed an inconsistency with wall heat transfer. In case of restart, the mesh temperature is taken from input script taken, not restart
  • Documentation updated: Indicated wrong definition of G* in cited paper. Thanks to user MiRa for reporting it.
  • Christoph Kloss (DCS) fixed an incorrect indexing for fix ave/euler. Thanks to Daniel Queteschiner (JKU) for bringing this up.
  • Arno Mayrhofer (DCS) fixed a bug in fix mesh/surface/* when auto_remove_duplicates was used
  • Christoph Kloss (DCS) fixed a bug with particle insertion when reset_timestep used together with restart. Thanks to Rouven Weiler (BASF) for a test case
  • Arno Mayrhofer (DCS ) made several improvements to Makefile.auto for increased robustness (new feature: postfix)
  • Arno Mayrhofer (DCS ) reworked the "Getting started" section in the documentation
  • Arno Mayrhofer (DCS ) reworked the handling of meshes in a periodic domain
  • Arno Mayrhofer (DCS ) fixed a bug when fixes were replaced. Thanks to user estefan31 for reporting it.
  • Christoph Kloss (DCS ) added a 'check_dist_from_subdomain_border' keyword to fix insert/pack and fix insert/rate/region which controls if the insertion leaves free space at processor boundaries to avoid overlapping particles (default = yes)
  • Alexander Podlozhnyuk (DCS ) added 'contactPoint' to compute pair/gran/local
  • Arno Mayrhofer (DCS ) wrote a new compute erotate that sums the rotational kinetic energies of different particle types
  • Arno Mayrhofer (DCS ) cleaned up the MAKE folder, only auto and two exemplary Makefiles remain.
    All other Makefiles are still in the distro (in the 'deprecated' folder, but will be removed with the next release
  • Arno Mayrhofer (DCS ) added a signal handler to LIGGGHTS. Sending SIGUSR1 will result in a restart file being written and SIGTERM or SIGINT in a quitting of LIGGGHTS after a restart file is written (a second signal will kill it immediately).
  • Arno Mayrhofer (DCS ) modified the fix move/mesh command that now has reference frames (for the center of rotation) that depend on parent fixes

LIGGGHTS® 3.6.0: (released by DCS Computing, 17 Jan 2016)

  • This release prepares LIGGGHTS® for a big event: Pegasus will land very soon: https://vimeo.com/199146136
  • Christoph Kloss (DCS Computing) modified rolling friction models cdt, epsd, epsd2, and epsd3. They now come with a 'torsionTorque' flag that allows to controll if norsion torque is used or not. The default is the previous behavior (off). There were a number of separate discussions regarding this topic. Thanks to Alice Hager (DCS Computing), Chen Mingyong (John Deere), Joshua Poganski (ÖAW), Alexander Podlozhnyuk (DCS Computing), and Sascha Behzad Soltanbeigi (University of Edinburgh) for the discussions!
  • Arno Mayrhofer (DCS Computing) implemented liquid bridges for particle-wall contacts and improved the nighbor list break-up handling. He also added a new fix mesh/surface/liquidtransfer.
  • Arno Mayrhofer (DCS Computing) also corrected the liquid bridge breaking in Washino and Easo models
  • Arno Mayrhofer (DCS Computing) re-worked the internals of fix mesh/surface, allowing flexible combination of mesh modules, such as liquid transfer modele and the stress module. There is full backward compatibility to the way how the commands were structured before.
  • Andreas Aigner (DCS Computing) reworked the style of the documentation to make it visually much nicer. You will see that the html documentation has really improved. For this release, we're not shipping a pdf of the manual, but we'll do so for the next one
  • Josef Kerbl (DCS Computing found and fixed a rare bug resulting in an internal error when using fix insert/stream in highly parallel simulation
  • Arno Mayrhofer (DCS Computing) re-wrote the replicate command and fixed several issues with it.
  • Sebastian Richter (DLR) found a copy/paste issue in the doc of fix insert/rate/region, which is now fixed
  • Arno Mayrhofer (DCS Computing) created a new Makefile auto, which . This is now the default recommended Makefile.
  • Josef Kerbl (DCS Computing) also improved the accuracy of particle distributions for specific cases.
  • Christoph Kloss (DCS Computing) fixed a bug with multisphere and fix massflow mesh and restart. Thanks to Rouven Weiler (BASF) for providing a test case
  • Christoph Kloss (DCS Computing) changed the implementation for fix massflow/mesh with multisphere. Now the constituent spheres of a multisphere particle are used for evaluating the mass-flow. This is more accurate and also removes a rare parallelization issue . Thank to Christian Richter (OVGU) for sharing a test-case for that issue
  • Christoph Kloss (DCS Computing) made a minor update for how STL mesh elements are imported. Stefan Radl (TUG) brought up a case where STL surface import failed when degenerated elements are present. This is fixed now.
  • Christoph Kloss (DCS Computing) fixed a small numeric issue with fix insert/rate/region which could pretend the command from working in some cases. Thanks to Chris Stoltz (PG) for a test case
  • Arno Mayrhofer (DCS Computing) added an example for the multicontact model (hydrogel compression)
  • Christoph Kloss (DCS Computing) fixed an issue with fix insert/stream in case restarted simulation has different time-step size
  • Some internal pointer clean-up. Thanks to Daniel Queteschiner (JKU) for the hint
  • Updated help functionality ( -h , -help), which now shows an up-to date overview of all compiled contact models, fixes, computes

LIGGGHTS® 3.5.0: (released by DCS Computing, 05 September 2016, patched on 6 Sept)

  • Multi-sphere goes parallel: release of the parallel computation module for mult-sphere simulation
  • LIGGGHTS® goes dense: Arno Mayrhofer (CFDEMresearch) added an implementation of the multi-contact model for dense packing with plastic deformation
  • Important update for particle insertion: Josef Kerbl (DCS Computing) added an isPrime() check, which is now used to enforce that the seeds for particle insertions (fix insert/*) are well-chosen
  • Also slightly changed the insertion behavior in case of no overlapcheck in parallel for fix insert/*
  • The group of Stefan Radl (TUG) added a fix multisphere/break command. Thanks for that contribution!
  • Andreas Aigner (DCS Computing) and Arno Mayrhofer (CFDEMresearch) overhauled the file read implementation for the VTK reader, it now can read files where one line contains multiple points
  • Added a new "error" flag to fix check/timstep/gran
  • A new fix massflow/mesh/sieve command was added by the group of Stefan Radl at TUG. Thanks for this contribution!
  • Added output for normal and tangential contact torque in compute pair/gran/local and compute wall/gran/local
  • Andreas Aigner (DCS Computing) found and fixed an indexing bug in liquid bridge models easo and washino
  • Andreas Aigner (DCS Computing) also found and fixed a small issue with the tangential models, where in some cases particle-wall torques were omitted for a second contact
  • Additional debugging information for fix mesh/surface when import fails due to geometry shift
  • Removed deprecated 'neigh_modify cluster' command
  • Robin Gibaud (Simap Grenoble) found a bug in the parallel algorithm of compute inertia/molecule, which was now fixed. Thanks for the bug report!
  • Some updates for the calculation of kinetic energy: "ke" calculation via the thermo keyword "ke" is now correct for multi-sphere by default, even when no compute ke/ms is used. Also added a patch so that ke is also computed for simulations with just one single particle present.Thanks to Arno Mayrhofer (CFDEMresearch) for the update!
  • Made a small modification to the doc for the fix insert/* commands, where the descripton for the 'set_property' functionality was a bit misleading. Thanks to Daniel Queteschiner (JKU) for the hint!
  • Robin Gibaud (Simap Grenoble) also noted that there might be a small difference between the compute pair/gran/local values for contact forces invoked at the end of the time-step for output and the contact force used for integration. This slight deviation is now mentioned in the doc.
  • Added the restriction that multisphere is not compatible with fix deform. Thanks for user ZYan for the hint that this was missing!
  • Christian Richter (OVGU Magdeburg) identified a copy/paste issue in the documentation of cohesion model washino/capillary/viscous, which was now fixed. Thanks for noting this.
  • Andreas Aigner (DCS Computing) commited a large patch that should remove loads of compilation warnings

VERSION CONSISTENCY AND BACKWARD COMPATIBILITY NOTES:

  • Seeds for particle insertion (fix insert/*) must now be large primes. LIGGGHTS will give a list of suggestions when you run into this issue.
  • Insertion in parallel different with no overlap check might give slightly different particle distribution
  • Kinetic energy calculation for multisphere simulations via thermo ke might give different results

LIGGGHTS® 3.4.1: (released 19.05.2016) - small update to 3.4.0

  • Fixed an issue with surface mesh - thanks to Daniel Queteschiner (JKU) for reporting this!
  • Fixed an issue with the documentation of 3.4.0, which did not correctly show the list of granular contact models
  • Fixe an issue with compilation using the ubuntuVTK makefile

LIGGGHTS® 3.4.0: (released 17.05.2016)

  • Made significant changes to the cohesion and liquid bridge models. Now the normal force from these models is "seen" by the tangential models, leading effectively to a larger macroscopic cohesion. The behavior before can be reproduced using the new "tangential_reduce" flag. Thanks a lot to Josef Kerbl and Andreas Aigner (both DCS Computing) for proposing this.
  • Added a new dump local/gran/vtk feature to viz force chain networks.
  • Refactoring of wall for implementation in preparation for non-intersection forces. Also added a driver for non-contacting wall forces. Thanks to Stefan Radl (TU Graz) for the suggesion.
  • Added a "region" filter to the family of fix mesh/surface/* commands to be able to filter the geometry for mesh import. Also changed the hard angle limit for tri import to a more permissive value.
  • Added a new fix property/atom/timetracer to calculate particle residence time in the simulation domain or a list of regions.
  • Stefan Radl (TUG) added implicit integration driver for the scalar transport equation. Thanks for that patch.
  • Changed the default enery calculation in a way that energy for simulations containing multi-spheres is now correct by default. Thanks to Arno Mayrhofer (CFDEMresearch) for the patch.
  • Added the "info" command from LAMMPS that prints information about the internal state of the simulation.
  • Improvements for fix massflow/mesh: Fixed an issue with fix massflow/mesh double counting. Thanks to Andreas Aigner and Alice Hager (DCS Computing). Also it supports now moving meshes
  • Fixed another issue with more than one fix massflow/mesh commands with multisphere. Thanks to Christian Richter (OVGU Magdeburg) for posting a patch.
  • Fixed a rare particle-triangle contact detection issue. Thanks to Christian Richter (OVGU Magdeburg) for sending over a test case.
  • Fixed a small memory leak in dump mesh/vtk when using interpolation.
  • Fixed an issue with meshes and PBCs in a very special context. Thanks to Chris Stoltz (PG) for sending a test case.
  • Compute pair/gran/local now supports output of non-contacting force. Thanks to T-MAPPP ESR Guiseppe Raso (JM) for the note that this was not working before.
  • Stefan Radl (TUG) motiviated the addition of a "compress_tags" feature to the fix insert/* commands.
  • User Christoph discovered a flaw in the reset_marker feature of compute nparticles/tracer/region, which is now fixed. Thanks a lot for finding this.

BACKWARD COMPATIBILITY NOTES:

  • Cohesion models have now a different default regarding how they handle tangential forces.
  • Energy calculation for simulations with multisphere particles might give different results

LIGGGHTS® 3.3.1: (released 14.01.2016) - happy new year from DCS Computing!

  • Surface mesh import via fix mesh/surface is now much faster. O(n) import of surface meshes, auto remove duplicates still O(n2)
  • Surface mesh import via fix mesh/surface has now stricter quality criteria and automated remedy: Now elements with a very small angle (< 0.18°) throw an error upon import. You can automatically take care of such cases by excluding such elements from the mesh using the element_exclusion_list feature. This exclusion list now works for both too many neighs per element and elements with angles < 0.18° which do not fulfill the quality criteria. See doc of fix mesh/surface for details
  • User HCSGXL found an issue with the finnie wear model . dt was missing at an essential point, which is now fixed. It should be noted that simulations using the model in versions before 3.3.1 also give qualitative results, which allow to compare different designs / geometries
  • Added all_in functionality to create_atoms command. Thanks to Chris Stoltz (PG) for the note that this might be useful
  • Added compute rigid, which can be used to compute / output body information for the multisphere method
  • Stefan Radl (TUG) prepared an update related to the ParScale coupling, which is also included in this version. Thanks a lot!
  • In preparation of a coupling to the MBD code "HOTINT", the library was added to the /lib directory. Thanks again to Stefan Radl's group at TUG.
  • A bugfix was made for the new all_in feature for region mesh/tet as used by fix insert/*
  • Raphael Schubert (IWM Fraunhofer) found a glitch in the fix move/mesh/linear/variable which could cause the code to stall at the start in some cases. This is now fixed - thanks for reporting!
  • Luca Orefice (TU Graz) found a small issue in the comm_I.h file, which caused a compilation error on some systems. This is now fixed
  • User Harald Unger discovered that 'thermo modify lost error' was no longer working. This is now fixed and the behavior of the error check was also slightly modified. The default setting for "thermo_modify lost" has been changed to "ignore". See doc for details
  • Fixed an issue with CFD coupling where the CFD forces were not taken into account for the first time-step of a LIGGGHTS run. Many thanks to Stefan Radl (TU Graz) for finding this!
  • Fixed a rare issue with contact history that could lead to a segfault. Many thanks to Etienne Guesnet (SIMAP Grenoble) for posting a testcase!
  • Added note to the doc and error message that liquid bridge models cohesion easo/capillary/viscous and washino/capillary/viscous may be used for particle-particle interaction only (not for walls). Thanks to Per Abrahamsson (Chalmers University) for the note
  • Moritz Höfert (BASF) discovered that parts of the doc for fix particletemplate/multsphere were outdated. This is now fixed
  • Made a couple of other small updates in the doc (removed fix langevin/* which is not supported in LIGGGHTS, small update for tangential model history)

