lammps

Changes since the release 3 March 2020:

• Update USER-COLVARS package to version 2020-02-27 (Giacomo Fiorin, Temple U) PR #1909
• Move fix bond/react into its own package, USER-REACTION (Jake Gissinger CU Boulder) PR #1822
• Update the molecule command to add support for setting molecule IDs and grouping atoms into fragments (Jake Gissinger CU Boulder) PR #1913
• Optimizations for the KOKKOS version of SNAP (Evan Weinberg, NVIDIA) PR #1922
• New pair styles with smeared coulomb charge interactions (coul/slater) (Evangelos Voyiatzis,Royal DSM) PR #1883
• New fix numdiff command to compute derivatives of per-atom data numerically from finite differences (Charles Sievers, UC Davis) PR #1891
• Require C++11 and CMake 3.10, simplify some CMake settings and remove checks for pre-C++11 compilers (Axel Kohlmeyer and Giacomo Fiorin, Temple U) PR #1915, PR #1930
• Improved CMake support for the "ninja" build tool (Christoph Junghans, LANL) PR #1940
• Improved CMake support for cross-compiling LAMMPS for Windows on Linux (Axel Kohlmeyer, Temple U) PR #1949 , PR #1950
• Avoid a clash between the KOKKOS and GPU package with -DCUDPP_OPT=on (default setting) (Axel Kohlmeyer, Temple U) PR #1918
• Changes to the binary restart file format, so that incomplete restart files can be detected. Also an option to fix halt to monitor file systems for available free space (Axel Kohlmeyer, Temple U) PR #1862
• Validate coordinate data when reading data files (Axel Kohlmeyer, Temple U) PR #1945
• Rename package USER-MESO to USER-MESODPD (Axel Kohlmeyer, Temple U) PR #1924
• Major documentation cleanup that removes artifacts from txt2rst processing, updates many external URLs, improves syntax highlighting settings, and uses more consistent typesetting for examples and throughout the text. USER-ATC documentation is now in rst format, too. (Axel Kohlmeyer and Richard Berger, Temple U) PR #1926, PR #1932, PR #1936, PR #1938, PR #1939 , PR #1948, PR #1945
• Various tweaks, bugfixes, and small updates (multiple authors) PR #1918 , PR #1923 , PR #1920 , PR #1928, PR #1929 , PR #1930, PR #1940, PR #1945

Backward compatibility note:

• As of this release LAMMPS requires a C++11 compatible compiler
• Building LAMMPS with CMake now requires CMake version 3.10 or later
• The USER-MESO package was renamed to USER-MESODPD. All styles included in the package have retained their names
• read_data will now stop with an error on invalid coordinates or image flag data (instead of using a 0) and on non-zero image flags for z in 2d-systems

IMPORTANT NOTE:
This is the last LAMMPS release that does NOT require a C++11 compiler for the core code and CMake version 3.10 or newer. The next patches and stable releases will require a C++11 compatible C++ compiler to compile all of LAMMPS, not just for some of the optional packages. For some older compilers (e.g. GCC 4.8.x on CentOS 7.x) a flag to enable C++11 support may be required (e.g. -std=c++11 or -std=gnu++11 in the case of GCC 4.8.x).

Below is a list of major changes since the last stable release 7 August 2019.

General changes

• The documentation file format has been changed from the custom txt2html markup to using reStructuredText (rst). This provides several advantages to have a better looking and more functional documentation. As a first step, all externally rendered equations have been converted to embedded math and all raw HTML markup converted. Most boxes with examples, will use syntax highlighting. Further improvements of typesetting and layout are forthcoming. Additional consistency checks are now applied when processing the documentation.
• Many file read operations are now monitored for read errors and thus allowing to detect damaged or incomplete files like potential parameter files and restart files
• Updates and improvements to the CMake build system to more closely match all the functionality provided by the conventional make based build system
• Support for threaded FFTs from FFTW and MKL, support for cuFFT and KISSFFT on the GPU with KOKKOS
• Expose neighbor lists to the library interface and the Python wrapper
• Documentation of installing LAMMPS via conda
• Many small improvements or bugfixes

Updates and new commands or styles

• new dump_modify options to embed ITEM: UNITS and ITEM: TIME with information about units and accumulated simulation time
• new pair styles cosine/squared, local/density, mesocnt
• new fix styles wall/reflect/stochastic, propel/self, 'npt/cauchy`
• new compute gyration/shape/chunk
• new compute hma for fast, high-precision computation of certain thermodynamic properties of solids
• new compute style centroid/stress/atom for computing per atom stress in a way allowing for more accurate heat flux computations with interactions involving more than two atoms
• new third_order command, a companion command to dynamical_matrix, for computing the third order force tensor from finite differences
• new options ratio and subset for create_atoms and similarly new options type/ratio and type/subset to the set command
• new tool for post-processing parallel tempering trajectories
• multiple updates, bugfixes, and improvements to fix bond/react
• improved implementation of minimization style fire
• support to run NWChem along with LAMMPS in client/server mode

