CHANGELOG.md

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog v1.1.0, and this project adheres to Semantic Versioning v2.0.0 since ESPResSo 4.0.0.

Numbers in brackets with a hashtag refer to ticket numbers on GitHub starting with release 4.0.0, or Savannah until release 3.3.1.

Unreleased

5.0.0 - 2026-02-26

This is a major release. New features were added and deprecated features were removed. The API has changed, and some of these changes are silent, i.e. warnings aren't necessarily emitted when running a script designed for ESPResSo 4.x that relies on features that have significantly changed in ESPResSo 5.0.

Highlights of the release include:

• rewrite of lattice-Boltzmann and electrokinetics using solvers backed by the waLBerla framework. For LB, this includes enhanced capabilities including vectorization support on the CPU and multi-GPU support, per-cell boundary conditions to build arbitrary geometries, per-particle friction coefficients, and Lees-Edwards boundary conditions.
• faster writing of simulation trajectories using the H5MD file format, particularly in parallel simulations.
• per-particle selection of equations of motion.
• the thermalized Stoner–Wohlfarth model for magnetodynamics, and the ability to obtain local magnetic fields at the particles' positions.
• virtual sites tracking the center of mass of a group of particles for umbrella sampling.
• initial support for shared-memory parallelism for some scenarios.
• several new tutorials, e.g., on the sedimentation of particles in a fluid, the Boltzmann inversion technique, electrode modelling, and machine-learned inter-atomic potentials.

Added functionality

• The original LB and EK methods have been completely replaced with equivalent implementations based on the high-performance waLBerla library (#4726, #5101). This is a major API change that requires adapting all LB and EK scripts to use the new classes and arguments.
• LB now support Lees-Edwards boundary conditions (#4977).
• LB and EK methods now support setting boundary slip velocities on individual nodes (#4252).
• LB now supports per-particle gamma (#4743). Only works for isotropic particles.
• Thermostats and integrators have been redesigned as unified propagators (#4820, #4603). Multiple combinations of thermostats and integrators are now supported to solve multiphysics problems. While the Python interface remains mostly unchanged, internally the user-selected integrator is now the "main" integrator. Alternative integration schemes can then be enabled on a per-particle basis using the new propagation flag. An important consequence is that all virtual site types can now be enabled in the same simulation.
• Magnetodynamics support was introduced with the thermal Stoner–Wohlfarth model (#5188). This is achieved through a virtual site that decouples the particle dipole from the particle quaternion.
• A new virtual site implementation was introduced to exert forces on molecules through their center of mass, for example to implement umbrella sampling (#5199).
• The OpenGL visualizer now uses different colors for arrows representing fluid velocities and slip velocities (#4252).
• ESPResSo now supports the ZnDraw visualizer (#4967, #5115, #5217).
• ESPResSo now has Atomic Simulation Environment (ASE) bindings (#4912), including calculators (#5162). One application is interfacing ESPResSo with machine-learned potentials.
• The magnetostatics.DipolarDirectSumCpu() feature now works in a MPI-parallel simulation (#4559).
• The magnetostatics.DipolarDirectSumCpu() feature now supports replicas via the new optional argument n_replicas (#4559).
• The magnetostatics.DipolarDirectSumGpu() feature now supports replicas via the new optional argument n_replicas (#5094).
• The magnetostatics.DipolarDirectSumCpu() and magnetostatics.DipolarDirectSumGpu() features can now calculate the total dipole field experienced by each particle (#4626, #5094). Requires feature DIPOLE_FIELD_TRACKING.
• Particle-based observables ParticleDirectors() and ParticleDipoleFields() were added (#4627, #4626).
• Particle bond energies can be calculated with system.analysis.particle_bond_energy() for a given bond and particle (#5040).
• Particle neighbor lists can be extracted with system.analysis.particle_neighbor_pids() (#4662). This feature will help prototyping simulations that interface with machine-learned potentials, which take a list of particle positions as input and output the force on the central particle.
• Observable PairwiseDistances and accumulator ContactTime were introduced to track the contact time, i.e. number of consecutive time steps during which two particles are closer than a cutoff value (#5032).
• The bond breakage feature now supports angle bonds (#4716).
• Tabulated interaction TabulatedNonBonded got a new method set_analytical() to automatically set the energy and force from the analytical expression of the potential using SymPy (#5019).
• Instrumentation tools Caliper, CUDA-GDB and kernprof are now natively supported (#4747).
• Instrumentation feature FPE (floating-point exceptions) is now natively supported on x86 and Armv8 architectures (#5020).

Changed requirements

• The project now requires C++20 and CUDA 12 (#3918, #4612, #4931).
• The build system now supports the Intel oneAPI C++ Compiler (#4532), the Cray Clang compiler (#5201), and the NVIDIA HPC SDK (#5257).
• The waLBerla library is now a dependency for all LB and EK methods (#2701, #4726). If not found, it is built from sources automatically.
• The heFFTe library is now a dependency for Coulomb P3M (#5063) and the EK FFT solver (#5101). If not found, it is built from sources automatically.
• The Kokkos and Cabana libraries are now dependencies for shared-memory parallelism (#5074). If not found, they are built from sources automatically.
• The OpenMP component of the FFTW3 library are now dependencies for shared-memory parallelism (#5086). This component is sometimes packaged separately from the MPI FFTW3 library on HPC clusters.
• The HighFive library is now a dependency for hdf5 file I/O (#5087). It is built from sources automatically. The h5xx library is no longer a dependency.
• The GNU GSL library is now a dependency for MMM1D (#5201).
• The minimal version of all dependencies was increased (#4532, #4612, #4717, #4931, #5093, #5201, #5223): CMake >= 3.27.6, Python >= 3.11, Cython >= 3.0.4, Boost >= 1.83, CUDA >= 12.0, OpenMPI >= 4.0, MPICH >= 3.4.1, GCC >= 12.2, Clang >= 18.1, AppleClang >= 17.0, CrayClang >= 17.0, Intel oneAPI C++ Compiler >= 2023.1, and Python packages versions are pinned on versions available in the Ubuntu 22.04 repository. CUDA 12.6 and later versions are now supported (#5129).

Feature configuration at compile time

• All project-specific CMake options have been renamed (#4612). This change was required to avoid name collisions with external projects. Please refer to the user guide chapter on installing ESPResSo to find out the new option names. Using the old option names will generate warnings, but CMake will carry on and use default values instead of the values you provided. Please don't ignore these warnings when adapting your build scripts.
• The CMake option ESPRESSO_CUDA_COMPILER was removed in favor of the environment variable CUDACXX (#4642).
• A config file is now available to build the project automatically in Codespaces (#5201, #4531).
• An AGENT.md is now available to guide agentic coding tools (#5220).

Improved documentation

• A Widom insertion tutorial was added (#4546)
• A lattice-Boltzmann sedimentation tutorial was added (#4570)
• A machine-learned potentials tutorial was added (#4982).
• An atomistic water simulation tutorial was added (#5174).
• An electrodes tutorial with ICC/ELC/ELC-IC was added (#4784).
• A Boltzmann inversion tutorial was added (#5187).
• A Grand Canonical Monte Carlo tutorial was added (#4670).
• The electrokinetics tutorial was completely rewritten and now features chemical reactions (#4782).
• All tutorials were re-designed for JupyterLab (#4830). Reliance on Jupyter extensions and plugins has been significantly reduced in an effort to improve compatibility with other Jupyter backends. In particular, VS Code Jupyter is still actively supported. Jupyter Notebook (Classic Notebook) should still be compatible, although it is not actively tested. IPython is no longer supported.
• Most tutorials adopted ZnDraw as the visualization backend (#4976, #4975).
• A high-throughput computing sample based on the Dask scheduler was added (#4781).
• All supported debuggers and profilers are now documented: Caliper, Valgrind, GDB, CUDA-GDB, kernprof, perf, UBSAN, ASAN (#4747).
• Installation instructions were improved with better sectioning (#5062).
• The CUDA 12 circular dependency in Ubuntu 24.04 packages is documented (#4642).

Interface changes

• The original LB classes LBFluid and LBFluidGPU were removed in favor of a unified LBFluid class for both CPU and GPU (#2701, #4726, #5230). Their arguments have also changed, e.g. dens became density and visc became viscosity. The pressure_tensor_neq property was removed.
• The original EK class Electrokinetics was removed in favor of a unified EKSpecies class for both CPU and GPU (#2701, #4726, #5101, #5230).
• Self-propelled particles (swimmers) have been completely re-implemented (#4745). The propulsion mechanism can now only be set up with a force. When coupling to a LB fluid, a real particle and a virtual site are used to create the dipole.
• The long-range actors API was completely redesigned (#4749).
• CPU and GPU algorithms now have a unified Python class (#5230). Pass optional argument gpu=True to the constructor to select the GPU backend. For example, class espressomd.electrostatics.P3MGPU was removed in favor of espressomd.electrostatics.P3M, which now manages both the CPU and GPU backends. Likewise, DipolarDirectSum replaces both DipolarDirectSumCpu and DipolarDirectSumGpu.
• The virtual sites API was completely redesigned (#4820, #4603).
• The collision detection API was completely redesigned (#4987).
• The Galilei transform API was completely redesigned (#4816).
• Class attributes expecting 3 boolean values no longer accept integer values (#4541). It is no longer possible to set properties system.periodicity, particle.fix and particle.rotation with e.g. [1, 1, 1] or [0, 0, 1].
• reaction_methods.ReactionAlgorithm.reaction() now takes steps instead of reaction_steps as argument for consistency with the MD integrator (#4666)
• io.mpiio.Mpiio() now takes a system as argument (#4950).
• analysis.pressure() and analysis.pressure_tensor() now take the DPD stress tensor into account into the total pressure, and got an additional member dpd (#5045).
• analysis.energy(), analysis.pressure() and analysis.pressure_tensor() got additional members kinetic_lin and kinetic_rot to separate linear and angular kinetic energy/pressure (#5043).
• cluster_analysis.ClusterStructure() now takes a system as argument (#4950).
• interactions.ThermalizedBond() parameter seed was moved to system.thermostat.set_thermalized_bond() (#4845). This change better reflects the fact there is only one global seed for all thermalized bonds; until now this global seed was overwritten by any newly created thermalized bond, whether it was added to the system or not.
• All P3M algorithms now accept an extra argument tune_limits to constrain the range of mesh values exploring during mesh size tuning (#5017).
• The check_complex_residuals optional argument of the P3M algorithm was removed (#5189).
• Python objects of type pathlib.Path can now be passed to functions that expect file paths (#5128).
• Bonds breakage can now be triggered manually (#4995). This is meant to be used in Lees-Edwards simulations with a time-dependent shear, since bonds extending across the shear boundary can increase in length without a change in particle positions when the simulation time increases.
• Analysis.particle_energy() was renamed to Analysis.particle_non_bonded_energy() to better reflect the calculated quantity, since kinetic, bonded, electrostatic and magnetostatic contributions are not part of this energy (#5226).

Removed functionality

• The lb.LBBoundaries() framework was removed (#4381). Shapes can now be passed directly to LB and EK objects.
• The magnetostatics.DipolarDirectSumWithReplicaCpu() method was removed, since the magnetostatics.DipolarDirectSumCpu() method now supports replicas (#4559).
• The electrostatics.MMM1DGPU() feature was removed (#4928).
• The magnetostatics.DipolarBarnesHutGpu() feature was removed (#4928).
• The MDAnalysis bindings were removed (#4535)
• The bind_three_particles collision mode was removed (#4823).
• LB populations are no longer accessible from the Python interface (#5075).

Improved testing

• The Armv8 architecture is now tested in CI (#5020).

Performance enhancements

• LB now supports multi-GPU acceleration (#5007).
• Observables are now fully MPI-parallel and show better performance than equivalent operations by hdf5 or MPI-IO writing on a SSD (#4748).
• Reaction methods are now fully MPI-parallel and now only invalidate the system state after a batch of particle changes have been applied (#4666).
• Performance of particle property getters and setters has improved, in particular vector quantities such as force and velocity are 25 times faster to read from and 3 to 4 times faster to write to (#5209, #5124, #5069).
• The RegularDecomposition cell system no longers uses a ghost layer when the simulation has only 1 MPI rank (and any number of OpenMP threads), which improves performance of a Lennard-Jones simulation by 11% for 1 thread (#5157).
• Shared-memory parallelism (OpenMP) is now supported in short-range force calculation (#4754, #5097), Coulomb and Dipolar P3M (#5086, #5189), LB and EK (#5083).

