https://github.com/noughtmare/breitnerbench
Benchmark of Joachim Breitner's Haskell Love talk
https://github.com/noughtmare/breitnerbench
Last synced: about 1 month ago
JSON representation
Benchmark of Joachim Breitner's Haskell Love talk
- Host: GitHub
- URL: https://github.com/noughtmare/breitnerbench
- Owner: noughtmare
- Created: 2020-08-05T20:23:22.000Z (almost 5 years ago)
- Default Branch: master
- Last Pushed: 2020-08-06T11:00:08.000Z (almost 5 years ago)
- Last Synced: 2025-03-23T05:25:43.939Z (4 months ago)
- Language: HTML
- Size: 783 KB
- Stars: 1
- Watchers: 2
- Forks: 0
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.md
Awesome Lists containing this project
README
# Benchmark of Joachim Breitner's Haskell Love Talk
I watched Joachim Breitner's Haskell Love talk today. He starts with a high level specification of a program that checks if a binary tree is ordered. During the talk he develops many different implementations. I thought that it would be interesting to see how fast all the different implementations really are.
Here is the link to the talk: https://vimeo.com/442720683#t=7067s
Breitner's talk starts around 1:57:47.
Currently, I only test the functions on balanced ordered trees with 10000
elements from 0 to 10000.
## Run these yourself
My machine may produce different results than yours.
Find out by cloning this repo and running it with cabal:
```
git clone https://github.com/noughtmare/breitnerbench.git
cd breitnerbench
cabal run -O2 breitnerbench -- -o report.html
xdg-open report.html
```
## On my machine
The fastest is `isOrdered12`, but `isOrdered5` can be specialized to `Int`, the tuple can be changed to two arguments and the max and min bounds can be used instead of using a maybe type.
The slowest is `isOrdered6`.
Here are all the results with GHC 8.10.1 and `-O2`:
Zoomed in:
