The latest versions of each distribution were benchmarked from their cloud platform and they included the 64-bit versions of Ubuntu 13.10, CentOS 6.5, Fedora 19, Debian 7.0, and Arch Linux 2013.05. During all of the instances being benchmarked the same droplet type of a quad-core CPU, 8GB of RAM, and 80GB of SSD storage was used.
Well, that was our plan at least for the latest DigitalOcean benchmarks... When running the five different droplets, we ended up with three different CPU configurations when creating all of the droplets on the same day and within a few hours of each other. While the same droplet type was selected, some of these virtual machines were running at 2.0GHz, others at 2.3GHz, and even a 2.4GHz droplet. All for the same cost, same droplet.
Getting slotted with measurably different hardware is also a problem we've run into with our Amazon EC2 benchmarking. Due to these different hardware configurations on the five different Linux cloud distributions, the ambitions from running a nice, multi-page, featured article about the DigitalOcean cloud Linux OS comparison was abandoned. I may repeat these results later and ensure the same CPU types across all droplets. I've also been meaning to run a DigitalOcean and EC2 Linux benchmarking article looking at performance fluctuations in the cloud.
Those interested in this five-way Linux OS comparison from DigitalOcean can find the performance benchmark results in full via 1404016-PL-DIGITALOC26. Using the Phoronix Test Suite you can also very easily run your own Linux cloud benchmarks.