BACKWARD COMPATIBILITY NOTES:

  • The default setting for "thermo_modify lost" has been changed to "ignore"
  • The surface mesh import quality criteria of fix mesh/surface has been changed as mentioned above. This can lead to LIGGGHTS rejecting meshes which were accepted in older versions. This can be remedied using theelement_exclusion_list feature
  • Finnie wear will now output different values; It should be noted that simulations using the model in versions before 3.3.1 also give qualitative results, which allow to compare different designs / geometries

LIGGGHTS® 3.3.0: (released 23.09.2015)

  • Release of the multisphere method for serial computation (on 1 process): LIGGGHTS®-PUBLIC can now simulate non-spherical clumped particles, with full support for the various insertion commands. See the doc for details.
  • Added two new liquid bridge models: cohesion washino/capillary/viscous and easo/capilary/viscous. Both include a viscous and a capilary term for liquid bridge force and have a simplistic model for liquid transfer due to bridge break-up. See the doc for details.
  • Added an optimization/improvement to the performance of particle insertion. This breaks down the overlap check upon insertion from O(n^2) to O(n), which means considerable improvement for insertion of large # of particles, as described here: https://github.com/ParticulateFlow/LIGGGHTS-PUBLIC/pull/1 Thanks to Richard Berger (JKU) for the contribution of this very nice feature!
  • Added a couple improvements to mesh import of fix mesh/surface. Two new keywords 'element_exclusion_list' and 'curvature_tolerant' were added to improve the import of surfaces meshes. See doc for details
  • Also made an important improvement to region mesh/tet. The all_in yes feature was implemented, and some algorithms necessary for particle insertions have been improved considerably. You should see a speed-up when using tet meshes for insertion.
  • Added a new fix buoyancy that exerts a buoyancy term on particles, assuming a stationary background fluid with a certain fill level of fluid
  • Niels Dallinger (TU Chemnitz) sent an update for the viblin/vibrot funtionality of fix move/mesh. Thanks for this contribution!
  • Added property "density" to be accessible by the variable command
  • Updated the cMake and Windows build options. Thanks to Richard Berger (JKU) and Andreas Aigner (DCS Computing) for their efforts!
  • Fixed an issue within variable handling of fix ave/euler. Thanks to Daniel Queteschiner (JKU) for the hint that there was an issue.
  • Fixed rare issues with fix mesh/surface, where LIGGGHTS® could crash if a mesh is unfixed while a fix wall/gran is still active
  • Fixed an issue with fix massflow/mesh when the same mesh was used for multiple fix massflow/mesh commands . Thanks to Andreas Ainger (DCS Computing) for providing a testcase.
  • Fixed a minor issue with argument parsing in fix massflow/mesh. Thanks to Daniel Queteschiner (JKU) for the hint!
  • Fixed a small issue with fix property/atom/tracer where the fix group was not taken into account for the evaluations

LIGGGHTS® 3.2.1-patched: (released 27.08.2015)

  • Made a small patch for the data coupling to ParScale. Thanks to Stefan Radl (TUG) for the note!

LIGGGHTS® 3.2.1: (released 21.07.2015)

  • Added a new "EPSD3" rolling friction model that is similar to EPSD1, but allows the user explicit control over the rolling stiffness. See doc for more detail. Thanks to Thomas Puffitsch (TUG) for the contribution.
  • The default length unit for commands "fix ave/spatial", "fix deform", "fix move", "fix dt/reset", "fix wall", "change_box", "region", "velocity", and "create_atoms" was changed from "lattice" to "box" units
  • Added two more outputs for compute pair gran local: "force_normal" and "force_tangential" can now also be output
  • In preparation of species coupling using ParScale, added a new fix couple/cfd/speciesConvection. Thanks to the group of Stefan Radl (TUG) for the collaboration on this topic.
  • Richard Berger (JKU) added a "modify_timing" command that can give fine-grained information about the run-time of each fix (rather than lumping all times together in "Other"). Thanks for this nice feature!
  • Added number based distribution definition to fix particletemplate/discrete. Thanks to Rouven Weiler (BASF) for the suggestion
  • Fixed a rare bug where restart data was not correctly assigned to fixes after restart. Thanks to Andreas Aigner (DCS) for providing a testcase.
  • Fixed a bug in fix ave/euler concerning the calculation of average velocity. Thanks to Daniel Queteschiner (JKU) for posting a bugfix!
  • Fixed an issue with the "dynamic" feature of the region command
  • Added support for fix group for fix massflow/mesh, i.e. the command now only counts the particles in the specified fix group
  • Added a new tangential model "off", which corresponds to no tangential model. For consistency, this model is now the default if no tangential model is specified

BACKWARD COMPATIBILITY 1: The default behavior for the tangential model was changed (in case no tangential model is specified).
BACKWARD COMPATIBILITY 2: The default unit system for some commands was changed to box units as outlined above

LIGGGHTS® 3.2: (released 29.04.2015)

Version 3.2 represents a major facelift for LIGGGHTS®. The focus of LIGGGHTS® is modelling of macroscopic phenomena (granular or sph), not atomic or molecular systems. Version 3.2 removes a lot of legacy features and documentation from LAMMPS which made LIGGGHTS® code or documentation misleading or ambiguous. Since this legacy code and documentation was not maintained or updated in the past, it did not make sense to keep it in the release

This makes the LIGGGHTS® 3.2 release concise and focused to its core capabilities: modelling of macroscopic phenomena (granular or sph). Most important changes in that respect are:
+ Most features for modelling atomic or molecular systems were removed from the code and doc
+ The manual was re-worked and updated
+ Now there is always the latest version of the documentation online under http://www.cfdem.com/media/DEM/docu/Manual.html (note: this is autogenerated by our release systems and may take up to 24 hrs to update after a fresh release)
+ The header files were re-worked, and are now in a uniform state, also representing the copyrights correctly
+ Changed the default thermo output for "one" and "multi" styles to "step atoms ke cpu"
+ Removed thermo keywords "temp", "press", since the calculation was meaningless (or at least, if correctly interpreted, the calculation was incorrect) for granular systems
+ Removed fix pour/legacy. Changed also the movingMeshGran example to reflect that

There were also a couple changes related to models and features in LIGGGHTS®:
+ Extended compute pair/gran/local command so now can output particle-particle or particle-wall contact overlap
+ Fixed a small issue with contact history
+ Improved the handling of "dirty" wall meshes, i.e. meshes where the mesh elements are not located perfectly edge-to-edge. Thanks to Christian Richter (OVGU) for posting a test case!.
+ Fixed an issue with the binary option in conjunction to the multifile option of dump mesh/stl. Thanks to Christian Richter (OVGU Magdeburg) for the note!
+ Fixed a small copy/paste issue in the set command for option "vz". Thanks to user Sirop for the note.
+ Added two new features to fix ave/euler: "parallel" and "basevolume_region". The first one is used to determine if each process uses a local grid or all procs contribute to one global grid. The latter option is used to correct volume fraction and pressure calculations in near-wall regions
+ Added a "set_property" option to the fix insert/* commands that can initialize scalars stored in a fix property/atom upon insertion (e.g. Temperature). This now makes it possible to to pour streams of particles with different temperatures into the simulation
+ Added a "title none" option to fix print
+ Note for developers: We changed naming of functions inside the contact model framework to make things more clear and more general. collision() is now called surfacesIntersect(), noCollision() is now called surfacesClose().
BACKWARD COMPATIBILITY: Restart files of older versions will not be compatible

LIGGGHTS® 3.1.0: (released 18 Feb 2015)

+ Added a new dump custom/vtk command that directly dumps per-particle properties LIGGGHTS output into VTK files. It automatically detects vectors and scalars and supports vtk, vtu, vtp formats and can write in both ASCII and binary formats. Thanks to Daniel Queteschiner (DCS Computing) for the implementation!
+ Added the hash code of the git commit into the log file. This allows the user to identify what exact version of LIGGGHTS® was used when running a case, and to go back to that version using the "git checkout" command
+ Added a functionality to fix mesh/surface/* that the fix group is used to determine which particles "see" a granular mesh. Thanks to Christian Richter (OVGU Magdeburg) for the suggestion
+ Fixed an issue with fix move/mesh/*/variable in case variables were changed inbetween runs. Thanks to Joshua P for running into the problem and to Christian Richter (OVGU Magdeburg) for pinpointing it.
+ Fixed a sign issue for compute wall/gran local in the calculation of the contact point. Thanks again to Christian Richter (OVGU) for noting this and tracking it down!
+ Added a safety check to ensure a fix move/mesh can only be applied to a fix of type fix mesh/*. Thanks to Robin Gibaud (Simap Grenoble) for sending a testcase.
+ Josef Kerbl (DCS Computing) and Daniel Queteschiner (JKU) fixed a small memory issue with the MPI coupling in cfd_datacoupling.cpp
+ Detected and fixed a rare issue with particle-particle contact history that led to outdated values in case a "delete_atoms" command was used in-between two "run" commands. This bug was also touching LAMMPS and reported to the LAMMPS developers (who fixed it with a different bugfix as part of the 6 Feb 2015 LAMMPS patch)
+ Fixed a minor issue with fix couple/cfd/convective where in rare cases the heat flux onto ghost particles was not correctly calculated
+ Fixed a missing parenthesis in the doc for fix check/timestep/gran. Thanks to Moritz Höfert (BASF) for the note.
+ Fixed an issue where fix massflow/mesh was not working correctly when used with "pair_style none". Thanks to Stefan Radl (TUG) for identifying the problem and sending a testcase!
+ Added an additional error check to fix insert/pack in case its region would be deleted during the simulation

LIGGGHTS® 3.0.7 patched: (released 14 January 2015)

+ Added a missing file necessary for compilation for the compilation of the new ParScale-LIGGGHTS® coupling functionality (cfd_datacoupling_simple.h). Thanks to Riccardo Maione for flagging this.

LIGGGHTS® 3.0.7: (released 5 January 2015)

The CFDEM®project team is proud to announce the release of a new particle-scale simulation engine: ParScale
More information be released soon...
+ The 3.0.7 release of LIGGGHTS adds the ParScale-LIGGGHTS® coupling functionality to LIGGGHTS®
+ Also fixed a critical bug in parallel mesh handling in connection to moving mesh.

LIGGGHTS® 3.0.6: (released 11 November 2014)

"Carnival edition" of LIGGGHTS® (in Germany and Austria, Carnival starts 11 Nov, 11:11 am)
+ Fixed a critical bug in parallel mesh handling. Thanks to Christian Richter (OVGU Magdeburg) and Rouven Weiler (BASF) for almost simultaneously flagging the issue!

LIGGGHTS® 3.0.5: (released 31 October 2014)

+ Fixed a critical bug introduced into fix mesh/surface/stress/servo by yesterday's 3.0.4 release. This was reveiled by a nightly test automation run. Thanks to Andreas Aigner (DCS Computing) for detecting and fixing the problem!
+ Andreas Aigner (DCS Computing) also included an improvement to the hooke/hysteresis pair style, which was originally contributed by Christian Richter (OVGU Magdeburg). Thanks to both!

LIGGGHTS® 3.0.4: (released 30 October 2014)

+ Initial preparation commit for a new package "PASCAL", to couple with a brand-new intra-particle-scale simulation code. This PaScal code will be released end of this year on www.CFDEM.com. Stay tuned!
+ Fixed 2 issues with restart of simulations that contain meshes: (i) fixed an issue which could cause crashing upon restart of simulations that used moving meshes and (ii) fixed an issue with restart of global mesh properties (such as for fix mesh/surface/stress/servo)
+ Andreas Aigner (DCS Computing) fixed a bug in the calculation for the effective radius for particle-wall contacts for normal model hooke/hysteresis and rolling friction models cdt, epsd, epsd2. Thanks to user "Sirop" for noting this
+ Richard Berger (JKU) detected an issue in fix mesh/surface/stress, which could lead to an unitialized variable in some cases. Thanks for this bugfix!
+ Fixed a couple typos in the documentation
+ Ensured compilability under Windows. Thanks to Daniel Queteschiner (JKU) for patching!
+ Stefan Radl (Graz University of Technology) added an "every" option via fix_modify to the fix transportequation/scalar command. Thanks!

