Skip to content

Latest commit

 

History

History
80 lines (53 loc) · 2.32 KB

README.md

File metadata and controls

80 lines (53 loc) · 2.32 KB
Raw

ArbDecimals.jl

===========

reliable and performant extended precision floating point math

Copyright © 2016 by Jeffrey Sarnoff. Released under the MIT License on 15-Sep-2016.

===========

Overview

ArbDecimals exports a floating point type realized for the decimal digit resolutions named:
Dec25, Dec50, Dec75, Dec100, Dec150, Dec200, Dec250, Dec300, Dec350, Dec400, Dec450, Dec500.

Most of Julia's arithmetic and elementary functions work. These functions are exported too:
pow, root, sincos, sinhcosh, rgamma, polylog, agm.

All math is performed at twice the nominal precision. This is managed internally, and the user only need know that there is very strong numerical backstopping in support of delivering highly reliable results with desireable performance.

Results for most scalar and vector computations are quite reliable.
With matrices, det, chol, lu work well; eigenvals and svdvals need interval algorithms to work.

Installation

ArbDecimals depends on ArbFloats, which needs Nemo. It is best to add these packages one at a time:

Pkg.add("Nemo")   # expect much compilation
using Nemo
quit()

Pkg.add("ArbFloats")
using ArbFloats
quit()

Pkg.add("ArbDecimals")
using ArbDecimals
quit()

===========

Overview

ArbDecimals exports a floating point type realized for the decimal digit resolutions named:
Dec25, Dec50, Dec75, Dec100, Dec150, Dec200, Dec250, Dec300, Dec350, Dec400, Dec450, Dec500.

Most of Julia's arithmetic and elementary functions work. These functions are exported too:
pow, root, sincos, sinhcosh, rgamma, polylog, agm.

All math is performed at twice the nominal precision. This is managed internally, and the user only need know that there is very strong numerical backstopping in support of delivering highly reliable results with desireable performance.

Results for most scalar and vector computations are quite reliable.
With matrices, det, chol, lu work well; eigenvals and svdvals need interval algorithms to work.

Installation

ArbDecimals depends on ArbFloats, which needs Nemo. It is best to add these packages one at a time:

Pkg.add("Nemo")   # expect much compilation
using Nemo
quit()

Pkg.add("ArbFloats")
using ArbFloats
quit()

#Pkg.add("ArbDecimals")
Pkg.clone("https://github.com/JuliaArbTypes/ArbDecimals.jl")
using ArbDecimals
quit()