This project contains the benchmarks used to ensure Feedjira stays fast.
Its important for benchmarks to be repeatable and verifiyable by another party. Getting setup to run the benchmarks requires that you first clone down this repo and then also clone down the feedjira one:
$ git clone git://github.com/feedjira/feedjira-benchmarks
$ cd feedjira-benchmarks
$ git clone git://github.com/feedjira/feedjira
This will give you a directory structure like this:
feedjira-benchmarks
|-- feed_xml
|-- feedjira
|-- fetch
|-- others
+-- parse
Next bundle
to grab the project's requirements.
Lastly, producing the charts does rely on having R installed. If you're on a Mac, this should work:
$ brew tap homebrew/science
$ brew install r
This benchmark is run against each version of the gem - run it like this:
$ parse/run
This command will output three files:
- parse/results.txt - the raw numbers from the benchmark
- parse/results.csv - the totals from the raw numbers in CSV
- parse/parse-benchmark.png - that CSV converted to chart form
This benchmark compares the speeds of various fetching options - run it like this:
$ fetch/run
This command will output three files:
- fetch/results.txt - the raw numbers from the benchmark
- fetch/results.csv - the totals from the raw numbers in CSV
- fetch/fetch-benchmark.png - that CSV converted to chart form
This benchmark compares Feedjira with other Ruby feed libraries in terms of parse speed:
$ others/run
This command will output three files:
- others/results.txt - the raw numbers from the benchmark
- others/results.csv - the totals from the raw numbers in CSV
- others/others-benchmark.png - that CSV converted to chart form