All notable changes to this project will be documented in this file. The format is inspired by Keep a Changelog. This project does not adhere to Semantic Versioning, but it uses versioning inspired by it. Given a version number X.Y.Z,

X is incremented for major changes in particular large backwards incompatible updates,
Y is incremented for minor changes like an external pull-request that adds one feature and
Z for the indication of a bug fix or other very small changes that are not backwards incompatible.

The major categories to group changes in this log are:

Input / Output for all, in particular breaking, changes, fixes and additions to the in- and output files.
Added for new features.
Changed for changes in existing functionality.
Fixed for any bug fixes.
Removed for now removed features.

Also possible, but for this project less relevant, is Deprecated for soon-to-be removed features.

SMASH-2.2

Date: 2022-05-10

Input / Output

Added option Minimum_Nonempty_Ensembles to define a minimum number of ensembles in which interactions took place
Added HepMC_treeroot output
Added option to Include_Weak_And_EM_Decays_At_The_End

Added

Light nuclei (A=3) production via 4-to-2 reactions
Possibility to perform weak decays at the end of the calculation
Possibility to impose transverse momentum or rapidity cut when extracting initial conditions for hydrodynamics
Option to add a velocity field of the form u_r = u_0 * r / R in radial direction to the sphere mode
Interface functions to initialize SMASH conveniently (see new library.h)
Extend benchmarks to include collisions at high energy to test strings

Fixed

The --version option of smash now returns a version also when SMASH is downloaded as source package without VCS information
The saturation density of the woods-saxon distribution is calculated such that its integral returns the number of nucleons in the nucleus
Initial conditions output now includes unformed particles

Changed

Minimum supported CMake version is now version 3.9
Renamed USE_ROOT, USE_HEPMC and USE_RIVET CMake options to TRY_USE_ROOT, TRY_USE_HEPMC and TRY_USE_RIVET, respectively
Upgraded to Pythia 8.307
Renamed version macro to SMASH_VERSION (from MAJOR_VERSION)
Pythia installation directory can also be specified with -DPYTHIA_ROOT_DIR and by the environment variables PYTHIA_ROOT_DIR or PYTHIA8, with /usr as default
Big endian architectures are not officially supported and compilation on them is, by default, disabled
Specific parameters for Xenon nuclei are used instead of default values

Removed

A possibly available system installation of YAML library is not considered anymore

Link to diff from previous version

SMASH-2.1.4

Date: 2022-02-24

Fixed

Properly reset for each event whether projectile and target interacted

Link to diff from previous version

SMASH-2.1.3

Date: 2022-02-15

Changed

Include unformed particles in initial conditions output
Add definition of endianness macro to FindSMASH.cmake module

Link to diff from previous version

Deprecated

Output option HepMC to enable HepMC ASCII output (equivalent to the newer HepMC_asciiv3)

SMASH-2.1.2

Date: 2022-02-04

Changed

Reflect new Cuba version in FindSMASH.cmake

Link to diff from previous version

SMASH-2.1.1

Date: 2022-01-31

Changed

Fixed typo in PgdCode class

Link to diff from previous version

SMASH-2.1

Date: 2021-12-21

Input / Output

Added option to specify a Fixed_Min_Cell_Length to control the grid size for the stochastic criterion.
Computation of thermodynamic quantities optionally restricted to participants only using the option Only_Participants

Added

5-to-2 reactions for NNbar annihilations via the stochastic collision criterion
2-to-5 reactions for NNbar annihilations to 5 pions
Thermodynamic lattice output
Parallel ensembles
VDF potentials
New smearing modes: discrete + triangular
New option: finite difference derivatives for calculation of density gradients on lattice
New option: sample particles according to quantum statistics
Non-relativistic electromagnetic potentials in a simple form meant for low collision energies
New option: start a box with an external particle list
Support of Apple silicon M1 chips
Optional N(1520) Dalitz decay with constant form factor
Pre-built docker container on Github

Changed

Evaluation of failed string processes. BBbar pairs are now forced to annihilate
Updated Dockerfile and Singularity definition file (matching pre-built container on Github)

Link to diff from previous version

SMASH-2.0.2

Date: 2021-06-23

Input / Output

Added interface with Rivet (the particle-physics MC analysis toolkit) that can process the data of the simulations at runtime and produce YODA output files with the analysis results
Refactoring of the HempMC3 output now available for final particles and for the collision history in asciiv3 format

