Ever wondered what it would take to have stable and reliable measurements for your performance? Or wished to know the evolution of it across releases and rubies? Or simply to know how fast or slow is some code?
Perfer is a benchmarking tool I developed as a GSoC project intending to solve these, and to provide an easy way to store, transform and compare the results and to make a graph of it.
The talk will start with some benchmarking best practices, explaining why using the benchmark stdlib is usually not sufficient.
I will have a look at alternatives and finally I will explain Perfer's approach and show how to use it and how it can be useful in real-world examples.