Performance of the divide-and-conquer approach used as an initial guess

Abstract Computational costs with the use of the initial guesses formed by the divide-and-conquer approach are examined. For conventional density functional theory calculations, their computational costs rapidly increase as the system sizes increase, therefore, an efficient methodologies to construct a fine initial molecular orbital guess or a fine density matrix (DM) is quite variable in order to enhance the fast convergence. We examined the DM construction approaches by changing the subsystem buffer sizes and number of divisions, and their computational costs were numerically investigated in order to clarify their characteristic features.

[1]  K. Kitaura,et al.  Definition of molecular orbitals in fragment molecular orbital method , 2002 .

[2]  Hiromi Nakai.  Themed issue: Fragment and localized orbital methods in electronic structure theory , 2012 .

[3]  P. Mezey,et al.  Diagonalization-free initial guess to SCF calculations for large molecules , 2006 .

[4]  Tai-Sung Lee,et al.  Linear-scaling quantum mechanical calculations of biological molecules: The divide-and-conquer approach , 1998 .

[5]  Weitao Yang,et al.  A density‐matrix divide‐and‐conquer approach for electronic structure calculations of large molecules , 1995 .

[6]  Shinya Honda,et al.  10 residue folded peptide designed by segment statistics. , 2004, Structure.

[7]  Fumitoshi Sato,et al.  Convergence process with quasi-canonical localized orbital in all-electron SCF calculation on proteins , 2003 .

[8]  Walter Kohn,et al.  Nobel Lecture: Electronic structure of matter-wave functions and density functionals , 1999 .

[9]  K. Kitaura,et al.  Fragment molecular orbital method: an approximate computational method for large molecules , 1999 .

[10]  Yang,et al.  Direct calculation of electron density in density-functional theory. , 1991, Physical review letters.

[11]  R. Parr.  Density-functional theory of atoms and molecules , 1989 .

[12]  Tjerk P. Straatsma,et al.  NWChem: A comprehensive and scalable open-source solution for large scale molecular simulations , 2010, Comput. Phys. Commun..

[13]  Hiroshi Kashiwagi,et al.  All‐electron density functional calculation on insulin with quasi‐canonical localized orbitals , 2005, J. Comput. Chem..

[14]  Masato Kobayashi,et al.  Divide-and-conquer self-consistent field calculation for open-shell systems: Implementation and application , 2010 .

[15]  Tetsuya Sakurai,et al.  Parallel Fock matrix construction with distributed shared memory model for the FMO‐MO method , 2010, J. Comput. Chem..

[16]  Yuichi Inadomi,et al.  FMO-MO Method as an Initial Guess Generation for SCF Calculation: Case of (-)-Epicatechin Gallate , 2008 .

[17]  Wei Pan,et al.  Parallel implementation of divide‐and‐conquer semiempirical quantum chemistry calculations , 1998 .

[18]  Yuichi Inadomi,et al.  Fragment Molecular Orbital (FMO) and FMO-MO Calculations of DNA: Accuracy Validation of Energy and Interfragment Interaction Energy , 2009 .