IDR(s): A Family of Simple and Fast Algorithms for Solving Large Nonsymmetric Systems of Linear Equations

  title={IDR(s): A Family of Simple and Fast Algorithms for Solving Large Nonsymmetric Systems of Linear Equations},
  author={Peter Sonneveld and Martin B. van Gijzen},
  journal={SIAM J. Sci. Comput.},
We present IDR($s$), a new family of efficient, short-recurrence methods for large nonsymmetric systems of linear equations. The new methods are based on the induced dimension reduction (IDR) method proposed by Sonneveld in 1980. IDR($s$) generates residuals that are forced to be in a sequence of nested subspaces. Although IDR($s$) behaves like an iterative method, in exact arithmetic it computes the true solution using at most $N + N/s$ matrix-vector products, with $N$ the problem size and $s… 

A block IDR(s) method for nonsymmetric linear systems with multiple right-hand sides

Algorithm 913: An elegant IDR(s) variant that efficiently exploits biorthogonality properties

This article derives a new IDR(s) variant, that imposes (one-sided) biorthogonalization conditions on the iteration vectors, and shows that the new variant is more stable and more accurate than the originals, and that it outperforms other state-of-the-art techniques for realistic test problems.

Exploiting BiCGstab(ℓ) Strategies to Induce Dimension Reduction

Through numerical experiments it is shown that IDRstab can outperform both IDR($s$) and BiCGstab($\ell$), and the relation between hybrid Bi-CG methods and IDR and the new concept of the Sonneveld subspace as a common framework is introduced.

IDR(s) for solving shifted nonsymmetric linear systems

Fast solution of nonsymmetric linear systems on Grid computers using parallel variants of IDR(s)

This paper reformulated the efficient and stable IDR(s) algorithm in such a way that it has a single global synchronisation point per iteration step and a methodology is presented for a–priori estimation of the optimal value of s using only problem and machine–based parameters.


IDR (Induced Dimension Reduction) is a family of efficient iterative methods for the numerical solution of large non-symmetric systems Ax = b of linear equations. Examples of IDR methods are

Minimizing synchronization in IDR (s)

This paper reformulated a recently proposed IDR (s) algorithm that is highly efficient and stable is reformulated in such a way that it has a single global synchronization point per iteration step.



Variants of BICGSTAB for Matrices with Complex Spectrum

The author presents for real nonsymmetric matrices a method BICGSTAB2 in which the second factor may have complex conjugate zeros, and versions suitable for complex matrices are given for both methods.

GPBi-CG: Generalized Product-type Methods Based on Bi-CG for Solving Nonsymmetric Linear Systems

This paper proposes a unified way to generalize a class of product-type methods whose residual polynomials can be factored by the residualPolynomial of Bi-CG and other polynmials with standard three-term recurrence relations.

Necessary and Sufficient Conditions for the Existence of a Conjugate Gradient Method

We characterize the class $CG(s)$ of matrices A for which the linear system $A{\bf x} = {\bf b}$ can be solved by an s-term conjugate gradient method. We show that, except for a few anomalies, the

Gmres: a Generalized Minimum Residual Algorithm for Solving

Overton, who showed us how the ideal Arnoldi and GMRES problems relate to more general problems of minimization of singular values of functions of matrices 17]. gmres and arnoldi as matrix

Solution of Systems of Linear Equations by Minimized Iterations1

In an earlier publication [14] a method was described which generated the eigenvalues and eigenvectors of a matrix by a successive algorithm based on minimizations by least squares. The advantage of

ML(k)BiCGSTAB: A BiCGSTAB Variant Based on Multiple Lanczos Starting Vectors

A variant of the popular BiCGSTAB method for solving nonsymmetric linear systems that produces a residual polynomial which is of lower degree after the same number of steps, but which also requires fewer matrix-vector products to generate, on average requiring only 1+1/k matvecs per step.

QMR: a quasi-minimal residual method for non-Hermitian linear systems

A novel BCG-like approach, the quasi-minimal residual (QMR) method, which overcomes the problems of BCG is presented and how BCG iterates can be recovered stably from the QMR process is shown.

Bi-CGSTAB: A Fast and Smoothly Converging Variant of Bi-CG for the Solution of Nonsymmetric Linear Systems

Numerical experiments indicate that the new variant of Bi-CG, named Bi- CGSTAB, is often much more efficient than CG-S, so that in some cases rounding errors can even result in severe cancellation effects in the solution.

Maintaining convergence properties of BiCGstab methods in finite precision arithmetic

This paper will propose a strategy for a more stable determination of the Bi-CG iteration coefficients and by experiments it will show that this indeed may lead to faster convergence.

CGS, A Fast Lanczos-Type Solver for Nonsymmetric Linear systems

A Lanczos-type method is presented for nonsymmetric sparse linear systems as arising from discretisations of elliptic partial differential equations. The method is based on a polynomial variant of