class DFT_workflow

This class manages the workflow of DFT calculation using TurboRVB built-in program prep.x.

Module usage

from turboworkflows.workflow_prep import DFT_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.

DFT parameters

argument	type	default value	description
dft_rerun	bool	False	if True, force rerun even if pickle file exists.
dft_grid_size	list	[0.1, 0.1, 0.1]	specify grid sizes [x,y,z] as 3 floats
dft_lbox	list	(None)	specify box sizes [x,y,z] as 3 floats (angstrom)
dft_smearing	float	0.0	smearing parameter (Ha)
dft_maxtime	int	172800	maximum time (sec.)
dft_memlarge	float	False	use more memory to speed up
dft_maxit	int	50	maximum number of iterations
dft_conv_thr	float	1.0e-5	SCF convergence threshold.
dft_h_field	float	0.0	magnetic field putting on each grid.
dft_magnetic_moment_list	list	[]	magnetic moment list, for all atoms.
dft_xc	str	lda	exchange correlation functionals. options: lda or lsda
dft_twist_average	bool	False	flag for twist average
dft_independent_kpoints	bool	False	flag for independent k-point calculation
dft_thr_lindep	float	1.0e-13	inverse of the condition number of basis sets. prep cuts the redundancy.
dft_kpoints	list	[1, 1, 1, 0, 0, 0]	Monkhorst-Pack k-grids, [kx,ky,kz,nx,ny,nz], kx,ky,kz for grids, nx,ny,nz for shift=0, noshift=1.

Description

In the DFT_workflow class, the DFT (prep) workflow is executed asynchronously (via async_launch()). The workflow runs a single DFT calculation using the prep.x binary from TurboRVB, with support for twist-averaged boundary conditions and k-point sampling.

If the pkl file already exists and dft_rerun is False, the calculation is skipped. Otherwise, the workflow:

• Builds a DFT_genius instance from the constructor parameters (grid size, box, smearing, XC, twist averaging, etc.) and generates prep.input.

• Submits the job (prep-mpi.x or prep-serial.x), waits for submission and completion using asyncio.sleep(), then fetches output files (e.g. out_prep, occupationlevels.dat, EDFT_vsk.dat, and optionally fort.10_new or twist-averaging scratch files).

• Persists state in a pkl file under the pkl directory.

On success, the method returns (status, list of output file paths under the root directory, and an output-values dict with "energy" and "error" set to None).

See also

• turbogenius.prep_genius.DFT_genius — DFT/prep driver used for input generation.