Zolotarev Quadrature Rules and Load Balancing for the FEAST Eigensolver

The FEAST method for solving large sparse eigenproblems is equivalent to subspace iteration with an approximate spectral projector and implicit orthogonalization. This relation allows to characterize the convergence of this method in terms of the error of a certain rational approximant to an indicator function. We propose improved rational approximants leading to FEAST variants with faster convergence, in particular, when using rational approximants based on the work of Zolotarev. Numerical experiments demonstrate the possible computational savings especially for pencils whose eigenvalues are not well separated and when the dimension of the search space is only slightly larger than the number of wanted eigenvalues. The new approach improves both convergence robustness and load balancing when FEAST runs on multiple search intervals in parallel.

[1]  Eric Polizzi,et al.  Efficient modeling techniques for atomistic-based electronic density calculations , 2008 .

[2]  Vladimir Druskin,et al.  Optimal finite difference grids and rational approximations of the square root I. Elliptic problems , 2000 .

[3]  Irene A. Stegun,et al.  Pocketbook of mathematical functions , 1984 .

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

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

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

[7]  Daniel Kressner,et al.  An Error Analysis of Galerkin Projection Methods for Linear Systems with Tensor Product Structure , 2013, SIAM J. Numer. Anal..

[8]  H.K. Kim.  Filtering in the time and frequency domains , 1978, Proceedings of the IEEE.

[9]  Tetsuya Sakurai,et al.  CONTOUR INTEGRAL EIGENSOLVER FOR NON-HERMITIAN SYSTEMS: A RAYLEIGH-RITZ-TYPE APPROACH , 2010 .

[10]  John Todd,et al.  Applications of Transformation Theory: A Legacy from Zolotarev (1847–1878) , 1984 .

[11]  J. N. Lyness,et al.  A Numerical Method for Locating the Zeros of an Analytic Function , 1967 .

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

[13]  Philip Rabinowitz,et al.  Methods of Numerical Integration , 1985 .

[14]  Van Valkenburg,et al.  Analog Filter Design , 1982 .

[15]  P. Petrushev,et al.  Rational Approximation of Real Functions , 1988 .

[16]  Lukas Krämer,et al.  On the parallel iterative solution of linear systems arising in the FEAST algorithm for computing inner eigenvalues , 2015, Parallel Comput..

[17]  W. Cauer,et al.  Ein Interpolationsproblem mit Funktionen mit positivem Realteil , 1934 .

[18]  J. Dicapua.  Chebyshev Polynomials , 2019, Fibonacci and Lucas Numbers With Applications.

[19]  Lloyd N. Trefethen,et al.  Numerical Algorithms Based on Analytic Function Values at Roots of Unity , 2014, SIAM J. Numer. Anal..

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

[21]  Ben Silver,et al.  Elements of the theory of elliptic functions , 1990 .

[22]  VLADIMIR DRUSKIN,et al.  Near-Optimal Perfectly Matched Layers for Indefinite Helmholtz Problems , 2015, SIAM Rev..

[23]  T. J. Rivlin.  The Chebyshev polynomials , 1974 .

[24]  Lloyd N. Trefethen,et al.  The Exponentially Convergent Trapezoidal Rule , 2014, SIAM Rev..

[25]  Lukas Krämer,et al.  The FEAST algorithm for large eigenvalue problems , 2011 .

[26]  V. I. Lebedev,et al.  On a Zolotarev problem in the method of alternating directions , 1977 .

[27]  Tetsuya Sakurai,et al.  Efficient Parameter Estimation and Implementation of a Contour Integral-Based Eigensolver , 2013 .

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

[29]  Mitsuhisa Sato,et al.  A parallel method for large sparse generalized eigenvalue problems using a GridRPC system , 2008, Future Gener. Comput. Syst..

[30]  Eugene L. Wachspress.  The ADI minimax problem for complex spectra , 1990 .

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

[32]  John Sabino,et al.  Solution of Large-Scale Lyapunov Equations via the Block Modified Smith Methods , 2006 .

[33]  Lukas Krämer,et al.  Dissecting the FEAST algorithm for generalized eigenproblems , 2012, J. Comput. Appl. Math..

[34]  V. I. Lebedev,et al.  Variable time steps optimization of Lω -stable Crank–Nicolson method , 2005 .

[35]  C. K. Yuen,et al.  Digital Filters , 1979, IEEE Transactions on Systems, Man, and Cybernetics.

[36]  S. Güttel.  Rational Krylov approximation of matrix functions: Numerical methods and optimal pole selection , 2013 .

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

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

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