Parallel iterative linear solvers for multistep Runge-Kutta methods

This paper deals with solving stiff systems of differential equations by implicit Multistep Runge-Kutta (MRK) methods. For this type of methods, nonlinear systems of dimension sd arise, where s is the number of Runge-Kutta stages and d the dimension of the problem. Applying a Newton process leads to linear systems of the same dimension, which can be very expensive to solve in practice. With a parallel iterative linear system solver, especially designed for MRK methods, we approximate these linear systems by s systems of dimension d, which can be solved in parallel on a computer with s processors. In terms of Jacobian evaluations and LU-decompositions, the k-steps-stage MRK applied with this technique is on s processors equally expensive as the widely used k-step Backward Differentiation Formula on 1 processor, whereas the stability properties are better than that of BDF. A simple implementation of both methods shows that, for the same number of Newton iterations, the accuracy delivered by the new method is higher than that of BDF.     

Existence-uniqueness and iterative methods for third-order boundary value problems

In this paper we shall provide necessary and sufficient conditions for the existence and uniqueness of solutions of third-order nonlinear differential equations satisfying three-point boundary conditions. For the linear case, we propose a constructive method which is a variation of the method of chasing. For the nonlinear problems sufficient conditions are provided to ensure the convergence of a general class of iterative methods. Several examples are also included.     

Parallel linear system solvers for Runge-Kutta methods

If the nonlinear systems arising in implicit Runge-Kutta methods like the Radau IIA methods are iterated by (modified) Newton, then we have to solve linear systems whose matrix of coefficients is of the form I-A hJ with A the Runge-Kutta matrix and J an approximation to the Jacobian of the righthand side function of the system of differential equations. For larger systems of differential equations, the solution of these linear systems by a direct linear solver is very costly, mainly because of the LU-decompositions. We try to reduce these costs by solving the linear systems by a second (inner) iteration process. This inner iteration process is such that each inner iteration again requires the solution of a linear system. However, the matrix of coefficients in these new linear systems is of the form I - B hJ where B is similar to a diagonal matrix with positive diagonal entries. Hence, after performing a similarity transformation, the linear systems are decoupled into s subsystems, so that the costs of the LU-decomposition are reduced to the costs of s LU-decompositions of dimension d. Since these LU-decompositions can be computed in parallel, the effective LU-costs on a parallel computer system are reduced by a factor s ³ . It will be shown that matrices B can be constructed such that the inner iterations converge whenever A and J have their eigenvalues in the positive and nonpositive halfplane, respectively. The theoretical results will be illustrated by a few numerical examples. A parallel implementation on the four-processor Cray-C98/4256 shows a speed-up ranging from at least 2.4 until at least 3.1 with respect to RADAU5 applied in one-processor mode.     

Parallel implementation of block boundary value methods for ODEs

The parallel solution of initial value problems for ODEs has been the subject of much research in the last thirty years, and different approaches to the problem have been devised. In this paper we examine the parallel methods derived by block boundary value methods (BVMs), recently introduced for approximating Hamiltonian problems. Here we restrict the analysis of the methods when applied to linear problems, since their nonlinear parallel implementation deserves further study. However, for linear problems, the methods can reach a high parallel efficiency.Some of these solvers can also be adapted for approximating continuous two-point boundary value problems. Numerical tests carried out on a distributed memory parallel computer are reported.     

Block iterative solvers for higher order finite volume methods

Recently, new higher order finite volume methods (FVM) were introduced in [Z. Cai, J. Douglas, M. Park, Development and analysis of higher order finite volume methods over rectangles for elliptic equations, Adv. Comput. Math. 19 (2003) 3-33], where the linear system derived by the hybridization with Lagrange multiplier satisfying the flux consistency condition is reduced to a linear system for a pressure variable by an appropriate quadrature rule. We study the convergence of an iterative solver for this linear system. The conjugate gradient (CG) method is a natural choice to solve the system, but it seems slow, possibly due to the non-diagonal dominance of the system. In this paper, we propose block iterative methods with a reordering scheme to solve the linear system derived by the higher order FVM and prove their convergence. With a proper ordering, each block subproblem can be solved by fast methods such as the multigrid (MG) method. The numerical experiments show that these block iterative methods are much faster than CG.     

