PFEAST: A High Performance Sparse Eigenvalue Solver Using Distributed-Memory Linear Solvers

The FEAST algorithm and eigensolver for interior eigenvalue problems naturally possesses three distinct levels of parallelism. The solver is then suited to exploit modern computer architectures containing many interconnected processors. This paper highlights a recent development within the software package that allows the dominant computational task, solving a set of complex linear systems, to be performed with a distributed memory solver. The software, written with a reverse-communication-interface, can now be interfaced with any generic MPI linear-system solver using a customized data distribution for the eigenvector solutions. This work utilizes two common “black-box” distributed memory linear-systems solvers (Cluster-MKL-Pardiso and MUMPS), as well as our own application-specific domain-decomposition MPI solver, for a collection of 3-dimensional finite-element systems. We discuss and analyze how parallel resources can be placed at all three levels simultaneously in order to achieve good scalability and optimal use of the computing platform.

[1]  Andreas Stathopoulos,et al.  PRIMME: preconditioned iterative multimethod eigensolver—methods and software description , 2010, TOMS.

[2]  Hiroto Tadano,et al.  A numerical method for nonlinear eigenvalue problems using contour integrals , 2009, JSIAM Lett..

[3]  Ping Tak Peter Tang,et al.  FEAST As A Subspace Iteration Eigensolver Accelerated By Approximate Spectral Projection , 2013, SIAM J. Matrix Anal. Appl..

[4]  E. Polizzi,et al.  Non-linear eigensolver-based alternative to traditional SCF methods. , 2012, The Journal of chemical physics.

[5]  D. Sorensen.  Numerical methods for large eigenvalue problems , 2002, Acta Numerica.

[6]  Eric Polizzi,et al.  A High-Performance Numerical Library for Solving Eigenvalue Problems: FEAST Solver v2.0 User's Guide , 2012, ArXiv.

[7]  Eric Polizzi,et al.  A threaded SPIKE algorithm for solving general banded systems , 2011, Parallel Comput..

[8]  Braegan S Spring.  Enhanced Capabilities of the Spike Algorithm and a New Spike-OpenMP Solver , 2014 .

[9]  S. Laux.  Solving complex band structure problems with the FEAST eigenvalue algorithm , 2012 .

[10]  G. G. Stokes.  "J." , 1890, The New Yale Book of Quotations.

[11]  W. Kohn,et al.  Self-Consistent Equations Including Exchange and Correlation Effects , 1965 .

[12]  Eric Polizzi,et al.  FEAST fundamental framework for electronic structure calculations: Reformulation and solution of the muffin-tin problem , 2011, Comput. Phys. Commun..

[13]  Ping Tak Peter Tang,et al.  Feast Eigensolver for Non-Hermitian Problems , 2015, SIAM J. Sci. Comput..

[14]  Ahmed H. Sameh,et al.  A parallel hybrid banded system solver: the SPIKE algorithm , 2006, Parallel Comput..

[15]  Endong Wang,et al.  Intel Math Kernel Library , 2014 .

[16]  Eric Polizzi,et al.  Spectral-based propagation schemes for time-dependent quantum systems with application to carbon nanotubes , 2010 .

[17]  Ping Tak Peter Tang,et al.  Zolotarev Quadrature Rules and Load Balancing for the FEAST Eigensolver , 2014, SIAM J. Sci. Comput..

[18]  Eric Polizzi,et al.  Enhancing the performance and robustness of the FEAST eigensolver , 2016, 2016 IEEE High Performance Extreme Computing Conference (HPEC).

[19]  Tetsuya Sakurai,et al.  A block Arnoldi-type contour integral spectral projection method for solving generalized eigenvalue problems , 2014, Appl. Math. Lett..

[20]  Yvan Notay,et al.  JADAMILU: a software code for computing selected eigenvalues of large sparse symmetric matrices , 2007, Comput. Phys. Commun..

[21]  Chao Yang,et al.  ARPACK users' guide - solution of large-scale eigenvalue problems with implicitly restarted Arnoldi methods , 1998, Software, environments, tools.

[22]  Andrea Toselli,et al.  Domain decomposition methods : algorithms and theory , 2005 .

[23]  Eric Polizzi,et al.  A Density Matrix-based Algorithm for Solving Eigenvalue Problems , 2009, ArXiv.

[24]  Ed Anderson,et al.  LAPACK Users' Guide , 1995 .

[25]  Edoardo Di Napoli,et al.  Efficient estimation of eigenvalue counts in an interval , 2013, Numer. Linear Algebra Appl..

[26]  T. Sakurai,et al.  CIRR: a Rayleigh-Ritz type method with contour integral for generalized eigenvalue problems , 2007 .

[27]  O. Schenk,et al.  ON FAST FACTORIZATION PIVOTING METHODS FOR SPARSE SYMMETRI C INDEFINITE SYSTEMS , 2006 .

[28]  Jack Dongarra,et al.  ScaLAPACK Users' Guide , 1987 .

[29]  Ping Tak Peter Tang,et al.  A new highly parallel non-Hermitian eigensolver , 2014, SpringSim.

[30]  Lloyd N. Trefethen,et al.  Computing Eigenvalues of Real Symmetric Matrices with Rational Filters in Real Arithmetic , 2015, SIAM J. Sci. Comput..

[31]  Patrick Amestoy,et al.  MUMPS : A General Purpose Distributed Memory Sparse Solver , 2000, PARA.

[32]  Merico E. Argentati,et al.  Block Locally Optimal Preconditioned Eigenvalue Xolvers (BLOPEX) in hypre and PETSc , 2007, SIAM J. Sci. Comput..

[33]  Konstantin Arturov,et al.  Asynchronous Approach to Memory Management in Sparse Multifrontal Methods on Multiprocessors , 2013 .

[34]  Ville Havu,et al.  All-electron density functional theory and time-dependent density functional theory with high-order finite elements. , 2009, The Journal of chemical physics.

[35]  B. Parlett.  The Symmetric Eigenvalue Problem , 1981 .

[36]  SkjellumAnthony,et al.  A high-performance, portable implementation of the MPI message passing interface standard , 1996 .

[37]  T. Sakurai,et al.  A projection method for generalized eigenvalue problems using numerical integration , 2003 .

[38]  Tetsuya Sakurai,et al.  A filter diagonalization for generalized eigenvalue problems based on the Sakurai-Sugiura projection method , 2008, J. Comput. Appl. Math..