Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

A look-ahead variant of TFQMR

Conference ·
OSTI ID:219556
 [1];  [2]
  1. AT&T Bell Labs., Murray Hill, NJ (United States)
  2. Oak Ridge National Lab., TN (United States)

Recently, Freund proposed a Krylov subspace iteration, the transpose-free quasi-minimal residual method (TFQMR), for solving general nonsingular non-Hermitian linear systems. The algorithm relies on a version of the squared Lanczos process to generate the basis vectors for the underlying Krylov subspace. It then constructs iterates defined by a quasi-minimization property, which leads to a smooth and nearly monotone convergence behavior. The authors investigate a variant of TFQMR that uses look-ahead to avoid some of the problems associated with breakdowns in the underlying squared Lanczos procedure. They also present some numerical examples that illustrate the properties of the new method, as compared to the original TFQMR algorithm.

Research Organization:
Front Range Scientific Computations, Inc., Boulder, CO (United States); USDOE, Washington, DC (United States); National Science Foundation, Washington, DC (United States)
OSTI ID:
219556
Report Number(s):
CONF-9404305--Vol.2; ON: DE96005736
Country of Publication:
United States
Language:
English