[dft]: DFT related inputs

List of parameters for the DFT calculation in charge-self-consistent calculations. The parameters are ignored in one-shot calculations.

dft_code

type = string

choose the DFT code interface, for now Quantum Espresso and Vasp are available. Possible values:

• ‘vasp’

• ‘qe’

dft_exec

type = string; default = ‘vasp_std’

command for the DFT executable

mpi_env

type = string; default ‘

selection for mpi env for DFT / VASP in default this will only call VASP as mpirun -np n_cores_dft dft_exec Possible values:

• ‘default’

• ‘openmpi’

• ‘openmpi-intra’

• ‘mpich’

n_cores

type = int, None

number of cores for the DFT code (VASP). Mandatory for charge-self-consistent calculations

n_iter

type = int; default = 4

only needed for VASP. Number of DFT iterations to feed the DMFT charge density into DFT, which generally takes multiple Davidson steps. For every DFT iterations, the charge-density correction is recalculated using newly generated projectors and hoppings from the previous DFT run

n_iter_first

type = int; default = dft.n_iter

number of DFT iterations in the first charge correction because this first charge correction usually changes the DFT wave functions the most.

plo_cfg

type = str; default = ‘plo.cfg’

config file for PLOs for the converter

projector_type

type = string; default = ‘w90’

plo: uses VASP’s PLO formalism, requires LOCPROJ in the INCAR w90: uses Wannier90 (for VASP and QuantumEspresso)

store_eigenvals

type = bool; default = False

stores the dft eigenvals from LOCPROJ (projector_type=plo) or wannier90.eig (projector_type=w90) file in h5 archive

w90_exec

type = string; default =’wannier90.x’

the command to start a single-core wannier run

w90_tolerance

type = float; default = 1e-6

threshold for mapping of shells and checks of the Hamiltonian