| Identifiant de l'article : |
 |
ensl-00150406, version 1 |
|
|
| Identifiant arXiv : |
|
arXiv:0705.4369 |
|
|
| Domaine : |
|
|
|
|
| Titre : |
|
Computing Integer Powers in Floating-Point Arithmetic |
|
|
| Auteur(s) : |
|
Peter Kornerup1, Vincent Lefèvre2, Jean-Michel Muller2 |
|
|
| Laboratoire : |
|
| 1 : |
IMADA - Department of Mathematics and Computer Science |
|
| 2 : |
LIP - Laboratoire de l'Informatique du Parallélisme |
|
|
|
| Équipe de recherche : |
|
[ARENAIRE - Arithmétique des ordinateurs] |
| Résumé : |
|
We introduce two algorithms for accurately evaluating powers to a positive integer in floating-point arithmetic, assuming a fused multiply-add (fma) instruction is available. We show that our log-time algorithm always produce faithfully-rounded results, discuss the possibility of getting correctly rounded results, and show that results correctly rounded in double precision can be obtained if extended-precision is available with the possibility to round into double precision (with a single rounding). |
|
|
|
Langue du texte
intégral : |
|
Anglais |
|
|
| Mots-clés : |
|
Floating-point arithmetic – Integer powers – Computer arithmetic – Correct rounding |
|
|
| Commentaire : |
|
Laboratoire LIP : CNRS/ENS Lyon/INRIA/Université Lyon 1 |
|
|
| Classification : |
|
ACM B.2.4; G.1.0 |
|
|
| Référence interne : |
|
Rapport de recherche LIP n°2007-23 |
|
|
|
Récupérer au format
