The benchmarked application was upgraded from Rails 3.2 to Rails 4.2 on March 30th 2015. That significantly impacted its performances.
The Application was running on a Vpsdime VM before the upgrade and it stayed on that same VM after the upgrade. Rails is the only thing that changed between the before and the after upgrade trial.
That's a 30% performance degradation.
From a memory usage perspective, each Rails process is now using about 120MB (rss), that's about 25MB more that the same application running on Rails 3.2.
I guess we should have stuck with 3.2! But what's done is done and there's no point in fighting "progress".
VPSBenchmarks strives to report benchmarks where all competitors are judged on the same basis. It would be unfair to compare 2 providers running on 2 different versions of Rails, therefore the chart on the home page will only show trials that were run on Rails 3.2 (all run before March 30th 2015) until enough providers are tested on the new Rails version.