Skip to main content

Remark on Algorithm 539: A Modern Fortran Reference Implementation for Carefully Computing the Euclidean Norm

Hanson, Richard J., Hopkins, Tim (2018) Remark on Algorithm 539: A Modern Fortran Reference Implementation for Carefully Computing the Euclidean Norm. ACM Transactions on Mathematical Software, 44 (3). pp. 1-23. ISSN 0098-3500. E-ISSN 1557-7295. (doi:10.1145/3134441) (KAR id:67205)

PDF Pre-print
Language: English
Download (283kB) Preview
Official URL
https://doi.org/10.1145/3134441

Abstract

We propose a set of new Fortran reference implementations, based on an algorithm proposed by Kahan,

The principal advantage of these routines over the current offerings is that, rather than losing accuracy

for vectors of length N < Nmax where Nmax depends upon the precision of the floating point arithmetic

detect occurrences of non-finite numbers in the input data and return suitable values as well as setting

portable access to the new routines.

floating point exception flags signal, do we fall back on Kahan’s algorithm. Since most input vectors are

while the use of compensated summation ensures high accuracy.

implementation and existing codes for both accuracy and the appropriate settings of the IEEE arithmetic

status flags.

Item Type: Article
DOI/Identification number: 10.1145/3134441
Uncontrolled keywords: Design, Algorithms, Performance, Portability, Reliability
Divisions: Faculties > Sciences > School of Computing
Depositing User: Tim Hopkins
Date Deposited: 05 Jun 2018 14:28 UTC
Last Modified: 29 May 2019 20:36 UTC
Resource URI: https://kar.kent.ac.uk/id/eprint/67205 (The current URI for this page, for reference purposes)
  • Depositors only (login required):