CPHF

References: DelBen2013

Parameters influencing the solution of the Z-vector equations in MP2 gradients calculations. [Edit on GitHub]

Keywords

• DO_POLAK_RIBIERE

• ENFORCE_DECREASE

• EPS_CONV

• MAX_ITER

• RECALC_RESIDUAL

• RESTART_EVERY

• SCALE_STEP_SIZE

• SOLVER_METHOD

Keyword descriptions

DO_POLAK_RIBIERE: logical = F

Lone keyword: T

Usage: ENFORCE_DECREASE T

Use a Polak-Ribiere update of the search vector in CG instead of the Fletcher Reeves update. Improves the convergence with modified step sizes. Ignored with other methods than CG. [Edit on GitHub]

ENFORCE_DECREASE: logical = F

Lone keyword: T

Usage: ENFORCE_DECREASE T

Restarts if residual does not decrease. [Edit on GitHub]

EPS_CONV: real = 1.00000000E-004

Usage: EPS_CONV 1.0E-6

Convergence threshold for the solution of the Z-vector equations. The Z-vector equations have the form of a linear system of equations Ax=b, convergence is achieved when |Ax-b|<=EPS_CONV. [Edit on GitHub]

MAX_ITER: integer = 30

Aliases: MAX_NUM_ITER

Usage: MAX_ITER 50

Maximum number of iterations allowed for the solution of the Z-vector equations. [Edit on GitHub]

RECALC_RESIDUAL: logical = F

Lone keyword: T

Usage: RECALC_RESIDUAL T

Recalculates residual in every step. [Edit on GitHub]

RESTART_EVERY: integer = 5

Usage: RESTART_EVERY 5

Restart iteration every given number of steps. [Edit on GitHub]

SCALE_STEP_SIZE: real = 1.00000000E+000

Usage: SCALE_STEP_SIZE 1.0

Scaling factor of each step. [Edit on GitHub]

SOLVER_METHOD: enum = POPLE

Usage: SOLVER_METHOD POPLE

Valid values:

• POPLE Pople’s method (Default).

• CG Conjugated gradient method (equivalent to Pople).

• RICHARDSON Richardson iteration

• SD Steepest Descent iteration

Chose solver of the z-vector equations. [Edit on GitHub]