Section WF_CORRELATION

• Sets up the wavefunction-based correlation methods as MP2, RI-MP2, RI-SOS-MP2, RI-RPA and GW (inside RI-RPA).

• Section path: CP2K_INPUT / FORCE_EVAL / DFT / KG_METHOD / XC / WF_CORRELATION

• This section can be repeated.

• This section cites the following references: [DelBen2012] [DelBen2013] [DelBen2015] [DelBen2015b] [Rybkin2016] [Wilhelm2016] [Wilhelm2016b] [Wilhelm2017] [Wilhelm2018]

• Subsections

  • CANONICAL_GRADIENTS
  • INTEGRALS
  • LOW_SCALING
  • MP2
  • PRINT
  • RI
  • RI_MP2
  • RI_RPA
  • RI_SOS_MP2

• Keywords

  • E_GAP
  • E_RANGE
  • GROUP_SIZE
  • MEMORY
  • SCALE_S
  • SCALE_T

• Keyword descriptions

  E_GAP
  	E_GAP {Real}
  	Gap energy for integration grids in Hartree. Defaults to -1.0 (automatic determination). Recommended to set if several RPA or SOS-MP2 gradient calculations are requested or to be restarted. In this way, differences of integration grids across different runs are removed as CP2K does not include derivatives thereof.
  	This keyword cannot be repeated and it expects precisely one real.
  	Default value: -1.00000000E+000

  E_RANGE
  	E_RANGE {Real}
  	Energy range (ratio of largest and smallest) energy difference of unoccupied and occupied orbitals for integration grids. Defaults to 0.0 (automatic determination). Recommended to set if several RPA or SOS-MP2 gradient calculations are requested or to be restarted. In this way, differences of integration grids across different runs are removed as CP2K does not include derivatives thereof.
  	This keyword cannot be repeated and it expects precisely one real.
  	Default value: -1.00000000E+000

  GROUP_SIZE
  	GROUP_SIZE {Integer}
  	Group size used in the computation of GPW and MME integrals and the MP2 correlation energy. The group size must be a divisor of the total number of MPI ranks. A smaller group size (for example the number of MPI ranks per node) accelerates the computation of integrals but a too large group size increases communication costs. A too small group size may lead to out of memory.
  	This keyword cannot be repeated and it expects precisely one integer.
  	Default value: 1
  	Alias names for this keyword: NUMBER_PROC

  MEMORY
  	MEMORY {Real}
  	Maximum allowed total memory usage during MP2 methods [MiB].
  	This keyword cannot be repeated and it expects precisely one real.
  	Default value: 1.02400000E+003

  SCALE_S
  	SCALE_S {Real}
  	Scaling factor of the singlet energy component (opposite spin, OS) of the MP2, RI-MP2 and SOS-MP2 correlation energy.
  	This keyword cannot be repeated and it expects precisely one real.
  	Default value: 1.00000000E+000

  SCALE_T
  	SCALE_T {Real}
  	Scaling factor of the triplet energy component (same spin, SS) of the MP2 and RI-MP2 correlation energy.
  	This keyword cannot be repeated and it expects precisely one real.
  	Default value: 1.00000000E+000