Bug fixes

• UTF-8 strings are now supported in all features (#5128).
• Updating an active non-bonded interactions via e.g. system.non_bonded_inter[0, 0].lennard_jones.set_params() now uses the default arguments when optional arguments are missing and raises an error when required arguments are missing (#4558). In previous ESPResSo versions, missing optional and required arguments would be recycled from the previous state of the non-bonded interaction (#4569).
• Thermalized LB simulations are now fully decorrelated (#4845, #4848). In previous ESPResSo versions, the LB thermostat seed argument was actually used as the RNG counter, thus ensemble runs would produce almost the same trajectory.
• Particle coordinates are now properly folded in the histogram and RDF classes to avoid off-by-one errors (#5109).
• The particle_data.ParticleHandle() and io.writer.h5md.H5md() writer now use properly folded particle coordinates (#4940, #4944). In previous ESPResSo versions, cached coordinates would be used, which could be out-of-date when large Verlet list skin values were used.
• Lees-Edwards now applies the offset in the correct direction and no longer requires the user to provide a shear_velocity multiplied by -1 (#5081).
• Lees-Edwards boundary conditions now support the regular decomposition cell system via the new fully_connected_boundary argument (#4958).
• The isotropic NpT algorithm was completely rewritten and now supports two barostats: Andersen (#5053) or MTK (#5077).
• It is no longer possible to change the reaction constant of an existing reaction with a gamma value less or equal to 0 (#4666).
• When setting up a reaction method with two or more reactions, a runtime error is raised if a reaction accidentally overwrites the default charge of a specific type with a different value (#4666).
• It is no longer possible to add an angle bond or dihedral bond with a list partner particle ids containing duplicate entries, since the angle would be undefined (#5012).
• Adding the same object twice in an ObjectList now raises a runtime error; removing an unknown object from an ObjectList now raises a runtime error (#4779). In previous ESPResSo versions, adding the same object twice in a list could have unintended side-effects.
• The SimplePore distance function was corrected and no longer generates NaN values (#5016).
• The FENE bond now breaks when compressed beyond its stretching limit (#5195).
• Coulomb and Dipolar P3M algorithms no longer emit warnings nor trigger assertions when particles are close to the box boundaries, when running simulations on a CPU that supports extended precision floating-point numbers (#5136).
• OpenMPI 5.0 now longer triggers random PRRTE errors at system exit (#5093).
• Default-constructed Utils::Vector and Utils::Array objects are now properly zero-initialized (#5257). In previous releases, the default constructor could accidentally leave the underlying data uninitialized when using the NVHPC compiler toolchain and building at the -O2 or -O3 optimization level (#5263).
• Thole corrections are no longer part of the energy calculated by Analysis.particle_non_bonded_energy() (#5226)

Under the hood changes

• Most Cython files have been converted to Python files (#4541, #4713).
• Cython 3 is now supported (#4845).
• Sources of NaN, float overflow, and most float underflow were addressed (#5020).
• GPU algorithms no longer leak device memory (#4741, #4764).
• CPU implementations of the P3M algorithm no longer leak memory (#4947).
• The CPU implementation of the P3M Coulomb algorithm was entirely rewritten using the heFFTe library (#5063). The method is now easier to modify and extend, supports shared-memory parallelism (#5086, #5189), and uses real-to-complex transforms (#5204).
• Project-specific compiler diagnostics are no longer propagated to external projects like waLBerla (#4642).
• The build system now relies on CMake's native CUDA support (#4642).
• The build system now installs the object-in-fluid Python module when espressomd is installed (#4931).
• The script interface was massively simplified (#4816).
• Most global variables were removed (#4741, #4783, #4816, #4845, #4950).
• The ESPResSo repository can now be cloned without git flag --recursive (#5031). ESPResSo developers are now expected to integrate new third-party libraries using the CMake FetchContent mechanism instead of git submodules.
• The build system now properly handles linking of ESPResSo against static and shared libraries, sets the correct runpaths, and avoids cyclic dependencies during the linking stage (#5221, #5173). These changes are most relevant to cluster admins and package maintainers.

4.2.2 - 2024-05-22

Improved documentation

• Installation instructions now mention the FFTW3 MPI dependency of long-range solvers and provide recommended version numbers for Jupyter Notebook dependencies (#4790).
• Installation instructions now mention Python environments (#4922).
• Observables not properly document return values, array shapes, and use a more consistent mathematical notation (#4898).

Bug fixes

• Fatal runtime errors due to MPI global variables lifetime were addressed (#4858). Older ESPResSo releases built with Boost 1.84 or later might randomly crash when exiting the Python interpreter.
• Virtual sites no longer contribute to the kinetic energy of the system (#4839). The regression was introduced in April 2021 and affected the 4.2 branch of ESPResSo.
• Inertialess tracers are now integrated along the z-axis (#4714). The regression was introduced in February 2022 and affected the 4.2 branch of ESPResSo.
• Inertialess tracers now throw an exception when attempting to use LB GPU with 2 or more MPI ranks (#4714). Before, tracers on non-root MPI ranks would be silently ignored by the CUDA kernels, and would have a constant velocity, either 0 if the particle never visited the fluid domain on the root rank, or the last known velocity if the particle was once on the root rank. This bug affected all ESPResSo versions.
• Particles close to the faces of the simulation box are now properly coupled to the LB fluid (#4827). Due to numerical instability, it was previously possible for particles to be outside the box simulation by a tiny amount and skip LB particle coupling. The probability of this bug occurring was low, but could be enhanced in simulations that purposefully placed particle near the faces of the simulation box: polymers sheared by Lees-Edwards boundary conditions, raspberry particles (colloids, bacteria, etc.) when crossing a periodic boundary, or cell membranes placed close to a periodic boundary.
• Resizing the box now throws a runtime error if there are constraints present (#4778), since constraint preconditions might no longer be fulfilled. For example, a wall constraint might end up outside the box boundaries when the box shrinks.
• Resizing the box via system.box_l = new_box_l now throws a runtime error if there are particles present, because particle position folding cannot be guaranteed to be correct (#4901); use system.change_volume_and_rescale_particles() instead, which properly rescales particle positions.
• The velocity Verlet NpT propagator doesn't apply friction and noise on angular velocities. ESPResSo now throws an error when NpT encounters a rotating particle (#4843). This bug affected all ESPResSo versions.
• The Brownian thermostat can no longer be configured with act_on_virtual=True due to an unresolved bug (#4295) that will be addressed in the next minor release.
• Restrictions on the number of MPI ranks have been lifted from the checkpointing mechanism (#4724). It is now possible to use checkpointing again in MPI-parallel simulations when the system contains LB boundaries or Union shape-based constraints. These restrictions had been introduced in 4.2.0 for technical reasons that have since been resolved.
• When passing an invalid value to a function that expects an input parameter of type list of size 3, an exception is now raised (#4911). Previously, some functions would print an error message and continue their execution with uninitialized data.
• The per-type and per-mol_id contributions from system.analysis.energy(), system.analysis.pressure() and system.analysis.pressure_tensor() now return the correct values (#4788). Older version of ESPResSo were confusing the particle mol_id with the particle type. The total pressure was unreliable when mol_id properties were set to non-zero values.
• The OpenGL visualizer now extracts the correct non-bonded potential parameter sigma when feature WCA is compiled in but LENNARD_JONES isn't (#4720). The regression was introduced in 4.2.1.
• Method OifCell.elastic_forces() no longer throws a TypeError (#4813).
• Benchmark scripts were adjusted to support large particle numbers (#4753).

Under the hood changes

• Several Clang 16 and GCC 13 compiler diagnostics have been addressed (#4715).
• A non-critical GCC C++20 deprecation warning in Cython-generated code was disabled (#4725).
• Several deprecation warnings emitted by CMake 3.27 have been silenced (#4792).
• Add support for setuptools version 67.3.0 and above (#4709).
• Add support for Python 3.12 in testsuites run by CTest (#4852).
• Python requirements have been updated (#4924).
• CI pipeline URLs have been fixed (#4736).

4.2.1 - 2023-04-17

Added functionality

• P3M and DipolarP3M can now be used with the hybrid decomposition cell system with 1 MPI rank (#4678).
• Lattice-Boltzmann can now be used with the N-square and hybrid decomposition cell systems with 2 or more MPI ranks (#4676).

Changed requirements

• The nbconvert version requirement was bumped to 6.5.1 to patch an XSS vulnerability (#4658).

Improved documentation

• The user guide now documents how to improve the reproducibility of simulations that have checkpointing enabled (#4677).
• The user guide now reflects that the lattice-Boltzmann profile observables can be used in parallel (#4583).
• The active matter tutorial now uses an adequate engine dipole for the swimmer particle (#4585).
• The error analysis tutorials have been improved (#4597).
• The tutorials can now be used in VS Code Jupyter (both the desktop and web versions) and the mathematical formula are now correctly displayed (#4531).
• All ESPResSo-specific CMake options are now documented in the installation chapter of the user guide (#4608).
• Python package installation instructions no longer feature package version numbers; instead, requirements.txt is used as a constraint file (#4638).
• MMM1D algorithms now properly document their parameter names (#4677).
• Reaction methods now cite the relevant literature (#4681).
• Caveats for chain analysis methods are now documented (#4698).
• Minor formatting issues in Sphinx and typos in Python docstrings were addressed (#4608).

Interface changes

• A new boolean property System.virtual_sites.override_cutoff_check was introduced to allow disabling the cutoff range checks from virtual sites (#4623).

Removed functionality

• The unused and untested Analysis.v_kappa() method was removed (#4534).

Improved testing

• Improve unit testing of core functionality: P3M, MMM1D, OIF, virtual sites, script interface factory (#4631).

Bug fixes

• The checkpointing mechanism now properly restores the particle quaternion and all derived quantities (#4637). Release 4.2.0 introduced a regression that caused checkpoint files to overwrite the particle quaternion/director by a unit vector pointing along the z direction, when the DIPOLES feature was part of the myconfig file. This lead to incorrect trajectories when reloading a simulation from a checkpoint file, if the particle director played a role in the simulation (ex: relative virtual sites, Gay-Berne potential, anisotropic particles, active particles, etc.). In addition, the angular velocity in body frame was restored with the wrong orientation. Since the default myconfig file contains DIPOLES, most ESPResSo users were affected.
• The checkpointing mechanism now properly restores LB boundaries (#4649). Release 4.2.0 introduced a regression where reloading LB populations would accidentally reset LB boundary flags.
• The checkpointing mechanism now restores P3M and DipolarP3M solvers without triggering a re-tune (#4677). In previous releases, the checkpointing code would automatically re-tune these algorithms during a reload, causing tiny deviations in the forces that were problematic for trajectory reproducibility.
• Brownian dynamics now integrates the rotational dynamics of rotatable particles whose position is fixed in 3D space (#4548).
• Langevin dynamics now properly integrates particles with anisotropic friction (#4683, #4690).
• A regression that caused virtual sites to incorrectly count their image box when crossing a periodic boundary has been fixed (#4564, #4707).
• Particles can no longer be created or updated with a negative mass or a null mass (#4679).
• Particles created without a user-specified type can now participate in reactions (#4589).
• When a Monte Carlo displacement move is rejected, the original particle velocity is now restored (#4589).
• Reaction methods now raise an exception when accidentally calling method.reaction(steps=20) instead of method.reaction(reaction_steps=20) (#4666). Since 4.2.0 the steps argument was ignored, in which case the default value reaction_steps=1 would used by the core. Note that in the next minor release of ESPResSo, the reaction_steps argument will be renamed to steps.
• Reaction methods now rebuild the list of free particle ids every time WidomInsertion::calculate_particle_insertion_potential_energy() and ReactionAlgorithm::do_reaction() are called (#4609). This was needed to allow multiple concurrent reactions, as well as avoiding subtle bugs when both the user and a reaction method tried to create a new particle with an id that used to belong to a deleted particle.
• When all particles are cleared, the reaction methods type map is now also cleared (#4645). In the past, it was possible to attempt a reaction on particles that had just been cleared from the system, which would raise an exception. This bug affected all ESPResSo releases since 4.0.
• The System.part.pairs() method now returns the correct particle pairs when particle ids aren't both contiguous and starting from 0 (#4628). The regression was introduced in release 4.2.0.
• The auto-exclusions feature no longer adds spurious exclusions to particle ids in the range [1, distance] (#4654). This bug would potentially break the physics of the system and potentially raise an exception in a system with non-contiguous particle ids. This regression was introduced in release 2.2.0b.
• The structure factor analysis code no longer double-counts particles when the same particle type is provided twice (#4534).
• The minimal distance distribution analysis code no longer has an arbitrary cutoff distance when the simulation box is aperiodic (open boundaries); this would cause spurious artifacts to appear in the histogram at r = np.sum(system.box_l) when particles were further apart than this arbitrary distance (#4534).
• The cluster analysis functions are now disabled for systems with Lees-Edwards periodic boundaries, since the cluster analysis position wrapping code doesn't properly handle the shear offset (#4698).
• The chain analysis methods now raise an error when the number of chains or beads per chain is invalid (#4708).
• The observable tests now longer rely on deprecated numpy options that were removed in numpy 1.24 (#4635).
• The visualizer *_arrows_type_materials options now have an effect on arrow materials (#4686).
• The visualizer exception handling mechanism has been made less brittle (#4686).
• The visualizer no longer raises exception when the optional dependency freeglut isn't installed (#4691).
• The visualizer can randomly freeze when using collision detection or bond breakage; a temporary workaround has been introduced that fixes the issue for simulations that use only 1 MPI rank (#4686).
• The __dir__() method of script interface objects no longer raises an exception (#4674).
• Compilation and testsuite issues involving missing or incorrect feature guards were addressed (#4562, #4648).
• The build system no longer silently ignores invalid external feature definitions in myconfig.hpp and CMake files (#4608). This issue would only affect feature developers, as well as users of very old compilers, and would lead to ESPResSo builds missing features.

