A massively parallel exponential integrator for advection-diffusion models

This work considers the Real Leja Points Method (ReLPM), [M. Caliari, M. Vianello, L. Bergamaschi, Interpolating discrete advection-diffusion propagators at spectral Leja sequences, J. Comput. Appl. Math. 172 (2004) 79-99], for the exponential integration of large-scale sparse systems of ODEs, generated by Finite Element or Finite Difference discretizations of 3-D advection-diffusion models. We present an efficient parallel implementation of ReLPM for polynomial interpolation of the matrix exponential propagators exp(@DtA)v and @f(@DtA)v, @f(z)=(exp(z)-1)/z. A scalability analysis of the most important computational kernel inside the code, the parallel sparse matrix-vector product, has been performed, as well as an experimental study of the communication overhead. As a result of this study an optimized parallel sparse matrix-vector product routine has been implemented. The resulting code shows good scaling behavior even when using more than one thousand processors. The numerical results presented on a number of very large test cases gives experimental evidence that ReLPM is a reliable and efficient tool for the simulation of complex hydrodynamic processes on parallel architectures.

[1]  Ronnie Kosloff,et al.  Low-order polynomial approximation of propagators for the time-dependent Schro¨dinger equation , 1992 .

[2]  Marco Caliari,et al.  Accurate evaluation of divided differences for polynomial interpolation of exponential propagators , 2007, Computing.

[3]  Marco Vianello,et al.  The ReLPM Exponential Integrator for FE Discretizations of Advection-Diffusion Equations , 2004, International Conference on Computational Science.

[4]  J. Crank,et al.  A practical method for numerical evaluation of solutions of partial differential equations of the heat-conduction type , 1947, Mathematical Proceedings of the Cambridge Philosophical Society.

[5]  Marco Vianello,et al.  Efficient approximation of the exponential operator for discrete 2D advection–diffusion problems , 2003, Numer. Linear Algebra Appl..

[6]  Mayya Tokman,et al.  Efficient integration of large stiff systems of ODEs with exponential propagation iterative (EPI) methods , 2006, J. Comput. Phys..

[7]  L. Bergamaschi,et al.  Interpolating discrete advection-diffusion propagators at Leja sequences , 2004 .

[8]  Marco Vianello,et al.  A Parallel Exponential Integrator for Large-Scale Discretizations of Advection-Diffusion Models , 2005, PVM/MPI.

[9]  H. Tal-Ezer,et al.  An accurate and efficient scheme for propagating the time dependent Schrödinger equation , 1984 .

[10]  Marlis Hochbruck,et al.  Exponential Integrators for Large Systems of Differential Equations , 1998, SIAM J. Sci. Comput..

[11]  Marco Vianello,et al.  Comparing Leja and Krylov Approximations of Large Scale Matrix Exponentials , 2006, International Conference on Computational Science.

[12]  B. Parlett,et al.  Accurate Computation of Divided Differences of the Exponential Function , 1984 .

[13]  I. Moret,et al.  THE COMPUTATION OF FUNCTIONS OF MATRICES BY TRUNCATED FABER SERIES , 2001 .

[14]  Message Passing Interface Forum MPI: A message - passing interface standard , 1994 .

[15]  Roger B. Sidje,et al.  Expokit: a software package for computing matrix exponentials , 1998, TOMS.

[16]  S. Cox,et al.  Exponential Time Differencing for Stiff Systems , 2002 .

[17]  I. Moret,et al.  An interpolatory approximation of the matrix exponential based on Faber polynomials , 2001 .

[18]  Marco Vianello,et al.  Interpolating discrete advection-diffusion propagators at spectral Leja sequences , 2004 .

[19]  C. Gonzáleza,et al.  A second-order Magnus-type integrator for nonautonomous parabolic problems , 2005 .

[20]  L. Bergamaschi,et al.  The LEM exponential integrator for advection-diffusion-reaction equations , 2007 .

[21]  Y. Saad Analysis of some Krylov subspace approximations to the matrix exponential operator , 1992 .

[22]  Marlis Hochbruck,et al.  Explicit Exponential Runge-Kutta Methods for Semilinear Parabolic Problems , 2005, SIAM J. Numer. Anal..

[23]  F. Leja Sur certaines suites liées aux ensembles plans et leur application à la représentation conforme , 1957 .

[24]  C. Lubich,et al.  On Krylov Subspace Approximations to the Matrix Exponential Operator , 1997 .

[25]  Youcef Saad,et al.  A Basic Tool Kit for Sparse Matrix Computations , 1990 .

[26]  S. Krogstad Generalized integrating factor methods for stiff PDEs , 2005 .

[27]  L. Reichel Newton interpolation at Leja points , 1990 .

[28]  HochbruckMarlis,et al.  Explicit Exponential Runge-Kutta Methods for Semilinear Parabolic Problems , 2005 .