Domain Decomposition Techniques for Large Sparse Nonsymmetric Systems Arising from Elliptic Problems with First-Order Terms

  • >David E. Keyes
  • William D. Gropp
  • Ali Ecder

Abstract

Parallel block-preconditioned domain-decomposed Krylov methods for sparse linear systems are described and illustrated on two-dimensional model problems of algebraic dimension up to 65,025. Four convective-diffusive transport problems typical of implicit upwind finite-difference discretizations of heat and mass transfer applications (pure conduction, a plug flow, a jet flow, and a recirculating flow) are tested for practicality of parallel solution under the domain decomposition paradigm. The discrete operators corresponding to the latter two lack constant coefficients and symmetry, and there is little iterative convergence theory to guide their solution, but much practical progress can be made. We describe techniques depending only on the sparsity structure and approximate diagonal dominance of the linear operator and thus of broad applicability. Results of tests run on an Encore Multimax with up to 16 processors demonstrate their utility in the coarse-granularity parallelization of hydrocodes.

Keywords

Convection Petrol Hull Peri Ditioned 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [1]
    C. Ashcraft,Domain Decoupled Incomplete Factorizations, Technical Report ETA-TR-49, Boeing Computer Services, April 1987.Google Scholar
  2. [2]
    O. Axelsson & G. Lindskog,On the Eigenvalue Distribution of a Class of Preconditioning Methods, Numer. Math., 48 (1986), pp. 479 – 498.CrossRefGoogle Scholar
  3. [3]
    J. H. Bramble, J. E. Pasciak & A. H. Schatz,The Construction of Preconditioners for Elliptic Problems by Substructuring, I, Math. Comp., 47 (1986), pp. 103 – 134.CrossRefGoogle Scholar
  4. [4]
    T. F. Chan & D. Goovaerts,Domain Decomposition Beneficial Even Sequentially, Technical Report 88-18, UCLA Comp. and App. Math., June 1988.Google Scholar
  5. [5]
    T. F. Chan & T. Y. Hou,Domain Decomposition Preconditioners for General Second Order Elliptic Operators, Technical Report 88-16, UCLA Comp. and App. Math., June 1988.Google Scholar
  6. [6]
    T. F. Chan & D. Resasco,A Survey of Preconditioners for Domain Decomposition, Technical Report 414, Computer Science Dept., Yale University, September 1985. In Proceedings of the IV Coloquio de Matemáticas del CINVESTAV, Workshop in Numerical Analysis and its applications, Taxco, Mexico, Aug. 18–24, 1985.Google Scholar
  7. [7]
    T. F. Chan & D. Resasco,A Domain-Decomposed Fast Poisson Solver on a Rectangle, SIAM J. Sci. Stat. Comp., 8 (1987), pp. sl4–s26.Google Scholar
  8. [8]
    T. F. Chan & D. Resasco, A Framework for the Analysis and Construction of Domain Decomposition Preconditioners, R. Glowinski, G. H. Golub, G. A. Meurant & J. Periaux (eds.),First International Symposium on Domain Decomposition Methods for Partial Differential Equations, SIAM, 1988, pp. 217–230.Google Scholar
  9. [9]
    R. W. Cottle,Manifestations of the Schur Complement, Lin. Alg. Appl., 8 (1974), pp. 189 – 211.CrossRefGoogle Scholar
  10. [10]
    A. R. Curtis, M. J. Powell & J. K. Reid,On the Estimation of Sparse Jacobian Matrices, J. Inst. Math. Appl., 13 (1974), pp. 117 – 119.Google Scholar
  11. [11]
    M. Dryja & W. Proskurowski,Capacitance Matrix Method using Strips with Alternating Neumann and Dirichlet Boundary Conditions, Appl. Numer. Math., 1 (1985).Google Scholar
  12. [12]
    M. Dryja & O. B. Widlund,An Additive Variant of the Schwarz Alternating Method for the Case of Many Subregions, Technical Report 339, Courant Institute of Mathematical Sciences, NYU, December 1987.Google Scholar
  13. [13]
    T. Dupont, R. Kendall & H. H. Rachford,An Approximate Factorization Procedure for Solving Self-Adjoint Elliptic Difference Equations, SIAM J. Num. Anal., 5 (1968), pp. 559 – 573.CrossRefGoogle Scholar
  14. [14]
    A. Ecder& D. E. Keyes,A Parametric Study of the Spectra and Convergence Rates of Preconditioned Generalized Minimum Residual Schemes for Convective-Diffusive Problems, 1988. (in preparation).Google Scholar