Under the hood changes

• The Clang 14 and AppleClang 14 compilers are now supported (#4601).
• Several Clang 14 compiler diagnostics have been addressed (#4606).
• Boost 1.81 and later versions are now supported (#4655).
• Compiler errors on non-x86 architectures were addressed (#4538).
• Test tolerances were adjusted for non-x86 architectures (#4708).
• The pypresso script now prints a warning when running with MCA binding policy "numa" on NUMA architectures that are not supported in singleton mode by Open MPI 4.x (#4607).
• The config file generator has been rewritten to properly handle external features and compiler errors (#4608).
• Security hardening for GitHub Workflows (#4577, #4638) and Codecov (#4600).
• Deployment of the user guide to GitHub Pages now relies on cloud providers to fetch JavaScript dependencies (#4656).

4.2.0 - 2022-06-30

This is a feature release, i.e., new functionality is added to ESPResSo. New thermostats, cell systems and boundary conditions have been introduced to simulate systems with Stokesian Dynamics, Brownian Dynamics, strongly inhomogeneous particle sizes or translation-invariant shear flow. The interface underwent (non-silent) changes, therefore scripts will have to be slightly adapted. Most notably, particle access by id and particle slices have a new syntax, and electrostatic/magnetostatic layer correction and reaction methods have a different setup. All errors are also now emitted as Python exceptions and are recoverable with minimal effort.

An additional focus of this release is the simplification of both the C++ core and the Python script interface to facilitate future extensions of ESPResSo. The testing of ESPResSo's functionality has been extended considerably.

Added functionality

• P3MGPU now supports energy and pressure calculation via the CPU kernels (#4506).
• ELC now works with P3MGPU (#4506).
• The LB grid now supports slicing operations (#4195) and LB slices are equality comparable (#4268).
• Lees-Edwards boundary conditions can be used for particle-based simulations (#4457). Lattice-Boltzmann support will be added in the 4.3.0 release.
• The non-bonded energy of a single particle can be calculated (#4401).
• The list of close neighbors of a single particle can be extracted (#4401).
• Brownian Dynamics simulations can be carried out with the newly added Brownian integrator and Brownian thermostat (#1842).
• Stokesian Dynamics simulations can be carried out with the newly added Stokesian integrator and Stokesian thermostat (#3790, #3987).
• Bonded interactions can now be automatically broken when the bond length exceeds a critical value (#4456). This feature can be combined with collision detection to model reversible bonds (#4464).
• A new cell system HybridDecomposition was introduced to speed up simulations with inhomogeneous particle interaction ranges (#4373).
• Shapes can be merged into meta-shapes (#3493, #3538).
• The HollowConicalFrustum can now be sliced open, made thick and rotated to model quarter pipes in any orientation (#4179). The main application is in the construction of complex microchannel geometries via LBBoundaries.
• A parametric weight function was added to the DPD interaction (#3570).
• H5MD output files now support a unit system (#3751).
• H5MD output files now support custom specifications to control which particle and box properties to write to disk (#4480).
• The H5md class is now checkpointable and usable in an interactive Python session (#3751).
• MDAnalysis integration now provides bond information (#3801).

Changed requirements

• The minimal version of all dependencies was increased (#3375, #3687, #3878, #3984, #3994, #4115, #4312, #4337, #4489): Python 3.8, Cython 0.29.14, CMake 3.16, Boost 1.69, Sphinx 2.3.0, and Python packages versions are pinned on versions available in the Ubuntu 20.04 repository.
• CMake no longer emits a warning about the deprecated distutils Python package, which is also no longer a requirement (#4433).
• CUDA 11 support was added (#3870).
• CUDA 8 and CUDA 9 support was removed (#3984).
• AMD GPU support via ROCm (HCC and HIP-Clang compilers) was removed (#3966).
• library libcuda is no longer a dependency in CUDA builds (#4095).
• Installation instructions for ESPResSo on Microsoft Windows via WSL are now available (#4348).
• LaTeX is no longer a requirement for building the Sphinx documentation and running the tutorials (#3256, #3395).

Feature configuration at compile time

• GPU support is now opt-in (#3582). Pass the CMake flags -DWITH_CUDA=ON to compile CUDA code and optionally -DWITH_CUDA_COMPILER=<compiler> to select the CUDA compiler: NVCC (default), Clang.
• Optional features HDF5, ScaFaCoS and Stokesian Dynamics are now opt-in (#3735, #4112). If they are requested with their -DWITH_<FEATURE>=ON flag and their dependencies are not found, CMake will raise an error. In the older 4.1 build system, CMake would silently ignore these features when their dependencies were not found, causing confusion as to what was exactly compiled.
• Experimental support for fast-math mode was added (#4318). Some features might break depending on the compiler used to build ESPResSo. Please quantify the numerical stability of your simulations before enabling fast-math mode in production.
• The LANGEVIN_PER_PARTICLE feature was renamed to THERMOSTAT_PER_PARTICLE (#4057).
• The magnetostatic extension DLC now depends on feature DIPOLES instead of DP3M, since FFTW is not a dependency of DLC (#4238).
• The electrostatic extension ICC now depends on feature ELECTROSTATICS instead of P3M, since FFTW is not a dependency of ICC (#4238).
• The MMM1D_MACHINE_PREC feature was added to enable Chebychev series for MMM1D on CPU without the need to define the (now removed) BESSEL_MACHINE_PREC macro (#4311).
• The EXPERIMENTAL_FEATURES feature was removed (#4482).

Improved documentation

• Tutorials have been renamed and organized by difficulty level (#3993).
• Tutorials Lennard-Jones, electrostatics, lattice-Boltzmann, raspberry electrophoresis and constant-pH have been improved (#3408, #3881, #3914, #3893, #4302, #4262).
• Tutorial lattice-Boltzmann was split into three tutorials: polymer diffusion, Brownian motion and Poiseuille flow (#4052, #4329).
• The active matter tutorial was rewritten into a Jupyter notebook (#3395, #4304).
• An error analysis tutorial was added (#4174).
• Tutorials now use the exercise2 plugin to hide solutions (#3872); since this plugin only exists for the classic Jupyter Notebook, a conversion script is provided for JupyterLab users (#4522).
• The user guide now includes a button on Python code samples to hide terminal output and Python prompt symbols (>>> and ...), so as to facilitate copy-pasting examples directly in the terminal (#4386).
• The user guide now uses a responsive theme for mobile/tablet users (#4504).
• The user guide chapter on thermostats was moved to the chapter on integrators, since they are tightly coupled (#4080).
• Mentions to non-existent functions were removed from the user guide (#4482).
• Scientific publications referenced in comment lines in the core have been converted to BibTeX citations and integrated into Doxygen blocks to make them accessible in the Doxygen HTML documentation (#3304).
• The Reaction Field electrostatic method is now documented (#4218).
• The H5MD feature is now better documented (#4480).
• A Gibbs ensemble sample was added to simulate the exchange of particles between two ESPResSo systems via the multiprocessing Python module (#4243).
• A reaction ensemble sample was added to simulate a complex chemical reaction involving 5 chemical species (#3778).

Interface changes

• The system.set_random_state_PRNG() method was removed (#3482).
• The []-operator on system.part was removed (#4402). Use system.part.by_id(1) to fetch a specific particle, system.part.by_id([1, 3]) to fetch a group of particles, or system.part.all() to fetch all particles. This change was necessary to resolve the ambiguity of particle slices containing non-contiguous particle ids.
• The domain decomposition cell system was renamed to regular decomposition (#4442). The system.cell_system.set_domain_decomposition() function was renamed to system.cell_system.set_regular_decomposition().
• Bonds are now immutable (#4350). Bonds added to the list of bonds can no longer be overwritten by a bond of a different type, as it could lead to undefined behavior when the number of bonded partners was higher in the overwriting bond. Bonds can now be removed from the list of bonds, after they have been removed from particles.
• Observable parameters are now immutable (#4206, #4211).
• The Electrokinetics actor parameters are now immutable (#4327).
• The LBFluid, LBFluidGPU, Electrokinetics and Species methods print_*() have been renamed to write_*() (#4049).
• The ELC actor is no longer an electrostatics extension (#4125, #4506). The ELC actor now takes a P3M or a P3MGPU actor as argument and modifies it. Only the ELC actor needs to be added to the system list of actors. The ELC actor can now be removed from the list of actors.
• The DLC actor is no longer a magnetostatic extension (#4506). The DLC actor now takes a magnetostatic actor as argument and modifies it. Only the DLC actor needs to be added to the system list of actors.
• The NpT thermostat now uses the Philox random number generator and requires a random seed on first instantiation (#3444).
• The analysis module energy() function now returns the lower triangle of the non-bonded interaction matrix, to be consistent with pressure() and stress_tensor() (#3712).
• The analysis module energy(), pressure() and pressure_tensor() functions now return only two slots for electrostatics and magnetostatics: short-range contribution in the first slot and long-range contribution + layer correction in the second slot (#3770).
• The analysis module pressure() and pressure_tensor() functions no longer provide a velocity-compensation flag to compute the pressure at half the time step in NpT simulations (#3756).
• The espressomd.reaction_ensemble module was renamed to espressomd.reaction_methods (#4482).
• The argument temperature in reaction methods was renamed to kT for clarity (#4305).
• All reaction methods now take keyword arguments instead of positional arguments (#4451).
• The constant pH method now implements a symmetric proposal probability instead of an asymmetric proposal probability (#4207).
• The reaction method parameter exclusion_radius was renamed to exclusion_range (#4469).
• Reaction method now take an optional parameter exclusion_radius_per_type for better control of the exclusion radius in simulations involving different particle sizes (#4469).
• The WidomInsertion.measure_excess_chemical_potential() method was replaced by WidomInsertion.calculate_particle_insertion_potential_energy(), which returns the instantaneous value of the excess chemical potential instead of the accumulated mean and standard error (#4374). The mean value and standard error of the excess chemical potential must be now be calculated by WidomInsertion.calculate_excess_chemical_potential().
• Reaction method constraints can now be safely changed from cylindrical to slab and can be removed (#4310). They will also raise an error when created with invalid parameters.
• The mpiio global variable was removed (#4455). The MPI-IO feature is now used by creating a local instance with mpiio = espressomd.io.mpiio.Mpiio().
• The MPI-IO feature now raises an exception from which the user can recover when the simulation script runs on 1 MPI rank, instead of an unrecoverable fatal error (#4455). This change is meant to help debugging read/write errors in simulation scripts. On 2 or more MPI ranks, exceptions still lead to a fatal error.
• The H5md class takes new arguments during instantiation (#3785).
• The system.cell_system.get_pairs_() method was renamed to system.cell_system.get_pairs() and now supports filtering particle pairs by type (#4035).
• The polymer setup code was moved from the core to Python (#3477). The espressomd.polymer.positions() function was renamed to espressomd.polymer.linear_polymer_positions() and the espressomd.diamond.Diamond class was converted to function espressomd.polymer.setup_diamond_polymer(). For diamond polymers, counter-ions must now be added manually by the user.
• The particle director can be set from the Python interface (#4053).
• The particle method vs_auto_relate_to() can take a particle as argument instead of a particle id (#4058).
• Particles can be serialized and deserialized in the Python interface with particle_dict = p.to_dict() and system.part.add(particle_dict) (#4060).
• It is no longer necessary to manually reshape the output of Observable objects. The Observable classes now return multi-dimensional numpy arrays and the documentation clearly indicates the shape and size of the calculated data (#3560). The same applies to accumulators and time series (#3578).
• Accumulator and Correlator classes now return the data in suitably shaped multi-dimensional numpy arrays; dependent properties such as lag times and sample sizes need to be obtained separately by calling methods lag_times() resp. sample_sizes() (#3848).
• Profile observables provide methods bin_centers() and bin_edges() to facilitate plotting (#3608).
• The observable ComForce was renamed to TotalForce, so as to better reflect what it actually calculates (#3471).
• The RDF feature was removed from the analysis module and converted to an Observable class (#3706). Time averages can be obtained using the TimeSeries accumulator.
• All occurrences of "Stress Tensor" in the analysis module, LB module and EK module were renamed to "Pressure Tensor" to better reflect what is actually calculated (#3723, #4228).
• The MeanVarianceCalculator interface changed (#3996).
• Observables now check their input parameters (#4211, #4255) and raise an exception when an invalid value is detected (e.g. min_x > max_x in profile-based observables).
• Cylindrical observable classes have an extra transform_params argument to change the orientation of the cylindrical coordinates systems and control the origin of the phi angle (#4152).
• Incompatible thermostat/integrator combinations raise an exception (#3880).
• The system.cuda_init_handle.list_devices() feature is now a function, and the system.cuda_init_handle.list_devices_properties() function disabled in 4.0.0 was restored (#4095).
• CUDA errors now halt the flow of the program by throwing a Python exception with a clear error message (#4095).
• Parameter particle_scales of coupling-based fields PotentialField and ForceField now takes a dict object instead of a list of tuples (#4121).
• The System class no longer has a globals member (#4276). Global variables are still accessible from other members of the System class.
• Methods from the cluster analysis class Cluster no longer returns False when a string passed to call_method() doesn't match the name of a core method; instead None is returned (#4234).
• Methods from the cluster analysis class ClusterStructure, integrator classes and interaction classes no longer returns True when the corresponding core method doesn't return a value; instead None is returned (#4234, #4516).
• Several parameters of the ICC class are no longer optional: epsilons, normals, areas, sigmas (#4162).
• The electrostatic actors charge neutrality check tolerance can be modified via actor.charge_neutrality_tolerance; this is mostly relevant to actors coupled to ICC, whose induced charges can have values spanning several orders of magnitude (#4506).
• Electrostatic and magnetostatic methods that support tuning now have a timings argument to control the number of integration loops to run during tuning (#4276).
• The Drude helpers (global variables and free functions) have been gathered into a checkpointable class DrudeHelpers, which now relies on particle handles instead of particle ids (#4353).
• ScaFaCoS integration now supports activating an electrostatics ScaFaCoS actor at the same time as a magnetostatics ScaFaCoS actor (#4036).
• The list of actors can no longer end up in an invalid state: updating an electrostatic or magnetostatic actor with an invalid parameter now automatically restores the original parameter; inserting an invalid electrostatic or magnetostatic actor in the list of actors now automatically removes the actor from the list of active actors, even when the exception happens during tuning (#4506).
• The LBBoundaries slip velocity check was lowered to Mach 0.35, or 0.2 in LB units (#4376).
• The OpenGL visualizer allows changing the radius of LB velocity arrows, documents all LB-related keyword arguments, and no longer suffers from a division-by-zero error that used to trigger a runtime warning for fluid inside boundaries (#4376).
• The Electrokinetics class got an optional ext_force_density parameter for consistency with other LB implementations (#4203).
• MDAnalysis integration now checks if the MDAnalysis package version is supported (#4386).