Updates for packages

• KOKKOS now supports minimization (limited to minimization styles cg and quadratic), updates to the KOKKOS library, general improvements and bugfixes, more styles ported to KOKKOS
• several new SNAP potentials
• tweaks to USER-INTEL to compile correctly with too aggressively optimizing compilers, corrections and improvements
• bugfixes and improvements for the GRANULAR package
• several new SNAP potentials and a new compute snap for training SNAP potentials
• bugfixes and improvements to the KIM package, new command kim_param for accessing KIM model parameters
• updates to the USER-COLVARS package with updates to the included Lepton library
• updates to the USER-PLUMED package with bugfixes and support for recent releases
• updates to the USER-CGDNA package including a new coarse grained RNA model (oxRNA2)
• improvements and updates to the SPIN package, L-BFGS minimizer added
• updates, extensions, and improvements to the USER-MEAMC package
• memory leaks fixed when using styles from the GPU with multiple run commands, new pair style lj/cut/tip4p/long/gpu
• updates to fix rigid variants to add gravity to rigid objects with overlapping (granular) particles
• support for rerun and read_dump with USER-ADIOS

Backward compatibility notices

• building the USER-COLVARS package with all features enabled now requires a C++-11 compatible compiler
• due to changes in the random number generator initialization, results when using fix pour, fix deposit, fix evaporate or create_atoms random will not match with previous versions of LAMMPS. see PR #1569
• the gjf keyword for fix langevin no longer supports the option yes. Instead you have to use either vhalf or vfull
• the FIRE minimizer has been replaced with an improved implementation. The previous implementation is available as fire/old.
• new documentation has to be written in reStructuredText format

This patch contains only a few updates to the documentation ( PR #1910 ) and a CMake tweak ( PR #1912 ) to make it a stable release.

This is the second release candidate for the Winter 2019/2020 stable release of LAMMPS.

Changes since the patch release 18 February 2020:

• Further refactoring of the manual with significant corrections and updates to the introductory, build, and package sections. Also, all remaining mathematical expressions typeset in LaTeX and included as pictures are now included as embedded math and will be rendered as needed during the creation of the documentation in either HTML or PDF or ePUB format. The embedded math in the HTML pages can now be viewed without internet access. Fixes to the embedded math to avoid errors with ePUB format output. Various updates and fixes to build procedures for USER-QMMM. (Axel Kohlmeyer and Richard Berger, Temple U) PR #1888 and PR #1903
• Updates to QM/MM compilation with Quantum Espresso version 6.3 and higher (Mariella Ippolito, CINECA) PR #1886
• bugfix for a race condition in temper commands in combination with timer timeout (Stephen Farr, University of Cambridge) PR #1892
• remove conditional C++11 check for stable release. will become an unconditional check after the release. (Axel Kohlmeyer, Temple U) PR #1885
• avoid divergence and energy spikes in in pair style thole of the USER-DRUDE package (Agilio Padua and Kateryna Goloviznina, ENS de Lyon) PR #1899
• new example in the COUPLE folder for dynamically loading LAMMPS as a plugin into a C code (Axel Kohlmeyer, Temple U) PR #1900
• several small bugfixes and updates (various authors) PR #1897 , PR #1893 , PR #1896 , PR #1901 , PR #1894 , PR #1905 , PR #1908

Backward compatibility note:

• No known backward compatibility issues

This is a release candidate for the Winter 2019/2020 stable release of LAMMPS.

Changes since the patch release 4 February 2020:

• GPU kernel optimizations for pair style lj/cut/tip4p/long/gpu (Vsevolod Nikolskiy, HSE Univ. Moscow) PR #1876
• Various small updates and bugfixes (multiple authors) PR #1875, PR #1877, PR #1882
• Update of USER-PLUMED to support Plumed 2.6 (Axel Kohlmeyer, Temple U) PR #1879
• Further refactoring of the manual by converting external math expressions to be embedded math in .rst files (Axel Kohlmeyer and Richard Berger, Temple U) PR #1869

Backward compatibility note:

• No known backward compatibility issues

Changes since the patch release 24 January 2020:

• Improved version of the FIRE minimizer option with better convergence and more options to adapt its behavior (Julien Guénolé, RWTH Aachen and Erik Bitzek, FAU Erlangen-Nuernberg) PR #1052
• Update Colvars library to version 2020-01-27 (Giacomo Fiorin and multiple Contributors to the Colvars project) PR #1859
• new SNAP potential parameter files for single element systems containing Cu, Ge, Li, Mo, Ni, or Si (Yunxing Zuo, UCSD) PR #1866
• Support threaded FFTs with MKL and refactoring of FFT support in KOKKOS (Axel Kohlmeyer, Temple U and Stan Moore, SNL) PR #1858
• remove all legacy documentation sources. From now on writing reStructuredText is the only option. The txt2rst tool remains available, though. (Axel Kohlmeyer, Temple U), PR #1855
• add support to run NWChem along with LAMMPS in client/server mode (Jeff Boschen, Iowa State) PR #1703
• various small changes, bugfixes and tweaks (multiple authors), PR #1857, PR #1860
• cleanup of the ReStructuredText source for the manual, convert more .tex file to embedded math, and corrections for issues due to inconsistencies in the conversion from the original txt2html markup or due to undetected errors in the original txt2html sources. (Axel Kohlmeyer, Temple U), PR #1861, PR #1864, PR #1867

Backward compatibility notice:

• the FIRE minimizer (min_style fire) has been replaced with an improved version. The original implementation remains available as fire/old.