  15. [15]
    H. C. Elman & E. Agrón,Ordering Techniques for the Preconditioned Conjugate Gradient Method on Parallel Computers, Comp. Phys. Comm. (1989, to appear).Google Scholar
  16. [16]
    R. Glowinski & M. F. Wheeler, Domain Decomposition and Mixed Finite Element Methods for Elliptic Problems, R. Glowinski, G. H. Golub, G. A. Meurant & J. Periaux (eds.),First International Symposium on Domain Decomposition Methods for Partial Differential Equations, SIAM, 1988, pp. 144–172.Google Scholar
  17. [17]
    G. H. Golub & D. Mayers, The Use of Preconditioning over Irregular Regions, R. Glowinski & J. L. Lions (eds.),Computing Methods in Applied Sciences and Engineering VI, North Holland, 1984, pp. 3–14.Google Scholar
  18. [18]
    W. D. Gropp & D. E. Keyes,Complexity of Parallel Implementation of Domain Decomposition Techniques for Elliptic Partial Differential Equations, SIAM J. Sci. Stat. Comp., 9 (1988), pp. 312 – 326.CrossRefGoogle Scholar
  19. [19]
    D. E. Keyes, Domain DecompositionMethods for the Parallel Computation of Reacting Flows, Comput. Phys. Comm. (1989, to appear).Google Scholar
  20. [20]
    D. E. Keyes & W. D. Gropp,A Comparison of Domain Decomposition Techniques for Elliptic Partial Differential Equations and their Parallel Implementation, SIAM J.Sci.Stat.Comp.,8 (1987),pp. sl66–s202.Google Scholar
  21. [21]
    D. E. Keyes & W. D. GroppDomain Decomposition Techniques for the Parallel Solution of Nonsymmetric Systems of Elliptic PDEs, Appl. Num. Meths. (1989, to appear).Google Scholar
  22. [22]
    T. A. Manteuffel, The Tchebychev Iteration for Nonsymmetric Linear Systems, Numer. Math., 28 (1977), pp. 307 – 327.CrossRefGoogle Scholar
  23. [23]
    J. A. Meierink & H. A. Van der Vorst,Guidelines for the Usage of Incomplete Decompositions in Solving Sets of Linear Equations as they Occur in Practical Problems, J. Comp. Phys., 44 (1981), pp. 134 – 155.CrossRefGoogle Scholar
  24. [24]
    J. S. Przemieniecki,Matrix Structural Analysis of Substructures, AIAA J., 1 (1963), pp. 138–147.CrossRefGoogle Scholar
  25. [25]
    Y. Saad, ChebyshevAcceleration Techniques for Solving Nonsymmetric Eigenvalue Problems, Math. Comp., 42 (1984), pp. 567 – 588.CrossRefGoogle Scholar
  26. [26]
    Y. Saad & M. H. Schultz,Parallel Implementations of Preconditioned Conjugate Gradient Methods, Technical Report YALEU/DCS/RR-425, Computer Science Dept., Yale University, October 1985.Google Scholar
  27. [27]
    Y. Saad & M. H. Schultz, GMRES:A Generalized Minimum Residual Algorithm for Solving Nonsymmetric Linear Systems, SIAM J. Sci. Stat. Comp., 7 (1986), pp. 856 – 869.CrossRefGoogle Scholar
  28. [28]
    H. Schlichting,Boundary Layer Theory, 7th ed., McGraw-Hill, New York, 1979.Google Scholar
  29. [29]
    H. F. Walker,Implementation of the GMRES Method Using Householder Transformations, SIAM J. Sci. Stat. Comp., 9 (1988), pp. 152–163.CrossRefGoogle Scholar
  30. [30]
    J. W. Watts, III,A Conjugate Gradient-Truncated Direct Method for the Iterative Solution of the Reservoir Simulation Pressure Equation, Soc. Petrol. Engin. J., 21 (1981), pp. 345 – 353.Google Scholar
  31. [31]
    O. B. Widlund, Iterative Substructuring Methods: Algorithms and Theory for Elliptic Problems in the Plane, R. Glowinski, G. H. Golub, G. A. Meurant & J. Periaux (eds.),First International Symposium on Domain Decomposition Methods for Partial Differential Equations, SIAM, 1988, pp. 113–128.Google Scholar
  32. [32]
    O. B. Widlund,Some Domain Decomposition and Iterative Refinement Algorithms for Elliptic Finite Element Problems, Technical Report 386, Courant Institute of Mathematical Sciences, NYU, June 1988.Google Scholar

Copyright information

© Plenum Press, New York 1989

Authors and Affiliations

  • >David E. Keyes
    • 1
  • William D. Gropp
    • 2
  • Ali Ecder
    • 1
  1. 1.Department of Mechanical Engineering, Research Center for Scientific ComputationYale UniversityNew HavenUSA
  2. 2.Department of Computer Science, Research Center for Scientific ComputationNew HavenUSA

Personalised recommendations