Removed functionality

• The ENGINE shear torque calculation feature deprecated in 4.1.1 was removed (#3277).
• The MEMBRANE_COLLISION and OifOutDirection features were removed (#3418).
• The AFFINITY feature was removed (#3225).
• The unused and untested UMBRELLA feature was removed (#4032, #4079).
• The unused and untested VIRTUAL_SITES_COM feature was removed (#3250).
• The unused and untested EK_DOUBLE_PREC feature was removed (#4192).
• The unused and untested MD metadynamics feature was removed (#3563).
• The unused and untested Stomatocyte shape was removed (#3730).
• The PdbParser feature deprecated in 4.1.1 was removed (#3257).
• The incorrectly implemented and untested HarmonicDumbbellBond interaction was removed (#3974, #4079).
• The layered cell system was removed (#3512).
• The unused Wang-Landau reaction ensemble algorithm was removed (#4288).
• The reaction ensemble tutorial deprecated in 4.1.1 was removed (#3256).
• The per-particle temperature feature was removed (#4057).
• The Current observable was removed in favor of the FluxDensityProfile observable (#3973).
• The incorrectly implemented analysis function cylindrical_average was removed in favor of the CylindricalDensityProfile observable (#3470).
• The minimize_energy member of the System class was removed (#3390, #3891). The steepest descent algorithm is now a regular integrator that is set up via the system.integrator.set_steepest_descent() method.
• The MMM2D electrostatics feature was removed (#3340). Electrostatics in slab geometries can still be achieved by ELC, with significantly better performance.
• The dipolar direct sum with replica method is now disabled on periodic systems with zero replica, as it does not apply minimum image convention (#4061).
• The analysis module min_dist2() function was removed and the dist_to() function was merged into system.distance_vec() (#3586).
• The analysis module nbhood() function slab search mode was removed (#4516) since it was incorrect (all ESPResSo versions were affected).
• The number of cells for the link cell algorithm can no longer be constrained to a range of values (#3701).
• The global Mersenne Twister RNG was removed (#3482). All thermostats are now Philox-based. Local Mersenne Twister RNGs are still used in the linear polymer position generator (now with proper warmup) and in the ReactionAlgorithm class.
• It is no longer possible to checkpoint an ESPResSo system instance that contains Union shape-based constraints when the simulation is running with 2 or more MPI ranks. An error will be raised (#4287, #4510).
• It is no longer possible to checkpoint an Electrokinetics instance (#4327).
• The unmaintained lj-demo.py sample was removed (#4482).
• The unmaintained mayaviLive visualizer was removed (#4515).

Improved testing

• The C++ core of ESPResSo is covered by unit tests and integration tests at 98% coverage (#4426, #4479, #4489).
• The structure factor code is tested against simple lattices (#4205).
• The MMM1D GPU code is tested (#4064).
• The reaction method core classes are unit tested (#4164).

Performance enhancements

• The Particle memory footprint was reduced and the MPI serialization code was improved (#4414). The structure size is now 584 bytes instead of 640 bytes on maxset configuration (10% reduction). All substructures in Particle are bitwise serializable and dynamic vectors are compact vectors. The performance gain is about 9% for a LJ liquid on both maxset and empty configurations, for both 1 000 and 10 000 particles per core.
• Particle creation happens in constant time instead instead of monotonically increasing with the number of particles already in the system (#4493).
• When only one MPI rank is used, the maximum cutoff of bonded interactions is ignored when initializing the cell properties, since the bond partners are always accessible on the same node, regardless of the cell size; if the system also doesn't have short-range interactions, the short-range loop is skipped (#4452).
• The ReactionAlgorithm::do_reaction() function used by reaction methods now caches the potential energy of the system and only updates it after a successful reaction trial move (#4374).
• Reaction methods can delegate the particle neighbor search to the cell system when evaluating the exclusion range of inserted particles (#4401). This leads to better performance only on 2 or more MPI ranks.

Bug fixes

• The transform_vector_cartesian_to_cylinder() now calculates the correct phi angle (#4094). The bug was present since ESPResSo 4.0.0 and affected observables CylindricalVelocityProfile, CylindricalFluxDensityProfile, CylindricalLBVelocityProfile, CylindricalLBVelocityProfileAtParticlePositions, CylindricalLBFluxDensityProfileAtParticlePositions.
• Several memory leaks were fixed in the TabulatedBond interactions (#3961), electrostatics and magnetostatics tuning functions (#4069), lattice-Boltzmann code (#4108) and Barnes-Hut code (#4404).
• The system.actors.clear() method was broken and would only remove half of the actors since 4.0.0. This is now fixed (#4037).
• The ClusterStructure feature did not properly handle box periodicity since 4.0.0 and would under rare circumstances calculate a center of mass to be outside a fully periodic simulation, and would incorrectly fold coordinates in aperiodic systems. This is now fixed (#4363).
• Adding a LB thermostat when any other thermostat was already active would silently fail since 4.0.0. This is now fixed (#4116).
• Setting the NpT or steepest descent integrators with incorrect parameters no longer leaves the system in an undefined state (#4026).
• The OpenGL visualizer had a tendency to slow down after pausing and resuming the simulation, or freezing when using the steepest descent integrator. This was due to a race condition between two threads that has been fixed (#4040).
• The OpenGL visualizer no longer raises an exception when activating the LB_draw_boundaries option without any other LB_draw_* option (#4479).
• The OpenGL visualizer now correctly updates bond information when the collision detection and bond breakage features are used (#4502).
• It is no longer possible to accidentally set a non-cubic NpT integrator with P3M (#4165).
• The NpT integrator used to work with P3MGPU even though it didn't implement long-range energy calculation and therefore couldn't contribute to the virial; now the long-range energy is calculated and added to the virial (#4026, #4506).
• Illegal LB node access is now properly caught by exceptions (#3978).
• EK node access no longer accepts floating-point values for node indices (#4228), and always requires exactly three integers (#4482).
• Accessing the flux property of EK species no longer throws an error (#4106).
• Accessing the boundary field of LB nodes from a LBFluid actor when LB_BOUNDARIES is not compiled in now returns 0 instead of a random integer (#4479).
• The LB grid in the GPU implementation is now automatically resized when the simulation box size changes (#4191).
• The LB code now throws an error when adding a LB boundary to the LBFluid actor when LB_BOUNDARIES is not compiled in, or to the LBFluidGPU actor when LB_BOUNDARIES_GPU is not compiled in (#4472).
• The lattice-Boltzmann Python interface no longer ignores runtime errors, nor converts them to cryptic system errors (#4355).
• The script interface no longer silently ignores runtime errors when converting Python objects to C++ data types (#4387, #4492).
• The system now throws an error when a non-bonded interaction cutoff is too large for the local box size in MPI-parallel simulations; in older releases the error was queued and deferred to the integration loop (#4479).
• The system now throws an error when a virtual site tracks a real particle too far away for the local box size in MPI-parallel simulations; in older releases the error was queued and deferred to the integration loop (#4479).
• It is no longer possible for a virtual site to track itself (#4479).
• It is no longer possible for a particle to exclude itself (#4493).
• It is no longer possible to accidentally add the same bond twice on the same particles (#4058).
• Fatal errors triggered by stale references in virtual sites, invalid particle ids and null quaternions have become runtime exceptions (#4479).
• Virtual sites now contribute to the rotational kinetic energy of the system (#4198).
• Particle creation no longer raises numpy.VisibleDeprecationWarning (#4493).
• The EK feature now generates VTK files that are compliant with the VTK 2.0 standard (#4106).
• The ELC and DLC actors now throw an error when a particle enters the gap region (#4051).
• The ELC actor is now updated when the box size changes in the z-direction (#4231).
• The DLC actor now raises an exception when tuning fails instead of causing a fatal error (#4238).
• The MMM1D actor now raises an exception for incorrect periodicity or cell system instead of causing a segfault (#4064).
• The DipolarP3M checkpointing mechanism was fixed (#3879).
• The DipolarP3M method now recalculates the energy correction when the box length changes (#4506).
• P3M-based actors now sanitize the user-provided alpha and accuracy parameters and no longer allow constraining the alpha parameter during tuning (alpha was always derived from the other parameters at the end of tuning) (#4118).
• A buffer overflow in the DipolarP3M tuning function lead to random failures during tuning, this is now fixed (#3879).
• A buffer overflow in the LB code could lead to incorrect results in grids of size 9x9x9 or larger with open boundaries, this is now fixed (#4078).
• Providing incorrect parameters to the ScaFaCoS actors no longer cause ESPResSo to crash (#4068).
• FENE, harmonic and quartic bonds now throw an error when the bond length is zero and the equilibrium bond length is non-zero, since the direction of the force cannot be determined (#4471).
• Immutable parameters default_scale, particle_scales and gamma of coupling-based fields PotentialField, ForceField, FlowField and HomogeneousFlowField now throw an error when an attempt is made to change their value via the class setter, instead of silently ignoring the new value (#4121).
• The CylindricalLBFluxDensityProfileAtParticlePositions observable now measures the correct quantity (#4152).
• The Boost 1.74 bug was patched (#3978).
• A bug involving an access out of bounds was fixed in the structure factor code (#4205).
• A bug in the collision detection feature that lead to a harmless warning being printed to the terminal upon collision was fixed (#4484).
• Calling collision_detection.set_params() with invalid arguments no longer leaves the collision detection feature in an indeterminate state; the previous state is automatically rolled back (#4484).
• Setting the collision detection mode glue_to_surface or bind_at_point_of_collision when feature VIRTUAL_SITES_RELATIVE is not compiled in now generates the correct error message (#4484).
• Passing a particle chain-based observable object (ParticleDistances, BondAngles, BondDihedrals, CosPersistenceAngles) that doesn't have enough particle ids for the calculation (e.g. only 1 particle id when 2 are needed for the bond distance calculation) to a Correlator object no longer causes a memory overflow (#4255).
• Calculating the energy of the system when an IBM object is present no longer terminates ESPResSo, instead a warning is issued (#4286).
• The Sphere shape no longer returns NaN values in the distance vector for particles located exactly in its center (#4384).
• Runtime errors raised when the maximal bonded interaction range becomes larger than the simulation box are no longer ignored when dihedral bonds are added to the list of interactions (#4383).
• Runtime errors about incorrectly initialized electrostatic/magnetostatic methods are no longer silently ignored at integration start (#4383).
• Runtime errors about incorrectly initialized GPU dipolar direct sum and Barnes-Hut are no longer silently ignored when the actors are instantiated (#4404).
• A bug that could potentially lead to stale references in the script interface was fixed (#4476).
• TabulatedNonBonded.is_active() now returns False instead of None when the interaction is inactive (#3586).