LIGGGHTS® 3.0.3: (released 28 Aug 2014)

+ Added the capability to use variables in the set command when used to set the values for properties defined via a fix property/atom.
+ Added a "vel" keyword to compute pair/gran/local and compute wall/gran/local so that velocities can now be output as well.
+ Thanks to Richard Berger (JKU), the compilation of LIGGGHTS® requires now less RAM and is a bit faster .
+ Added more information when "verbose" is used for fix insert/pack. Now it is also stating the progress of the insertion.
+ Fix an issue with communication in the "atom_style molecule". Thanks to Eric Parteli (U Erlangen) for reporting a test case.
+ Richard Berger (JKU) fixed a bug concerning the hybrid/overlay pair style. Thanks to Eric Parteli (U Erlangen) for reporting a test case.
+ Fixed a rare issue where restarting with mesh contact history did not work in some cases when fix balance was used at the same time. Thanks to Rouven Weiler (BASF) for providing a test case
+ Corrected the doc of pair gran with respect to the order of cohesion and rolling_friction keyword, if both are used. Thanks to Chris Stoltz (PG) and Christian Grohs (Plansee) for the hint!
+ Fixed some internal issues with arg parsing, e.g. for the "tangential_damping" flag. Thanks to Rouven Weiler (BASF) for the hint!
+ Fixed a bug in the "cu" thermo keyword to display the Cundall number as thermo output

BACKWARD COMPATIBILITY: Please note that due to the additional keyword for compute pair/gran/local and compute wall/gran/local the output of these commands has changed. Please see doc for details

LIGGGHTS® 3.0.2: (released 12 June 2014)

Important note for CFDEM®coupling users: There has been a slight patch in CFDEM®coupling to make it compile with LIGGGHTS® 3.0.2. Please perform a "git pull" for CFDEM®coupling.
+ Added a new 'contact_area' keyword to the fix heat/gran/conduction and fix wall/gran command, that serves to provide different choices of how particle-particle and particle-wall contact area for heat conduction is calculated
+ Added a feature to fix couple/cfd/force so that LIGGGHTS® can send an arbitrary addtional property to CFDEM®coupling if needed
+ The group of Stefan Radl (TUG) added an implementation which removes the "force artifact" that normal models show attractive forces during the repusion phase of the collision. Please note that in the current implementation the coefficient of restitution will not be recovered. Thanks for providing this!
+ Richard Berger (JKU) made some changes in the inner working of contact models. Please see the LIGGGHTS® developers wiki on https://github.com/CFDEM®project/LIGGGHTS®-PUBLIC/wiki for details
+ Richard Berger (JKU) made a small update to the LIGGGHTS® MPI wrappers, which could give some performance improvements for large-scale simulations
+ Richard Berger (JKU) fixed issues with argument parsing of fix wall/gran
+ Christian Richter (OVGU Magdeburg) found a bug in fix gravity, when using it in the variable-style. Thanks for flagging this!
+ Tijmen Ton (Twister BV) sent a test case revealing that for some cases memory overflow was not correctly caught for mesh contact history. This is now fixed.
+ Fixed a small issue with dumping mesh temperature via the dump mesh/vtk keyword
+ Fixed several minor memory leaks

LIGGGHTS® 3.0.1: (released 10 April 2014)

+ We now have a wiki containing updated installation tutorials and also we're starting to put together some coding style guide. See https://github.com/CFDEM®project/LIGGGHTS®-PUBLIC/wiki. Thanks to Richard Berger (JKU) for starting to fill this with content!
+ Richard Berger (JKU) made a large patch to the code base removing most of the compiler warnings, so LIGGGHTS® should compile much cleaner now.
+ Richard Berger (JKU) also made some fine-tuning and performance optimizations for handling of mesh walls (fix neighlist/mesh and fix contacthistory/mesh)
+ Extended the area correction feature of fix heat/gran/conduction, the feature is also now generalized to work with all types of contact models. Also fixed an issue with causing this feature not to work correctly with LIGGGHTS® 3.X. Thanks to Rouven Weiler (BASF) for sending a testcase.
+ Also fixed a sign issue in particle-wall heat transfer. Thanks again to Rouven Weiler (BASF) for sending a testcase.
+ Fixed an internal issue with contact history which could occur for periodic boundary conditions in some rare cases. Thanks again to Rouven Weiler (BASF) for sending a testcase.
+ Fixed an issue that could lead to a crash/segfault for cases where the pair_style was changed during simulation or after restarting the simulation.
+ Disallowed multiple fix wall/gran/* commands of mesh type.
+ Philippe Seil (JKU) added some code to fix couple/cfd so that also torques on particles can be exchanged with a fluid model
+ Fixed an issue with the delete_atoms functionality of fix Massflow/mesh, which was sometimes not deleting the particles as it should. Thanks to Christian Richter (OVGU) for a testcase
+ Fixed an issue with the documentation links for the epsd type rolling friction models. Thanks to Christian Richter (OVGU) for the hint.

LIGGGHTS® 3.0.0: (released 23 January 2014)

Changes from LIGGGHTS® 3.0.0-beta to LIGGGHTS® 3.0.0:

+ Introduced a new contact_distance_factor (replacing contactHistoryDistanceFactor) in neigh_modify, which can be used to increase the range of granular neighbor lists.
+ Improved a couple of things "under the hood" as compared to LIGGGHTS® 3.0.0-beta
+ LIGGGHTS® 3 features major mesh optimizations vs. 2.3.8, particle-mesh neighborlist building has been optimized by Richard Berger (JKU)
+ Andreas Aigner (JKU) updated the fix mesh/surface/stress servo implementation, with improved stability and the ability to operate in direction of an arbitrary axis
+ There were also a couple bookkeeping changes in the epsd rolling friction model, and the epsd2 model which was missing in the beta release was added. Thanks again to Andreas Aigner (JKU) for this.
+ Fixed a communication bug that could arrise when using tricilinc boxes: thanks to Stefan Radl (TUG) and Yile Gu (Princeton) for detecting this
+ Added a more restrictive check for the #atom types defined in create_box vs. length of vector in and fix property/global. Thanks to Rouven Weiler (BASF) and Tom Simmons (TUBS) for bringing this up
+ Fixed an issue with fix property/atom/tracer that prohibited restarting of a simulation with the same fix parameters. Thanks again to Rouven Weiler (BASF) for the note.
+ Fixed a bug where changing pair style upon restart had no effect.
+ Fixed some omissions/typos in the manual. Thanks to Chris Stoltz (PG) for the note.
+ Fixed an issue when using fix insert/stream and reset_timestep could lead to divergence of the simulations.
+ Made the "invalid fix style" error message more verbose, telling the user which fix command causes the issue.

Backward comaptibility:

No restart backward compatibility to LIGGGHTS® 2.X
Input script syntax was changed. Details to be found in doc/LIGGGHTS®_3.X_tutorial.html

LIGGGHTS® 3.0.0-beta preamble: (released 03 December 2013)

Dear LIGGGHTS® user community,

as you might know, LIGGGHTS® has not been designed from scratch, but rather grew organically, with inputs from fields as diverse as molecular modelling, fluid mechanics, chemical engineering, mechanical enigineering, bulk solids handling, high performance computing and some more. We have seen applications across a multitude of disciplines of science and engineering, and we're very happy about that.

As LIGGGHTS® grew organically, there is continuous need for adaptions as the code and its applications grow and diversify. Things that were excellent a couple years ago are not necessarily good in a couple years from now. We need to be ready to tackle issues like maintainability and be able to react on paradigm shifts in hardware architecture

We have been working pretty hard in the last couple months to bring LIGGGHTS® one step further. The major LIGGGHTS® 3 update consists of the following "puzzle pieces":

1) A re-write of the way granular interaction forces has been implemented

A new template-based mechanism was introduced, and particle-particle and particle-wall implementations are now unified. LIGGGHTS® grew considerably over the last couple years, and anticipating further development, the old way of doing things would not have been sustainable, not even for the next year or 2.

2) Merge with the latest version of LAMMPS

Bridging communities, making latest development in the LAMMPS molecular modelling community available also for LIGGGHTS® users.

3) Testing and bug-fixing

There is now a much more profound regression test running behind the scenes, so LIGGGHTS® 3 is more rigorously tested better than all the versions before. Daniel Queteschiner (JKU) was working full-time for almost two months preparing a set of regression tests. Thanks for this!

Just to give you an impresison of the sheer numbers:

+ the git diff patch (just for the C++ source code) is roughly 415,000 lines
+ developing the new granular interaction force framework, merging, testing and bug-fixing took roughly one man-year
+ during the tests, 5 bugs in the LAMMPS codebase were detected and fixed (thanks to Steve Plimpton/Sandia and Axel Kohlmeyer/ICTP for the great collaboration!)

All this was accomplished and is now provided to the community without any direct funding for this! LIGGGHTS® 3.0 beta is not perfect yet. So we need your help to double-check, test, and improve it. The discussion boards now have a new forum dedicated to feedback about LIGGGHTS® 3 (http://CFDEM.dcs-computing.com/?q=forum/17). We're looking forward to your feed-back and suggestions!

Christoph Kloss

Managing partner
DCS Computing GmbH

Senior scientist,
JKU Linz Department of Particulate Flow Modelling

Richard Berger

Senior developer,
JKU Linz Department of Particulate Flow Modelling

Preliminary LIGGGHTS® 3.0.0-beta release notes: (released 03 December 2013)

IMPORTANT NOTE: the 3.0.0-beta release is in a separate repository,
which can be reached at https://github.com/CFDEM®project/LIGGGHTS®-3-beta-PUBLIC

+ All-new modular granular pair styles and wall implementation
+ Implementations for granular pair and wall are no longer duplicate/redundant
+ Included all LAMMPS updates till 23 Nov 2013
+ Contact history has now a much more efficient storage for polydisperse systems
+ Fixed various minor memleaks and fixed a number of LAMMPS bugs
+ Added a clear error message when mesh was unfixed while dump mesh/stl was active
+ Fixed some bugs in fix massflow/mesh
+ Fixed an issue where the sjkr2 cohesion mode was not applicable to granular walls.

Known issues:

+ Restart functionality is still showing some issues

Backward comaptibility:

No restart backward compatibility to LIGGGHTS® 2.X
Input script syntax was changed. Details to be found in doc/LIGGGHTS®_3.X_tutorial.html

LIGGGHTS® 2.3.8: (released 10 October 2013)

+ Major MPI communication optimization for granular systems, reducing the number of ghost atoms on neighboring processors for poly-disperse systems.
+ Stefan Radl (TU Graz) added a Crank-Nicolson integration scheme for fix couple/cfd/force/implicit. Many thanks for this contribution!
+ Slight revision of the mesh handling in LIGGGHTS®, where some mesh topology details were not detected correctly for different node orientation of adjacent triangles (which is usually not the case). Thanks to user rkumar310 for posting a test case.
+ Fixed a small issue with writing restart files of simulations using fix massflow/mesh.
+ Fixed a logic bug in the Finnie wear model of fix mesh/surface/stress, which could lead ot negative wear values in rare cases.
+ Added a sanity check for characteristic velocity, which now needs to be > 1e-2 for SI units.
+ Andreas Aigner (JKU) fixed a bug that could result in incorrect neighborlist build in rare cases for growing particles. Thanks for this!
+ Stefan Amberger (JKU) fixed a bug in the bounding box calculation for region wedge. Thanks for this!
+ Fix insert now outputs # of inserted particles and inserted mass. Also fixed a small bug in the calculation of calc final insertion step.

BACKWARD COMPATIBILITY:
The Finnie model of fix mesh/surface/stress may output different values.

However, there are some restrictions in 2.3.8 (see below) which will be relaxed over time:
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.3.7: (released 6 September 2013)

+ Added a warning message to prompt the user to use "neigh_modify delay 0" with granular pair styles.
+ Fixed some bugs in fix massflow/mesh which could lead to a crash; thanks to Rouven Weiler (BASF) and Stefan Radl (TU Graz) for noting/fixing this.
+ Fixed a sign bug in fix mesh/surface/stress/servo. Thanks to Andreas Aigner (JKU) for fixing this.
+ Fixed an issue where the new epsd2 rolling friction model requested the user to define a viscousdampingratio, which was not used and needed by the model. Thanks to Christian Richter (University of Magdeburg) for noting this.
+ Fixed an issue where the #particles to insert for the "volumefraction" option combined with "all_in yes" of fix insert/pack was not correctly calculated in case the particle radius was in the same order of magnitude than the length of the insertion volume. Thanks to Paul Winkler (University of Weimar) for noting this.
+ Fixed a a communication bug in the ellipsoidal particle implementation which could lead to a crash. Thanks to Daniel Queteschiner (JKU) for fixing this.
+ Fixed an issue with fix transportequation/scalar that could lead to the fix ignoring very small initial values for the stored property in some cases. Thanks to Josef Kerbl (DCS Computing) for finding and fixing this.
+ Fixed a problem where restarting simulations using the Finnie wear model was not possible. Thanks to Kirk Fraser (Roche) for posting a testcase.
+ Andreas Eitzlmayr (TU Graz) added new options to compute reduce for SPH and fixed a bug in his new functionality in fix ave/spatial.
+ Included some LAMMPS patches for pair hybrid/overlay. Thanks to Evan Smuts (U Cape Town) for merging these in.

However, there are some restrictions in 2.3.7 (see below) which will be relaxed over time:
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.3.6: (released 2 July 2013)

+ Added post-processing capability to fix heat/gran/conduction, which now also calculates and stores directional heat flux. Thanks to Stefan Radl (TU Graz).
+ Some enhancements to fix massflow/mesh, which is now also working with curved geometries. Thanks again to Stefan Radl (TUG).
+ Added a more clear error message in case mesh elements are lost due to moving out of the domain.
+ Fixed a rare bug with read_data associtated to initialization of atom data

However, there are some restrictions in 2.3.6 (see below) which will be relaxed over time:
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.3.5: (released 6 June 2013)

+ Included a commit and couple of fixes by Richard Berger (JKU) to ensure Windows compilation. Thanks to Richard and also to Christian Richter (OVGU Magdeburg) for an initial proposition.
+ Added an error check to the fix mesh/* commands to check for illegal rotation axis (such a 0,0,0)
+ Andreas Aigner (JKU) added an 'epsd2' rolling friction model, which is a variant to the recently-added EPSD model. Thanks to Andreas and to Christian Richter (OVGU Magdeburg) for helpful discussions.
+ Included a patch of Stefan Radl's group (TU Graz) who improved the compute nparticles/tracer/region command so that it computes and outputs more detailed information.
+ Fixed a bug with the newly introduced improvement of wall contact tracking for non-history wall styles (e.g. wall/gran/hooke, wall/gran/hooke/stiffness) which could lead to a program crash. Thanks to Daniel Queteschiner (JKU) for sending a test case that showed the issue.
+ Fixed a bug when restarting the simulation and a fix insert/* command should continue insertion. Somethimes this did not happen. Thanks to C. Toppin (Overland Conveyor) for posting a test case.

