ReadSave

Checkpoint

Module designed to save relevant parameters and configurations throughout the simulation to allow restarting.

build_states(mc_params::MCParams, ensemble::AbstractEnsemble, temp::TempGrid, potential::AbstractPotential)

For use initialising states and outputs when NOT restarting, but beginning from files. Builds empty Output struct named results and a vector of mc_states using either: one configuration stored in config.data OR a series of configurations stored in config.i. NB if config.i doesn't exist the default will be config.1. In this way states can be initialised with different starting configurations.

checkpoint(index::Int, mc_states::MCStateVector, results::Output, ensemble::AbstractEnsemble, rdfsave::Bool)
checkpoint(index::Int, mc_states::MCStateVector, results::Output, ensemble::NPT, rdfsave::Bool)

Function to save relevant information about the current state of the system at step index. Saves the configurations in each mc_state save_configs as well as the histograms stored in results. Optionally stores the volume histograms if using the NPT ensemble and the radial distribution functions if desired.

checkpoint_config(savefile::SaveFile, state::MCState{T, N, BC, Ptype, Etype}) where {T, N, BC <: SphericalBC, Ptype, Etype}
checkpoint_config(savefile::SaveFile, state::MCState{T, N, BC, Ptype, Etype}) where {T, N, BC <: CubicBC, Ptype, Etype}
checkpoint_config(savefile::SaveFile, state::MCState{T, N, BC, Ptype, Etype}) where {T, N, BC <: RhombicBC, Ptype, Etype}

Function writes a single config in the standard xyz format. N atoms, the comment line contains the boundary condition information (implemented for Spherical BC and both types of Periodic BC) as well as max_displ information determining the stepsize used at the current step of the monte carlo simulation. The comment row is followed by 1 as a placeholder for the atom type to be implemented in future and the positions x,y,z in order.

read_checkpoint_config(xyzdata)

Function designed to take a single xyz-style checkpoint file and return the configuration and max displacement data associated with a saved mc_state. This is used to reconstruct an MC simulation from checkpoints.

read_config(xyzdata)

Designed to read in one xyz-style file with one configuration and return this for starting a simulation from files without restarting.

read_init()

Function to reinitialise the fixed parameters of the MC simulation as saved by the save_init function.

read_params(paramsvec)
read_params(paramsvec,restart)

Function to turn a delimited paramsvec into an MCParams and TempGrid struct. Second method includes a bool restart to determine whether or not to set eq_cycles=0 or 0.2*cycles if not restarting

readensemble(ensemblevec)

Function to convert delimited file contents ensemblevec and convert them into an ensemble.

readpotential(potinfovec)

Function to convert delimited file contents potinfovec into a potential. Implemented for:

• ELJPotentialEven
• ELJPotential
• ELJPotentialB
• EmbeddedAtomPotential

rebuild_states(n_traj::Int, ensemble::AbstractEnsemble, temps::TempGrid, potential::AbstractPotential)

Function to rebuild the MCStates vector and results struct from checkpoint information. The ensemble temps and potential along with n_traj are reconstructed elsewhere, but required to accurately recreate the states.

save_configs(mc_states::MCStateVector)

Function to save the configuration of each state in a vector of mc_states. Writes each configuration according to checkpoint_config into a file config.i where i indicates the order of the states.

save_histparams(results)

Initialises and populates a data file containing the information necessary to interpret histogram data.

save_init(potential,ensemble,params,temp)

Function to write all static parameters into a single parameters file. If a params file does not exist, it is created in ./checkpoint as params.data. This contains the mc_params ensemble and potential data.

setresults(histparams,histdata,histv_data,r2data)

Function to re-initialise the results struct on restarting a simulation.

writeensemble(savefile::SaveFile, ensemble::NVT)
writeensemble(savefile::SaveFile, ensemble::NPT)
writeensemble(savefile::SaveFile, ensemble::NNVT)

Function to write the ensemble data into a savefile including the move types. First method is for the NVT ensemble which does not include volume changes, second method is NPT ensemble and does inclue volume moves.

writeparams(savefile,params,temp)

Function to write the mc_params and temp data into a savefile. These are static parameters that define how the simulation is to proceed such as the number of cycles, trajectories and the temperatures to be covered.

writepotential(savefile::SaveFile, potential::Ptype) where Ptype <: AbstractDimerPotential
writepotential(savefile::SaveFile, potential::Ptype) where Ptype <: AbstractDimerPotentialB
writepotential(savefile::SaveFile, potential::Ptype) where Ptype <: EmbeddedAtomPotential
writepotential(savefile::SaveFile, potential::Ptype) where Ptype <: AbstractMachineLearningPotential

Function to write potential surface information into savefile. implemented methods are the Embedded Atom Model, Extended Lennard-Jones and ELJ in Magnetic Field. This does not work for machine learning potentials.