Changes since the patch release 9 January 2020:

• new pair style mesocnt for mesoscopic modeling of (carbon) nanotubes (Philipp Kloza, U Cambridge), PR #1841
• new fix npt/cauchy for improved handling of variable cells with large deformation (Ronald Miller, Carlton U, Fabio Pavia, and Subrahmanyam Pattamatta, UMN) PR #1096
• new fix propel/self for self-propelling particles (Stefan Paquay, Brandeis U) PR #1847
• new fix wall/reflect/stochastic for multiple variants of stochastic reflecting walls (Quy-Dong To, Université Paris Est Marne la Vallée), PR #1691 , PR #1838
• added support for threaded FFTW3 in all versions of pppm and support for using cuFFT or KISSFFT on the GPU with KOKKOS when CUDA is enabled (Stan Moore (SNL), Sam Mish (U.C. Davis)), PR #1520
• new orient option to fix deposit (Tongtong Shen), PR #1685
• new quit option to message command for client/server coupling with the MESSAGE package (Steve Plimption, SNL), PR #1835
• new examples for using rerun and read_dump (Steve Plimpton, SNL) PR #1829
• new options ratio and subset for create_atoms plus type/ratio and type/subset for set to populate an exact ratio or exact number of lattice points of atoms when creating atoms or change the atom type with set. This also includes a significant performance improvement when creating a very large number of atoms with create_atoms (Jake Gissinger, UC Boulder and Steve Plimpton, SNL), PR #603
• add support for read_dump and rerun to USER-ADIOS package (Norbert Podhorszki, ORNL) PR #1820
• updated ILP potential files and new files for bulk systems, improve energy conservation of ILP pair styles with dangling bonds (Wengen Ouyang, Tel Aviv U) PR #1843
• disallow pair_modify pair special for styles from the GPU and USER-INTEL package (Axel Kohlmeyer, Temple U) PR #1834
• fix a logic error in reading potential files for pair style bop and updates to the potential files that were missing data. (Don Ward, SNL and Axel Kohlmeyer, Temple U) PR #1831
• added tool to check consistency of lists of command styles in documentation (Axel Kohlmeyer, Temple U) PR #1832
• various small code updates, improvements, and bug fixes (multiple authors) PR #1811, PR #1830, PR #1833, PR #1836, PR #1837, PR #1845, PR #1848, PR #1850, PR #1852, PR #1853, PR #1857
• various updates to documentation, examples, readme files (multiple authors) PR #1828, PR #1837

Backward compatibility notice:

• no known backward compatibility issues

Changes since the patch release 20 November 2019:

• More flexible handling of certificates in KIM package (Yaser Afshar, UMN) PR #1794
• Expose LAMMPS neighbor lists to the library interface and to the Python wrapper (Richard Berger, Temple U) PR #1674
• Addition of oxRNA2 model for RNA to the USER-CGDNA package (Oliver Henrich, Strathclyde U) PR #1772 , PR #1804 , PR #1812
• New compute snap command to support training SNAP package potentials (Aidan Thompson, SNL) PR #1779 , PR #1796
• New pair style lj/cut/tip4p/long/gpu for the GPU package (Vsevolod Nikolskiy, Vladimir Stegailov, Higher School of Economics, Moscow) PR #1776
• Build the documentation directly from .rst files. Various updates to the process and integration of formula as direct rendered graphics via MathJAX (Richard Berger, Temple U) PR #1786
• Updates, bugfixes, and new examples for the SPIN packages (Julien Tranchida, SNL) PR #1771
• Multiple updates and bugfixes to the USER-INTEL package (Mike Brown, Intel Corp. and Axel Kohlmeyer Temple U), PR #1813
• Detect parallel NetCDF when building with CMake (Johannes Hörmann, Uni Freiburg) PR #1819
• Updates and fixes to the GPU package including some memory leak fixes (Trung Nguyen, Northwestern U) PR #1752 , PR #1807
• Updates to the KOKKOS package (Stan Moore, SNL) PR #1803 , PR #1816
• Updates to fix rigid variants to support adding gravity to overlapping extended particles (Dan Bolinteanu and Steve Plimption, SNL) PR #1801 , PR #1823
• Updates, extensions, and improvements to the USER-MEAMC package (Sebastian Huetter, Otto van Guericke U and Sungkwang Mun, MS State U) PR #1798
• Updates and bugfixes to fix bond/react (Jake Gissinger, UC Boulder) PR #1795
• Various small bugfixes and improvements (multiple authors) PR #1784, PR #1787, PR #1792, PR #1802, PR #1806 , PR #1783 , PR #1824 , PR #1817 , PR #1825
• Fixes and updates to the documentation (multiple authors) PR #1789, PR #1787, PR #1800

No known backward compatibility issues.