Under the hood changes

• The Python code is now checked with Pylint to prevent the introduction of unused code and dangerous anti-patterns (#3293, #3203).
• The CMakeLists.txt files are now formatted automatically with cmake-format (#3622).
• The Python code and C++ code were checked with LGTM to detect and fix coding errors and anti-patterns (#3851, #3856, #4300).
• Compiler warnings and diagnostics from GCC 11, 12, from Clang 10, 12, 13, 14 and from Intel 19.0.4 were addressed (#4084, #4426, #4510, #4526).
• The Particle struct was moved to a dedicated header file Particle.hpp to improve separation of concerns in the core (#3251, #3164).
• The Observable_stat structs were moved to a dedicated header file Observable_stat.hpp and decoupled from the pressure/energy/coulomb/dipolar frameworks (#3712) and made stateless (#3723).
• Observables based on particle ids have been rewritten using particle traits to decouple the Particle struct from Observable classes (#3667).
• The Python Integrator class was split into multiple classes, one for each integrator, with a structure similar to actor and interaction classes (#3390). This layout better reflects the structure of integrators in the core and will make it easier to include new integrators in the future. This change doesn't break the API.
• The ghost communication infrastructure was simplified (#3216, #3399).
• The LB coupling for the regular decomposition scheme was rewritten (#4470).
• Thermostats are now fully object-oriented in the core to reduce code duplication (#3438, #3444, #3461).
• Bonded interactions are now fully object-oriented in the core to facilitate the development of new interactions (#4161).
• Bonded interactions are now communicated between MPI processes automatically and transparently by the script interface (#4350).
• The custom MpiCallbacks framework has been simplified and the callbacks made more homogeneous (#4383).
• The custom MpiCallbacks framework is being progressively replaced by boost::mpi communication (#4506, #4511).
• The local_particles global variable is no longer accessible directly (#3501).
• The Python tests now use specialized assertions to generate more helpful error messages (#3419).
• The tutorial tests were simplified using AST to parse Jupyter notebooks (#3408).
• The CMake logic for tutorials has been simplified (#3408, #3486).
• The Cython interface was thoroughly cleaned up from unused imports (#3496, #3510).
• The ScriptInterface framework was rewritten (#3794).
• The ScriptInterface framework is now the preferred way to implement new features. Existing features were converted to ScriptInterface objects: bonded interactions (#4350), bond breakage (#4464), collision detection (#4484), reaction methods (#4451), MPI-IO (#4455), H5MD (#4480), cell system (#4511), actors, scafacos, electrostatics and magnetostatics (#4506). The corresponding Cython files were converted to Python files.
• It is now possible to extend the list of available specifications in the H5MD feature at the C++ level (#4480).
• The duplicated functions between P3M and DipolarP3M were factored out (#3879).
• Statistical tests are no longer executed in coverage and sanitizers builds (#3999).
• The Utils::Mpi::gather_buffer() function was fixed (#4075). The bug didn't affect ESPResSo.
• Parameters can be passed to CTest at configuration time via the new CTEST_ARGS CMake option (#3862). This replaces the deprecated and non-portable ARGS Makefile variable expansion.
• A superfluous and non-portable CMake target_compile_options() statement was removed (#3852).

4.1.4 - 2020-10-19

Bug fixes

• Fix a bug in the LB CPU implementation that lead to incorrect LB shear stress tensors in thermalized fluids since 4.1.0 (#3847)
• Fix a bug in the LB CPU and GPU implementations that lead to gamma_bulk being used in place of gamma_shear in the second order term for forces in all previous ESPResSo versions (#3885)
• Fix a bug that always set the epsilon value to zero in P3M and P3MGPU actors since 4.0.0 (#3869)
• Fix an issue in the Python script interface that rejected integer epsilon values in the P3M and P3MGPU actors, and the 'metallic' epsilon value in the P3M actor (#3869)
• Fix the exception mechanism in the P3M, P3MGPU and DipolarP3M code to forward errors to the Python interface instead of silencing them or running infinite loops (#3869)
• Fix range checks in the OIF code that failed to raise TypeError exceptions (#3846)

Documentation and tutorials corrections and improvements

• Explain the discrepancy between the Gay-Berne formula from the user guide and from the original paper (#3839)
• Add note explaining the P3M algorithm works with non-metallic epsilon values only when the box is cubic (#3869)

Build system and platform-related corrections and improvements

• Fix several CMake warnings raised by CMake 3.17 and above (#3830, #3859)

Improved testing

• Add a test to check the off-diagonal elements of the LB stress tensor in long simulations of thermalized fluids (#3847)

Under the hood changes

• Remove unnecessary memory allocation on GPU from MPI worker nodes (#3911)

4.1.3 - 2020-07-08

Feature configuration at compile time

• The number of features which need to be defined at compile time in myconfig.hpp has been reduced. Features without performance impact are now always present. These are:
  • OIF_LOCAL_FORCES
  • OIF_GLOBAL_FORCES

Bug fixes

• Many bonded interactions were not considered in the bond cutoff calculation: umbrella, OIF local, OIF global, IBM tribend, IBM volcons, angle harmonic, angle cosine, angle cossquare, tabulated angle, bonded Coulomb, subtracted bonded Coulomb, subtracted LJ, quartic, harmonic dumbbell. This can lead to sub-optimal skin values when such bonds are used with an inter-particle distance that is longer than other bonded (FENE, harmonic bond, rigid bond, thermalized distance, tabulated bond, tabulated dihedral, dihedral, IBM triel), non-bonded (LJ, Morse, Buckingham, etc.) and long-range (electrostatics, magnetostatics) interactions in the same system. All bonded interactions are now considered in the cutoff calculation (#3443).
• Fix a bug in a rotation function that resulted in improper treatment of rotation vectors with norm different from unity (#3559); all observable classes inheriting from CylindricalProfile are affected
• Fix a bug in the LB GPU implementation that lead to incorrect velocity interpolation near LB boundaries (#3593)
• Fix a bug in the LB CPU implementation that lead to incorrect grid sizes (#3678)
• Object-in-fluid bugfixes have been backported from the OIF development branch; in particular, the bending force between two triangles is now torque-free (#3385)
• Rewrite the linear polymer position generator, which was inefficient and frequently rejected valid positions (#3402, #3484, #3491)
• Fix an error in the distance calculation of the SpheroCylinder shape (#3629)
• Fix a sign flip in the surface normal calculation of the Torus shape (#3728)
• Fix an IndexError when running system.number_of_particles() without a value for the argument type (#3496, #3494) and fix the range check (#3536)
• Fix a NameError when running system.analysis.rdf() without a value for the argument r_max (#3496, #3494)
• Fix a NameError raised by the OpenGL visualizer when drawing bonds in periodic images of the unit cell (#3511)
• Correctly calculate the orientation of bonds cut by the faces of the simulation box in the OpenGL visualizer (#3511)
• Fix a memory leak in the OpenGL visualizer when drawing shapes containing cylindrical elements (Cylinder, SpheroCylinder, SimplePore, Slitpore) and drawing bonds between particles (#3533)
• Fix an issue in the OpenGL visualizer that drew the channel of the Slitpore shape at the center of the box, instead of using the dividing_plane attribute (#3728)
• Fix a bug in the ELC algorithm that ignored the Coulomb prefactor (#3731). The same bug is also present in MMM2D but could not be fixed.
• Correctly check the P3M parameter mesh (#3676)
• The LB checkpointing argument binary now takes a boolean value (#3541); integers values 0 and 1 are still accepted (integers are implicitly cast to boolean values)
• Reinitialize the P3M and dipolar P3M solvers when the box size or skin changes (#3717)
• Clarify error messages in the Steepest Descent integrator (#3763)
• Fix an incorrect formula in the tensor_product mode of the Correlator class that always returned an array of 0's since 4.1.0 (#3781)
• Fix a runtime error when calling the get_params() method of a ScaFaCoS-based actor (#3784)

Documentation and tutorials corrections and improvements

• Fix paragraph formatting in Jupyter notebooks and update Sphinx bibliography (#3395).
• The Sphinx documentation generation doesn't run in parallel any longer due to plugin sphinxcontrib.bibtex throwing a warning when executed with more than one thread in Sphinx v2.3.1 (#3393). The slowdown is not significant (less than a second).
• Fix compatibility issues with Sphinx 2.4.0 (#3480), 3.0.1 (#3642, #3659) and 1.6.7 (#3743)
• Clarify the quaternion formalism used in ESPResSo (#3392, #3748)
• The p3m.py sample showcased an incorrect usage of the ELC actor (the gap region was missing). The actor was removed and a new, stand-alone sample visualization_elc.py was created (#3468)
• The visualization_constraints.py sample showcased an incorrect usage of the Slitpore and Wall shapes that lead to a discontinuous potential; this is now fixed (#3728)
• Correct errors in the documentation of the constructor parameters for shape classes Cylinder, SpheroCylinder, Rhomboid (#3567), for class System (#3542) and for cylindrical observables (#3569)
• Correct an error in the formula of the electrostatic prefactor in the electrostatics documentation, give the full expression of the electrostatic prefactor in tutorials and samples (#3673)
• Improve documentation of the Slitpore shape and document the Torus shape (#3728)
• Improve installation instructions (#3673, #3699, #3732)
• Document BoxGeometry-related functions (#3747)
• Explain release workflow and how to obtain released versions of ESPResSo (#3745)
• Improve citation instructions with examples (#3745)
• General improvements (#3740, #3743)

Build system and platform-related corrections and improvements

• The benchmarks can now be run with any MPI library supported by ESPResSo (#3412)
• The CMake logic was simplified (#3574, parts of #3582). The minimal required Cython version is now checked. CMake now generates an error message if WITH_CLANG_TIDY is ON but no Clang-Tidy can be found with a version matching the Clang version. The CUDA library installed via the Ubuntu package nvidia-cuda-toolkit is now correctly detected.
• Add support for ROCm versions 3.0 (#3386), 3.1 (#3574) and 3.3 (#3623)
• Fix compiler errors with HDF5 > 1.10.2 (#3604)
• Fix compiler errors with Boost 1.73 (#3725)
• Fix a deprecation warning from the collections.abc that will become an error in the upcoming Python 3.9 interpreter (#3568)
• Fix a compatibility issue with pint 0.10.1 in tutorial "12 - constant pH" (#3423)

Improved testing

• Fix a tolerance value that was incorrectly divided by 100, causing unit tests to fail on i586 architectures (#3427)
• Compile CUDA code in the Travis-CI image to detect more compiler errors (#3699). GPU tests are skipped on Travis-CI.
• Add a test for the Utils::get_n_triangle function used in OIF (#3391)
• Add a test for sample visualization_constraints.py (#3533)
• Add missing LENNARD_JONES and GPU feature guards in Python tests (#3403, #3676)
• Fix a few non-functional Python tests (#3419) and sample tests (#3791)
• Improve testing of ELC (#3731)
• Improve testing of the Slitpore shape (#3728)
• Fix an issue in a core test (#3677)
• Add cleanup function in the checkpointing tests (#3699)
• Add a test for fold_position() (#3747)
• Improve testing of observables, correlators and accumulators (#3781, #3783, #3784)

