Elixir synthetic performance test

Almost five years have passed, I can’t find the Ruby-Vert.x-based repo to repeat the test.

But the main point in our case was not about JVM vs. ErlangVM. It was about the easy transition of Ruby developers to some faster tech stack. Since with Vert.x, you just write Ruby, and Elixir has ruby-ish syntax.

We found out that in both cases, Ruby developers have to adjust to a new mindset. The syntax is secondary.

If one wants a raw performance, they should go entirely with C++ or Rust (according to Round 20 of TechEmpower benchmark, which is latest for the moment). Or drop some NIFs into critical parts of their app.

We have Ruby, Elixir, and Java (Spring, I believe) projects in our company. Clients come with some preset, and we never had to convince them to change the stack because of inability to scale.

If we start a new project, the choice is based on the available and planned resources.

I still feel that Elixir allows me to design better solutions in terms of architecture.

2 Likes