Changes since the patch release 30 October 2019:

• new command kim_param to access KIM model parameters, (Yaser Afshar and the OpenKIM team at UMN), PR #1745
• new compute style gyration/shape/chunk to compute gyration tensor eigenvalues and shape parameters for chunks, (Evangelos Voiatzis, Royal DSM),PR #1739
• new compute style centroid/stress/atom plus support for collecting the per atom stress in a way that leads to more accurate heat flux computations (Donatas Surblys, Tohoku University), PR #1704, PR #1782
• document installation of pre-compiled LAMMPS binaries via conda (Ryan Elliot, UMN), PR #1768
• Convert documentation to use .rst files as source instead of .txt. Retain .txt and tool to convert to .rst for transition period. Remove legacy documentation tools and makefile targets. (Richard Berger, Temple U and others), PR #1643, PR #1761, PR #1762, PR #1765, PR #1766, PR #1775
• add a min and max limit to the number of particles inserted by fix gcmc (Jared Wood, U of Sydney), PR #1764
• Various small fixes and code improvements (multiple authors), PR #1756, PR #1758, PR #1759, PR #1755, PR #1766, PR #1763, PR #1767, PR #1740, PR #1769

Backward compatibility notice:

• Documentation for the LAMMPS manual is now taken directly from .rst files and while the txt2rst converter tool is still available, .txt files will be incrementally removed as changes to the corresponding .rst files are applied that could not have done automatically via the txt2rst converter tool. Changes to the .rst files directly are encouraged and once those are done, the corresponding .txt (if it still exists) will have to be removed as those changes will not be backported.

Changes since patch release 19 September 2019:

• new third_order command, a companion command to dynamical_matrix, that calculates the third order force constant tensor from finite differences, (Charles Sievers, UC Davis), PR #1690
• add angle and Arrhenius constraints to fix bond/react, (Jake Gissinger, UC Boulder), PR #1697, PR #1713, PR #1734
• support quasi-unique base pairing in USER-CGDNA (Oliver Henrich, U of Strathclyde), PR #1730
• update to the Lepton library in the USER-COLVARS package, (Giacomo Fiorin, Temple U), PR #1724
• updates to the gjf option in fix langevin to give access to either half-step or full-step velocities, (Charles Sievers, Niels Gronbech-Jensen, UC Davis), PR #1673, PR #1707, PR #1725
• add missing option to add single impropers with the create_bonds command (Thomas Farmer, ISIS Neutron and Muon Source), PR #1706
• add max_norm option to all minimizers and conjugate gradient and L-BFGS minimizers to the SPIN package. (Aleksei Ivanov, U of Iceland, Julien Tranchida, SNL), PR #1598
• bugfix for pppm/tip4p when using non-orthogonal simulation cells, PR #1719
• bugfix for restarting SPIN package runs and documentation clarification (Julien Tranchida, SNL), PR #1692
• multiple bugfixes and documentation corrections for pair style granular (Jibril B. Coulibaly, Northwestern U), PR #1631
• avoid correlations between multiple runs with random seeds incremented by 1 when using random placement of atoms (Jun-Chieh Wang (Applied Materials), PR #1569
• fix a memory leak in the GPU package when doing multiple run commands in the same input, (Trung Nguyen, Northwestern U), PR #1735
• Optimizations for using SNAP with KOKKOS (Stan Moore, SNL and Rahul Gayatri, NERSC), PR #1693
• new examples folder "UNITS" with examples highlighting how to run inputs for the same system with different unit sets and convert output between different them (Steve Plimpton, SNL), PR #1709
• capture return value of calls to fread by using a wrapper, utils::sfread() and thus enable error reporting on reading restarts, (Axel Kohlmeyer, Temple U/ICTP) PR #1731
• A couple of singularity definition files to build container images for compiling/testing/debugging LAMMPS on different Linux variants, PR #1726, PR #1738
• various small changes, updates, and bugfixes (multiple authors), PR #1701, PR #1699, PR #1716, PR #1714, PR #1727, PR #1728, PR #1733, PR #1732, PR #1746, PR #1750, PR #1744, PR #1751

Backward compatibility notice:

• The gjf keyword for the fix langevin no longer supports yes as an option. Instead you have to use either vhalf or vfull.
• due to changes in the random number generator initializations, results when using fix pour, fix deposit or fix evaporate and when using create_atoms random will not match with previous versions of LAMMPS. see PR #1569
• Building the Colvars library with all features included now requires a C++11 compiler (see http://colvars.github.io/README-c++11.html).