However, there are some restrictions in 2.3.5 (see below) which will be relaxed over time:
+ fix breakparticle/force is not yet available

BACKWARD COMPATIBILITY: Compute nparticles/tracer/region does now compute a vector, not a scalar

LIGGGHTS® 2.3.4: (released 17 May 2013)

+ Fixed a bug in yesterday's release of the new EPSD (elastic-plastic spring-dashpot) rolling friction model, implemented by Andreas Aigner (JKU). See Ai et al, Powder Technology, 206 (3), 269–282 for details on the model.

However, there are some restrictions in 2.3.4 (see below) which will be relaxed over time:
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.3.3: (released 16 May 2013)

+ Added an initial version of the EPSD (elastic-plastic spring-dashpot) rolling friction model, implemented by Andreas Aigner (JKU). See Ai et al, Powder Technology, 206 (3), 269–282 for details on the model.
+ Andreas Aigner (JKU) also made multiple improvements to the servo walls (fix mesh/surface/stress/servo) - thanks for this.
+ Implemented wall contact tracking for non-history wall styles (e.g. wall/gran/hooke, wall/gran/hooke/stiffness). This will result in better edge/corner contact detection and handling for non-history wall styles.
+ Added documentation for fix ave/euler and dump euler/vtk. Also fixed a small bug in fix ave/euler where a group ID other than "all" was not processed correctly.
+ Fixed an unwanted behavior in fix mesh/surface where non-normalized axis lead to extra scaling/rotation. Thanks to Pete Falkingham for the note and a testcase.
+ Fixed an issue which could arise when an insertion region for fix insert/pack is highly divided across many processors.
+ Fixed a rare memory issue with dump mesh/VTK for dumping certain parameters.
+ Also a small bugfix in dump mesh/stl where files had missed out the footer line in some cases. Thanks to the guys from Overland Conveyor for the note.
+ Improved the primitive wall neighlists that were actually too large in some cases. Thanks to Sam Bateman (USNRL) for the note.
+ Fixed a small issue in compute pair/gran/local when heat transfer properties were dumped and the group used was not "all".

However, there are some restrictions in 2.3.3 (see below) which will be relaxed over time:
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.3.2: (released 29 March 2013)

+ Patched a memory issue that came in with yesterday's release of shear history handling for primitve walls

However, there are some restrictions in 2.3.2 (see below) which will be relaxed over time:
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.3.1: (released 28 March 2013)

+ Released a new fix couple/cfd/force/implicit for use with CFDEM®coupling that uses a dragforce that varies with particle velocity rather than a constant one on the DEM side.

A couple of small updates for fix insert/*:
+ Added output of fix-ID to insertion info, which can be helpful in case multiple insertions are used.
+ Fixed an issue where in some cases no particles were inserted for cases with large # of procs (e.g. 128). Thanks to Sam Bateman (USNRL) for bringing this up.

A couple updates for wall handling:
+ Added a fix_modify option to fix mesh/surface/stress/servo which enables the user to change the target value during the simulation. Thanks to Andreas Aigner (JKU) for adding this functionality
+ Fixed an issue with the new fix mesh/surface/stress/servo where an unfix operation led to a misleading error message. Thanks to Xuetao Wang (Züblin Stuttgart) for flagging this.
+ Fixed an issue with consistency of the wiggle and riggle styles of fix move/mesh across simulation restart and multiple runs. Thanks again to Sam Bateman (USNRL) for posting this issue.

+ Fixed a bug where the conveyor feature did not set the surface velocity correctly for all ghost elements in parallel. Thanks to Christian Richter (OVGU Magdeburg) for flagging this issue.
+ Fixed 2 issues with store_force yes for fix wall/gran/* where values where not correctly computed/initialized in some cases. Thanks to François Guillard (Universités à Aix en Provence) for sending a patch.
+ Fixed an issue with shear history in relationship to primitive walls where the particle-contact releationship could get messed up for some particles upon leaving the domain. Thanks again to François Guillard (Universités à Aix en Provence) for noting this.

Further updates:
+ Included a patch from Andreas Aigner (JKU) that resolves an arg parsing issue with pair/gran/hooke/history/hysteresis.
+ Fixed a small communication issue with fix ave/euler and a dump problem in parallel with dump euler/vtk.
+ Added another patch for the CMake build system. Thanks to Anton Gladky (TU Freiberg) for the hint.
+ Fixed a couple minor typos in the doc and added a doc file for fix wall/region/sph.

However, there are some restrictions in 2.3.1 (see below) which will be relaxed over time:
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.3: (released 26 February 2013)

+ Added a new fix mesh/gran/surface/stress/servo to represent force-controlled boundaries with a PID controller. Thanks to Andreas Aigner (JKU) for implementing and testing the PID controller.
+ Added new fix property/atom/tracer and fix property/atom/tracer/stream commands along with a new compute nparticles/tracer/region command for the calculation of residence time distributions. Thanks to Rouven Weiler (BASF) for the suggestion as well as helpful discussions and providing test cases.
+ Added a new fix massflow/mesh command that counts #particles and mass penetrating thru an STL surface mesh. Thanks to Rouven Weiler and Moritz Höfert (BASF) for the suggestion as well as helpful discussions and providing test cases.
+ Added sanity checks for the values of Young's modulus, Poisson's ratio and coefficient of restitution as required by pair/gran/hooke, pair/gran/hooke/history and pair/gran/hertz/history.
+ Added output possibility for # of inserted particles and inserted mass to the fix insert/* commands.
+ Added an 'nparticles INF' option to fix insert/stream and fix insert/rate/region to be able to pour an infinite stream of particles.
+ Added a new keyword 'mix' to compute coord/atom to quantifiy species mixing. Thanks to Moritz Höfert (BASF) for contributing a code snippet for this.
+ Added an error check to fix mesh/surface/* to check for illegal args, such as a zero-rotation axis.
+ Fixed a typo bug in the SPH spiky2D kernel. Thanks to Markus Schörgenhumer and Andreas Aigner (both JKU) for tracing this back and providing a fix.
+ Fixed a typo in the doc of compute pair/gran/local. The "torque" output is the torque divided by the particle radius, not diameter. Thanks to user mojgan for bringing this up.
+ Added files for fix wall/gran/hooke/simple that were unintentionally left out in the last releases.
+ Fixed a small issue with fix adapt that could have led to a crash when unfixed before the end of the simulation.
+ Fixed a problem that could lead to a crash upon simulation restart in some cases when simulations using pair hybrid/overlay with gran/* was used . Thanks to Evan Smuts (U Cape Town) for sending a test case.
+ Fixed some broken macros in the source code. Thanks to Richard Berger (JKU) for patching this.
+ Added a patch by Richard Berger (JKU) that is necessary to compile the latest version of LIGGGHTS® with cygwin gcc.
+ Fixed a problem with the cross-plaform Cmake makefile contribution by Anton Gladky (TU Freiberg). Thanks to Anton and Richard Berger (JKU) for the note that this was broken.
+ Fixed a small memory issue for a (rarely used) functionality of fix property/global.

However, there are some restrictions in 2.3 (see below) which will be relaxed over time:
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.2.4: (released 24 January 2013)

+ Added an 'sjkr2' cohesion model for pair gran and fix wall/gran. It calculates the sphere-sphere contact area differently, i.e. A = 2*Pi * normal_overlap * (2 R), where R = ri*rj/(ri+rj).
+ Added a 'force off' keyword to pair gran to turn off all force computation of granular pair styles. This can be useful if you want to use pair granular in the frame of a hybrid pair style and just want to turn off the granular part.
+ Added a feature into pair gran to vary the coefficient of restitution depending on the local Stokes number. Thanks to Joris Heyman (EPFL) for this contribution.
+ Fixed a bug that could lead to a hang instead of an error message when the insertion region was too small to insert particles. Thanks to Andrew Hobbs (Astec Inc) for sending an example case
+ Fixed an issue with fix move/mesh superposition. Thanks to Christian Richter (OVGU Magdeburg) for sending a test case.
+ Fixed an issue with constistency of shear history across multiple runs or restart if compute pair/gran/local or compute/wall/gran/local was used.
+ Fixed an issue where in some rare cases there was not enough memory allocated for shear history when restarting a simulation.
+ Fixed an issue where the wall velocity used for the damping term for particle-wall collisions was slightly off in some rare cases.
+ Fixed an issue where particles were not inserted correctly when two fix insert/stream commands were used at the same time. Thanks to Chris Stoltz (PG) for sending a test case.

However, there are some restrictions in 2.2.4 (see below) which will be relaxed over time:
+ fix breakparticle/force is not yet available

BACKWARD COMPATIBILITY: You may not be able to use restart files with 2.2.4 that were generated by older version of LIGGGHTS®

LIGGGHTS® 2.2.3: (released 29 November 2012)

+ Extended fix adapt so it can handle atom-style variables for particle growth/shrinkage. Also added the example "packing" that was already there in 1.5.
+ Update for region/mesh/tet with improvements regarding volume calculation.
+ Fixed an issue where unfixing a fix insert/stream command mid-simulation could lead to a crash. Thanks to Andrew Hobbs (Astec) for sending a testcase that showed this problem.
+ Fixed a bug with random generation, affecting "gaussian","lognormal", and "uniform" distributions with fix particletemplate/sphere . Thanks to David Heine (Corning Inc) for providing a testcase.

However, there are some restrictions in 2.2.3 (see below) which will be relaxed over time:
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.2.2: (released 14 November 2012)

+ One more commit for new dump atom/vtk functionality (see yesterday's release).

However, there are some restrictions in 2.2.2 (see below) which will be relaxed over time:
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.2.1: (released 13 November 2012)

+ Added a dump atom/vtk command which now enables LIGGGHTS® to write to VTK files directly. However, this requires LIGGGHTS® to link against VTK libraries, see the new makefile Makefile.ubuntuVTK for an example how to do this. Thanks a lot to Anton Gladky (TU Freiberg) for this contribution.
+ Added a CMake ("Cross Platform Make", www.cmake.org) build system for LIGGGHTS®. Thanks again to Anton Gladky (TU Freiberg) for setting this up.
+ Added the region mesh/tet functionality as it was implemented in LIGGGHTS® 1.5.3, also added a tutorial "mesh/tet".
+ Added an auto-check for option atom_modify map array when LIGGGHTS® is used in coupling mode. Thanks to Moritz Höfert (BASF) for the motivation for this.
+ Fixed a mesh restart bug introduced in 2.2. Thanks to Tapan Sabuwala (U Illinois) for noting this.
+ Fixed a problem with the shear option for cylinder. thanks to Chris Stoltz (PG) and Kevin Brennan (Perdue) for letting us know about this.
+ Fixed an issue with fix mesh/surface/stress. Thanks to Christian Richter (OVGU Magdeburg) for sending a test-case.
+ Fixed a small issue with shear history continuity across planar faces, when the faces were no direct neighbors.
+ First commit for an implicit integration driver, which should allow to run implicit integration schemes in the future.

However, there are some restrictions in 2.2.1 (see below) which will be relaxed over time:
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.2: (released 8 October 2012)

A couple of updates and improvements for surface mesh import and handling:
+ LIGGGHTS® is now more stringent in how STL meshes are checked upon loading by fix mesh/surface. More quality checks and according error/warning messages were added to check for corrupt/numerically bad meshes upon import into LIGGGHTS®. Criteria that are now checked for are duplicate elements, coplanar overlapping elements, element aspect ratio, number of neighbors per element, mesh elements leaving the domain and a curvature sanity check. For all these changes, see doc for fix mesh/surface for further details.
+ Also, for each issue, LIGGGHTS® now indicates the exact line number in the STL file where the problem happens (LIGGGHTS® 1.5 did give aproximate locations in the file).
+ Added a 'heal auto_remove_duplicates' function to fix mesh/surface that tries to repair a geometry that suffers from duplicate elements.
+ Keyword 'accuracy' for fix mesh/surface was renamed to 'precision'.
+ A new keyword 'verbose' for fix mesh/surface to control the verbosity of warnings and outputs.
+ Added a new fix mesh/surface/planar with additional functionality for planar meshes.
+ Added a "region" keyword to dump mesh/stl that allows the user to dump just these mesh element where the element center is in the specified region.
+ Niels Dallinger (Chemnitz UT) contributed "viblin" and "vibrot" styles for fix move/mesh. These mimic mesh movements which are are typically used e.g. for vibratory conveyors. Thanks for this contribution!
+ Fixed an inconsistency in the doc for fix move/mesh/rotate/variable. Keyword 'period' is renamed to keyword 'omega' as the number specified after this keyword is actually an angular velocity (as was stated correctly in the doc). Thanks to Stefan Amberger (JKU) for the note.
+ Fixed a bug in the fix neighlist/mesh routines that could - under rare circumstances - result in LIGGGHTS® to hang. Thanks to M. Sandlin (Georgia Tech) for posting an example that showed this problem.

New features and improvements for particle insertion:
+ Added the missing 'all_in yes' functionality for fix insert/stream, which uses the aforementioned fix mesh/surface/planar. The new functionality should perform much better for large insertion meshes than the old method in LIGGGHTS® 1.5.
+ Based on a contribution of Stefan Radl (Graz UT, thanks!), "vel uniform" and "vel gaussian" velocity options were added for the fix insert/pack, fix insert/stream and fix insert/rate/region commands.
+ Made random seeds for particle insertion processor-dependant so particle size distributions are covered more accurately for large # of processors for species that contain just few particles. Thanks to David Heine (Corning Inc) for the suggestion.

