GAMESS-UK

Author

M. Dupuis, D. Sprangler and J. J. Wendoloski (NRCC, USA). Extended and modified by M. F. Guest, J. H. van Lenthe, J. Kendrick, K. Schoffel, P. Sherwood, and R. J. Harrison, with contributions from R. D. Amos, R. J. Buenker, M. Dupuis, N. C. Handy, I. H. Hillier, P. J. Knowles, V. Bonacic-Koutecky, W. von Niessen, V. R. Saunders, and A. J. Stone.

Origin

The UK version of GAMESS is derived from the original GAMESS code due to M. Dupuis, D. Spangler and J. Wendoloski, NRCC Software Catalog, Vol. 1, Program No. QG01 (GAMESS), 1980. Distributed by Computing for Science (CFS) Ltd., Daresbury Laboratory, under licence.

Version

GAMESS-UK 8.0 & 7.0

Official Web Site

http://www.cfs.dl.ac.uk

Brief Program Description

General purpose ab initio molecular electronic structure program for performing SCF- and MCSCF-gradient calculations together with a variety of techniques for post-Hartree-Fock calculations. The program utilises the Rys Polynomial or Rotation techniques to evaluate repulsion integrals over s, p and d type Cartesian Gaussian orbitals. Open- and closed-shell SCF treatments are available within both the RHF and UHF framework, with convergence controls provided through a hybrid scheme of level shifters and the DIIS method. In addition, GVB-SCF, CASSCF, MCSCF, DFT, MP2 and MP3 calculations may be performed.

The analytic energy gradient is available for each class of wavefunction above. Gradients for s and p Gaussians are evaluated using the algorithm due to Schlegel, while gradients involving d Gaussians utilise the Rys Polynomial Method. Geometry optimisation is performed using a quasi-Newton rank-2 update method, while transition state location is available through either a synchronous transit or trust region or "hill walking" method. Force constants may be evaluated analytically or by numerical differentiation.

Coupled Hartree-Fock (CPHF) treatment is available for the computation of molecular properties, including polarizabilities and hyperpolarizabilities and, through the calculation of dipole moment and polarizability derivatives, the computation of IR and Raman intensities.

Ab initio core potentials are provided in both semi-local and non-local formalism for valence-only MO treatments.

Conventional CI (using the table-driven selection algorithms within the framework of MR-DCI calculations), Direct-CI and Full-CI correlation treatments are available. Both CCSD and CCSD(T) calculations are available for closed shell systems. The direct calculation of molecular valence ionisation energies may be performed via Green's function techniques, using either the outer-valence Green's function (OVGF) or the two-particle-hole Tamm-Dancoff method (2ph-TDA).

The calculation of electronic transition energies and corresponding oscillator strengths may also be performed using either the Random Phase Approximation (RPA) method or the Multiconfigurational Linear Response (MCLR) procedure. Direct RPA is also available.

A variety of wavefunction analysis methods are provided, including population analysis and distributed multipole analysis, localised orbitals, graphical analysis and calculation of 1-electron properties.

New functionality available in this version includes:

• conventional and direct-RPA (random phase approximation) calculations of transition energies and oscillator strengths

• a MCLR (multi-configurational linear response) module

• inclusion of coupled cluster CCSD and CCSD(T) capabilities

• enhancements to the DFT functionality, including second derivatives

• improvements to the memory allocation/deallocation

Areas of Application

• Theoretical studies of reaction surfaces, SCF and MCSCF calculations of equilibrium and transition state geometries, and the evaluation of force constants and vibrational frequencies at these stationary points.

• Energy assignments in the entire field of electronic spectroscopy. Calculation of spectroscopic properties of highly excited molecular and ionic states. Calculation of transition moments.

• Graphical and numerical analysis of SCF and MCSCF wavefunctions.

• Generation of zero-order wavefunctions required in the more extensive treatments of dynamical correlation energy.

• Theoretical mechanistic studies of chemisorption in heterogeneous catalysis.

Implementation and Access

GAMESS-UK version 8.0 runs in parallel only on Magellan. To access the package the command rungamess8 is used as shown:

$CHEM/rungamess -p nprocs [ -k file ] jobname

where nproc is the number of processors and jobname is the name of the input datafile postfixed by .in (i.e. datafile.in). The results of the calculation will be found in a file whose name is generated from the datafile post-fixed by .out (i.e. datafile.out). -k specifies the file to be saved. Please refer to the man page (see below) for further details on the options available and the local implementation.

GAMESS-UK version 7.0 is still the default version and can be accessed using the rungamess command.

Machines

Available on Magellan.

Documentation

A short Unix-style man page, available online accessible by typing: man rungamess8, gives details on the local implementation of the program. The GAMESS User Guide and Reference Manual, containing a detailed description of the program, of the data input, and of the options available is also online in the directory $CHEM/doc/gamess8. The Manual is also available from the web site: http://www.dl.ac.uk/CFS/ by clicking on Documentation.

For version 7.0, please refer to the rungamess man page.

Literature References

References to all the methods used are available in the online documentations above.

Specialist Support

For queries regarding the usage of the code, and any problems and errors encountered whilst using GAMESS-UK, please use the gamess_uk_users mailing list.

Program Restrictions and Comments

Program restrictions:

• Gaussian functions of s, p, d, f and g may be used.

• The maximum number of contracted orbitals depends in general on the functionality requested. Conventional SCF and DFT energy and gradient calculations are now effectively open-ended (the default being 1024 functions), while much of the CI, MCSCF and CASSCF functionality remains limited to 255 functions (a function of the integral packing currently in place), Note that lifting this restriction to an arbitrary number in direct-SCF and direct-MP2 calculations is trivially performed by modifying the FORTRAN INCLUDE file.

• The maximum number of primitive functions within a contraction is 25.

• The maximum number of centres is 1000, allowing for the use of extensive point-charge arrays.