PatchyParticles

PatchyParticles is a code for perfoming Monte Carlo simulations of hard particles decorated with four patches modelled through the Kern-Frenkel pair interaction potential. The code accompanies the Colloquium paper How to simulate patchy particles, Eur. Phys. J. E (2018) 41: 59, and it is educational in the sense that it is meant to be read as much as meant to be used.

Compilation

PatchyParticles does not require any dependencies but a C compiler, the standard C library and GNU's make. The default compiler, specified in the makefile file, is gcc, but the code can be compiled with icc as well. The compilation has been tested with gcc >= 4.8 and icc >= 13.

In order to compile the code it is sufficient to launch make within the main folder:

$ make

Depending on the platform, it might be necessary to explicitly pass the makefile to make. This can be done by issuing

$ make -f makefile

The code can be compiled with icc with the command

$ make -f makefile.icc

The default behaviour is to compile PatchyParticles with full optimisations enabled (that is, with the options -O3 -ffast-math -DNDEBUG or -fast -DNDEBUG if the code is compiled with gcc or icc, respectively). The code can also be compiled with optimisations and debug flags (make g=1) or with debug flags only (make dbg=1).

At the end of the compilation stage there will be two executables: PatchyParticles and generator. Note that if you don't use dbg=1, by default the -march=native flag is used. If you want to compile and run the code on computers with different architectures, remove that flag from the makefile.

Usage

The PatchyParticles program takes one mandatory argument, the input file. The syntax of the file is quite simple, as it is just a list of "key = value" lines. The next section has a list of mandatory options. The Examples folder contains some simple input files that can be used as starting points.

The generator binary can be used to generate initial configurations of N particles at density ρ, with ρ < 0.7. It requires two arguments (an integer for N and a decimal number for ρ). The program prints the configuration in a new file named generated.rrr.

Input file

Here is a list of mandatory options. Please refer to the input files in the Examples folder for common values.

Simulation options

Dynamics = <int>: use 0 for rototranslations, 1 for VMMC and 2 for AVB.
Ensemble = <int>: use 0 for NVT, 1 for Grand Canonical (muVT), 3 for Successive Umbrella Sampling and 5 for NPT.
Temperature = <float>: temperature of the simulation, in units of the patch-patch bond.
P = <float>: pressure of the simulation, in units of the patch-patch bond over the cube of the particle diameter. Used in NPT simulations only.
Steps = <int>: length of the simulation, in Monte Carlo steps.
GC_N_max = <int>: maximum number of particles, above which the simulation will be stopped. Meaningful only for Grand Canonical (or SUS) simulations

Input/output options

Initial_conditions_file: initial configuration.
Print_every = <int>: output frequency for energy, density and acceptances.
Save_every = <int>: output frequency for configurations.
Energy_file = <string>: name of the output file for the energy.
Density_file = <string>: name of the output file for the density.
Configuration_last = <string>: name of the last configuration file (printed with frequency Save_every and at the end of the simulation)
Confguration_folder = <string>: name of the folder where configuration files will be stored.

Kern-Frenkel options

KF_delta = <float>: Radial width of the Kern-Frenkel patches.
KF_cosmax = <float>: Angular width of the Kern-Frenkel patches.

Monte Carlo moves options

Disp_max = <float>: maximum trial displacement for translations.
Theta_max = <float>: maximum trial angular displacement for rotations, in radians.
vmmc_max_move = <float>: maximum allowed displacement for VMMC moves: if a vmmc move attempts to move a particle for more than this value, the move will be rejected.
vmmc_max_cluster = <int>: maximum cluster size for VMMC moves: if a vmmc move attempts to move more than this number of particles, the move will be rejected.
rescale_factor_max = <float>: maximum trial displacement for volume changes. The proposed new volume is computed as ln_final_V = log(initial_V) + (R - 0.5)*rescale_factor_max, where R is a [0, 1] number extracted from a uniform distribution. Used in NPT simulations only.
Lx_move = <int>: if set to 1, enable a move that changes the Lx side of the box while keeping the volume constant (with Ly = Lz).
Lx_change_max = <float>: maximum trial changes for the Lx box side.
Lyz_min = <float>: this value times the initial Ly is the minimum value allowed for Ly and Lz. It may not be larger than 1.
Lyz_max = <float>: this value times the initial Ly is the maximum value allowed for Ly and Lz. It may not be smaller than 1.

