Sparse matrix multiplications for linear scaling electronic structure calculations in an atom‐centered basis set using multiatom blocks

A sparse matrix multiplication scheme with multiatom blocks is reported, a tool that can be very useful for developing linear‐scaling methods with atom‐centered basis functions. Compared to conventional element‐by‐element sparse matrix multiplication schemes, efficiency is gained by the use of the highly optimized basic linear algebra subroutines (BLAS). However, some sparsity is lost in the multiatom blocking scheme because these matrix blocks will in general contain negligible elements. As a result, an optimal block size that minimizes the CPU time by balancing these two effects is recovered. In calculations on linear alkanes, polyglycines, estane polymers, and water clusters the optimal block size is found to be between 40 and 100 basis functions, where about 55–75% of the machine peak performance was achieved on an IBM RS6000 workstation. In these calculations, the blocked sparse matrix multiplications can be 10 times faster than a standard element‐by‐element sparse matrix package. © 2003 Wiley Periodicals, Inc. J Comput Chem 24: 618–622, 2003

[1]  Vanderbilt,et al.  Generalization of the density-matrix method to a nonorthogonal basis. , 1994, Physical review. B, Condensed matter.

[2]  Christian Ochsenfeld,et al.  Locality and Sparsity of Ab Initio One-Particle Density Matrices and Localized Orbitals , 1998 .

[3]  Warren J. Hehre,et al.  AB INITIO Molecular Orbital Theory , 1986 .

[4]  M. Head‐Gordon,et al.  Curvy steps for density matrix based energy minimization: tensor formulation and toy applications , 2003 .

[5]  Martin Head-Gordon,et al.  Fractional tiers in fast multipole method calculations , 1996 .

[6]  Eric Schwegler,et al.  Linear scaling computation of the Fock matrix , 1997 .

[7]  Eric Schwegler,et al.  Linear scaling computation of the Fock matrix. II. Rigorous bounds on exchange integrals and incremental Fock build , 1997 .

[8]  Daw.  Model for energetics of solids based on the density matrix. , 1993, Physical review. B, Condensed matter.

[9]  Li,et al.  Density-matrix electronic-structure method with linear system-size scaling. , 1993, Physical review. B, Condensed matter.

[10]  I. Duff,et al.  Direct Methods for Sparse Matrices , 1987 .

[11]  Matt Challacombe,et al.  A simplified density matrix minimization for linear scaling self-consistent field theory , 1999 .

[12]  Michael J. Frisch,et al.  Achieving Linear Scaling for the Electronic Quantum Coulomb Problem , 1996, Science.

[13]  Stefan Goedecker,et al.  Low Complexity Algorithms for Electronic Structure Calculations , 1995 .

[14]  Yihan Shao,et al.  An improved J matrix engine for density functional theory calculations , 2000 .

[15]  Michael J. Frisch,et al.  A linear scaling method for Hartree–Fock exchange calculations of large molecules , 1996 .

[16]  S. Goedecker.  Linear scaling electronic structure methods , 1999 .

[17]  Gustavo E. Scuseria,et al.  Semiempirical methods with conjugate gradient density matrix search to replace diagonalization for molecular systems containing thousands of atoms , 1997 .

[18]  Trygve Helgaker,et al.  Direct optimization of the atomic-orbital density matrix using the conjugate-gradient method with a multilevel preconditioner , 2001 .

[19]  Gustavo E. Scuseria,et al.  An O(N) tight-binding study of carbon clusters up to C8640: the geometrical shape of the giant icosahedral fullerenes , 1996 .

[20]  Martin Head-Gordon,et al.  Chebyshev expansion methods for electronic structure calculations on large molecular systems , 1997 .

[21]  Martin Head-Gordon,et al.  A J matrix engine for density functional theory calculations , 1996 .

[22]  Benny G. Johnson,et al.  Linear scaling density functional calculations via the continuous fast multipole method , 1996 .

[23]  M. Challacombe.  A general parallel sparse-blocked matrix multiply for linear scaling SCF theory , 2000 .

[24]  David R. Bowler,et al.  Density matrices in O(N) electronic structure calculations: theory and applications , 1998 .

[25]  Trygve Helgaker,et al.  Direct optimization of the AO density matrix in Hartree-Fock and Kohn-Sham theories , 2000 .

[26]  Christian Ochsenfeld,et al.  A reformulation of the coupled perturbed self-consistent field equations entirely within a local atomic orbital density matrix-based scheme , 1997 .

[27]  Michael J. Frisch,et al.  Achieving linear scaling in exchange-correlation density functional quadratures , 1996 .

[28]  Eric Schwegler,et al.  Linear scaling computation of the Fock matrix. IV. Multipole accelerated formation of the exchange matrix , 1999 .

[29]  Christian Ochsenfeld,et al.  Linear and sublinear scaling formation of Hartree-Fock-type exchange matrices , 1998 .

[30]  Benny G. Johnson,et al.  THE CONTINUOUS FAST MULTIPOLE METHOD , 1994 .

[31]  David E. Manolopoulos,et al.  Canonical purification of the density matrix in electronic-structure theory , 1998 .

[32]  Yousef Saad,et al.  Iterative methods for sparse linear systems , 2003 .

[33]  Gustavo E. Scuseria,et al.  Linear scaling conjugate gradient density matrix search as an alternative to diagonalization for first principles electronic structure calculations , 1997 .

[34]  Yuefan Deng,et al.  New trends in high performance computing , 2001, Parallel Computing.

[35]  Gustavo E. Scuseria,et al.  Linear Scaling Density Functional Calculations with Gaussian Orbitals , 1999 .

[36]  Yihan Shao,et al.  Efficient evaluation of the Coulomb force in density-functional theory calculations , 2001 .

[37]  David Feller,et al.  Basis Set Selection for Molecular Calculations , 1986 .