Under the hood changes

• Remove unused code (#3556, #3738)
• Remove the unused FindPythonModule CMake module (#3736)
• Update the espresso-ci bot scripts (#3613)

4.1.2 - 2019-12-13

Bug fixes

• Remove correlation between the rotational noise and translational noise in the Langevin thermostat (#3355)
• Fix a bug that may cause the wrong temperature to be set by the Langevin and DPD thermostats in the first time step after the system was altered from the Python level, e.g., by changing particles or interactions (#3341)
• Fix a bug that caused the DPD thermostat to generate an incorrect velocity distribution when used together with the Langevin thermostat (#3352)
• Fix a bug in MMM2D and ELC with potential boundary conditions, where one of the correction factors was over-counted resulting in wrong energies (#3310)
• Fix a bug that caused the wrong bonds to be deleted when removing particles from the system (#3356)
• Fix an ambiguity in ParticleSlice: the values in the square brackets refer to particle ids, not array indices (#3367). This means the ill-defined syntax system.part[0:-1] is no longer valid. See the User Guide section on Setting up particles for more information.
• Remove the mass prefactor in the ComForce observable and use the correct Particle ids in the ParticleAngularVelocities and ParticleBodyVelocities observables (#3380)
• Fix a rounding error that caused debug builds of ESPResSo running with multiple MPI threads to crash when a particle was placed exactly on the boundary between two cells (#3377)
• Fix espressomd.has_features() for the corner case where the list of all compiled-in features is passed as argument, returning False instead of True (#3318)
• Refactor the random number generator code (#3349)
• Minor fixes (#3351, #3336)

Documentation and tutorials corrections and improvements

• Improve documentation of Monte Carlo methods (#3254, #3330)
• Minor fixes (#3342, #3334)

Build system and platform-related corrections and improvements

• List all Python dependencies in requirements.txt with the supported version numbers (#3300). Please note that most of them are optional.
• Add MPIEXEC_PREFLAGS and MPIEXEC_POSTFLAGS to the command lines of parallel tests (#3221)
• Add the -oversubscribe flag to the command lines of parallel tests running with OpenMPI v2.X to avoid exiting early from a Python test configured without MAX_NUM_PROC on a machine with a hyperthreaded CPU where OpenMPI is configured such that the number of threads cannot exceed the number of cores (#3335)
• Refactor the CI, maintainer, Doxygen and pypresso shell scripts to make them more portable and support filepaths containing whitespaces (#3326, #3373)
• Fix a nvcc compiler warning on the empty config (#3329)

Improved testing

• Add a test for ELC and MMM2D using analytic expressions of the force and energy (#3331)
• Sped-up seven Python tests (#3319)
• Fix a test that broke on s390x architectures with Fedora 31 (#3312)
• Fix tests that broke on i586 architectures with OpenSUSE Tumbleweed (#3327, #3358)

4.1.1 - 2019-11-13

Interface changes

• Integrator.set_isotropic_npt(): input value direction=[0,0,0] now throws an error instead of being silently changed to [1,1,1]
• ParticleHandle.swimming: deprecated value 'rotational_friction' is now disabled

Bug fixes

• Restore checkpointing mechanism for the steepest descent and NPT integrators, LB and NPT thermostats (#3245)
• Increase the minimum MPI version to 3.0; OpenMPI versions 1.6.5 and lower are no longer supported (#3236)
• Fix Integrator.set_isotropic_npt(): remove the silent conversion of the incorrect input parameter direction=[0,0,0] to [1,1,1] in the core; the function now throws an exception for fixed-volume boxes; this change is unlikely to break pypresso scripts since not providing a value to direction or providing [1,1,1] were the two standard ways to set up a box with all directions allowed to rescale (#3253)
• Fix Integrator.set_vv(): this function failed to set the velocity Verlet integrator if the NPT integrator was active; this is now resolved (#3274)
• Fix the random segmentation fault triggered by the removal of a particle with a bond or a virtual site relationship to another particle (#3288)
• Fix system.part.writevtk(): the function now writes down all particles when using types="all" (#3290)
• Disable the deprecated and broken ENGINE shear torque calculation feature; the feature will be completely removed from the core in the upcoming 4.2 release (#3277)
• Fix unit conversion for the LB fluid viscosity (#3287)

Documentation and tutorials corrections and improvements

• Add more detailed installation instructions for ESPResSo and its Python dependencies on MacOS X (#3236)
• Add links to Dockerfiles providing installation instructions for ESPResSo and its Python dependencies on CentOS 7, Fedora 30, Debian 10 and OpenSUSE Leap 15.1 (#3244)
• Add instructions to read PDB files with MDAnalysis, which is one of the recommended tools to read/write molecular dynamics topologies and trajectories in ESPResSo; the PdbParser feature will be removed in the upcoming 4.2 release (#3257)
• Add a new tutorial on the constant pH method; the reaction ensemble tutorial will be removed in the upcoming 4.2 release (#3184)

Build system and platform-related corrections and improvements

• Fix a PYTHONPATH error when ESPResSo is built in a directory containing whitespace characters (#3238)
• Fix several issues with the command make install that lead to import errors in Python (incorrect runtime path, missing shared objects, name collision for submodule cluster_analysis) and deprecate the make install DESTDIR=/path/to/espresso command in favor of the standard cmake .. -DCMAKE_INSTALL_PREFIX=/path/to/espresso command (#3228), install espressomd module in a platform-dependent python path, typically lib{,64}/python3.X/{dist,site}-packages (#3258)
• Fix an issue in mpiio that triggered an assertion in systems with no bonds when ESPResSo is built with stdlibc++ range checking enabled (#3234)
• Fix the pypresso script to correctly parse filepaths containing whitespaces passed after a pypresso flag, such as --gdb, and make conditional statements cross-platform (#3292)

Improved testing

• Test checkpointing of integrators and thermostats (#3245)
• Fix and improve check_cmake_install test (#3228, #3258) and add a new CI job to test an installed version of ESPResSo (#3228)
• Test engine LB (#3277)
• Add more LB tests (#2748)

4.1.0 - 2019-10-01

This is a feature release, i.e., new functionality is added to ESPResSo. An additional focus of this release is quality assurance and modernization. The testing of ESPResSo's functionality has been extended considerably. Also, sample and tutorial scripts are now automatically tested. Moreover, a large effort was put into modernizing the C++ simulation core. Work has been done, e.g., on particle sorting, the MPI communication infrastructure, and the lattice-Boltzmann implementations. Electrostatic and magnetostatic methods now have a clear and common interface. These changes will facilitate future extensions of ESPResSo and make the code more understandable to new developers.

Changed requirements

• Python 2 support has been dropped. ESPResSo now requires Python 3.
• ESPResSo now needs a C++14-capable compiler, such as GCC 4.9 and later or Clang 4 and later.
• It is discouraged to use ESPResSo with Boost versions below 1.67.

Added functionality and documentation

• The distance between a shape (such as sphere) and a position can now be queried via shape.calc_distance().
• The lattice nodes of a lattice-Boltzmann fluid can now be iterated using LBFluid.nodes().
• A tutorial on magnetic fluids has been added.
• The stress created by the dissipative particle dynamics interaction (DPD) can now be obtained via the DPDStress observable.
• The stress of a lattice-Boltzmann fluid can now be obtained via the LBFluidStress observable.
• A torus shape has been added.
• Two new accumulators for observables have been added: MeanVarianceCalculator and TimeSeries.
• An ElectricPlaneWave constraint was added.
• Experimental support for AMD GPUs via HIP. The future of this feature is unclear. Please do not base hardware buying decisions on its presence.
• Visualization of slit pores in the OpenGL visualizer.
• A Weeks-Chandler-Anderson short-range potential has been added (WCA).
• The external force density applied to a lattice-Boltzmann fluid can now be changed during the simulation.
• Sanity checks for Mach limits and unequal MD and lattice-Boltzmann time steps have been added.
• The system.cell_system.tune_skin() method now has a keyword argument adjust_max_skin. If set to True, the maximum skin to be tested will be reduced such that it is compatible with the local box size.
• For the CPU lattice-Boltzmann implementation, the limit on the Verlet list skin (less than agrid / 2) has been lifted.
• A new observable CosPersistenceAngles has been added for the bond angles of a polymer (needed, e.g., for determining the persistence length).

Feature configuration at compile time

• The number of features which need to be defined at compile time in myconfig.hpp has been reduced. Features without performance impact are now always present. These are:
  • PARTIAL_PERIODIC
  • LB
  • LB_GPU (for builds with CUDA)
  • IMMERSED_BOUNDARY
  • BOND_ANGLE
• For most compilers, it is checked that only known features are declared in myconfig.hpp.
• The default feature configuration applied when no myconfig.hpp is present has been extended significantly. In particular, all tutorials can now be run with the default feature configuration. For production simulations, it is still recommended to use a custom myconfig.hpp containing only necessary features. This is true in particular, if particle rotation is not needed.
• Features that do not have automated tests now require EXPERIMENTAL_FEATURES to be defined in myconfig.hpp.

Interface changes

• Several parts of ESPResSo now use a method-specific seed for random number generation. For the following individual methods a random number seed has to be passed using the keyword argument seed:

  • Langevin thermostat
  • DPD thermostat
  • LB Thermostat
  • Thermalized bond
  • Wang-Landau reaction ensemble
  • Constant pH ensemble
  • Widom insertion method

• Changes in the lattice-Boltzmann (LB) interface:

  • By default the LB fluid is not thermalized. A temperature can be set using the LBFluid's keyword parameter kT. If kT > 0, an additional seed keyword parameter has to be provided.
  • The LB thermostat gets its temperature from the LB fluid.
  • The frictional coupling coefficient gamma is now a keyword parameter of the LB thermostat. For more detailed information on how to set up a LB fluid and thermostat, please see chapter lattice-Boltzmann.

• The method for polymer creation has been replaced. Now espressomd.polymer.positions() can be used to obtain particle positions for one or more polymer chains. Based on these positions, polymers can be created. For an example please see Setting up polymer chains.

Changed and removed functionality

• The remove_total_momentum() method for lattice-Boltzmann fluids has been removed. The overall velocity of a fluid can be changed using the lb_fluid.nodes() iterator.
• The CATALYTIC_REACTIONS feature has been removed.
• The method for creating a polymer has been replaced. espressomd.polymer.positions() can now be used to obtain particle positions for one or more polymer chains.
• Checkpointing has been added for the electrokinetics method.
• The global random number seed has been partly replaced by method-specific ones. These are specified when activating the relevant feature such as the Langevin, DPD and lattice-Boltzmann thermostats via a seed keyword argument.
• The random number generator has been switched to Philox for most algorithms requiring random numbers.
• Limitations on the exclusion radius have been relaxed in the reaction ensemble method.
• A new observable CosPersistenceAngles has been added for the bond angles of a polymer (needed, e.g., for determining the persistence length).
• ELC has been disabled for non-neutral systems with constant potential.
• The calculation of the linear particle momentum included the forces of the last time step. The function system.analysis.linear_momentum() now returns the sum of the product of mass and velocity of all particles, if no lattice-Boltzmann fluid is coupled.

Performance enhancements

• Speedup in the short-range force calculation in situations where the short-range cutoff varies strongly for different pairs of particles, e.g., in a bidisperse fluid.
• Speedup in particle resorting triggered when particles have moved by more than a skin.
• Significantly faster back-transfer of particle forces from GPU-based methods such as the GPU implementations of lattice-Boltzmann and P3M.

Bug fixes

• Lattice-Boltzmann boundaries, constraints and auto_update_accumulators are now included in checkpointing (#2915).
• Collision detection is now checkpointed (#2342).
• The rhomboid shape was fixed (#2756).
• Deadlocks on certain GPUs have been resolved for the dipolar Barnes-Hut method (#2719).
• The visualization of dihedrals has been fixed (#2677).
• The ENGINE implementation for CPU LB has been fixed (#3025).
• The external force density in lattice-Boltzmann fluids is no longer ignored in the first integration step after setting the force density (#3144).
• The positions of virtual sites and the charges of ICC particles are now updated before observable calculation (#3128).
• The forces and torques for the Gay-Berne potential have been corrected (#3091).
• Remove undocumented behaviour in the case of using a cylindrical sampling area in the reaction ensemble, constant pH ensemble, Wang-Landau ensemble, Widom-Insertion method (#3174).
• The ELC tuning error calculation has been rearranged to produce correct results for higher accuracies (#3123).

New tutorials

• Tutorials for simulating ferrofluids and for using the constant-pH method have been added.

