Section WF_CORRELATION

• Sets up the Wavefunction-based Correlation parameters if requested

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

• This section can be repeated.

• This section cites the following references: [DelBen2012] [DelBen2013] [DelBen2015] [DelBen2015b] [Rybkin2016]

• Subsections

  • CPHF
  • DIRECT_CANONICAL
  • ERI_MME
  • INTERACTION_POTENTIAL
  • MP2_INFO
  • OPT_RI_BASIS
  • RI_LAPLACE
  • RI_MP2
  • RI_RPA
  • WFC_GPW

• Keywords

  • CALC_COND_NUM
  • COL_BLOCK
  • ERI_METHOD
  • GROUP_SIZE
  • MEMORY
  • METHOD
  • RI_METRIC
  • ROW_BLOCK
  • SCALE_S
  • SCALE_T

• Keyword descriptions

  CALC_COND_NUM
  	CALC_COND_NUM {Logical}
  	Calculate the condition number of the (P|Q) matrix for the RI methods.
  	This keyword cannot be repeated and it expects precisely one logical.
  	The lone keyword behaves as a switch to .TRUE.
  	Default value: .FALSE.
  	Alias names for this keyword: CALC_CONDITION_NUMBER

  COL_BLOCK
  	COL_BLOCK {Integer}
  	Size of the column block used in the SCALAPACK block cyclic data distribution.Default is (COL_BLOCK=-1) is automatic. A proper choice can speedup the parallel matrix multiplication in the case of RI-RPA and RI-SOS-MP2-Laplace.
  	This keyword cannot be repeated and it expects precisely one integer.
  	Default value: -1
  	Alias names for this keyword: COL_BLOCK_SIZE

  ERI_METHOD
  	ERI_METHOD {Keyword}
  	Method for calculating periodic electron repulsion integrals (MME method is faster but experimental, forces and 3-center integrals not yet implemented).
  	This keyword cannot be repeated and it expects precisely one keyword.
  	Default value: GPW
  	List of valid keywords: GPW Uses Gaussian Plane Wave method [DelBen2013] MME Uses MiniMax-Ewald method (ERI_MME subsection)

  GROUP_SIZE
  	GROUP_SIZE {Integer}
  	Group size used in the computation of the integrals. Default is to use all processors (GROUP_SIZE=-1).A smaller group size (for example the node size), might a better choice if the actual MP2 time is large compared to integral computation time. This is usually the case if the total number of processors is not too large.
  	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

  METHOD
  	METHOD {Keyword}
  	Which method should be used to compute the MP2 energy
  	This keyword cannot be repeated and it expects precisely one keyword.
  	Default value: DIRECT_CANONICAL
  	List of valid keywords: DIRECT_CANONICAL Use the direct mp2 canonical approach MP2_GPW Use the GPW approach to MP2 NONE Skip MP2 calculation OPTIMIZE_RI_BASIS Optimize RIMP2 basis set RI_MP2_GPW Use the GPW approach to RI-MP2 RI_RPA_GPW Use the GPW approach to RI-RPA RI_SOS_LAPLACE Use the GPW approach to RI-SOS-Laplace-MP2
  	This keyword cites the following references: [DelBen2012] [DelBen2013]

  RI_METRIC
  	RI_METRIC {Keyword}
  	Define which RI metric should be used for calculating B^P_ia.
  	This keyword cannot be repeated and it expects precisely one keyword.
  	Default value: COULOMB
  	List of valid keywords: COULOMB Use Coulomb metric. OVERLAP Use overlap metric, recommended in combination with imaginary time for O(N^3) RPA.
  	Alias names for this keyword: RI

  ROW_BLOCK
  	ROW_BLOCK {Integer}
  	Size of the row block used in the SCALAPACK block cyclic data distribution.Default is (ROW_BLOCK=-1) is automatic. A proper choice can speedup the parallel matrix multiplication in the case of RI-RPA and RI-SOS-MP2-Laplace.
  	This keyword cannot be repeated and it expects precisely one integer.
  	Default value: -1
  	Alias names for this keyword: ROW_BLOCK_SIZE

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