Algorithmic fault tolerance using the Lanczos method

124. D. L. Boley, R. P. Brent, G. H. Golub and F. T. Luk, Algorithmic fault tolerance using the Lanczos method, SIAM J. Matrix Anal. Appl. 13 (1992), 312-332 (Gene Golub 60th birthday special issue). MR 93f:65034.

Abstract: dvi (3K), pdf (66K), ps (27K).

Paper: dvi (39K), pdf (191K), ps (93K).


We consider the problem of algorithm-based fault tolerance, and make two contributions. First, we show how very general sequences of polynomials can be used to generate the checksums, so as to reduce the chance of numerical overflows. Second, we show how the Lanczos process can be applied in the error location and correction steps, so as to save on the amount of work and to facilitate actual hardware implementation.

Go to next publication

Return to Richard Brent's index page