Changes since stable release 7 August 2019:

• new pair style cosine/squared, (Eugen Rozic, UCL), PR #1626
• new pair style local/density for a generalized density dependent potential, (Tanmoy Sanyal, UCSB), PR #1666
• support for minimization with KOKKOS (min_style cg with quadratic line search only), (Stan Moore, SNL) PR #1650
• support for dihedral style harmonic in KOKKOS, (Stan Moore, SNL) PR #1669
• new compute hma for faster or higher precision computation of certain thermodynamic properties of solids, (A Schultz, A Purohit, D Kofke, U Buffalo), PR #1503
• add support for setting fix external callbacks to the library interface with python interface included (Richard Berger, Temple U), PR #1641
• add support for Morse potential to wall fixes (Eugen Rozic, UCL and Axel Kohlmeyer, Temple U), PR #1628, PR #1640
• new tool to post process parallel tempering trajectories, (Tanmoy Sanyal, UCSB), PR #1660
• small updates to the KIM package, (Ryan Elliot, UMN), PR #1625, PR #1632, PR #1638
• more cmake tweaks, (Christoph Junghans, LANL, Richard Berger, Temple U), PR #1633, PR #1644, PR #1634
• granular package bug fixes (Dan Bolinteanu, SNL), PR #1637
• extend compute bond/local to provide access to force components, not just the total force, (Edward Smith, U of Brunel), PR #1667
• bugfix and refactoring of handling of custom computed properties in compute pair/local for hybrid pair styles, (Axel Kohlmeyer, Temple U), PR #1662
• extend compute orientorder/atom to optionally output third order invariants, (Aidan Thompson, SNL), PR #1648, PR #1677
• new SNAP potential for W-Be systems, (Mary Alice Cusentino, SNL), PR #1686
• add dump_modify units yes option to have an 'ITEM: UNITS' entry added at the beginning of LAMMPS native text mode dump files. (Axel Kohlmeyer, Temple U), PR #1672
• small bugfix for fix bond/react and documentation updates, (Jake Gissinger, UC Boulder), PR #1645
• more detailed error messages when using known fix, compute, or variable IDs in variable expressions, (Axel Kohlmeyer, Temple U) PR #1621
• document example of using compute chunk/spread/atom for binning per-chunk computed data, (Steve Plimpton, SNL) PR #1642
• updates and additions to machine specific makefiles for the conventional build system. PR #1654, PR #1661, PR #1653
• tweaks to USER-INTEL code and Makefiles for improved performance and to work around issues with some Intel compilers, (Mike Brown, Intel Corp.) PR #1683
• support direct GPU communication for HPE MPI in KOKKOS (Gabriele Jost, NASA Ames Lab, Stan Moore, SNL), PR #1668
• include preprocessor test to detect non-C++11 compilers with support for define and CMake option to disable it. (Richard Berger, Axel Kohlmeyer, Temple U), PR #1684
• small documentation updates. PR #1664, PR #1681

Backward compatibility notice:
We are phasing out support for C++ compilers that do not support the C++11 standard. For older compilers (e.g. GCC 4.8.x) some flag to enable C++11 support may be required (e.g. -std=c++11 or -std=gnu++11 in the case of GCC 4.8.x).

This is a special edition stable release on occasion of the bi-annual LAMMPS users workshop on August 13-15 2019 in Albuquerque. Below is a list of major changes since the last stable release 5 June 2019:

General changes:

• further tweaks, refactoring, and improvements to building LAMMPS with CMake
• refactoring of header file handling for improved consistency and removal of cross dependencies
• refactoring of include keyword processing to lift some limitations and to enable KIM simulator models (see below)

No new packages

Updates for packages:

• major update of the KIM package to use KIM-API v2.1.x with support for Simulator models and new commands to generate variables for transparent unit conversion in input files
• optimizations and corrections to the KOKKOS packages, especially when used with CUDA, update of KOKKOS library included in LAMMPS to version 2.9.0
• updates and bugfixes of the USER-COLVARS package bringing it up to version 2019-08-05
• new PPPM and Ewald solvers for electric dipoles and nuclear spins
• optimizations and improvements to pair style and computes in the SNAP package
• updates, bugfixes and portability improvements for USER-PLUMED package (supports plumed 2.5.2)
• update to USER-CGDNA package in preparation of supporting an RNA coarse grain force field
• new keyword hybrid/pair for compute pressure
• new compute gyration/shape to compute eigenvalues and shape parameters from gyration radius tensor
• new compute momentum command

In addition there are many small bug fixes, corrections for memory leaks, and memory management inconsistencies, and other general improvements.

