https://github.com/mikolalysenko/robust-sum

Exactly computes the sum of a pair of non-overlapping sequences of floats.
https://github.com/mikolalysenko/robust-sum

Last synced: about 2 months ago
JSON representation

Exactly computes the sum of a pair of non-overlapping sequences of floats.

• Host: GitHub
• URL: https://github.com/mikolalysenko/robust-sum
• Owner: mikolalysenko
• License: mit
• Created: 2013-09-18T17:09:49.000Z (over 11 years ago)
• Default Branch: master
• Last Pushed: 2014-05-28T17:18:10.000Z (over 10 years ago)
• Last Synced: 2024-10-20T14:28:21.766Z (2 months ago)
• Language: JavaScript
• Size: 176 KB
• Stars: 4
• Watchers: 3
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
robust-sum

==========

Computes the sum of two increasing non-overlapping sequences of floats as an increasing non-overlapping sequence.  This can be used to perform exact arithmetic calculations on floating point values.

For more information, see:

* Jonathan Richard Shewchuk, ["Adaptive precision floating-point arithmetic and fast robust geometric predicates"](http://www.cs.cmu.edu/afs/cs/project/quake/public/papers/robust-arithmetic.ps). 1997

[![testling badge](https://ci.testling.com/mikolalysenko/robust-sum.png)](https://ci.testling.com/mikolalysenko/robust-sum)

[![build status](https://secure.travis-ci.org/mikolalysenko/robust-sum.png)](http://travis-ci.org/mikolalysenko/robust-sum)

## Install

		npm install robust-sum

## Example

```javascript

var robustSum = require("robust-sum")

var seq = robustSum([1, 64], [1e-64, 1e64])

console.log("result = ", seq)

```

## API

### `require("robust-sum")(a, b)`

Computes the sum of two non-overlapping increasing sequences of floats exactly as a non-overlapping increasing sequence of floats.

* `a` is a non-overlapping sequence of floats that is increasing in magnitude

* `b` is a non-overlapping sequence of floats that is increasing magnitude

**Returns** A non-overlapping increasing sequence that encodes the result of `a+b`

## Credits

Based on JRS' robust geometric predicates for floating point arithmetic.

Implementation (c) 2013 Mikola Lysenko. MIT License