Under the hood changes

• Automated testing has been enhanced. It now also includes samples and tutorials. The overall test coverage for the simulation core has increased by ~12% since ESPResSo 4.0.2.
• The CPU LB and LB-particle coupling have been refactored.
• Particle resorting has been simplified and sped-up.
• The MPI callback mechanism has been simplified. Furthermore, reduction operations such as summing values from all MPI ranks can now be performed.
• Nearly all manual memory management and C-style arrays have been removed.
• The rotation-related code has been simplified.
• Long-range electrostatic and magnetostatic methods now have a common interface.
• The kernels for short-range and bonded interactions have been simplified.
• The CMake build system has been refactored and dependencies between different parts of the code have been made clear.
• Python code formatting: the autopep8 version now matches the one in Ubuntu 18.04 (autopep8 v1.3.4 with pycodestyle v2.3.1).

4.0.2 - 2019-04-26

Bug fixes

• A sign error in tabulated interactions was corrected such that the force equals the negative gradient of the potential (#2519, #2520).
• The flow field of the CPU lattice-Boltzmann implementation was deleted when aspects of the molecular dynamics cell grid were changed; e.g., when interactions, the skin or the parallelization setup were changed. ESPResSo now terminates with an error, when this happens. To avoid this, please setup the CPU lattice-Boltzmann after all other aspects of the system. The GPU LB is not affected in the 4.0 release, but was affected in the current development branch (#2728, #2736).
• Corrected the force acting on LB Boundaries for the case of agrid and density not equal to 1 (#2624).
• Corrected the cutoff calculation for the soft sphere interaction. In the previous implementation, the offset parameter was ignored (#2505).
• The "three point coupling" of particles to the lattice-Boltzmann method has been removed. While it works in most environments, for some compilers the calculation gives wrong values. This is likely caused by undefined behavior. A corrected implementation is available in ESPResSo's development branch. It cannot be safely backported to 4.0.2, because the code has diverged too far (#2516, #2517). Users who did not explicitly activate this coupling via couple="3pt" are not affected.
• The velocity of existing particles no longer changes when setting or changing the simulation time step (#2480).

Further changes

• Fixed the electrokinetic Python interface (#2486)
• Correction to the installation instructions for mac (#2510)
• Corrected file permissions (#2470)
• Minor corrections and extensions to the test suite (#2477, #2552)
• Fixed a dead-lock in the dipolar Barnes Hutt method on the GPU for recent NVIDIA cards such as RTX 2080 (#2719).
• Restored Mayavi visualizer's API-compatibility with OpenGL visualizer (#2751)

4.0.1 - 2019-01-25

Bug fixes

• The GPU lattice-Boltzmann method produced incorrect results when EXTERNAL_FORCES was not declared in myconfig.hpp. This issue was present since around June 2018 (#2241).
• The temperature fluctuations for the GPU lattice-Boltzmann implementation were larger than the ones for the CPU lattice-Boltzmann implementation. The cause was likely weak or incorrect random number generation. It is not clear since when this issue existed. It has been resolved by using library code for a counter-based random number generator (Philox) rather than the existing custom code.
• Particles which were moved with the configuration changing moves (MC) implemented in the reaction ensemble module did not get assigned a random velocity. This was not a problem if you were looking at observables which do not depend on velocity.
• Particles which were created in the Reaction Ensemble module were assigned a random velocity which was not distributed according to the Maxwell-Boltzmann distribution. This was not a problem if you were looking at observables which do not depend on velocity. If you looked at velocity-dependent observables but used a thermostat for thermalization before taking a sample you are also fine (#2377).
• Under some conditions, the torque on self-propelled particles in a lattice-Boltzmann fluid was incorrect due to a sign error in the ENGINE feature. This was the case since the introduction of the feature (#2383).
• The SimplePore shape was incorrect (#2379).
• The parameters passed from Python to some features were narrowed to single precision. I.e., the values used were only accurate up to the 7th-8th significant digit. Further calculations with those values still were done using double precision. Classes backed by the script interface were affected. This includes shapes, LB boundaries, pair criteria and the collision detection. This issue was likely present since the introduction of Python support for the relevant features (#2379).
• Forces on LB boundaries retrieved via the LbBoundary.get_force() method in a Python script were incorrect for the CPU LB implementation. The GPU implementation was not affected. It is not clear, when the issue was introduced (#2366).
• In highly dense systems, the Widom insertion scheme reported wrong values (#2294)
• Fixed the Gay-Berne potential for sigma != 1. Downgraded the GAY_BERNE feature to an experimental feature due to insufficient testing (#2424).

Further general corrections and improvements

• Fix restoring checkpointed simulations which contain exclusions (#2418)
• Allow neutral systems in reaction ensemble (#2378)
• Support re-tuning the dipolar P3M method (#2351, #1937)
• Support checkpointing of the collision detection parameters (#2342)
• Virtual sites tracers: don't complain about missing lb for non-virtual part (#2341)
• Prevent access to non-existent lb nodes from Python
• OpenGL wildcard imports + mac fix (#2295)
• Fix segfaults in virtual sites based collision detection schemes on more than one processor core (#2195)
• Fixed find_current_cell() and particle deletion (#2410, #2441)
• Fixed coordinate folding for corner cases (0 - epsilon on 32 bit architectures) (#2415)

Documentation and tutorials corrections and improvements

• Mention how to read h5md files in the docs (#2353)
• Correct docs for setting the skin (#2340)
• Corrections to the Lennard-Jones tutorial (#2339)
• Fixed Gay-Berne equation in the user guide (#2234)
• Update PDF of tutorial 6 with bibliography (#2285)

Build system and platform related corrections and improvements

• Fix Cython detection on Fedora 29 (#2371)
• Properly detect CUDA libraries if multiple are available (#2352)
• Guard tests and compilation for different feature sets in myconfig.hpp (#2344, #2326, #2350, #2349, #2335, #2337, #2266)
• install libH5mdCore (#2277)
• Fix MPI on Ubuntu 18.04 with CUDA (#2271)
• Fix PARTIAL_PERIODIC on big-endian platforms (#2259)
• Fixes for OpenSUSE rpm build (#2255)
• Add ARGS option to custom make check target (#2252)
• Add workaround for CMake bug MPI_COMPILE_FLAGS is list instead of a string (#2244)
• Remove some superfluous executable bits (#2242)
• Silence arm32 warnings (#2267)
• Fix unit test without SSE instructions (#2265)
• Turn off ScaFaCoS support by default. Has to be enabled explicitly using -DWITH_SCAFACOS=on in CMake (#2417).
• Fix builds with the GNU Scientific Library in non-standard locations (#2429)
• Resolve some floating point rounding issues on I386 32-bit (#2454)

Improved testing

• Test the LB thermostat (#2391, #2362)
• Better stability of the LB test (#2360)
• testsuite: fix checkpoint test (#2336)
• Test the installation routine (#2268)
• Improved testing of the cylindrical lb profile observables (#2272)
• testsuite: Check planar shear profile of LB (#2263)
• Test builds without optional dependencies (#2406)
• Test builds on non-Amd64 architectures (#2401)
• Improved test coverage of the Wang-Landau reaction ensemble (#2404)
• Testing of the electrostatic pressure (#2409)

4.0.0 - 2018-09-06

ESPResSo 4.0 is the first release of ESPResSo with a Python interface. I.e., the language in which simulation scripts are written is now Python, and support for the Tcl-interface has been dropped. We recommend all users switch to ESPResSo 4.0, there will be no more fixes to Tcl-based versions.

Transitioning from Tcl-based versions

The following can serve as a starting point for the new interface:

• The "Introduction" chapter in the user's guide
• The tutorials provided with ESPResSo Both can be found at https://espressomd.github.io/

Please note that, starting from ESPResSo 4.0, enabling additional features in myconfig.hpp no longer changes the behavior of a simulation scripts. Features such as particle rotation and virtual sites have to be activated explicitly in the simulation scripts.

New features

• Reaction ensemble methods for the modelling of chemical reactions, including constant pH and Wang-Landau schemes
• Polarisable particles via the Drude-oscillator scheme
• Steepest-descent energy minimization scheme
• Methods for active particles (swimmer_reaction, dipolar swimmers)
• GPU-accelerated P3M method for electrostatic interactions
• GPU-accelerated direct summation and Barnes-Hutt schemes for dipolar interactions under open boundary conditions
• Support for the electrostatic and magnetostatics method provided by the ScaFaCoS library
• Immersed boundary method for soft immersed objects in an LB fluid
• Virtual sites acting as inertialess tracers in an LB fluid usable as part of the immersed boundary method or independently
• Online cluster analysis based on criteria for pairs of particles
• New bonded quartic and Coulomb interactions
• Possibility to tune the skin parameter
• Support for saving to the h5md file format for molecular dynamics data.
• Connection to the MD Analysis Python package
• A rotate_system command was added, which rotates the particle positions around the system's center of mass

Visualisation

ESPResSo contains two online visualizers, one based on Mayavi, one on the OpenGL Python module. Support for online visualization in VMD has been dropped, but particle data can be stored in file formats which VMD can read (vcf/vtf, h5md).

User-visible changes

• When the ROTATION feature is compiled in, the rotation of a particle around each of its Cartesian axes in the body-fixed frame can be enabled
• The rotational degrees of freedom of virtual sites is now integrated. Virtual sites no longer use the quaternion attributes to store their relative orientation
• Alternatively, the orientation of the virtual site can be specified relative to that of the particle it is derived from.
• The random number generator was changed to Mersenne Twister as implemented by C++11.
• ESPResSo now depends on the Boost library including Boost-Mpi and a C++11 capable compiler
• The build system is now based on CMake.

Removed features

The following features are not part of the ESPResSo 4.0 release, but support is expected to be re-added in a subsequent release:

• Generalized hybrid Monte Carlo thermostat
• Virtual sites located at the center of mass of a group of particles
• Metadynamics, umbrella sampling and parallel tempering
• Non-equilibrium molecular dynamics (shear boundary conditions implemented by moving slabs of particles at the boundaries)
• The memd/maggs electrostatic solver
• The com_force feature
• The Shan-Chen lattice-Boltzmann extension
• Lees-Edwards (shear) boundary conditions
• The chain rdf analysis

The following functionality is removed permanently:

• Blockfile support, which is replaced by support for the h5md format as well as checkpointing based on the pickle Python module.
• Some of the convenience functions for particle creation such as salt and counterions, as these can be replaced by a very few lines of Python in the simulation script.
• The lj_angle and angledist (not-quite)-pair potentials
• The following analysis methods: diffusion_profile, cwvac, p_inst, ..._mol, cel_gpb, dipmom_normal, MSD, angularmomentum, cluster_size_dist, mol, lipid_orient_order, get_lipid_orients, current, wall_stuff, necklace, bilayer, modes2d
• The uwerr error estimator

Known issues and limitations

• Lattice Boltzmann interpolated velocities are only accurate up to one lattice cell next to a boundary (#1865)
• Adding boundaries in the electrokinetics method resets species densities to their initial values (#1899)
• The collision detection feature does not support checkpointing (#1943)
• Changing interaction parameters does not automatically update the value of the force property stored on the particles. Forces are re-calculated in the next integration step (#1973).
• The domain decomposition cell system does not always choose the node grid yielding the best performance in parallel simulations (#1970)
• The dipolar P3M method can only be tuned once in a simulation script, unless method parameters are explicitly reset to 0 before the second tuning (#1937)
• In rare cases, P3M produces an rs mesh overflow warning (#2067)
• The H5MD file writer cannot handle bounds that involve more than two bond partners (#1715)
• Pickle checkpoints are not guaranteed to be compatible between different ESPResSo versions (#1181)
• Switching from the GPU-Based MMM1D and MMM2D methods to a different electrostatic method within a single simulation script may not work (#396)
• The membrane_collision interaction of the object in fluid framework may use an outward direction which is outdated by one time step (#2214)

3.3.1 - 2015-09-01

Bug fixes

• Fix a bug in the particle coupling close to a LB boundary on GPU.

Changes visible for developers

• Improved compiler support and added missing feature checks.

3.3.0 - 2014-08-11

New features

• SHANCHEN: a bicomponent lattice-Boltzmann fluid, with support for rigid boundaries and coupling to particle dynamics.
• ELECTROKINETICS: An algorithm to treat species of ions on a mean-field level. Implementation takes advantage of a GPU.
• MMM1DGPU: The MMM1D electrostatics algorithm is now available on the GPU.
• Support of P3M on GPU.
• Provide the preliminary Python interface. Configure with --with-python-interface.
• Allow P3M to dump the mesh in x, y, and z (before it only dumped the mesh in x). This allows for the dumping of the P3M in blockfile format in the case of non-cubic boxes. It will also now be possible to tune the settings of a non-cubic mesh.
• Stomatocyte LB-boundary and constraint implemented.
• MEMD electrostatics can now handle spatially varying dielectric constants.
• Pore constraint / lbboundary can have a two outer radii now to create nozzles.
• Slitpore constraint and dielectric for IL based supercaps implemented.
• New command time_integration to get the runtime of the integration loop.
• New harmonic well that runs on the GPU.