A splitting preconditioner for the iterative solution of implicit Runge-Kutta and boundary value methods

We study preconditioned iterative methods for the linear systems arising in the numerical integration of ODEs and time-dependent PDEs by implicit Runge-Kutta and boundary value methods. A preconditioning strategy based on a Kronecker product splitting of the coefficient matrix is proposed, and some useful properties of the preconditioned matrix are established. Numerical examples are presented to illustrate the effectiveness of this approach.     

Direct and iterative solvers for finite-element problems

Parallelization promises to allow increasing performance demand from users of scientific computation codes. In this article we discuss implementation of parallelization techniques in industrial computation codes for applications in Fluid Mechanics (N3S) and Structural Mechanics (Code_ASTER). We outline the gains afforded by parallelization on a Cray C98 machine in production mode, using a single version of the program running in sequential or in parallel mode, with minimum modifications to the initial code. Parallelization is applied at the discretion of the user, for whom no specialist knowledge of this technique is assumed. This revised version was published online in August 2006 with corrections to the Cover Date.     

Monotone iterative scheme for nonlinear hyperbolic boundary value problem

Monotone iterative scheme for nonlinear hyperbolic boundary value problem is established.     

Parallel implementation of hybrid iterative methods for nonsymmetric linear systems

In this paper, we study efficient parallel implementation for hybrid iterative methods BiCGSTAB and BiCGSTAB (?) with ? = 2 on the CRAY C90, and the efficiency of their parallel performance is evaluated. Numerical experiments suggest that on the CRAY C90 a parallel inner product algorithm called PDOTB be used for the parallelization of hybrid iterative methods containing sensitive values of inner products. Lastly, it is shown that the number of iterations in which parallel hybrid iterative methods satisfy a certain convergence criterion depends on the number of processors to be used.     

Parallel multisplitting iterative methods for singular M‐matrices

In this paper, we study parallel (two‐stage) multisplitting methods for singular M‐matrices. Some theoretical analysis in consistency and convergence of methods are presented, which also affirms that the Bru, Canto and Climents' conjecture holds. Copyright © 2001 John Wiley & Sons, Ltd.     

A comparison of Rosenbrock and ESDIRK methods combined with iterative solvers for unsteady compressible flows

In this article, we endeavour to find a fast solver for finite volume discretizations for compressible unsteady viscous flows. Thereby, we concentrate on comparing the efficiency of important classes of time integration schemes, namely time adaptive Rosenbrock, singly diagonally implicit (SDIRK) and explicit first stage singly diagonally implicit Runge-Kutta (ESDIRK) methods. To make the comparison fair, efficient equation system solvers need to be chosen and a smart choice of tolerances is needed. This is determined from the tolerance TOL that steers time adaptivity. For implicit Runge-Kutta methods, the solver is given by preconditioned inexact Jacobian-free Newton-Krylov (JFNK) and for Rosenbrock, it is preconditioned Jacobian-free GMRES. To specify the tolerances in there, we suggest a simple strategy of using TOL/100 that is a good compromise between stability and computational effort. Numerical experiments for different test cases show that the fourth order Rosenbrock method RODASP and the fourth order ESDIRK method ESDIRK4 are best for fine tolerances, with RODASP being the most robust scheme.     

Galerkin-wavelet methods for two-point boundary value problems

Summary Anti-derivatives of wavelets are used for the numerical solution of differential equations. Optimal error estimates are obtained in the applications to two-point boundary value problems of second order. The orthogonal property of the wavelets is used to construct efficient iterative methods for the solution of the resultant linear algebraic systems. Numerical examples are given.This work was supported by National Science Foundation     

Newton-like methods for two-point boundary value problems

A class of Newton-like methods for discrete two-point boundary value problems is constructed from the sum equation formulation of the problem. Each step of the Newton-like method can be described as first solving a system of linear algebraic equations. The solution vector of this system gives boundary values to a number of discrete boundary value problems which can be solved explicitly.     

Monotone iterative sequences for nonlinear boundary value problems of fractional order

In this paper we extend the maximum principle and the method of upper and lower solutions to boundary value problems with the Caputo fractional derivative. We establish positivity and uniqueness results for the problem. We then introduce two well-defined monotone sequences of upper and lower solutions which converge uniformly to the actual solution of the problem. A numerical iterative scheme is introduced to obtain an accurate approximate solution for the problem. The accuracy and efficiency of the new approach are tested through two examples.     

Monotone iterative technique for periodic boundary value problems with causal operators

In this paper, we develop the monotone iterative technique for periodic boundary value problems with causal operators.