AUXILIARY_DENSITY_MATRIX_METHOD

References: Guidon2010

Parameters needed for the ADMM method. [Edit on GitHub]

Keywords

Keyword descriptions

ADMM_PURIFICATION_METHOD: enum = MO_DIAG

Valid values:

  • NONE Do not apply any purification

  • CAUCHY Perform purification via general Cauchy representation

  • CAUCHY_SUBSPACE Perform purification via Cauchy representation in occupied subspace

  • MO_DIAG Calculate MO derivatives via Cauchy representation by diagonalization

  • MO_NO_DIAG Calculate MO derivatives via Cauchy representation by inversion

  • MCWEENY Perform original McWeeny purification via matrix multiplications

  • NONE_DM Do not apply any purification, works directly with density matrix

Method that shall be used for wavefunction fitting. Use MO_DIAG for MD. [Edit on GitHub]

ADMM_TYPE: enum = NONE

Valid values:

  • NONE No short name is used, use specific definitions (default)

  • ADMM1 ADMM1 method from Guidon2010

  • ADMM2 ADMM2 method from Guidon2010

  • ADMMS ADMMS method from Merlot2014

  • ADMMP ADMMP method from Merlot2014

  • ADMMQ ADMMQ method from Merlot2014

References: Guidon2010, Merlot2014

Type of ADMM (sort name) as refered in literature. This sets values for METHOD, ADMM_PURIFICATION_METHOD, and EXCH_SCALING_MODEL [Edit on GitHub]

BLOCK_LIST: integer

Keyword can be repeated.

Usage: LIST {integer} {integer} .. {integer}

Specifies a list of atoms. [Edit on GitHub]

EPS_FILTER: real = 0.00000000E+000

Usage: EPS_FILTER

Define accuracy of DBCSR operations [Edit on GitHub]

EXCH_CORRECTION_FUNC: enum = DEFAULT

Valid values:

  • DEFAULT Use PBE-based corrections according to the chosen interaction operator.

  • PBEX Use PBEX functional for exchange correction.

  • NONE No correction: X(D)-x(d)-> 0.

  • OPTX Use OPTX functional for exchange correction.

  • BECKE88X Use Becke88X functional for exchange correction.

  • PBEX_LIBXC Use PBEX functional (LibXC implementation) for exchange correction.

  • BECKE88X_LIBXC Use Becke88X functional (LibXC implementation) for exchange correction.

  • OPTX_LIBXC Use OPTX functional (LibXC implementation) for exchange correction.

  • DEFAULT_LIBXC Use PBE-based corrections (LibXC where possible) to the chosen interaction operator.

  • LDA_X_LIBXC Use Slater X functional (LibXC where possible) for exchange correction.

Exchange functional which is used for the ADMM correction. LibXC implementations require linking with LibXC [Edit on GitHub]

EXCH_SCALING_MODEL: enum = NONE

Valid values:

  • NONE No scaling is enabled, refers to methods ADMM1, ADMM2 or ADMMQ.

  • MERLOT Exchange scaling according to Merlot (2014)

Scaling of the exchange correction calculated by the auxiliary density matrix. [Edit on GitHub]

METHOD: enum = BASIS_PROJECTION

Valid values:

  • BASIS_PROJECTION Construct auxiliary density matrix from auxiliary basis.

  • BLOCKED_PROJECTION_PURIFY_FULL Construct auxiliary density from a blocked Fock matrix, but use the original matrix for purification.

  • BLOCKED_PROJECTION Construct auxiliary density from a blocked Fock matrix.

  • CHARGE_CONSTRAINED_PROJECTION Construct auxiliary density from auxiliary basis enforcing charge constrain.

Method that shall be used for wavefunction fitting. Use BASIS_PROJECTION for MD. [Edit on GitHub]

OPTX_A1: real = 1.05151000E+000

OPTX a1 coefficient [Edit on GitHub]

OPTX_A2: real = 1.43169000E+000

OPTX a2 coefficient [Edit on GitHub]

OPTX_GAMMA: real = 6.00000000E-003

OPTX gamma coefficient [Edit on GitHub]