However, there are some restrictions in 2.2 (see below) which will be relaxed over time:
+ region tetmesh is not yet available
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.1.2: (released 21 September 2012)

+ Andreas Aigner (JKU) added a first version of a hysteretic pair/gran/hooke/history/hysteresis style with cohesion. There is also a new tutorial case 'hysteresis' for that. Doc is to follow...
+ Andreas Aigner (JKU) also committed a patch for fix ave/euler/stress, improving support for triclinic boxes
+ Improved numerical round-off for reading STL files. Added an 'accuracy' keyword to fix mesh/surface so the user can control this. Also added an 'nneighs' keyword to dump mesh/vtk to check if neighbor topology is governed correctly. Thanks to Chris Stoltz (P&G) for sending cases that showed that there was room for improvement.
+ Re-work for particle distributions as defined by fix particletemplate/sphere. Now it supports 'constant', 'uniform mass', 'uniform number', 'gaussian number' and 'lognormal number' distributions.
+ Fixed a bug in compute wall/gran/local. Thanks to Tapan Sabuwala (U Illinois) for posting a testcase.
Improvements for fix insert/stream:
+ Firstly, a bug in the parallelization was fixed. Thanks to Rouven Weiler (BASF) for noting the problem.
+ Secondly, a 'parallel' keyword has been added to fix insert/stream which can improve insertion performance if the fix insert/stream extrusion extends across many processes. However, as this is normally not the case, this option if off per default.

However, there are some restrictions in 2.1.2 (see below) which will be relaxed over time:
+ region tetmesh is not yet available
+ 'all_in yes' is not yet available for fix insert/stream
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.1.1: (released 7 September 2012)

+ Due to recurring problems in the latest releases, there is a complete re-work of edge and corner detection in mesh and update to the particle-mesh distance algorithm, with simpklifications, bugfixes and corrections for the parallel case. Thanks ot Chris Stoltz (PG) for sending lots of test cases that helped us find the source of the issues.
+ Added "aedges", "acorners" keywords to the dump mesh/vtk command to be able to visualize active/inactive edges and corners in a mesh.
+ Other small bug-fixes in parallel mesh handling.
+ Re-work of the fix insert/stream command so it can cope with mesh extrusion extending over multiple subdomains.
+ Fixed a couple small typos and omission in the doc. Thanks to Moritz Höfert (BASF) for the note.
+ Added a few missing files for the SPH upgrade in 2.1 that were unintentionally left out. Thanks to Andreas Aigner (JKU) for the note.
+ Modified the fix nve/limit command so the limit can be given in terms of particle radius, which is more convenient for relaxing granular simulations.
+ Some preparations for a coarsegraining approach

However, there are some restrictions in 2.1.1 (see below) which will be relaxed over time:
+ region tetmesh is not yet available
+ 'all_in yes' is not yet available for fix insert/stream
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.1: (released 30 August 2012)

+ Added a big SPH patch from Andreas Aigner (JKU) with improved efficiency and numerics for SPH. Also added a new SPH example
+ Included some re-work by Philippe Seil (JKU) for the particle-mesh distance algorithm where corner contacts were not detected correctly in rare cases. Thanks to Tapan Sabuwala (U Illinois) for posting an example that showed this.
+ Added support to specify the list of meshes to be dumped by the dump/mesh/stl and dump/mesh/vtk commands, rather than dumping all meshes in the simulation automatically.
+ Added an improved restart logic to fix move/mesh/linear/variable and fix move/mesh/rotate/variable.
+ Fixed an issue with superposition of two rotation commands that work with the same rotation axis. Thanks to Christian Richter (OVGU) for pointing out there was an issue.
+ Fixed a small glitch in the new "output" option of the dump mesh/VTK command. Thanks to Philippe Seil (JKU) and Christian Richter (OVGU) for noting this.

However, there are some restrictions in 2.1 (see below) which will be relaxed over time:
+ region tetmesh is not yet available
+ 'all_in yes' is not yet available for fix insert/stream
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.0.7: (released 23 August 2012)

+ Added a fix ave/euler/stress that calculates the correct stress for dense granular system and averages it over a Eulerian Grid (compute stress/atom neglects the contribution of the average velocity in the molecular diffusion term of the stress tensor). Also added a dump euler/vtk to dump the stress to VTK files. Thanks to Simon Schneiderbauer (JKU) for helpful discussions and to Andreas Aigner (JKU) for bugfixes and for adding code to make this feature work for triclinic boxes.
+ Added a new "verbose" option to all the fix insert commands to turn off printing info about inserted particles during a run that can potentially make thermo output unreadable.
+ Added a new "output" option to dump mesh/VTK command to output per-face values in an interpolated manner. Thanks to Felix Kleinfeldt and Christian Richter (both OVGU Magdeburg) for this contribution.
+ Fixed a small issue with how fix insert/rate/region was parsing args where wrong command syntax could result in mis-leading error messages.
+ Corrected the fix heat/gran doc with respect to the area_correction flag. Thanks to Moritz Höfert (BASF) for the note.
+ Fixed a bug where particle-triangle interactions were not calculated correctly for some cases where the triangle hs obtuse angles. Thanks to Chris Stoltz (P&G) for a test case and to Philippe Seil for help with fixing this.
+ Added fixgroup support for fix wall/gran/*. Now the wall is only applied to particles in the fix group. Thanks to Tap Sab and Mark Bentley (Space Reseach Institute, ÖAW) for noting that this was missing.
+ Re-added some missing part in compute wall/gran/local where the triangle ID was always zero in the output. Thanks again to Mark Bentley (Space Reseach Institute, ÖAW) for the note.

However, there are some restrictions in 2.0.7 (see below) which will be relaxed over time:
+ region tetmesh is not yet available
+ 'all_in yes' is not yet available for fix insert/stream
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.0.6: (released 8 August 2012)

+ Fixed an issue with the recently added fix move/mesh/linear/variable command. Thanks to Christian Richter (OVGU Magdeburg) for finding this bug.
+ Added an error message if a user would provide a vector with length 0 for the fix gravity command. Thanks to user jooj for reporting this.
+ Fixed an issue that would prevent the correct particle-wall restitution coefficient to be pictured for particle-triangle contacs at the edge of two neighboring elements. Thanks to Chris Stoltz (P&G) for noting this deficiency.
+ Fixed an issue with distributing the remainder of non-integer-number for particles to insert of a particle distribution. Now particle distributions should be matched more exactly. Thanks again to Chris Stoltz (P&G) for noting this.
+ Fixed a bug in particle-wall heat transfer. Thanks to Kirk A. Fraser (Roche Consulting Group) for letting us know.
+ Fixed a bug with detection of mesh topology (edges and nodes) in some parallel cases. Thanks to Eric Parteli (U Erlangen) for reporting the bug and sending a case that showed this.
+ Some more small internal bugfixes.

However, there are some restrictions in 2.0.6 (see below) which will be relaxed over time:
+ region tetmesh is not yet available
+ 'all_in yes' is not yet available for fix insert/stream
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.0.5: (released 3 August 2012)

+ Fixed an issue with fix insert/pack where option "n_particles_in_region" with "all_in yes" did not work correctly due to a round-off issue when the insertion volume was the whole box.
+ Performed a small re-work in container classes to make mesh communication more efficient.
+ Fixed an internal bug in mesh communication.
+ Fixed a bug in particle-wall heat transfer. Thanks to Kirk A. Fraser (Roche Consulting Group) for letting us know.
+ Fixed a bug in fix move/mesh that resulted in wrong mesh position for consecutive runs. Thanks to Chris Stoltz (P&G) for providing a script that showed the issue.
+ Improved random generation for parallel particle insertion. Now random seeds are different for each process, making the insertion look "more random".
+ Couple of low-level bug-fixes.

However, there are some restrictions in 2.0.5 (see below) which will be relaxed over time:
+ region tetmesh is not yet available
+ 'all_in yes' is not yet available for fix insert/stream
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.0.4: (released 20 July 2012)

+ Andreas Aigner (JKU) made an an update to the SPH models: he added 2D-sph-kernels and fixed a bug in the wendland kernel
+ Christian Richter (OVGU Magdeburg) contributed variable-style mesh movmement for fix move/mesh: linear/variable and rotate/variable
+ Fixed a bug in fix move/mesh where the mesh was was not rotated correctly when the rotation origin was not 0/0/0.
+ Fixed a typo in the fix wall/gran doc. Thanks to Sam Bateman (USNRL) for noting this.
+ Fixed a bug with torque calculation via fix mesh/surface/stress for moving meshes. The issue was that the torque reference point was not updated. Thanks to Ryan Houlihan (Temple U) for the note.
+ Fixed a bug in the set command where the per-particle mass was not correctly inititalized when changing the diameter
+ Fixed an issue with overlapping particles at inter-processor boundaries when fix insert/pack or fix insert/rate/region was used. Thanks to Rikio Soda (Abo University) for providing a test case that showed this
+ Fixed one more typo bug in particle insertion. Thanks to Chris Stoltz (P&G) for drawing attention to this.
+ Fixed an issue with mesh parallelization where mesh elements could be lost under some circumstances which would lead to a simulation crash

However, there are some restrictions in 2.0.4 (see below) which will be relaxed over time:
+ region tetmesh is not yet available
+ 'all_in yes' is not yet available for fix insert/stream
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.0.3: (released 13 July 2012)

+ Finally released a first version for fix mesh/surface/stress to perform mesh stressanalysis. Re-added the "chute_wear" example into the release, doc was also updated accordingly
+ Implemented general reverse communication for mesh properties
+ Fixed an issue with shear history transfer between co-planar faces revaled by a test harness run
+ Improved recognition of shared edges and nodes in the surface mesh implementation
+ Added a new option 'area' to dump mesh/vtk to dump the area
+ Added the 'shear' option for primitive fix wall/gran styles that was missing. Removed the 'wiggle' option from the doc which is no longer available in 2.X. Please note that you can do all this and more with fix move/mesh. Thanks to Eric Partelli (U Erlangen) for noting there was a problem.

Fixed a couple issues with particle insertion:
+ Fixed a round-off glitch in particle insertion which could make the insertion hang when run in parallel.
+ Fixed a bug in fix insert/rate/region where the 'nparticles' flag was not correctly recognized. Thanks to Chris Stoltz (P&G) for the hint.
+ Fixed a bug where fix insert/stream insertion was not performed correctly if there was a run between the fix mesh/surface and the fix insert/stream command. Thanks to Daniel Schiochet Nasato (JKU) for noting this and providing a test case.

LIGGGHTS® 2.0.2: (released 06 July 2012)

+ Added a check to fix property/global for style 'atomtypepair' that checks if the values are symmetric. Note that now style 'matrix' is not necessarliy symmetric, but 'atomtypepair' is.
+ Fixed a bug in random number generation for parallel cases when fix insert/rate/region or fix insert/pack were used. Thanks to Chris Stoltz (P&G) for sending a test case.
+ Added some new option 'random_distribute'='exact' to all the fix insert/* commands (which is now default). This allows to match the requested # of particles to be inserted at each insertion step more exactly, but introduces some small MPI overhead. Option 'random_distribute'='uncorrelated' will reporduce the way things worked in LIGGGHTS® 2.0 and 2.0.1.
+ Fixed an issue with unfixing multiple fix move/mesh commands which operate on different meshes. Thanks to Joe Satch and Peter Falkingham (Royal Veterinary College, University of London) for posting a cases that showed this issue
+ Fixed a bug with pair gran/hooke/history/stiffness where LIGGGHTS® would exit with a segfault afer finishing the simuation. Thanks again to Joe Satch for letting us know.
+ Fixed some more minor machine-dependant compilation issues and typos in the doc. Thanks to Stefan Amberger (JKU) for the notes.
+ Fixed a formatting issue in the fix mesh/surface doc. Thanks to Andreas Aigner (JKU) for the note.
+ Applied a whitespace patch that should help making LIGGGHTS® code more readable. Thanks to Steve Plimpton, Paul Crozier (both Sandia) and Axel Kohlmeyer (Temple) for collaboration on that subject.
+ Fixed a small bug in fix transportequation/scalar that would only arise if that fixed is used for other purpose than for heattransfer.

However, there are some restrictions in 2.0.2 (see below) which will be relaxed over time:
+ region tetmesh is not yet available
+ fix mesh/gran/stressanalysis is not yet available
+ 'all_in yes' is not yet available for fix insert/stream
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.0.1: (released 29 June 2012)

+ Fixed an issue where neighborlist build was not triggered correctly for some cases with fix move/mesh. Thanks to Peter Falkingham (Royal Veterinary College, University of London) for posting a case that showed this issue.
+ Added restart functionality for fix mesh and fix move/mesh that was missing in the 2.0 version.
+ Fixed a general bug with restart where not all atom properties were restarted correctly.
+ Fixed a two typos and omissions in the doc. Thanks to Niels Dallinger (TU Chemnitz) for noting this.
+ Fixed an issue with on-the-fly reallocation of contact history.

