2015 IEEE 22nd Symposium on Computer Arithmetic 2015
DOI: 10.1109/arith.2015.22
|View full text |Cite
|
Sign up to set email alerts
|

Code Generators for Mathematical Functions

Abstract: Abstract-A typical floating-point environment includes support for a small set of about 30 mathematical functions such as exponential, logarithms and trigonometric functions. These functions are provided by mathematical software libraries (libm), typically in IEEE754 single, double and quad precision.This article suggests to replace this libm paradigm by a more general approach: the on-demand generation of numerical function code, on arbitrary domains and with arbitrary accuracies.First, such code generation o… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
22
0
1

Year Published

2015
2015
2021
2021

Publication Types

Select...
3
3
2

Relationship

3
5

Authors

Journals

citations
Cited by 30 publications
(23 citation statements)
references
References 22 publications
0
22
0
1
Order By: Relevance
“…The implementation of a mathematical function, such as exp(x), sin(x) or atan(x), in the floating-point environment provided by IEEE754-2008 has been extensively studied and described in the literature [4], [5], [6], [7], [8], [9], [10]. Classically, a mathematical function on a floating-point argument is computed in four or five steps.…”
Section: Implementation Of Mathematical Functionsmentioning
confidence: 99%
See 1 more Smart Citation
“…The implementation of a mathematical function, such as exp(x), sin(x) or atan(x), in the floating-point environment provided by IEEE754-2008 has been extensively studied and described in the literature [4], [5], [6], [7], [8], [9], [10]. Classically, a mathematical function on a floating-point argument is computed in four or five steps.…”
Section: Implementation Of Mathematical Functionsmentioning
confidence: 99%
“…We have optimized all our polynomials coefficients using the techniques described in [3]. None of our polynomials uses coefficients with higher precision than the working precision, whilst other library tend to use floating-point expansions at least for the coefficients of lower degree [8], [10].…”
Section: A a Bird's Eyes View On The Proposed Algorithmsmentioning
confidence: 99%
“…The "middle" and "large" errors are obtained by increasing the bound by one and two orders of magnitude respectively. input: t ∈ [1,3], w ∈ [-5, 5]…”
Section: Appendixmentioning
confidence: 99%
“…In this article, we deal with the vectorized implementation of the logarithm function, log(x), in floating-point arithmetic, with a particular focus on its automation through the MetaLibm framework [1], [2]. It enables to describe the implementation of a function using a meta-language and to generate C codes optimized for different micro-architectures.…”
Section: Meta-implementation Of Vectorized Logarithm Function In Binamentioning
confidence: 99%