Elixir vs Ruby Showdown - Phoenix vs Rails

The post compares Phoenix and Rails on the same web task: route matching, controller rendering, view rendering with partials, and returning the response. It says the setup isolates these pieces so throughput and latency can be compared on a more realistic framework workload.

The reported results favor Phoenix in both localhost and Heroku tests. On localhost, Phoenix reaches 10.63x more throughput and lower latency variation; on Heroku, it reaches 8.94x more throughput with 3.74x less CPU load, while Rails shows much higher latency and less consistent response times.

Reading notes

#

• The comparison uses Phoenix 0.3.1 with Cowboy and Erlang 17.1 against Rails 4.0.4 with Puma and MRI Ruby 2.1.0.
• The benchmark measures request routing, controller action rendering, layout and view rendering, partial rendering, and response delivery.
• No view caching was used, and request logging was disabled in both apps.
• The localhost test reports 12,120.00 req/s for Phoenix and 1,140.53 req/s for Rails.
• The Heroku test reports 2,691.03 req/s for Phoenix and 301.36 req/s for Rails.
• The author says benchmarks are only a good idea of performance and should be measured against your own system.
• The conclusion claims Elixir combines Ruby-like productivity with Erlang concurrency and fault tolerance, while Phoenix still needs a stronger ecosystem to match Rails.