Changed

Minimum supported cmake version is now version 3.1
Continuous integration now uses Github Actions

Link to diff from previous version

SMASH-2.0.1

Date: 2020-12-21

Input / Output

Fixed event number counting for intermediate OSCAR output

Link to diff from previous version

SMASH-2.0

Date: 2020-12-17

New major version of SMASH

Major changes:

Multi-particle interactions: There is infrastructure for 3<->2 and 3<->1 interactions and example processes are implemented.
The SMASH hybrid is available including SMASH in initial and final state as well as the vHLLE viscous hydrodynamics code.
Updated to Pythia 8.303 and optimised the function calls to allow SMASH runs at LHC and high RHIC energies.
No large backwards incompatible updates for this major version.

Input / Output

Unify event number counting: The event number counting now starts at 0 for all output formats.
Added charge chemical potential to the box modus as an option.
Added option to specify the Maximum_Cross_Section that is considered from the config file.
Clarify naming of empty events output flag: changing empty to scattering_projectile_target.
Newly available HepMC3 output.
Option to scale all cross sections by a global Cross_Section_Scaling factor from the config file.
New Additional_Elastic_Cross_Section option to add an additional constant contribution to the elastic cross section in the config file.
Option to Only_Warn_For_High_Probability in case of long production runs with the stochastic criterion.

Added

3-to-1 reactions for mesons via the stochastic collision criterion.
3-to-2 reactions for deuterons via the stochastic collision criterion.
Parallel ensembles technique to run simulations with mean field potentials faster
New covariant collision criterion.
Hadron Gas EoS extended by nQ and muQ.
Various tests for photons.
Various tests for the stochastic criterion.
Various tests for multi-particle reactions.
Command-line option to enable completely silent output.
Travis CI check to ensure zero doxygen warning about undocumented instances.
Added a new Process Type to tag for failed string processes.
Option to initialize the box with Bose or Fermi distributions.
More extensive documentation for the different collision criteria.
A map of Pythia objects to reduce number of Pythia initializations.
f-resonance to pi-pi-photon Bremsstrahlung cross-sections.
Tests for nucleon density normalization.

Fixed

Consider cross section scaling factor of incoming particles for photon production. This factor was previously neglected, resulting in exploding weights and overestimated photon production.
Use form factors for binary scattering photons also in the case of Nfrac=1.
The evaluation of the interaction point in the boxmodus is considering interactions through the wall properly.
Enforced a small time step if the box modus is used.
Bug in displaying the total energy per particle.

Changed

The Pythia version is increased to 8.303.
In collisions of unformed particles with equal formation time the outgoing particles now always inherit the smaller scaling factor.
The default collision criterion changed from "geometric" to "covariant".
The interaction point for the stochastic criterion is the coordinate of a random incoming particle instead of the middle point. This prevents density artifacts in the center of grid cells.

Removed

Integrator1dMonte as it was not used anymore.
Integrator2d with the GSL Monte-Carlo integration functions. Replaced by the Integrator2dCuhre which performs 2D integrations according to the Cuhre algorithm. The latter was now renamed to Integrator2d.

Link to diff from previous version

SMASH-1.8.1

Date: 2020-08-13

Changed

Improve version determination for Pythia.

Link to diff from previous version

SMASH-1.8

Date: 2020-04-07

Input / Output

Some column names in the extended ROOT output have changed to be agreement with those in the extended OSCAR output:
- coll_per_part -> ncoll
- formation_time -> form_time
- xsec_factor -> xsecfac
Add strangeness and baryon number to VTK particles output.
The Only_Final option for the Particles output has the new possibility if IfNotEmpty is entered. Then only output is written if the event is not empty i.e. there was a collision between projectile and target to save disk space. The other two options are now Yes or No (previously True or False).
Restructure photon configuration:
- Add new Photon subsection to Collision_Term section, with options 2to2_Scatterings: True/False, Bremsstrahlung: True/False and Fractional_Photons: Nfrac, where Nfrac is an arbitrary integer
- Add new Dilepton subsection to Collision_Term, with option Decays: True/False
- From now on, only the format and whether or not it shall be extended, can be set in the Output subsection Photons.

