Search

Cyclops (cyclic-operations) Tensor Framework (CTF) 1 is a distributed library for tensor contractions. CTF aims to scale high-dimensional tensor contractions such as those required in the Coupled Cluster (CC) electronic structure method to massively-parallel supercomputers. The framework preserves tensor structure by subdividing tensors cyclically, producing a regular parallel decomposition. An internal virtualization layer provides completely general mapping support while maintaining ideal load balance. The mapping framework decides on the best mapping for each tensor contraction at run-time via explicit calculations of memory usage and communication volume. CTF employs a general redistribution kernel, which transposes tensors of any dimension between arbitrary distributed layouts, yet touches each piece of data only once. Sequential symmetric contractions are reduced to matrix multiplication calls via tensor index transpositions and partial unpacking. The user-level interface elegantly expresses arbitrary-dimensional generalized tensor contractions in the form of a domain specific language. We demonstrate performance of CC with single and double excitations on 8192 nodes of Blue Gene/Q and show that CTF outperforms NWChem on Cray XE6 supercomputers for benchmarked systems.

We extend our assessment of the potential of perturbative coupled cluster (CC) expansions for a test set of open-shell atoms and organic radicals to the description of quadruple excitations. Namely, the second- through sixth-order models of the recently proposed CCSDT(Q-n) quadruples series [J. Chem. Phys. 140, 064108 (2014)] are compared to the prominent CCSDT(Q) and lambda-CCSDT(Q) models. From a comparison of the models in terms of their recovery of total CC singles, doubles, triples, and quadruples (CCSDTQ) energies, we find that the performance of the CCSDT(Q-n) models is independent of the reference used (unrestricted or restricted (open-shell) Hartree-Fock), in contrast to the CCSDT(Q) and lambda-CCSDT(Q) models, for which the accuracy is strongly dependent on the spin of the molecular ground state. By further comparing the ability of the models to recover relative CCSDTQ atomization energies, the discrepancy between them is found to be even more pronounced, stressing how a balanced description of both closed- and open-shell species - as found in the CCSDT(Q-n) models - is indeed of paramount importance if any perturbative CC model is to be of chemical relevance for high-accuracy applications. In particular, the third-order CCSDT(Q-3) model is found to offer an encouraging alternative to the existing choices of quadruples models used in modern computational thermochemistry, since the model is still only of moderate cost, albeit markedly more costly than, e.g., the CCSDT(Q) and lambda-CCSDT(Q) models., 18 pages, 3 figures, 1 supporting information (attached as an ancillary file)

We have calculated frequencies and intensities of fundamental and overtone vibrational transitions in water and water dimer with use of different vibrational methods. We have compared results obtained with correlation-corrected vibrational self-consistent-field theory and vibrational second-order perturbation theory both using normal modes and finally with a harmonically coupled anharmonic oscillator local mode model including OH-stretching and HOH-bending local modes. The coupled cluster with singles, doubles, and perturbative triples ab initio method with augmented correlation-consistent triple-zeta Dunning and atomic natural orbital basis sets has been used to obtain the necessary potential energy and dipole moment surfaces. We identify the strengths and weaknesses of these different vibrational approaches and compare our results to the available experimental results.

The problem of the generally inferior convergence behavior of higher-order coupled cluster methods, such as CCSDT and CCSDTQ, compared to CCSD is analyzed in terms of Møller-Plesset perturbation theory. A new structure for the CCSDT and CCSDTQ equations (and various approximations of these) is proposed which reorders contributions between the various cluster amplitudes and emphasizes lower-order corrections to the energy at each iteration. Numerical testing of the proposed method compared to the widely used direct inversion in the iterative subspace convergence acceleration technique shows significant improvement in the rate of convergence and total time-to-solution, especially for methods including quadruple excitations.

The theory of non-orthogonal spin-adaptation for closed-shell molecular systems is applied to coupled cluster methods with quadruple excitations (CCSDTQ). Calculations at this level of detail are of critical importance in describing the properties of molecular systems to an accuracy which can meet or exceed modern experimental techniques. Such calculations are of significant (and growing) importance in such fields as thermodynamics, kinetics, and atomic and molecular spectroscopies. With respect to the implementation of CCSDTQ and related methods, we show that there are significant advantages to non-orthogonal spin-adaption with respect to simplification and factorization of the working equations and to creating an efficient implementation. The resulting algorithm is implemented in the CFOUR program suite for CCSDT, CCSDTQ, and various approximate methods (CCSD(T), CC3, CCSDT-n, and CCSDT(Q)).