Basic Keywords

To compute a ground-state CCSD or CCSD(T) energy at a fixed geometry, the following keywords are common:

WFN = string

Acceptable values are ccsd, ccsd_t [for CCSD(T)], bccd (for Brueckner-orbital-based CCD), or bccd_t [for Brueckner-orbital-based CCSD(T)] There is no default.

REFERENCE = string

Acceptable values are reference = rhf, rohf, or uhf. There is no default.

JOBTYPE = string

Acceptable values are sp, opt, freq, oeprop, or response. There is no default.

CONVERGENCE = integer

Sets the order of magnitude on the convergence of the CC wave function, perturbed wave function, and/or lambda parameters. The root-mean-square of the difference in amplitude vectors from consecutive iterations is used to determine the convergence. The default is 7.

MAXITER = integer

The maximum number of iterations allowed for solving the CC amplitude or lambda amplitude equations. Defaults to 50.

MEMORY = (real MB)

Specified the amount of core memory to be used, in MB. Defaults to 256. Other units (e.g., KB or GB) are also allowed.

BRUECKNER_CONV = integer

Specifies the order of magnitude convergence required for the Brueckner orbitals. The convergence is determined based on the largest T1 amplitude.

AO_BASIS = string

Specifies the algorithm to be used in computing the contribution of the four-virtual-index integrals ( $\langle ab\vert\vert cd\rangle$) to the CC amplitude equations. If AO_BASIS=NONE, the MO-basis integrals will be used; if AO_BASIS=DISK, the AO-basis integrals, stored on disk, will be used; if AO_BASIS=DIRECT, the AO-basis integrals will be computed on the fly as necessary. NB: The AO_BASIS=DIRECT option is not fully implemented and should only be used by experts. Default is NONE. Note: The developers recommend use of this keyword only as a last resort because it significantly slows the calculation. The current algorithms for handling the MO-basis four-virtual-index integrals have been significantly improved and are preferable to the AO-based approach.

FREEZE_CORE = boolean

Specifies whether core orbitals (which are determined automatically) are to be excluded from the correlated calculations. Default is FALSE.

RESTART = boolean

Determine whether previous amplitude vectors may be used as guesses in a given CC calculation. Defaults to TRUE. For geometry optimizations, Brueckner calculations, etc. the iterative solution of the CC amplitude equations may benefit considerably by reusing old vectors as initial guesses. Assuming that the MO phases remain the same between updates, the CC codes will, by default, re-use old vectors, unless the user sets RESTART = false.

PRINT = integer

The desired print level for detailed output. Setting this to 2 is a good idea for larger calculations so that the progress of the calculation may be easily followed. Defaults to 0.

CACHELEV = integer

Sets the level of automated cacheing of four-index quantities in the CC modules. These modules are capable of keeping in core as much as possible, various four-index quantities categorized by the number of virtual/unoccupied-orbital indices they contain. Setting CACHELEV=0 will cache nothing (wise and sometimes necessary for very large CC calculations), CACHELEV=1 will keep quantities with up to one virtual index in core (e.g., integrals of the form $\langle ij\vert\vert ka\rangle$), CACHELEV=2 will keep quantities with up two two virtual indices in core (e.g., integrals of the form $\langle ij\vert\vert ab \rangle$ or $\hat{T}_2$ amplitudes), CACHELEV=3 will keep three-virtual-index quantities in core, and CACHELEV=4 will keep everything in core. Note that the cache behavior is tempered by the MEMORY keyword, and items will be deleted from the cache (in an order determined based on the CACHETYEP keyword) as additional memory is required in a given calculation.

CACHETYPE = string

Specifies the type of cache to be used, either LOW or LRU. If CACHETYPE=LOW, then elements are deleted from the cache based on a predefined order of priority. If CACHETYPE=LRU, then elements are deleted from the cache based on a ``least recently used'' criterion: the least recently used item is the first to be deleted. The LOW criterion has been developed only ccenergy codes. The default is LRU for all CC modules except ccenergy.

NUM_AMPS = integer

Specifies the number of wave function amplitudes to print at the end of the energy calculation. Defaults to 10.

PRINT_MP2_AMPS = boolean

Specifies if the initial guess (MP2) amplitudes should be printed in the output file. Defaults to FALSE.