However, there are some restrictions in 2.0.1 (see below) which will be relaxed over time:
+ region tetmesh is not yet available
+ fix mesh/gran/stressanalysis is not yet available
+ 'all_in yes' is not yet available for fix insert/stream
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.0: (released 26 June 2012)

+ Made a final syntax change - fix mesh/surface is now the syntax for a surface mesh in LIGGGHTS®. Also removed the backward compatibility to the fix mesh/gran syntax as this would have lead to troubles sooner or later.
+ Removed package "GRANULAR" as its files are already present in the core of LIGGGHTS® - this should help to avoid confusion and unwanted file destruction.
+ Fixed a machine-dependant compilation issue. Thanks to Christian Richter (U Magdeburg) and Mark Bentley (Space Reseach Institute, ÖAW) for pointing this out.
+ Fixed another compilation issue when compiling with the MPI stubs library. Thanks again to Mark Bentley (Space Reseach Institute, ÖAW) for noting this.
+ Added some files necessary for the SPH models which were accidentially left out in the 2.0RC version. Thanks to Andreas Aigner (JKU) for noting this.
+ Added a "region wedge" command written by Stefan Amberger (JKU).
+ Fixed a bug related to unfixing of fix move/mesh commands. Thanks for Moritz Höfert (BASF) and Philippe Seil (JKU) for noting this.
+ Fixed a bug related to the fix insert/rate/region command. Thanks to Chris Stoltz (P&G) for reporting this.
+ Fixed a couple small memory leaks.
+ Introduced a couple of cosmetic changes for internal use of vector functions and MPI wrapper functions.
+ Fixed an issue with fix transportequation/scalar (used by fix heat/gran) where the quantity to be calculated was not forward communicated to some ghosts correctly on reneigh steps.
+ Fixed a small issue with correct damping in pair gran and fix wall/gran for particles that are part of a fix rigid. Thanks to Steve Plimpton (Sandia) for code snippets and a helpful discussion.

However, there are some restrictions in 2.0 (see below) which will be relaxed over time:
+ restart for fix mesh is not available, but will be released shortly
+ region tetmesh is not yet available
+ fix mesh/gran/stressanalysis is not yet available
+ 'all_in yes' is not yet available for fix insert/stream
+ fix breakparticle/force is not yet available

LIGGGHTS® 2.0 RC: (released 15 June 2012)

Major feature upgrades include:
+ Merged in more than 2 years of LAMMPS updates & patches (since 10 March 2010), see http://lammps.sandia.gov/bug.html
+ New syntax for many commands, yielding a hopefully clearer input script. For details, see doc/LIGGGHTS®_2.X_tutorial.html in the LIGGGHTS® tarball
+ All-new mesh implementation, featuring:
+ + modular mesh design using property storage class
+ + fix mesh is now faster due to reworks in mesh topology
+ + fix mesh contains a new optimzed particle-mesh distance algorithm by Philippe Seil and Patrick Wijerama (JKU)
+ + mesh operations and moving mesh operations are more efficient as well
+ + it is now possible to superpose multiple fix move/mesh commands
+ particle insertion via the family of fix insert commands is now fully parallel, overlapcheck will scale in parallel

Further feature upgrades include:
+ thermo output now supports Cundall Number (thermo_custom keyword cu)
+ added a new dump decomposition/vtk command to dump the processor decompoistion for parallel runs into VTK files
+ doc/Manual.pdf is now always up-to-date with every release

Fixed a number of bugs:
+ some bugfixes related to fix property/atom
+ fixed a bug when using fix insert commands in parallel with keyword 'volume_fraction_region'

However, there are some restrictions in 2.0RC which will (see below) which will be relaxed later:
+ region tetmesh is not yet available
+ fix mesh/gran/stressanalysis is not yet available
+ restart functionality for fix mesh is not yet available
+ 'all_in yes' is not yet available for fix insert/stream
+ fix breakparticle/force is not yet available

We now ported to Github, meaning the location of the GIT server has changed for the LIGGGHTS® 2.X branch. For info how to access it, see here.

LIGGGHTS® 1.5.3 (released 22 May 2012)

+ Fixed an insertion bias in fix insert/stream. Now the stream is more continuous. Thanks to Christian Richter (OVGU Mageburg) for noting this
+ Fixed some small inconsistencies when checking for overlaps in fix insert/pack
+ Fixed an issue with region tetmesh when used with 'side out'
+ Added a new makefile by Christoph Goniva to ease compilation together with the CFDEM®coupling
+ Fixed a small issue with compute pair/gran/local. Thanks to Sergei Dorofeenko (Russian Academy of Sciences) for sending a testcase that showed this problem.
+ Fixed an issue with outputting per-body data from fix rigid. Thanks to Will Smith (U Mich) for reporting this

LIGGGHTS® 1.5.2 (released 13 March 2012)
CFDEM® COMPATBILITY: 1.5.2 is currently not able to couple to CFDEM®
+ Fixed two issues in the rolling friction model. Thanks to Matt Roesler (ETH Zürich) for detecting the issues in the code and to Prof. Jin Ooi (U Edinburgh) and Matt Roesler for a very helpful discussion discussion.
+ Fixed a bug in the 'region cone' command. Thanks to Maria Angelova (TUHH) for bringing this up and to Andreas Aigner (JKU) for tracking down the bug.
+ Added error messages thrown by missing quantities more informative by providing info which model requires them.
+ Fixed an issue with compute pair/gran/local and compute/wall/gran/local when the pair/fix they are referring to changes during the simulation. Thanks to Sergei Dorofeenko (Russian Academy of Sciences) for sending a testcase that showed this problem.
+ Added a 'dump_modify label' option to dump custom, to enable outputting the simulation time to the real-time viz in Paraview developed by Christian Richter (OVGU Magdeburg)

Improvements to particle insertion:
+ Resolved an issue with using new meshes and fix insert/stream commands in a restarted simulation that were not present in the simulation that generated the restart file.
+ Fix insert/stream now works also with 'all_in yes'.
+ Made some optimizations to make simulations faster if a large number of fix insert/stream commands are used in the simulation
+ Deactivated dynamic regions for usage with fix insert/pack and fix insert/rate/region. Thanks to Sam Bateman (US Naval Research Lab) for sending a script that showed there is a problem.

BACKWARD COMPATIBILITY: The fix for the rolling friction model will lead to different results now if the same parameters are used. Anyway, the results should be comparable to the results produced by versions older than 1.5
CFDEM® COMPATBILITY: 1.5.2 is currently not able to couple to CFDEM® - Please use the stable version 1.5 from 'git://CFDEM.git.sourceforge.net/gitroot/CFDEM®/LIGGGHTS®' instead

LIGGGHTS® 1.5.1 (released 2 Feb 2012)
CFDEM® COMPATBILITY: 1.5.1 is currently not able to couple to CFDEM®

+ Fixed a bug with the "riggle" for fix move/mesh/gran. Thanks to Niels Dallinger (TU Chemnitz) for noting this.
+ Fixed a bug in region block when used as wall. Thanks to Andreas Aigner (JKU) for debugging and to Moritz Höfert (BASF) for a script that showed the problem.
+ Fixed an inconsistency where the set command for property/atom had no effect when issued right after the fix. Thanks to Claudio Wolfer (NTB Buchs) for flagging this problem.
+ Fixed typos and inconsistencies in the doc of fix insert/pack, thanks to both Evan Smuts (U Cape Town) and Stefan Amberger (JKU) for bringing this up.
+ Stefan Amberger (JKU) changed the particle insertion to the newer "fix insert" commands, which replace the older fix pour and fix pour/dev commands.
+ Fixed a neighborlist issue with variable particle radius.
+ Updating the rolling friction model for pair gran/hooke was accidentally lef out in the 1.5 patch. Also fixed an issue where the rolling torque might be unitialized in rare cases. Thanks to Diego Peinado (Intrame) for noting this
+ Dump mesh/gran/VTK now can dump mesh temperature along with other quantities - thanks to user skyopener for suggesting this.

BACKWARD COMPATIBILITY: Examples will produce differrent results
CFDEM® COMPATBILITY: 1.5.1 is currently not able to couple to CFDEM® - Please use the stable version 1.5 from 'git://CFDEM.git.sourceforge.net/gitroot/CFDEM®/LIGGGHTS®' instead

LIGGGHTS® 1.5 (released 30 November 2011)

New features:
+ Added a new (simple) model for particle breakage via command fix breakparticle/force with new example 'breakage'. Doc for the command to come in the next release

Updates to the insertion commands:
+ Re-work for fix insert/pack, see doc for details
+ New insertion command fix insert/rate/region
+ New keywords "start" for fix insert/pack, fix insert/stream and fix insert/rate/region and new keyword 'duration' fix insert/stream.
+ Improvements to fix insert/stream which makes the stream more continuous for option overlapcheck = yes

Some bug fixes:
+ Fixed a bug in rolling friction model that actually violated equilibrium by applying different torques to the two particles for poly-disperse systems - thanks to Prof. Katterfeld (University of Magdeburg) for pointing this out.
+ Fixed a memory issue with fix particledistribution/discrete and fix template/sphere if distributions were used by multiple fix insert commands or templates were used by multiple particle distributions.
+ Fixed a typo in compute pair/gran/local. Thanks to Rami Jabakhanji for tracking this down.

BACKWARD COMPATIBILITY NOTE:
+ Changed keyword 'volumefraction' to 'volumefraction_region' for fix insert/pack - check doc for details.
+ The fix for the rolling friction model will lead to different results now if the same parameters are used.

LIGGGHTS® 1.4.6 (released 21 October 2011)

+ Extended the neighborlist build criterion to include variable particle radii as trigger for neigh list re-build
+ Added a warning to fix insert/stream if insertion area extends outside the simulation domain so that atoms might be lost during insertion. Thanks to Alice Hager (JKU) for providing a test script.
+ Fixed a rare bug in internal fix handling. Thanks to Mark Bentley (Space Research Institute, Austrian Academy of Sciences) for posting an offending input script.
+ Fixed a parsing issue with pair hybrid. Thanks to Moritz Höfert (BASF) for posting an offending input script.
+ Fixed a bug in compute pair/gran/local where the number of contacts was not counted correctly in parallel, causing LIGGGHTS® to crash. Also corrected the output of the periodicity flag for parallel computation. Thanks again to Mark Bentley (Space Research Institute, Austrian Academy of Sciences) for flagging these issues.

Some updates to fix mesh/gran:

+ Added an additional keyword "curvature" that lets the user define up to which angle between two faces the faces should be treated as belonging to the same surface (e.g. useful for bends)
+ Removed the force area scaling for particle-mesh interactions as introduced in 1.2.7 (fix_wall_gran.cpp:913). Thanks to Sergei Dorofeenko (Russian Academy of Sciences) for sending a testcase that showed a systematic deficiancy.
+ Fixed a small bug that prevented correct contact history transition from one face to another in some (rare) cases.

LIGGGHTS® 1.4.5 (released 12 October 2011)

+ Improved the fix heat/gran command to account for correction of the contact area by providing an original (unsoftened) value for the Young's modulus.
+ Added the possibility in fix wall/gran to store the per-particle wall force. Also fixed a small bug in arg error check for fix wall/gran. Thanks to Shantanu Maheshwari for posting an offending script.
+ Improved the accuracy for restart of contact history for granular wall - restarts should be more exact now.
+ Added a new keyword 'stresscomponents' to the dump mesh/gran/vtk command. Thanks to Christine Bertsch (Züblin Stuttgart) for the suggestion and for sending some code.
+ Hopefully fixed a bug with compute pair/gran/local. Thanks to Mark Bentley (Space Research Institute, Austrian Academy of Sciences) for collaboration on that subject. Also, changed output for periodic simulations - now the coordinate wrapped into the box is output.
+ Fixed a bug with particle type initialization when using fix insert. Thanks to Timo Kuljo (University of Oulu) for sending a testcase that showed this.
+ Re-added makefile.lib to the /src directory that got lost somehow. Thanks to Moritz Höfert (BASF Ludwigshafen) for noting this.
+ Fixed small memory issues in fix property/global and fix couple/cfd.
+ Fixed a bug in the fix insert/pack command using option 'all_in = yes'.
+ Fixed an issue with fix insert/stream. Thanks to Sergei Dorofeenko (Russian Academy of Sciences) and Simon Wallner (Vienna UT) for reporting this and to Alice Hager (JKU) for help with debugging. Also fixed a few issues with missing error-checks on the arguments of fix insert/stream parameters and updated the fix insert/stream doc.

BACKWARD COMPATIBILITY: You may not be able to use restart files with 1.4.5 that were generated by older version of LIGGGHTS®

LIGGGHTS® 1.4.4 (released 22 September 2011)

+ Changed the fix heat/gran implementation so that heat transfer between particles in the fix group and particles not in the fix group is included. Thanks to Mark Bentley (Space Research Institute, Austrian Academy of Sciences) for the suggestion
+ Fixed a bug in the FixPropertyGlobal routine discovered by Christine Bertsch (Züblin Stuttgart).
+ Introduced the option to include or exclude mass for damping in the frame of the wall/gran/hooke/history/stiffness command like already introduced for the pair/gran/hooke/history/stiffness style.
+ Some updates related to compute pair/gran/local: Fixed a segfault issue in compute pair/gran/local and added a new periodicity flag that states if an interaction is periodic(useful for post-processing). Also, added a new command compute wall/gran/local to be able to access particle-wall contact data as well
+ Fixed a bug in the fix wall/region command. Thanks to Tap Sab for reporting this.

