In this report we describe the implementations of the GMRES algorithm for both real and complex, single and double precision arithmetics suitable for serial, shared memory and distributed memory computers. For the sake of simplicity, exibility and eeciency the GMRES solvers have been implemented using the reverse communication mechanism for the matrix-vector product, the preconditioning and the...

This work is the follow-up of the experimental study presented in [3]. It is based on and extends some theoretical results in [15, 18]. In a backward error framework we study the convergence of GMRES when the matrixvector products are performed inaccurately. This inaccuracy is modeled by a perturbation of the original matrix. We prove the convergence of GMRES when the perturbation size is propo...

Recently, the complementary behavior of restarted GMRES has been studied. We observed that successive cycles of restarted block BGMRES (BGMRES(m,s)) can also complement one another harmoniously in reducing the iterative residual. In the present paper, this characterization of BGMRES(m,s) is exploited to form a hybrid block iterative scheme. In particular, a product hybrid block GMRES algorithm ...

We consider the behavior of the GMRES method for solving a linear system Ax = b when A is singular or nearly so, i.e., ill conditioned. The (near) singularity of A may or may not affect the performance of GMRES, depending on the nature of the system and the initial approximate solution. For singular A, we give conditions under which the GMRES iterates converge safely to a least-squares solution...

GMRES is a popular iterative method for the solution of linear system of equations with an unsymmetric square matrix. Range restricted GMRES (RRGMRES) is one GMRES version proposed by Calvetti et al in 2000. In this paper, a weighted implementation for RRGMRES is proposed. Numerical results prove this weighted RRGMRES is better than RRGMRES.

The GMRES and Arnoldi algorithms, which reduce to the CR and Lanczos algorithms in the symmetric case, both minimize p(A)b over polynomials p of degree n. The difference is that p is normalized at z 0 for GMRES and at z x for Arnoldi. Analogous "ideal GMRES" and "ideal Arnoldi" problems are obtained if one removes b from the discussion and minimizes p(/l)II instead. Investigation of these true ...

We describe a Krylov subspace technique, based on incomplete or-thogonalization of the Krylov vectors, which can be considered as a truncated version of GMRES. Unlike GMRES(m), the restarted version of GMRES, the new method does not require restarting. Our numerical experiments show that DQGMRES method often performs better than GMRES(m). In addition, the algorithm is exible to variable precond...

When solving a linear algebraic system Ax = b with GMRES, the relative residual norm at each step is bounded from above by the so-called ideal GMRES approximation. This worstcase bound is sharp (i.e. it is attainable by the relative GMRES residual norm) in case of a normal matrix A, but it need not characterize the worst-case GMRES behavior if A is nonnormal. Characterizing the tightness of thi...

This note describes the usage of the GMRES solver using reverse communication protocol. The GMRES control flow is outlined, and an example calling sequence explained.