RMC Execution Tutorial

Applies to: rmc_dft v0.6.13

This tutorial explains the basic procedure for running RMC_POT independently by executing rmc.py directly.

1. Overview

rmc.py runs RMC_POT on the current *.cfg structure.

RMC input parameters are specified in rmc_init.py.

The basic workflow is:

config.yml
    |
    +--> RmcPotDat
    |
rmc_init.py
    |
    +--> run_rmc_init()
            |
            +--> Initialize *.dat
            +--> Set RMC parameters
    |
rmc.py
    |
    +--> run_rmc()
            |
            +--> Execute RMC_POT
            +--> Save to logN_rmc

2. Required Files

At minimum, prepare the following files in the calculation directory:

config.yml
rmc.py
rmc_init.py
aaa.cfg

The aaa part depends on cfg_name in config.yml.

The RMC_POT executable, experimental data, and any other files referenced by config.yml or *.dat are also required.

3. rmc_init.py

When rmc.py is launched directly, the following function is called before execution:

run_rmc_init(
    config=config_,
    rmc_dat=rmc_dat_,
    logger=logger_,
)

RMC parameter changes should generally be defined in rmc_init.py, rather than in rmc.py.

For example, use setters:

rmc_dat.set_density(0.0570880440)

rmc_dat.set_maximum_moves(
    [
        0.01,
        0.01,
        0.01,
        0.01,
    ]
)

rmc_dat.set_sigma_factor(0.05)
rmc_dat.set_moveout_option(True)
rmc_dat.set_time_and_save(15, 1)

Free-format settings, including COORD, SNC/Qn, and angular constraints, are also configured in rmc_init.py.

4. rmc_dat.init()

run_rmc_init() first calls RmcPotDat.init().

The basic behavior in v0.6.13 is:

• If an existing *.dat file is present, use it as the initial input.

• If no *.dat file exists, create the initial input.

• Calling init() again on an already initialized RmcPotDat does not repeat initialization.

• Subsequent set_*() calls perform the actual parameter changes.

Thus, init() and the setters have distinct responsibilities.

5. Execution

Run:

python rmc.py

rmc.py applies rmc_init.py and then executes RMC_POT.

6. Log Directories

The RMC output directory is determined by next_log_dir_name("rmc").

For example:

log0_rmc
log1_rmc
log2_rmc
...

When RMC and DFT are combined, the numbering is a single shared sequence across RMC and DFT, not separate RMC and DFT counters.

7. Ctrl+C

In v0.6.13, rmc.py / run_rmc() wraps the RMC_POT execution in try/finally.

Consequently, even if RMC_POT is interrupted with Ctrl+C, postprocessing is performed on partially generated files such as *.cfg, and the files are archived in the log directory.

8. CONTINUE_NUMBERED_INIT

rmc_init.py contains:

CONTINUE_NUMBERED_INIT = False

This setting is effectively unused when rmc_init.py is used for standalone RMC execution.

It matters when rmc_vasp_script.py or rmc_vasp_script_sequence.py uses a numbered initialization file:

rmc_init_N.py

See rmc_dft.md and rmc_dft_sequence.md for details.

9. Typical Parameters to Edit

Normally, edit rmc_init.py rather than rmc.py to change RMC conditions.

Typical settings include:

density
cutoff
maximum move
sigma
moveout
RMC time / save interval
experimental range
r spacing
experimental data
FNC / topology
COORD
SNC / Qn
angle constraint

The basic design separates execution control from input-parameter configuration, keeping calculation conditions in the initialization file.