LIGGGHTS® 1.4.3 (released 9 September 2011)

+ Fixed an issue in the fix property/atom implementation revealed by a nightly testharness run

LIGGGHTS® 1.4.2 (released 8 September 2011)

+ Preparing for merging some basic LIGGGHTS® stuff into LAMMPS together with Steve Plimpton (Sandia). Therefore, some cosmetic changes in filenames and class names have been performed. E.g. renamed fix property/peratom class to fix property/atom, and changed fix property/global keywords 'peratomtype' and 'peratomtypepair' to 'atomtype' or 'peratomtype'. However, the old keywords are still supported.
+ Fixed a bug in the CFD coupling that slipped in one of the last releases. It only applied if using the data coupling via files, not the MPI coupling. Thanks to Christoph Goniva (JKU) for flagging this via his CFDEM® test harness.
+ Fixed a bug where the coordination number was calculated incorrectly by compute coord/gran. Thanks to Mark Bentley (Space Research Institute, Austrian Academy of Sciences) for finding this and providing a testcase
+ Added funtionality to allow compute pair/gran/local to access contact area and heat flux (for thermal simulation), enabled vis 'contactArea' and 'heatFlux' keywords. Thanks again to Mark Bentley (Space Research Institute, Austrian Academy of Sciences) for the suggestion.
+ Added a sanity check for the VTK check of the 'region tetmesh' command. Thanks to Fabian Uhlig (TU Freiburg / U Twente) for motivating this.
+ Corrected/updated the doc for the pair/gran/hertz/history/stiffness and pair/gran/hooke/history/stiffness styles. Also, you can now choose between two options for damping in the pair/gran/hooke/history/stiffness style. Thanks to Fabian Uhlig (TU Freiburg / U Twente) for pointing out there was an incosistency
+ Removed option 'gaussian' from the particletemplate/sphere command as it is no longer supported. Thanks to user willroc7 for flagging this.
+ Fixed an issue with the dump stl command writing the STL footer multiple times in parallel. Thanks to Timo Kulju (Oulu University Finland) for reporting this.
+ Finally added the missing doc for the region tetmesh, fix insert/stream and fix insert/pack commands added in version 1.4

BACKWARD COMPATIBILITY: Renamed keywords 'peratomtype' and 'peratomtypepair' to 'atomtype' and 'atomtypepair'. However, the old keywords are still supported.

LIGGGHTS® 1.4.1 (released 4 August 2011)

+ Fixed a bug in the region tetmesh command revealed by a testharness run by Stefan Amberger (JKU)

LIGGGHTS® 1.4 (released 3 August 2011)
+ Added a new SPH (Smoothed Particle Hydrodynamics) modelling capability based on the work of Andreas Aigner (JKU). See example here. Also added a new example "sph".
+ Fixed a rare bug in the probability distribution implementation. Thanks to Andre Katterfeld (Univ. Magdgebug, Cepartec GmbH) for sending an example that showed the bug.
+ Improved the accuracy of the fix check/timestep/gran command to help the user avoid "dangerous builds in triangle neighbor list" messages
+ Changed the output for the wear prediction feature from wear to wear/area
+ Fixed a bug in the way damping is calculated in case a fix rigid is active for granular particle-particle and particle-wall interactions.

BACKWARD COMPATIBILITY: The fix/mesh/gran/stressanalysis feature will output wear/area now instead of wear, so results will not be directly comparable with former versions

LIGGGHTS® 1.3.3 (released 12 July 2011)
UPGRADE FROM 1.3.2 RECOMMENDED
+ Fixed a bug in fix wall/gran - thanks to Stefan Beinert (TU Braunschweig) and Thomas Wagner (TU Freiberg) for reporting this in the forum
+ updated the restart2data tool
+ fixed a couple of minor memory issues

LIGGGHTS® 1.3.2 (released 11 July 2011)

+ Some optimization for fix wall/gran/* styles, they should now run a bit faster on most systems
+ Initial commit that will allow to track the energy (elastic, dissipated...) for several new pair styles we are working on

LIGGGHTS® 1.3.1 (released 15 June 2011)

Fixed a typo bug in the fix wall gran command

LIGGGHTS® 1.3 (released 14 June 2011)
New major pack of updates

Multiple updates to the fix mesh/gran command:
1) Speed up in the mesh import routine, is now ~factor 10 faster for large meshes
2) Added a new rotation option to mesh gran. Thanks to Evan Smuts (U Cape Town) for submitting this
3) Fixed a tolerance issue for calculating the damping force calculation in case of moving mesh
4) Now triangle meshes can also be imported from (legacy) ASCII VTK files.

Further updates and bug fixes:
1) Added a new cone-stype granular wall via the fix wall/gran command. Thanks to Chris Stoltz (P&G) for sending the code.
2) Fixed a typo bug in the check timestep gran routine and also fixed a typo in the doc. Thanks to Sergei D for finding this
3) Fixed an initialization bug in the pair/gran/hooke/history/simple and pair/gran/hertz/history/simple styles. Thanks to Christine Bertsch (Züblin) for running into this.

The following new features that ease up particle insertion are added in a beta mode. There is no full doc yet, feel free to try the new examples and browse the source code.
1) Added a new region tetmesh command, that reads tetrahedral meshes from VTK that can be used as regions. This is not only useful for particle insertion, but also for post-processing
2) Added a new fix insert/stream, which is similar to the fix pour commands, but takes a triangle mesh (fix mesh/gran) as boundary condition. This mesh is then extruded in the direction of the insertion velocity to give an insertion volume. The assumption 'gravity must point in -z direction' is now dropped.
3) Added a new fix insert/pack command that can now use arbitrary regions (also the new region tetmesh) for insertion instead of just box and z-cylinder

Added 3 new tutorials to show these new capabilities: region_tetmesh, insert_pack and insert_stream

LIGGGHTS® 1.2.9 (released 26 April 2011)

+ Added a new fix transportequation/scalar that solves general particle-based transport equations, generalizing the code from the heat transfer model
+ Some improvements to heat transfer model resulting in simulation speed-up for dilute cases
+ Initial commit of the LIGGGHTS® part for a convective heat transfer model in the frame of the CFD-DEM coupling, CFD part to come over the next weeks
+ Improved CFD-DEM coupling so that now particles can be inserted/deleted during a coupled run ("consecutive tags" error).
+ Added a doc and fixed some bugs with the new new pair styles pair/gran/hooke/history/stiffness, pair/gran/hertz/history/stiffness and according wall styles.
Thanks to Christine Bertsch (Züblin) for reporting this

LIGGGHTS® 1.2.8 (released 06 April 2011)

Test harness results vs. 1.2.4 can be seen here.

+ Initial release of the MPI datacoupling module for LIGGGHTS® side the CFD-DEM coupling. MPI coupling module soon to be released on the CFD side. So stay tuned... Thanks to Christoph Goniva and Alice Hager (both JKU) for all the cooperation on the subject!
+ Fixed an issue with shear history, discussed here. Thanks to George Marketos (Imperial) for detecting this and to Steve Plimpton (Sandia) for suggesting the remedy that is now implemented (in LAMMPS as well as LIGGGHTS®).
+ Added new pair styles pair/gran/hooke/history/stiffness, pair/gran/hertz/history/stiffness and according wall styles. These are largely identical to the LAMMPS granular pair styles / wall styles. Additionally, they support the full additional LIGGGHTS® functionality (triangular walls, multiple materials, cohesion model, rolling friction model,...). Thanks again to George Marketos (Imperial) for the suggestion. This is in beta stage, so there is no doc yet, so people are invited to have a look at the code (e.g. pair_gran_hooke_history_simple.h/cpp) and try it.
+ Added a new compute rdf/gran to evaluate the radial distribution function for granular systems. This may be useful for the development of continuum models based on the granular gas theory
It is basically equivalent to the existing compute rdf, but instead of evaluation up to the force cutoff, it evaluates up to force cutoff plus skin, which makes for sense for granular systems.
+ Fixed a rare bug where wall/gran fixes did not tidy up correctly upon unfixing. Thanks to Xuetao Wang (Züblin AG) for sending an offending script
+ Fixed a small issue in the trianlge neighbor list routine. It did not result in a wrong calculation, but it issued a warning message that was not correct. Thanks to Francois Guill-Guillard for digging into the code and seeing this.

LIGGGHTS® 1.2.7 (released 22 March 2011)
UPGRADE FROM 1.2.6 RECOMMENDED
+ Fixed a bug related to triangular wall neighborlist. This bug unfortunately came in when fixing a small issue.
For all users of triangular walls it is heavily recommended to update to 1.2.7!
+ Improved accuracy of the triangle wall representation in case the triangle area is smaller than particle size.
Now the contact forces are scaled down with the ratio of triangle size to the theoretical contact area size based on the overlap
+ Fixed a bug related to restart with granular walls, that made LIGGGHTS® crash under certain circumstances.
Thanks to Colman-Patrick Caroll (BASF) and user Somnath (U Texas) for calling attention to this
+ Added the possibility to dump the mesh velocity for moving mesh and the conveyor model
+ Added an additional timestep criterion to fix check/timestep/gran: Now you will be also warned if particles travel further than skin in a timestep, which would result in incorrect interaction calculation

LIGGGHTS® 1.2.6 beta (released 18 February 2011)

+ Fixed a bug in the contact history. Many thanks to Jean-Francois Jerier (EPFL) for tracking that down.
+ Fixed a bug in the triangle neighbor list calculation.
+ Fixed a bug in the create_atoms single command. Thanks to Christoph Goniva (JKU) for running into that issue.

LIGGGHTS® 1.2.5 beta (released 2 February 2011)

Fixed two restart issues:
+ One of them was discovered by Jean-Francois Jerier (EPFL), and happend for shrink-wrapped BCs were used together with neigh_modify_once yes
+ The other was discovered by Moritz Höfert (BASF), and prevented correct restart of heat transfer simulations
Both should be fixed now

LIGGGHTS® 1.2.4 beta (released 1 February 2011)

+ Updated the new shear history formulation - now dissipation comes from either sliding or damping, not from both.
The way it was done in 1.2.3 beta produced unreasonable results, e.g. for the conveyor test case
+ Fixed a bug with deleting a mesh wall during the simulation - thanks to Diego Peinado (Intrame) for catching this.
+ Fixed a typo bug in the new contact history. Thanks to Jean Francois Jerier (EPFL) for finding this tricky issue.
+ Fixed a restart bug - thanks to Ram Chand for reporting this
+ Fixed a bug in the communication of pair gran and compute pair/gran/local
+ Corrected the doc regarding rollingfriction - thanks to Christian Grohs (Plansee) for finding this

IMPORTANT NOTE: This version is in beta state.

LIGGGHTS® 1.2.3 beta (released 23 December 2010)

Introducing and update and a couple of fixes:
+ Added a triaxial compression example contributed by Tong Shan (Hong Kong University of Science & Technology).
+ Fixed a bug in the wear model, not correctly writing out the data to VTK in parallel. Thanks to Alexander Polson (Mecalysis) for seeing this.
+ Fixed a bug related to simulation restart with granular walls. Thanks to Ram Chand.
+ Fixed a fix rigid issue. Thanks to thanks to Thomas Wagner (TU Freiberg) and Colman-Patrick Carroll (BASF) for reporting this.
+ Updated the doc of pair/gran and fix wall/gran to include info on how to use the new rolling friction model
+ Updated the doc of fix check/timestep/gran (fixed a typo). Thanks to to Fabian Uhlig (TU Freiburg / U Twente) for noting this
+ Updated the doc of compute pair/gran/local (fixed a typo).

IMPORTANT NOTE: This version is in beta state.

LIGGGHTS® 1.2.2 beta (released 1 December 2010)

+ Fixed a bug related to simulation restart
+ Added git support for the 1.2 beta branch, via a new repositoty 'LIGGGHTS®dev'. For details please see git instruction in the download section

IMPORTANT NOTE: This version is explicity released in beta state.

LIGGGHTS® 1.2.1 beta (released 29 November 2010)
New features are introduced in 1.2.1:

Added the following commands:

1) Compute coord/gran to compute coordination numbers for granular simulations

2) Compute pair/gran/local to compute pairwise properties for granular simulations such as interparticle contact forces and shear history (which can be output using with the dump local command)

IMPORTANT NOTE: This version is explicity released in beta state.

IMPORTANT NOTE: 1.2.1 beta is currently only available as tarball, the git repository still contains the latest stable
version 1.1.10

LIGGGHTS® 1.2 beta (released 22. November 2010)
New features are introduced in 1.2:

1)Introduced some changes "under the hood" regarding the granular pair and wall styles.
Classes containing core functionality and modelsare now seperated, so the code should now be structured
in a way that allows easier implementation of modified contact laws

2) Changed the way how some fixes request other fixes that store data for them. The old way of doing this
would have sooner or later led to troubles. Many thanks to Steve Plimpton (main LAMMPS developer, Sandia NL)
for helpful discussion and the suggestion how to resolve the issue.

3) Introduced the concept of rolling friction into the existing pair styles. Thanks to Chris Stoltz (Procter & Gamble)
for sending the code for this.

4) Extended the shear history implementation towards a general concept of contact history. This now allows the
implementation of interesting pair styles like incremental history force models. Thanks to Sam Bateman
(US Navy Research Lab) for contributions this feature is based on. A class to dump the contact history is being
worked on.

5) Modified the way shear history is implemented for the existing pair gran styles. This should help picturing
segregation better. Thanks to George Marketos (Imperial) for helpful discussions on this.

6) Modified the way torques on particles are applied in pair and wall styles. The contact radius is now used instead
of the geometric radius. Thanks to Fabian Uhlig (TU Freiburg / U Twente) for suggesting this and to George Marketos
(Imperial) again for helpful discussions.

IMPORTANT NOTE: This version is explicity released in beta state. Selected test harness results vs. version 1.1.10 can
be found at the Quality Assurance page. Please be aware of that in these tutorial cases, the overlap is quite high.
In realistic simulations, the difference between the models will be less pronounced.

