Linear-scaling DFT-pseudopotential calculations on parallel computers

Abstract A parallel implementation of linear-scaling first-principles total-energy calculations is presented. The theoretical basis is density functional theory (DFT) and the pseudopotential approximation. The linear-scaling method is the one due to the present authors, but the parallelisation techniques are also relevant to other linear-scaling DFT methods. The theoretical and computational framework of the linear-scaling method is summarised, in order to identify the main classes of computer operation required. For each class of operation, the issues involved in distributing tasks and data between processors are discussed and a solution is proposed. Practical tests of the proposed implementation on a Cray T3D are presented, and it is shown that the scaling with respect to both the number of atoms and the number of processors is excellent for systems containing up to over 6 000 atoms.

[1]  Liu.  Self-consistent tight-binding method. , 1995, Physical review. B, Condensed matter.

[2]  Ladislas P. Kubin,et al.  Computer Simulation in Materials Science , 1991 .

[3]  M. Teter,et al.  Tight-binding electronic-structure calculations and tight-binding molecular dynamics with localized orbitals. , 1994, Physical review. B, Condensed matter.

[4]  Hernández,et al.  Self-consistent first-principles technique with linear scaling. , 1995, Physical review. B, Condensed matter.

[5]  Francesco Mauri,et al.  O(N) tight-binding molecular dynamics on massively parallel computers : an orbital decomposition approach , 1996 .

[6]  Martin,et al.  Structure and energetics of giant fullerenes: An order-N molecular-dynamics study. , 1996, Physical review. B, Condensed matter.

[7]  Galli,et al.  Electronic-structure calculations and molecular-dynamics simulations with linear system-size scaling. , 1994, Physical review. B, Condensed matter.

[8]  Michael P. Allen,et al.  Computer simulation in chemical physics , 1993 .

[9]  Kohn,et al.  Density functional and density matrix method scaling linearly with the number of atoms. , 1996, Physical review letters.

[10]  Car,et al.  Unified approach for molecular dynamics and density-functional theory. , 1985, Physical review letters.

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

[12]  White,et al.  Implementation of gradient-corrected exchange-correlation potentials in Car-Parrinello total-energy calculations. , 1994, Physical review. B, Condensed matter.

[13]  Vogl,et al.  Simple model for structural properties and crystal stability of sp-bonded solids. , 1987, Physical review. B, Condensed matter.

[14]  Jackson,et al.  Atoms, molecules, solids, and surfaces: Applications of the generalized gradient approximation for exchange and correlation. , 1992, Physical review. B, Condensed matter.

[15]  William H. Press,et al.  The Art of Scientific Computing Second Edition , 1998 .

[16]  Martin,et al.  Linear system-size scaling methods for electronic-structure calculations. , 1995, Physical review. B, Condensed matter.

[17]  Martin,et al.  Unconstrained minimization approach for electronic computations that scales linearly with system size. , 1993, Physical review. B, Condensed matter.

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

[19]  Leonard Kleinman,et al.  Efficacious Form for Model Pseudopotentials , 1982 .

[20]  Galli,et al.  Large scsle quantum simulations: C60 Impacts on a semiconducting surface. , 1994, Physical review letters.

[21]  Walter Kohn,et al.  Density functional theory for systems of very many atoms , 1995 .

[22]  O. Sankey,et al.  Ab initio multicenter tight-binding model for molecular-dynamics simulations and other applications in covalent systems. , 1989, Physical review. B, Condensed matter.

[23]  A. Becke,et al.  Density-functional exchange-energy approximation with correct asymptotic behavior. , 1988, Physical review. A, General physics.

[24]  Kim,et al.  Total-energy global optimizations using nonorthogonal localized orbitals. , 1995, Physical review. B, Condensed matter.

[25]  Hernández,et al.  Linear-scaling density-functional-theory technique: The density-matrix approach. , 1996, Physical review. B, Condensed matter.

[26]  William H. Press,et al.  Numerical recipes in C. The art of scientific computing , 1987 .

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

[28]  P. Giannozzi,et al.  Towards Very Large-Scale Electronic-Structure Calculations , 1992 .

[29]  Colombo,et al.  Efficient linear scaling algorithm for tight-binding molecular dynamics. , 1994, Physical review letters.

[30]  Satoshi Itoh,et al.  Order-N tight-binding molecular dynamics on parallel computers , 1995 .

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

[32]  Car,et al.  Orbital formulation for electronic-structure calculations with linear system-size scaling. , 1993, Physical review. B, Condensed matter.

[33]  E. Stechel,et al.  Order-N methods in self-consistent density-functional calculations. , 1994, Physical review. B, Condensed matter.

[34]  Williams,et al.  N-scaling algorithm for density-functional calculations of metals and insulators. , 1994, Physical review. B, Condensed matter.

[35]  T. Arias,et al.  Iterative minimization techniques for ab initio total energy calculations: molecular dynamics and co , 1992 .

[36]  Mike C. Payne,et al.  Large-scale ab initio total energy calculations on parallel computers , 1992 .