On the robustness of the 2Sum and Fast2Sum algorithms - Archive ouverte HAL Access content directly
Journal Articles ACM Transactions on Mathematical Software Year : 2017

On the robustness of the 2Sum and Fast2Sum algorithms

(1) , (2) , (3, 4)
1
2
3
4

Abstract

The 2Sum and Fast2Sum algorithms are important building blocks in numerical computing. They are used (implicitely or explicitely) in many compensated algorithms (such as compensated summation or compensated polynomial evaluation). They are also used for manipulating floating-point expansions. We show that these algorithms are much more robust than it is usually believed: the returned result makes sense even when the rounding function is not round-to-nearest, and they are almost immune to overflow.
Fichier principal
Vignette du fichier
FaithfulTwoSum-Final-Fev2017.pdf (310.83 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

ensl-01310023 , version 1 (01-05-2016)
ensl-01310023 , version 2 (22-02-2017)

Identifiers

  • HAL Id : ensl-01310023 , version 2

Cite

Sylvie Boldo, Stef Graillat, Jean-Michel Muller. On the robustness of the 2Sum and Fast2Sum algorithms. ACM Transactions on Mathematical Software, 2017, 44 (1). ⟨ensl-01310023v2⟩
849 View
1375 Download

Share

Gmail Facebook Twitter LinkedIn More