Preconditioned Conjugate Gradient Schemes
نویسنده
چکیده
The conjugate gradient method is a powerful algorithm to solve well-structured sparse linear systems that arise from partial diierential equations. We consider here three diierent conjugate gradient schemes for solving elliptic partial diierential equations that arise from 5-point diierence schemes: the classical CG, CG with a block diagonal-block incomplete Cholesky preconditioner and the reduced system approach. In this paper, we describe brieey the Cedar architecture, the algorithms and their implementation on Cedar. Numerical results on diierent Cedar conngurations are presented and examined.
منابع مشابه
Numerical Experiments with Algebraic Multilevel Preconditioners
This paper numerically compares different algebraic multilevel preconditioners to solve symmetric positive definite linear systems with the preconditioned conjugate gradient algorithm on a set of examples arising mainly from discretization of second order partial differential equations. We compare several different smoothers, influence matrices and interpolation schemes.
متن کاملA Class of Nested Iteration Schemes for Generalized Coupled Sylvester Matrix Equation
Global Krylov subspace methods are the most efficient and robust methods to solve generalized coupled Sylvester matrix equation. In this paper, we propose the nested splitting conjugate gradient process for solving this equation. This method has inner and outer iterations, which employs the generalized conjugate gradient method as an inner iteration to approximate each outer iterate, while each...
متن کاملSolving large systems arising from fractional models by preconditioned methods
This study develops and analyzes preconditioned Krylov subspace methods to solve linear systems arising from discretization of the time-independent space-fractional models. First, we apply shifted Grunwald formulas to obtain a stable finite difference approximation to fractional advection-diffusion equations. Then, we employee two preconditioned iterative methods, namely, the preconditioned gen...
متن کاملPressure Correction Scheme 1
Pressure correction schemes and the method of characteristics are combined to obtain numerical approximation procedures for incompressible Navier-Stokes ows on massive parallel computers. The projection step is carried out with a new more eecient parallel preconditioned conjugate gradient method and for the computation of the characteristics we use a fast local Crank-Nicholson solver for linear...
متن کاملSolving Initial – Value Problems by Explicit Domain Decomposition Approximate Inverses
A class of hybrid heterogeneous methods using time implicit backward differences and Crank-Nicolson approximating schemes in conjunction with explicit domain decomposition approximate inverse matrix techniques is introduced for computing various families of approximate inverses based on approximate LU-type factorization techniques. Explicit preconditioned conjugate gradient type schemes in conj...
متن کامل