class LRDMC_workflow

This class manages the workflow of LRDMC calculations.

Module usage

from turboworkflows.workflow_lrdmc import LRDMC_workflow

Constructor arguments

job parameters

argument	type	default value	description
server_machine_name	str	localhost	name of the server machine for job submission.
queue_label	str	(None)	queue label for job submission.
mpi	bool	False	use MPI parallelization.
version	str	stable	version of packages to use.
sleep_time	int	1800	interval of job status check in seconds.

lrdmc parameters

argument	type	default value	description
lrdmc_rerun	bool	False	if True, force rerun even if pickle file exists.
lrdmc_max_continuation	int	2	maximum number of continuation runs.
lrdmc_target_error_bar	float	2.0e-5	target error bar in Ha.
lrdmc_trial_steps	int	150	number of trial steps for initial run.
lrdmc_bin_block	int	10	binning length.
lrdmc_warmupblocks	int	5	number of warmup blocks to discard.
lrdmc_safe_trial_steps	bool	True	use safe minimum trial steps.
lrdmc_correcting_factor	int	10	correcting factor for error estimation.
lrdmc_trial_etry	float	0.0	trial energy (Ha)
lrdmc_alat	float	-0.20	lattice spacing (Bohr)
lrdmc_time_branching	float	0.10	interval between two branching steps. (a.u.)
lrdmc_nonlocalmoves	str	dla	treatment of locality approximation, choose from tmove, dla, dlatm
lrdmc_num_walkers	int	-1	number of walkers, -1 (default) = the number of MPI processes
lrdmc_twist_average	bool	False	flag for twist average
lrdmc_kpoints	list	[]	k Monkhorst-Pack grids, [kx,ky,kz,nx,ny,nz], kx,ky,kz for grids, nx,ny,nz for shift=0, noshift=1.
lrdmc_force_calc_flag	bool	False	if True, compute energy and force, if False, compute only energy
lrdmc_pw_regularization	float	0.0	if it is > 0.0, the Pathak-Wager regularization is turned on.
lrdmc_maxtime	int	172000	maximum time (sec.)

Description

In the LRDMC_workflow class, the LRDMC (Lattice Regularized Diffusion Monte Carlo) workflow is executed asynchronously (via async_launch()). The workflow runs LRDMC in a continuation loop (up to lrdmc_max_continuation runs). For each continuation index \(i_{\mathrm{cont}}\):

test run (icont = 0)

Uses fixed trial steps (lrdmc_trial_steps) to produce energy and error-bar data. Optionally enforces a minimum trial length when lrdmc_safe_trial_steps is True (40 blocks plus warmup).
No step estimation is performed.
Results are used by the next run to estimate required LRDMC steps.

production run (icont ≥ 1)

Loads the previous run's results from the corresponding pkl file and reads the number of MCMC steps from fort.12 (with twist-averaging adjustment when lrdmc_twist_average is enabled).
If the current error bar is below 1.1 × lrdmc_target_error_bar, the workflow exits early and returns the current energy and error (target achieved).
Otherwise, estimates the total steps required to reach lrdmc_target_error_bar via \(N_{\mathrm{total}} \propto (\sigma / E_{\mathrm{target}})^2\), then sets the additional steps for this run to \(\max(N_{\mathrm{total}} - N_{\mathrm{done}}, 1)\).
Estimates runtime and builds a LRDMC_genius instance with the (additional) lrdmcsteps, generates input, and submits the job.

For each run, the workflow waits for job submission and completion (using asyncio.sleep()), fetches output files (e.g. fort.11, fort.12, out_fn_*, parminimized.d), computes energy and forces with the configured bin block and warmup blocks, stores results, and persists state in pkl files under the pkl directory.

If all continuation pkl files already exist and lrdmc_rerun is False, the calculation is skipped. On success, the method returns (status, list of output file paths, and an output-values dict containing "energy" and "error").