Backward compatibility notices:

• SNAP parameter files that contain the diagonalstyle keyword will generate an error. Solution is to remove the diagonalstyle keyword
• changes in the pair_coeff syntax breaks backward compatibility with inputs using USER-CGDNA. pair style oxdna2/stk has been removed and aliased to oxdna/stk
• restart files using bonded hybrid styles from older LAMMPS versions will not be readable
• the KOKKOS package command keyword gpu/direct was renamed to cuda/aware

This patch has a couple of last minute changes and bugfixes included after the 6 August 2019 patch release to make the stable version. It includes the pull requests #1620 and #1622

This patch has only minor corrections and documentation updates to update the 2 August 2019 patch release to a stable version. In particular:

• Update of the CMake documentation and some information about developer features available through CMake, Richard Berger (Temple U) PR #1613, PR #1617
• Some tweaks to CMake and USER-INTEL code to improve Linux to Windows cross-compilation, Axel Kohlmeyer (Temple U) PR #1616
• bugfix which updates the USER-COLVARS package to version 2019-08-05, Giacomo Fiorin (Temple U / NIH) PR #1618
• fix for docs on kim_query command, Ryan Elliot (UMN) PR #1619

Changes since patch release 19 July 2019:

• Implement native LAMMPS support for KIM simulator models and multiple improvements in the KIM package. Update to KIM API 2.1.x, Ryan Elliot, Ellad Tadmor (UMN) Axel Kohlmeyer (Temple U), PR #1440
• major cleanup of include file ordering and removal/cleanup of how includes are used in headers. Added a file describing the current policy. Axel Kohlmeyer (Temple U), PR #1542, plus related bugfixes PR #1599, PR #1605
• a more robust version of handling the atom exchange data buffer size, Steve Plimpton (SNL), PR #1556
• bugfix in restart handling of bond, angle, dihedral, improper hybrid styles to correctly restart tabulated styles. Axel Kohlmeyer (Temple U), PR #1584
• bugfix for communication issue in a couple of interlayer potentials. reorganization of the corresponding folder with examples, Wengen Ouyang (Tel Aviv U) PR #1606
• bugfixes for USER-COLVARS and update to version 2019-08-01, Giacomo Fiorin (NIH and Temple U) and several collaborators PR #1607
• multiple bugfixes for USER-INTEL package, Mike Brown (Intel Corp.) PR #1610, PR #1608
• multiple bugfixes for code in the KOKKOS package, Stan Moore (SNL), Alfredo Metere (LLNL), Denis Taniguchi (Newcastle U), PR #1580, PR #1587
• various CMake bugfixes and improvements, Axel Kohlmeyer (Temple U), PR #1588, PR #1592, PR #1601
• derive default communication cutoff from bond length based heuristics, if not pair style is active and the cutoff is not explicitly given. warn, if heuristic cutoff is larger than current or when explicitly set cutoff is overridden by pair style based cutoff (which is a hard requirement), Axel Kohlmeyer (Temple U) Steve Plimpton (SNL), PR #1581
• new keyword hybrid/pair for compute pressure, Alain Dequidt Julien Devemy (U Clermond-Ferrand) PR #1511
• new compute gyration/shape to compute eigenvalues and shape parameters from gyration radius tensor. Evangelos Voyiatzis (Royal DSM), PR #1583
• new compute momentum command, Rupert Nash (EPCC), PR #1529
• add support for timer timeout to "temper" commands, Axel Kohlmeyer (Temple U) PR #1604
• multiple minor bugfixes and code or documentation improvements, Adrian Diaz (U Florida) PR #1576 , Jake Gissinger (U Colorado), PR #1548, Michal Kansky (Jagiellonian U) PR #1512, Axel Kohlmeyer (Temple U) PR #1592 #1605, Stan Moore (SNL) PR #1597, Richard Berger (Temple U) PR #1599 Evangelos Voyiatzis (Royal DSM) PR #1611, E. Anne Gunn (Sheridan College) PR #1603

Backward compatibility notice:

• restart files using bonded hybrid styles from older LAMMPS versions will not be readable
• the KOKKOS package command keyword gpu/direct was renamed to cuda/aware

Changes since patch release 18 June 2019:

• optimized version of SNAP potential in KOKKOS package, Rahul Gayatri (NERSC), Stan Moore (SNL), Aidan Thompson (SNL), Sarah Anderson (Cray), Evan Weinberg (NVIDIA): PR #1571, PR #1540
• several fixes for KOKKOS package issues, Stan Moore (SNL): PR #1555, PR #1560, PR #1563
• update KOKKOS library included in LAMMPS to version 2.9.00, Stan Moore (SNL): PR #1546
• change namespace scoping for some code in the GPU library to resolve issues with CLang 7.x and later, Axel Kohlmeyer (Temple U): PR #1564
• improved error detection when parsing various kinds of files, Axel Kohlmeyer (Temple U): PR #1575, PR #1552, PR #1559
• small refactoring for USER-MEAMC to move redundant code to an inline function, optimizations, Sebastian Hütter (OvGU): PR #1551
• refactor handling of nested input file handling and single command execution; this allows to use the include keyword inside of if/then/else and results in the correct order of commands, when LAMMPS commands request reading of include files; limit level of include file nesting, Axel Kohlmeyer (Temple U): PR #1537
• updates for USER-CGDNA package, Oliver Henrich (U Strathclyde): PR #1573
• implementation of inner/middle/outer compute methods for cutoff based r-RESPA in pair style lj/class2/coul/long, Evangelos Voyatzis (Royal DSM): PR #1535
• lots of small bugfixes and improvements to code, docs and cmake configuration scripts, various contributors: PR #1553, PR #1558, PR #1561, PR #1562, PR #1567, PR #1570, PR #1544, PR #1550, PR #1538, PR #1547, PR #1539, PR #1525, PR #1534, PR #1531, PR #1530

