Worst Cases and Lattice Reduction (original) (raw)
Rapport (Rapport De Recherche) Année : 2002
Résumé
We propose a new algorithm to find worst cases for correct rounding of an analytic function. We first reduce this problem to the real small value problem --- i.e. for polynomials with real coefficients. Then we show that this second problem can be solved efficiently, by extending Coppersmith's work on the integer small value problem --- for polynomials with integer coefficients --- using lattice reduction [4,5,6]. For floating-point numbers with a mantissa less than N, and a polynomial approximation of degree d, our algorithm finds all worst cases at distance < N [power](-d2/2d-1) from a machine number in time [OMICRON](N[power]((d-1)/(2d-1)+e)). For d=2, this improves on the [OMICRON](N- [power]((2/3)+e)) complexity from Lefèvre's algorithm [12,13] to [OMICRON](N[p- ower]((3/5)+e)). We exhibit some new worst cases found using our algorithm, for double-extended and quadruple precision. For larger d, our algorithm can be used to check that there exist no worst cases at distance < N[power]-k in time [OMICRON](N[power](1/2+[OMICRON](1/k))).
Mots clés
Connectez-vous pour contacter le contributeur
https://inria.hal.science/inria-00071999
Soumis le : mardi 23 mai 2006-19:30:10
Dernière modification le : mardi 4 novembre 2025-12:02:29
Archivage à long terme le : dimanche 4 avril 2010-22:48:19
Dates et versions
inria-00071999 , version 1 (23-05-2006)
Licence
Identifiants
- HAL Id : inria-00071999 , version 1
Citer
Damien Stehlé, Vincent Lefèvre, Paul Zimmermann. Worst Cases and Lattice Reduction. [Research Report] RR-4586, INRIA. 2002. ⟨inria-00071999⟩
564 Consultations
589 Téléchargements