Some useful non-mandatory options

Save_also_as_mgl = <int>: if not 0, PatchyParticles will output an additional mgl file, which can be directly visualised with cogli1, whenever configurations are printed. Defaults to 0.
Restart_step_counter = <int>: true by default, if set to false will restart the simulation from the time step found in the initial configuration file and append the energy, density and acceptance output to their respective files.
Log_file = <string>: by default PatchyParticles writes the output to the standard error. If the input file contains this option, PatchyParticles will redirect its output to the specified file.
Acceptance_file = <string>: by default acceptance probabilities are written to the acceptance.dat file. If given, the value found in this option will be used instead.

Output files

A simulation will produce at least two files, which by default are energy.dat and acceptance.dat. If Grand Canonical or SUS simulations are run, there will also be a density.dat file. Each line of these files contain the time step and the instantaneous energy, acceptance probabilities and density, respectively.

The moves to which acceptance probabilities refer depend on the chosen ensemble and dynamics:

dynamics = 0: one column for the rototranslation acceptance
dynamics = 1: one column for the VMMC acceptance
dynamics = 2: two columns for the rototranslations and AVB acceptances

If ensemble = 1 or ensemble = 3 the next-to-last and last columns contain the acceptance probabilities of particle additions and deletions, respectively.

Configuration files

PatchyParticles uses plain-text configuration files to store some information about the system and the positions and orientations of each particle. The first (header) line contains the time step of the configuration, the number of particles and the size of the box along the x, y and z directions.

The remaining part of the file contain the orientations and positions of the particles. The information about each particle takes up three lines. The first two store the two upper rows of the orientation matrix, while the third is the position of a particle's centre of mass.

Code organisation

We have tried to logically split the code up in modules. We have also tried to use naming conventions that make it easier to understand the purpose of each module, function and (to a lesser extent) variable.

The code is commented lightly. However, we have tried to highlight what is done in the most important sections of the code.

The main.c file contains calls to the initialisation functions, the calculation of the initial energy, the main loop and calls to the cleanup functions.
The general data structures used throughout the code, as well as some useful macros, are stored in the defs.h file.
The MC.c, MC.h pair contains the main logic of the code. It manages the calculation of the energy, the particle rototranslations, the simulation of the different ensembles, etc.
The avb.c, avb.h and vmmc.c, vmmc.h pairs contain the data structures and functions pertaining to the Aggregation-Volume-Bias and Virtual-Move-Monte-Carlo moves that can be optionally enabled to speed up the calculations.
The cells.c, cells.h pair contain the data structures and functions pertaining to the linked-lists used to keep track of the list of neighbours of each particle.
The utils.c, utils.h pair contains commonly-used functions to work with vectors and matrices.
The system.c, system.h pair contains the functions used to initialise and cleanup the main data structure (System).
The output.c, output.h pair contains the functions used to initialise and cleanup the data structure responsible for printing the output (Output), as well as the functions that actually print most of the output.
The parse_input.c, parse_input.h pair contains the data structures, logic and functions used by the code to parse the input file passed to PatchyParticles

Acknowledgements

The code has been developed by Lorenzo Rovigatti (Dipartimento di Fisica, Sapienza University of Rome), John Russo (School of Mathematics, University of Bristol) and Flavio Romano (Dipartimento di Scienze Molecolari e Nanosistemi, Università Ca' Foscari di Venezia). Feel free to contact us if you have any questions/comments, and if you use the code in your project please consider adding a reference to the PatchyParticles paper.

We thank Daniele Notarmuzi, Martin Oettel and Alessandro Simon for discussions, suggestions and bugfixing related to the code.

lorenzo-rovigatti / patchyparticles Goto Github PK