Backward compatibility notice:

• changes in the pair_coeff syntax breaks backward compatibility with inputs using USER-CGDNA. pair style oxdna2/stk has been removed and aliased to oxdna/stk.

Changes since patch release 5 June 2019:

• using PLUMED to compute committor with LAMMPS, PR #1516, Gareth Tribello, Queen's U Belfast)
• correctly support hybrid pair styles with computes from the USER-TALLY package, PR #1513 (Axel Kohlmeyer)
• flag computes in USER-TALLY package as not compatible with dynamic groups, PR #1518 (Axel Kohlmeyer)
• performance and reduced memory consumption improvements for pair style snap and related tools, PR #1507 (Aidan Thompson, SNL)
• added logfreq3() function for another variant of generating logarithmic spaced output, PR #1476 (Vishnu V. Krishnan, Tata Institute of Fundamental Research)
• add support for building neighbor lists on the GPU with hybrid pair styles containing some '/gpu' styles, PR #1430 (Trung Nguyen, Northwestern U)
• add PPPM and Ewald solvers for electric dipoles and spins, PR #1469 (Stan Moore & Julien Tranchida, SNL)
• use memmove() instead of memcpy() in some cases where both pointers may be overlapping or identical, PR #1485 (Adrian Diaz, UFL)
• update syntax file and installation instructions for vim, PR #1501 (Vishnu V. Krishnan, Tata Institute of Fundamental Research)
• refactoring of CMake configuration to reduce the size of the master CMakeLists.txt file, PR #1502 (Richard Berger, Temple U)
• optimizations for qeq/reax with KOKKOS on CUDA, PR #1496 (Kamesh Arumugam, NVIDIA)
• add full support for use in hybrid pair styles to pair style extep, PR #1494 (Michał Kański, Jagiellonian U)
• improve warning messages for fix bond/react, PR #1517, (Jake Gissinger, Colorado U, Bolder)
• optimizations of KOKKOS styles for small systems, PR #1422 (Stan Moore, SNL)
• turn off optimization for functions that build "style factories", also turns of variable tracking for GCC 4.4 and later, PR #1492 (Axel Kohlmeyer)
• teach cmake to check for immintrin.h, which is required for USER-INTEL (and thus makes it x86-only), PR #1497 (Christoph Junghans LANL)
• advance message about taking some time to gather package information, so that it does get printed before any time consuming i/o happens, PR #1492 (Axel Kohlmeyer)
• bugfix in variable expansion where new/delete was mixed with malloc()/free(), PR #1504 (Axel Kohlmeyer)
• update to info variable command to support "internal" style variables, PR #1504 (Axel Kohlmeyer)
• small MPI bugfix in bond style table, PR #1506 (Sam Niblett, LBNL)
• sanitize settings for Kspace meshes consistently, PR #1492 (Axel Kohlmeyer)
• documentation typesetting improvements, PR #1492 (Axel Kohlmeyer)
• improve support for Ninja build tool, PR #1518 (Axel Kohlmeyer)
• workaround for ICE with GCC 4.8, PR #1518 (Stan Moore)

Backward compatibility notice:
-SNAP param files that contain the diagonalstyle keyword will generate an error. Solution is to remove the diagonalstyle keyword.

Here is a list of major changes since the last stable release 12 December 2018:

General changes:

• further tweaks and improvements to building LAMMPS with CMake.
• improved error messages, e.g. when reading data files or when using optional styles, that are not installed
• generation of the PDF version of the LAMMPS manual with Sphinx and pdflatex instead of txt2html and htmldoc (Evgenii Fetisov, PNNL).
• refactoring of Install.py scripts in the lib folders to use the argparse module
• better checking for style compatibility with suffixed styles

New packages:

• USER-YAFF providing styles for force fields using QuickFF used for simulation of MOFs (Steven Vandenbrande, Ghent University)
• USER-ADIOS package to provide Exascale compatible I/O support for dump files through the ADIOS v2.x library (Norbert Podhorszki, ORNL)

Updates for packages:

