Section LBFGS

• Provides parameters to tune the limited memory BFGS (LBFGS) optimization

• Section path: CP2K_INPUT / MOTION / GEO_OPT / LBFGS

• This section cannot be repeated.

• This section cites the following reference: [BYRD1995]

• Subsections

  • none

• Keywords

  • MAX_F_PER_ITER
  • MAX_H_RANK
  • TRUST_RADIUS
  • WANTED_PROJ_GRADIENT
  • WANTED_REL_F_ERROR

• Keyword descriptions

  MAX_F_PER_ITER
  	MAX_F_PER_ITER {Integer}
  	Maximum number of force evaluations per iteration(used for the line search)
  	This keyword cannot be repeated and it expects precisely one integer.
  	Default value: 20

  MAX_H_RANK
  	MAX_H_RANK {Integer}
  	Maximum rank (and consequently size) of the approximate Hessian matrix used by the LBFGS optimizer. Larger values (e.g. 30) will accelerate the convergence behaviour at the cost of a larger memory consumption.
  	This keyword cannot be repeated and it expects precisely one integer.
  	Default value: 5

  TRUST_RADIUS
  	TRUST_RADIUS {Real}
  	Trust radius used in LBFGS. Not completely in depth tested. Negativ values means no trust radius is used.
  	This keyword cannot be repeated and it expects precisely one real.
  	Default value: -5.29177209E-001
  	Default unit: [angstrom]

  WANTED_PROJ_GRADIENT
  	WANTED_PROJ_GRADIENT {Real}
  	Convergence criterion (overrides the general ones):Requested norm threshold of the gradient multiplied by the approximate Hessian.
  	This keyword cannot be repeated and it expects precisely one real.
  	Default value: 1.00000000E-016
  	Default unit: [internal_cp2k]

  WANTED_REL_F_ERROR
  	WANTED_REL_F_ERROR {Real}
  	Convergence criterion (overrides the general ones):Requested relative error on the objective functionof the optimizer (the energy)
  	This keyword cannot be repeated and it expects precisely one real.
  	Default value: 1.00000000E-016
  	Default unit: [internal_cp2k]