Cluster Perturbation Theory

Cluster Perturbation Theory (CP) extends coupled cluster methods to include perturbative corrections to better capture correlation effects. These methods are particularly useful for systems where full coupled cluster calculations may be computationally prohibitive. These methods provide a range of accuracy and computational cost, allowing users to select the approach best suited to their specific system and computational resources.

Available Cluster Perturbation Theory Tasks

Task	Description
CPS(D)	Coupled Cluster Singles and perturbative Doubles.
CPSD(T)	Coupled Cluster Singles and Doubles with perturbative Triples.

General Keywords for CP Tasks

order
Description	Specifies the perturbation order of the CP calculation.
Type	int
Default	3

solver_thresh
Description	Convergence threshold for the CP solver.
Type	float
Default	1.0E-06

parent_thresh
Description	Convergence threshold for the CC parent state equations.
Type	float
Default	1.0E-06

max_iter
Description	Maximum number of iterations allowed in the CP/CC solver.
Type	int
Default	100

crop
Description	Whether CROP acceleration should be used in the CP/CC solver. If False, DIIS is used.
Type	bool
Default	True

diis_max
Description	Number of CROP/DIIS vectors stored at a time in the CP/CC solver.
Type	int
Default	10

restart
Description	Enables restarting the calculation from a checkpoint file in a .h5 format.
Type	bool
Default	True

frozencore
Description	Enables freezing of core orbitals to reduce computational cost.
Type	bool
Default	False

cholesky
Description	Enables Cholesky decomposition of AO electron repulsion integrals (ERIs).
Type	bool
Default	False
Availability	CPS(D) CPSD(T)

max_cholesky
Description	Upper limit of the number of Cholesky vectors allowed. In practical cases cholesky thresh defines the used amount of Cholesky vectors.
Type	float
Default	1.0E-09

cholesky_thresh
Description	Threshold for constructing Cholesky vectors.
Type	float
Default	solver_thresh / 1.0E-02

reduce_print
Description	Suppresses iteration information from the linear solver.
Type	bool
Default	True