• consolidation of "soft" pair styles in USER-FEP package and addition of CLASS2 pair style variants.
• KOKKOS package bugfixes and enhancements (Stan Moore, SNL)
• new generalized pair style granular (Dan S. Bolintineanu, Ishan Srivastava, Jeremy B. Lechman, SNL)
• new pair style drip for dihedral-angle-corrected registry-dependent interlayer potential (DRIP) (Mingjian Wen, U Minnesota)
• new pair style e3b for adding option to model the E3B water potential (Steven E. Strong, U Chicago)
• new fix style electron/stopping for modeling energy loss through interactions high speed atoms with electrons (K. Avchaciov (Tartu U and U Helsinki) Risto Toijala (University of Helsinki) and others)
• update to USER-COLVARS package to version 2019-04-26
• update of the KIM package to use KIM-API v2.0 (Ryan Elliot, UMN)
• update to USER-INTEL package to enable use of the package with -DLAMMPS_BIGBIG and some other small updates and corrections: (W. Michael Brown, Intel Corp.)
• update to USER-PTM package. examples added, license info updated, documentation updates (Peter Larsen, MIT)
• updates and additions to the SPIN package (Julien Tranchida, SNL)
• updates to hyper-dynamics support (Steve Plimpton, SNL)
• updates, improvements, and bugfixes to fix bond/react (Jake Gissinger, U Colorado)
• bugfix/refactoring of REBO pair style to correctly implement what is advertised in the documentation. Added consistency check on potential files for AIREBO, REBO, AIREBO-M (Cyril Falvo Université Paris Sud)

In addition there are many small bug fixes, corrections for memory leaks, and memory management inconsistencies and general improvements.

Backward compatibility notices:

• as announced with the last stable release, the packages REAX and MEAM are no longer included.
• fixes atom/swap, deposit, and gcmc had incompatible changes to the restart handling, so those fixes cannot be restarted from restart files written by older versions of LAMMPS. If those fixes are not included or not restarted, restart files should remain backward compatible.
• the syntax for pair style kim is changed with the update to KIM-API v2.0
• the syntax of the pair_style snap command has changed by removing a redundant element specifier
• pair style rebo no longer is compatible with the CH.airebo potential file, but now requires its own potential file CH.rebo. Tests are added to all AIREBO derived pair styles, that only compatible potential files are read in.
• the moltemplate package is no longer bundled with LAMMPS. it can be installed with pip or downloaded from www.moltemplate.org

Changes since patch release 31 May 2019:

• recover compatibility with CMake version 2.8.12 in CMakeLists.txt. PR #1489 (Axel Kohlmeyer, Temple U)
• topology overflow checks added in fix bond/react. PR #1477 (Jake Gissinger, U Colorado)
• document removed features in the manual. PR #1488 (Axel Kohlmeyer, Temple U)
• corrections for documentation of pair style eb3. PR #1486 (Steven Strong, U Chicago)

Changes since patch release 15 May 2019:

• More meaningful error messages when a style is not found. Now LAMMPS will report, which package an unrecognized style belongs to, and will also notify if a package has been installed, but the style is missing because it depends on another package, which is not installed. PR #1419 (Axel Kohlmeyer)
• Add option to compute coord/atom to restrict the coordination number by allowing only members of a specific group to be counted. PR #1464 (Axel Kohlmeyer, Temple U)
• Do not enable OpenMP (i.e. BUILD_OMP) by default with gcc 9.x due to incompatibilities in the OpenMP directive semantics. PR #1483 (Axel Kohlmeyer, Temple U)
• Updates for restart file processing of fixes atom/swap, deposit, and gcmc. Now it is checked that the time step was not reset and the number of total MC attempts and successful MC attempts are restarted, too. PR #1484 (Axel Kohlmeyer, Temple U)
• Properly and cleanly support pair style hybrid and hybrid/overlay for pair_style quip. PR #1479 (Axel Kohlmeyer)
• Make sure that format style variables (and format suffixes in immediate variables) only use valid conversions. PR #1475 (Axel Kohlmeyer, Temple U)
• Bugfix for KOKKOS pair styles eam/alloy/kk and eam/fs/kk. PR #1467 (Stan Moore, SNL)
• Fix for memory leak in KOKKOS neighbor list creation; PR #1474 (Stan Moore, SNL)
• Update list of examples in manual and examples/README. PR #1465 (Axel Kohlmeyer, Temple U)
• Documentation updates, remove some obsolete content, re-enable Sphinx 2.x for processing of the documentation, and correct some links to packages. PR #1475, PR #1465 (Axel Kohlmeyer, Temple U)
• Avoid segmentation fault on unexpected empty lines in Coeff sections of data files. PR #1465 (Axel Kohlmeyer, Temple U)

Backward compatibility notices:

• Restart files for simulations using fix atom/swap, fix deposit, or fix gcmc are not backward compatible due to changes in the processing of restart files for them. You can still restart the simulation, but you cannot correctly restart the fixes from restart files created with older versions of LAMMPS