Known bugs

• The implementation of dielectric contrasts in conjunction with ELC seems to have an error. Please do not use it if you do not know exactly what you are doing until we have fixe the problem.

Interface changes

• Added new arguments recalc_forces and reuse_forces to the command integrate. These arguments can be used to enforce or suppress the recalculation of the forces at the beginning of the call to integrate. This is important for doing checkpoints, where the forces have to be stored and reloaded.
• Removed command invalidate_system.
• Comfixed now works with periodic boundary conditions.
• The pressure contribution due to rigid bodies constructed by means of the virtual sites relative mechanism is included, both for pressure and stress tensor. Note that the corresponding contribution for rigid bonds is still not implemented.
• The configure option --with-myconfig has been removed. Instead, the configure variable MYCONFIG can be set to give the name of a myconfig file.
• Generic LJ can now be turned into a soft potential.
• Renamed torque to torque_{lab,body}, improved torque handling in blockfiles. Clearer distinction between the reference frames for torque and angular momentum.
• Lattice-Boltzmann now has a additional 3-point coupling scheme.
• The noise type in thermalized Lattice-Boltzmann is now selectable via Tcl.
• Interaction with a wall can be restricted to the outside.
• Removed compiler switch GRANDCANONICAL.
• Removed outdated ADRESS code.
• external_potential tabulated to include arbitrary potentials applied to all particles.
• thermostat inter_dpd ignore_fixed_particles 1 allows the user to swith on DPD with fixed particles.
• New Observable concept that includes running averages and other observables with history

Changes visible for developers

• Feature GHOSTS_HAVE_BONDS allows for ghost particles to also store the bonds.
• The code has been switched to using a C++ compiler.
• Introduced object-oriented interfaces for forces/constraints/external fields: SystemInterface and Actor.
• Improved Lattice structure
• Generic TclAppend function

3.2.0 - 2013-05-10

New features

• Added a new immersed boundaries code that allows to simulate extended, flexible objects in conjunction with the LB code. This is described in the UG in the new chapter "Object-in-fluid".
• The script tools/blockfile2vtf.tcl can be used to convert ESPResSo blockfiles into VTF files.
• Two new short-ranged nonbonded potentials have been added:
  • HAT: the classical conservative DPD interaction, a simple force ramp.
  • GAUSSIAN: a Gaussian potential.
• The feature CATALYTIC_REACTIONS and the command reaction model a simple chemical reaction: when a particle of the reactant type comes into the vincinity of a particle of the catalysator type, the reactant is transformed into a particle of the product type.
• The feature GHMC implements a generalized hybrid Monte-Carlo thermostat.
• The feature GRANDCANONICAL provided functions to simplify grandcanonical simulations in ESPResSo. To that end it provides functions to find and delete random particles of a certain type.
• Added new feature ROTATION_PER_PARTICLE that allows to choose whether a particle has rotational degrees of freedom or not.

Interface changes

• Removed non-working tools and samples.
• Bond angle potentials are now defined on the Tcl-level instead of on the feature-level, i.e. to choose the type of the bond-angle potential, you should use the corresponding Tcl command instead of a feature. To allow for any bond angle potential, activate the feature BOND_ANGLE.
• We have removed the forcecaps for the different interaction types. Instead, there is now a single global forcecap, plus it is possible to define individual forcecaps on a particle pair level.
• By default, CUDA is now automatically detected and activated when it is there and usable.
• The functions for galilei transformations have been overhauled and renamed. The old functions are deprecated.
• DPD (as thermostat) now works without any other interaction, so a pure DPD fluid is possible.

3.1.2 - 2013-03-08

• Fixed bugs #38475, #37725, #36434.
• Added workaround for bug in VTF reader plugin of VMD 1.9.1 that made all particles look the same when ELECTROSTATICS is turned on.
• Added Tcl-scripts of the tutorials to the distribution.
• Added forgotten part of the UG on the correlators.
• Removed buggy readline code tclline.tcl. Fixes #36432.
• Removed Ewald code from ESPResSo that never worked and was never documented.

3.1.1 - 2012-10-09

• The Espresso binary now outputs the header to stderr instead of stdout.
• Fixed bugs #36431, #37120, #37214, #37374, #37306, #35767.
• Fixed generation of myconfig-sample.h.
• Fixed AdResS example.
• Added new logo.
• The documentation sources are now left out of the distribution package, only the PDF files are included. This fixes problems when calling make doc in a distribution package (#35958).

3.1.0 - 2012-03-08

New features

• While so far, observables were typically computed and stored on the Tcl-level after a call to integrate, a new observable concept now allows to compute observables while the C-core is running. This is particularly useful for observables that have to be computed very frequently, as for example in the case of time-correlations (e.g. the RMSD of particles). The new concept is documented in the User's Guide in Section 8.4 ("Correlations and observables").
• The new feature COLLISION_DETECTION allows to dynamically add new bonds between colliding particles.
• Added new feature LANGEVIN_PER_PARTICLE that allows to set the Langevin parameters temperature and gamma per particle.
• Added new constraint and LB boundary condition "rhomboid".
• Lattice-Boltzmann:
  • non-zero velocity boundary conditions are available
  • forces on boundaries can be calculated
  • wall constraints work
• Added basic tutorial on Lennard-Jones liquid, made second tutorial (simple charged systems) better available (all in doc/tutorials/)
• The ICC* algorithm is now ready to use. It is used with the command iccp3m and allows to take into account dielectric boundaries of arbitrary shape. The command dielectric allows to create the boundaries in a similar fashion as constraints and lbboundaries.

Interface changes

• The blockfile C-library has been removed. In theory, there was a library usable from C that could be used to read and write blockfiles. Since several years, it was broken, so apparently it wasn't used anyway.
• The NPT barostat now works in many more combinations of algorithms than so far (e.g. MMM2D, ELC, ...)
• The Lattice-Boltzmann CPU implementation now works when Verlet lists are used.
• Removed unneccessary overhead of neutral particles in simulations with P3M.
• Removed unneccessary overhead of activated but ununsed interaction features.
• Some changes in the Verlet list construction yield a performance gain of up to 90% in systems where the interaction ranges differ significantly, or where some particle types do not interact at all (phantom particles).
• Renamed Coulomb method maggs to MEMD (inter coulomb maggs => inter coulomb memd).

Changes visible for developers

• code_info does now not only show the main package version, but also the exact git commit id and whether or not the code was modified ("dirty"). Also, the distribution package provides this information.
• The various features are now defined in the file src/features.def. From this file, a couple of Python scripts automatically generate myconfig-sample.h and what was config.h and config.c. Also, this allows to check whether all features are documented, tested and defined. Note that due to this change, ESPResSo development now requires Python.
• The domain decomposition, P3M, LB and MEMD now use MPI cartesian communicators. This will hopefully speed up the simulations on some platforms.
• Split interaction code into .c and .h files.
• Split off the Tcl interface (into src/tcl/) from the C core code (in src/)
• Added ESPResSo logos, cover issue images, and some other material to the repo (doc/).
• Started new LaTeX-Developer's Guide (DG), and moved all "Related pages" from the doxygen docs. The doxygen stuff still exists and is available via doc/doxygen/.

3.0.2 - 2011-10-04

• Fixed features ADRESS and VIRTUAL_SITES_COM that did not work for a while.
• Fixed bugs #33489, #34238.
• Fixed a few bugs that were never reported via the bug tracker, mainly in the documentation.

3.0.1 - 2011-05-25

• Fixed bugs #33375, #33376, #32005.
• Fixed a few bugs that were never reported via the bug tracker.

3.0.0 - 2011-04-19

Interface changes

• The lattice-Boltzmann implementation in ESPResSo has been significantly overhauled and is now documented.
• A CUDA implementation for NVIDIA GPUs of the lattice-Boltzmann algorithm is included.
• The Maggs algorithm (or Maxwell Equation Molecular Dynamics, MEMD) for computing electrostatic interactions has been included and documented. The algorithm is a fast and scalable alternative for P3M or other such algorithms.
• The P3M algorithm for electrostatic and dipolar (magnetostatic) interactions has been extended.
  • Electrostatic P3M allows for non-cubic boxes.
  • The electrostatic P3M pressure tensor is computed correctly now.
  • Dipolar P3M has been parallelized (electrostatics P3M was parallelized already!)
  • The MDLC method (magnetic dipolar layer correction) has been parallelized.
• The virtual sites feature has been significantly extended and documented. Virtual sites are particles that can interact with other particles but whose positions are not integrated via the normal integrator. Instead, the positions are determined by the positions of other, "real" particles. Virtual sites can either be set into the center of mass of a set of real particles (feature VIRTUAL_SITES_COM), or it can be set to an arbitrary location relative to another particle (VIRTUAL_SITES_RELATIVE). This allows to create rigid structures within ESPResSo.
• The User's Guide has been significantly updated and extended.
• Some constraints can now be made penetrable and reflecting.
• Espresso is now the actual binary instead of a wrapper shell script that automatically calls MPI. This means that it is necessary to call mpiexec or mpirun youself to run ESPResSo in an MPI environment.
• The directory where the scripts are installed is now compiled into the binary, i.e. it is not necessary anymore to set the environment variable ESPRESSO_SCRIPTS. Still, it will heed the environment variable if it is set.
• The build system has been overhauled and simplified.
  • Running make check will run the testsuite with a single number of processors, which gives a significant speedup. configure will try to determine the available number of CPUs/cores. If it can't, it will use 1 by default.
  • MPI recognition has changed. The build system first needs to find out how to compile an MPI binary. If it doesn't find out automatically, you can help it by setting the variable MPICC or LDFLAGS correctly. Furthermore, it needs to know how to run an MPI binary. If mpiexec is available, everything is fine. If it is not, you can provide a script mympiexec.sh that behaves like mpiexec.
  • So far, when you built in the source directory, all files ended up in a subdirectory obj.XXX, where XXX was some description of the processor you used. This non-standard behavior has been removed. Instead, the files are generated directly in the source directory when compiling there. If you want to compile several binaries from a single source directory, you can use build directories as described in the User's Guide.
• A number of functions have been deprecated, as they are not maintained anymore. Plase do not use them in your code and replace them with appropriate alternatives if you use them in your code! The following functions are deprecated: checkpoint_*, polyBlock*, calcOb*, calcObs*, plot*, polyConf*
• FFTW2 is no longer supported.

Organizational changes

• The home page of ESPResSo has been relocated to https://espressomd.org
• The mailing list and source code repository have been moved to GNU Savannah. Furthermore, we have a bugtracker now. The development homepage is https://savannah.nongnu.org/projects/espressomd
• The maintenance of ESPResSo has been relocated from Torsten Stühn from the Max-Planck-Institute for Polymer Research at Mainz to Olaf Lenz [email protected] from the Institute for Computational Physics at the University of Stuttgart.
• The old "ESPResSo license" was removed, as it was not really a license, but only contained additions that were either non-binding or already contained in the GPL.
• The license was upgraded to GPLv3.
• The package was adapted to GNU standards. RELEASE_NOTES was moved aside to old/RELEASE_NOTES, instead, we will have the files ChangeLog and NEWS. NEWS will contain a description of the most noteworthy changes since the last release.

Changes visible for developers

• The source code repository has been moved from CVS to git, and it has moved to GNU Savannah (see above) and GitHub. The main development code repository is https://github.com/espressomd/espresso
• We have an automated build server (Jenkins) at http://espressomd.org/jenkins/ that automatically builds and checks the ESPResSo package whenever new code is pushed to the main development repository. Every night, a number of additional tests are performed
• The .c/.h/.cu source files have been moved to the subdirectory src/. The "Espresso" binary is still placed in the top-level build directory. myconfig.h may be placed eitther in the src/ or top-level dirs.
• The automatically generated files configure and Makefile.in are not part of the repository anymore. A developer can generate them using the command bootstrap.sh, which requires the GNU autotools (autoconf and automake).
• All functions that represent the interface to Tcl should follow the same naming conventions. These are described here: 3d4d6f3
• Adding a new MPI function to communication.c has been significantly simplified. It is enough to add the function in the C file at a single location.

---

For older ESPResSo releases, see old/RELEASE_NOTES@bb2cd93.