Added

Photon production from pion-pion bremsstrahlung
Option to randomize the reaction plane (rotate all particles by a random angle around the z axis).
Test to verify that cross sections do not depend on particle order.
Various tests for deformed nuclei
Various tests for string processes
Proper CHANGELOG.md file
Option to set up an equilibration time for the box modus, after which the output is written.
Most resonance integrals are now cached on disk, reducing the time until the simulation starts.

Changed

Extend box examples in user guide to directly run.

Fixed

Fix user guide for baryon density dependent symmetry potential.
ASCII initial conditions output:
- Bugfix: write PDG ID instead of unique particle ID.
- Exclude spectators.
Add lower bound for initial conditions proper time.
Fix floating point exceptions that were raised if initial conditions output is enabled
Fix light nuclei not being affected by potentials

Link to diff from previous version

SMASH-1.7

Date: 2019-10-14

Input / Output

New output content: Initial conditions output in ASCII, Oscar, Binary and ROOT format
New option to explicitly specify the output times in terms of a list
Input option Included_2to2 now has a new possible value NNbar. This is a breaking change since this value now has to be in the included list of 2to2 reactions in order to use the resonances option for NNbar_treatment.
Input configuration of deformed nuclei has changed. Additional level of Orientation provided where the angles or a random rotation can be specified. This change is backwards incompatible to config files for SMASH-1.6 or older.
Most outputs now specify whether the projectile collided with the target
ROOT output now includes charge and optionally the extended output
Added particle ID and number of collisions to VTK particles output
Thermodynamic output can now also output electric, baryonic and strange currents

Added

New stochastic collision criterion (as in A. Lang, H. Babovsky, W. Cassing, U. Mosel, H. G. Reusch, and K. Weber, J. Comp. Phys. 106, 391 (1993)) for 2-to-2 reactions of one particle species with a fixed elastic cross-section.
Initial conditions for hydrodynamic simulations can be extracted and are provided in the initial conditions output. The proper time of the produced hypersurface can, if desired, be specified in the configuration file
Fermi motion for deformed nuclei
Random rotation of custom nuclei
Automatic deformation parameters for ruthenium and zirconium
Option to add net-baryon density dependence to symmetry potential
Option to add high momentum particle into a box calculation

Changed

Changed the NN transition region to 3.5-4.5 GeV to better describe the experimental data
Box now has minimum length given by the grid size
The final-state cross-section output provided by smash -S pdgID1,pdgID2 no longer includes resonances

Fixed

Collision finding with frozen Fermi motion: Use beam momentum for collision finding if nuclei have not interacted
Fix output time for output at the event end on the command line to always be correct.

Link to diff from previous version

SMASH-1.6

Date: 2019-05-24

In- and Output

Fix thermodynamic VTK output of the Landau velocity

Added

Particle properties updated to use PDG 2018 data (this includes 5 new N* and a new Delta*)
Lund fragmentation function for leading baryons in soft non-diffractive string processes
Support for customized nuclei: optionally reading in initial nucleon positions in collider modus
New density type for charge/isospin and possibility to output the charge currents without smearing
Option to place a single high-momentum particle in the center of the sphere modus (~jet)
SMASH compiles now also with GCC >= 9, Clang >= 7

Changed

No debug output in the default build type, which results in much better performance

Fixed

Documentation of formation time
Correct declaration of Woods-Saxon distribution for deformed nuclei

Link to diff from previous version

SMASH-1.5.2

Date: 2019-03-07

Version for JETSCAPE run

Added

New tests for density and potential
New density type for charge/isospin
Added option to propagate a high momentum hadron through the sphere

Changed

Exact Pythia version is now required
Improve third party infrastructure
Default build type amended with one flag
Unify functions for EM and hadronic widths of resonances

Fixed

Bug fixes related to formation of particles including
- Correcting the boost
- Adjust resonance decay times
- Add unformed particles to grid for collision finding, if they form during the time step

Link to diff from previous version

SMASH-1.5.1

Date: 2018-12-04

Changed

Documentation of formation time
Continuous particle formation for unformed particles from string fragmentation processes

Fixed

Correct generic radius for nuclei with A > 16, except for Au, Pb, U, Cu as they are/were exactly specified

Link to diff from previous version

SMASH-1.5

Date: 2018-11-27

First public version of SMASH

Known issues:

Simulations at high √s_NN above ~30 GeV are slow due to the initialization of Pythia for hard scatterings
Strangeness production at intermediate energies is off due to the string excitation and fragmentation (only tuned to pp elementary data)
No Coulomb potential is available