IMPORTANT NOTE: 1.2 beta is currently only available as tarball, the git repository still contains the latest stable
version 1.1.10

LIGGGHTS® 1.1.10 (released 8. November 2010)

Features added in 1.1.10:
1) Added a simple model for qualitative prediction of wear (a modified version of the Finnie model). See doc for fix mesh/gran/stressanalysis and the seperate finnie-wear.pdf for details. Thanks to Georg Aichinger (Siemens VAI) for motivating addition of add such a functionality
2) Added a "chute_wear" example to show this new functionality
3) Extended the fix adapt in order to be able to use more than one of them together in a simulation - thanks for user Xinze Lee for catching this
4) Added an example for a simple compression test. Many thanks to Jean-Francois Jerier (Ecole Polytechnique Fédéral de Lausanne) for contrbuting this.

Fixed issues in 1.1.10:

1) Fixed an issue with fix pour/dev where the particle distribution was not correctly governed for a very low number of inserted particles - thanks to Christian Grohs (Plansee) for catching this.
2) Fixed a bug in the assignment of material properties - thanks again to Christian Grohs (Plansee).
3) Fixed a tolerance issue related to wall velocity calculation with the moving mesh feature. Thanks to George Marketos (Imperial College) for seing this.

BACKWARD COMPATIBILITY: you will not be able to use restart files from earlier versions with 1.1.10

LIGGGHTS® 1.1.9 (released 19. October 2010)

1) Added the possibility to include particle-wall heat conduction with constant wall temperatures. This is both possible for both generic and CAD walls, see fix wall/gran and fix mesh/gran for details. Also fixed an issue with the heat transfer model, where particle based heat sources/sinks could have been double counted under certain circumstances
2) George Marketos (Imperial College) sent some code to improve 2d simulations. Now, particle inertia in 2d simulations is calculated for a disk rather than for a sphere
3) Improved the restart capability of fix pour/dev: It is now possible to restart a fix pour/dev that already completed insertion in the old run by specifying a higher number of particles to insert in restart script. Thanks to Ram Chand (Beijing Institute of Technology) for catching this
4) One more update to the mesh/gran command that should improve the correctness of damping for particle-triangle corner contacts
5) It is now possible to use the per-particle radius as an input for the variable command
6) Fixed an issue where variable particle mass could have led to problems in conjunction with periodic boundary conditions. Also, variable particle density is now accounted for in parallel
7) Fixed a typo in the doc for pair gran - thanks to Fabian Uhlig (TU Freiburg / U Twente) for catching this

LIGGGHTS® 1.1.8 (released 6. October 2010)

1) Added an important update to the CAD wall handling improving the correctness of particle-wall damping for edge and corner contacts. Many thanks to Chris Stoltz (Procter & Gamble USA) for the helpful testcases and discussions
2) Fixed a general issue with fixes defined after previous runs where some properties were not correctly initialized. Thanks to Carsten Schilde (TU Braunschweig) for running into this problem
3) Fixed an issue with LIGGGHTS® not accepting material parameters for atom types that are inserted in later runs. Thanks to Josef Fuchs (University of Leoben) for providing a script that suffered from this problem
4) Added an error check to the fix pour/dev command where under special circumstances a segfault could occur if the given massflowrate was too low. Thanks to Michael Aigner (JKU Linz) for providing a test script
5) Added a check for fix wall/gran to yield an error if the user would specify lo and hi values for the wall incorrectly.
6) Adapted the fix move/mesh/gran command so that the last argument (which was actually no longer used) can be omitted
7) Fixed a memory leak related to the heat transfer model. Thanks to Colman-Patrick Carroll (BASF) for a script revealing this
8) Fixed some LAMMPS issues with fix move and create_atoms that arose if these commands were used inbetween two runs
9) Added a possibility to use the set command to set omega of particles

LIGGGHTS® 1.1.7 (released 17. September 2010)

1) Fixed a LAMMPS bug where processor communication was not correct when using fix adapt under certain circumstances. Thanks to Steve Plimpton (Sandia) for providing a solution to this
2) Fixed a typo in calculating the particle contact area - this impacts on both the heat transfer and the cohesion model
3) George Marketos (Imperial) added a feature to add omega as a variable, as can be done with x,v,etc. Thanks for submitting this.
4) Fixed a bug where shear wall stress was incorrectly dumped to VTK
5) Fixed a quite tricky issue with the moving mesh feature where viscous damping was not calculated correctly for some cases. Many thanks to Chris Stoltz (Procter & Gamble) for doing extensive testing and help finding this bug.
6) This bug fix also motivated to change the wiggle oscillation command for granular walls from 1-cos to sin to be able to compare native oscillatory walls with oscillations done with fix move/mesh/gran
7) Also improved a detail with shear history on mesh surfaces.
8) Fixed typos in some formulas of the pair_gran.html doc (the actual implementation was correct). Thanks again to Chris Stoltz (P&G) for finding this.
9) Changed the 'cohesion' example in order to make a more visible comparison between cohesion and no cohesion - thanks to Quinn Reynolds (Mintek) for suggesting this

LIGGGHTS® 1.1.6 (released 11. August 2010)

Introducing a couple of book-keeping patches, so the code should be much cleaner now.
A major re-write of the LIGGGHTS® granular pair styles to allow for greater implementation flexibility is underway, but please be patient.

Fixed an issue where creating atoms with a restart pending lead to a crash - thanks to Thomas Wagner (TU Freiberg) for the bug report

Fixed two issues with the cohesion model - there was a bug in calculating the contact area. Also added a check to ensure the model is only used for 3d simulation. Thanks to George Marketos (Imperial College London) for calling attention to this.

As of version 1.1.6 a quality assurance check will be done by running automated testloops comparing various thermo outputs (e.g. energy) over time that are generated by different versions. Please see this page for details.

LIGGGHTS® 1.1.5 (released 8. July 2010)

The third patch for today, this time 2 minor issues:
Added a warning if the user wants to dump stress on wall, but forgot to use the appropriate fix/mesh/gran/stressanalysis. Also fixed a numerical round-off issue for CAD imported walls that could arise under unlikely conditions.
Thanks to Jean-Francois Jerier (Ecole Polytechnique Fédéral de Lausanne) for reporting these issues

LIGGGHTS® 1.1.4 (released 8. July 2010)

One more patch that ought to be included in 1.1 was unintentionally left out - This patch fixes it.
Sorry for the inconvenience - this version should be clean now.
Thanks to Christian-Andreas Winkler (BASF) for running into this problem

LIGGGHTS® 1.1.3 (released 8. July 2010)

Some piece of code related to the new "riggle" moving mesh functionality was unintentionally left out - This patch fixes it. Thanks to Chris Stoltz (Procter & Gamble) for noting this.

LIGGGHTS® 1.1.2 (released 5. July 2010)

Did some optimisation for pair and wall forces (slight performance improvements)
Fixed a bug in the stress analysis feature

LIGGGHTS® 1.1.1 (released 30. June 2010)
BACKWARD COMPATIBILITY:
You will be able to use restart files from 1.1
You may not be able to use restart files from versions older than 1.1

Version 1.1.1 changes the way the effective Shear modulus for two particles in contact (G*) is calculated. See doc/pair_gran.html for details on the formula.

The old calculation was based on Chemical Engineering Science 59 (2004) 525 – 541, whereas 1.1.1 uses the calculation as given in Comput. Methods Appl. Mech. Engrg. 197 (2007) 80–94.

Tests indicate that the second version is more suitable. Many thanks to Diego Peinado (Intrame AS) for discussing this detail!

LIGGGHTS® 1.1 (released 29. June 2010)
BACKWARD COMPATIBILITY: You may not be able to use restart files from previous version with 1.1!

Version 1.1 contains the following new features:

1) Added a feature to calculate stresses on CAD imported meshes to the main distribution. See doc for fix mesh/gran/stressanalysis for details.
2) Added a dump mesh command to post-process these stresses with paraview
3) Dump stl and dump mesh now have an option to choose which meshes/stresses should be dumped to STL/VTK
4) Added a new fix check/timestep/gran to estimate the accuracy of the chosen time-step by means of Rayleigh and Hertz time-step criteria.
5) Added a capability to generate particle packings by initially inserting particles with a low volume fraction with a new fix pour/dev/packing. Particles can then be successively grown over time until the desired particle size distribution and volume fraction is reached. This is achieved by a 'fix adapt' command which can modify properties over time in conjunction with variable scripting (see example 'packing')
6) Added a capability to exempt certain regions from the insertion using the fix pour/dev or the fix pour/dev/packing command.
7) Added a 'riggle' option to the fix move/mesh/gran, which gives the possibility to impose oscillatory rotational motion to a CAD imported mesh. Thanks to Christopher Stoltz (Procter & Gamble USA) for suggesting this

Version 1.1 contains the following improvements:

1) Performance optimizations done for CAD imported meshes, example meshGran now runs over 50% faster
2) Also added a size ratio and coplanarity check for triangular meshes which should improve result quality
3) Some data compaction has been achieved, so restart files should be smaller now for cases with CAD imported geometry
4) Improved the 'massflowrate' option for fix pour/dev command. The inserted mass is now tracking the expectancy value more accurately.

Version 1.1 contains the following bug fixes:

1) Fixed a cut-off issue that occured for some poly-disperse simulations. Thanks to Diego Peinado (Intrame AS) for finding this tricky bug and to Steve Plimpton (main LAMMPS developer, Sandia NL, USA) for providing a solution to this issue.
2) Fixed a minor issue related to computing particle-triangle interactions - thanks to Martin Schilling (BASF) for providing a test case
3) Fixed 2 issues for cases when the particle-triangle contact list was dynamically grown during simulation - thanks to Stefan Pirker (Johannes Kepler University Linz) for providing a test script that showed the problem
4) Fixed a bug in fix rigid - thanks to Christian Grohs (Planseee SE) for reporting this

Version 1.1 comprises the following changes in examples:

1) Added a new example 'packing' to highlight the packing feature (see above).
2) Added time-step checking to examples 'meshGran' and 'conveyor'

LIGGGHTS® 1.0.3 (released 03. June 2010)

BACKWARD COMPATIBILITY: You will not be able to use restart files from previous version with 1.0.3!

Version 1.0.3 comprises a new 'conveyor belt' wall model added to the fix mesh/gran command.
The mesh/gran can be assigned a constant velocity, but is not moving itself. This can be used to mimic a conveyor belt for meshes that are flat or curved in one direction . Also added an example for the conveyor belt model

Version 1.0.3 comprises the following new features for particle insertion.
A new fix pour/dev command has been added which will sooner or later replace fix pour. It is quite similar to fix pour, but some major re-works allow the following new capabilities:
1) With fix pour/dev, you can now define conditions for mass-flow rate and total insertion mass.
2) The insertion height probability has been re-worked to be in accordance with the velocity distribution
3) Moreover, fix pour/dev has now restart capability, which means that it is not necessary to change the fix parameters in a restarted simulation run.
4) The insertion now takes place in a way that allows higher volume fractions than the old fix pour command for poly-disperse simulations.
5) The fix pour/dev command can now takes discrete particle distributions, defined by a new fix particledistribution/discrete. These distributions are based on mass % (as opposed to the number density based distributions in the old fix pour, which are less common). The fix particledistribution/discrete can be fed with one or more fixes of type particletemplate/sphere (also new). Moreover, continuous distributions of diameter and density within a particle template distribution are possible.
Also added an example to show how 'fix pour/dev' works.

Version 1.0.3 contains the following bug fixes:
Fixed an issue with growing shear history
Fixed a bug that obstructed correct particle shear history simulation restart in some cases when CAD-imported geometry was used
Fixed a bug that for some cases CAD imported geometry was not recognized during the first run - thanks to Martin Köster (University of Düsseldorf)
Some debugging code was left in 1.0.2 so that some simulations crashed at time-step 320000. This issue is now fixed.
Fixed 2 issues with importing CAD geometry: STL file face normal vectors are now automatically normalized.
Fixed an issue with rotating imported geometry around the z-axis - thanks to Christian Grohs (Plansee SE).

LIGGGHTS® 1.0.2 (released 19. May 2010)
Improvements vs. LIGGGHTS® 1.0.1
+ Fixed a few issues with CAD geometry handling
+ Fixed a LAMMPS bug and a LIGGGHTS® issue related to the moving mesh feature

LIGGGHTS® 1.0.1 (released 5. May 2010)
Improvements vs. LIGGGHTS® 1.0
+ Speed improvement:
.---E.g. example meshGran runs 5% faster;
.---Other benchmarks indicate up to 30% speed-up for dense systems
+ Fixed a minor issue with exporting STL geometry during simulation run
+ Fixed an issue with handling of CAD walls

LIGGGHTS® 1.0 (released 19. April 2010)
New features vs. LIGGGHTS® beta 1:
+ Restart capability for all new features
+ Better documentation
+ Include LAMMPS patches and bug fixes since 24-Jan-2010
+ Improved "fix pour" command

Bug fixes for LIGGGHTS® beta 1:
+ Fixed a LAMMPS bug related to the 'unfix' command (see 10 March 2010 LAMMPS bug fix)
+ Fixed a bug in the moving mesh capability
+ Fixed a bug when reading CAD geometry data

LIGGGHTS® beta 1 (released 19. Feb. 2010)
Available for download in the Downloads section
Based on 24-Jan-2010 Version of LAMMPS

New features vs. LAMMPS:
+ Import and handling of complex wall geometries from CAD
+ A moving mesh feature to account for moving geometry
+ Re-write of contact force formulation
+ Cohesion forces are now possible
+ Heat conduction between particles in contact

The following items are being worked on (if you are interested, please contact us)
+ Improved Multi-Sphere feature
+ A 6 degree of freedom solver for rigid bodies
+ Bonded particle models
+ Coupling to CFD
+ Particle representation by triangular meshes