If your goal is to show that Elixir is an excellent language for writing distributed mathematical calculations, you’re going to fail at that goal. It is not. This isn’t because it’s bad at being distributed, but because it’s bad at doing high performance numerical calculations.

It can be GREAT at coordinating high performance numerical calculations, and the work and data necessary to make that happen. So for example at Cargosense we run a lot of machine learning stuff in which all the heavy lifting is done in C libraries wrapped by Python. But we use Elixir to coordinate those jobs, get the jobs their data, and manage their lifecycle.

Do keep in mind the things that Elixir is good at. If you’re trying to predict the weather, there are languages that will get you the most out of your CPU power better. If you’re trying to do something IO heavy, Elixir can be a good choice.

What kind of work are you trying to have distributed?

Tools are built for purposes. What purpose are you trying to use Elixir? It matters.

FYI: "Our goal is that applications run 0.75 x N times faster on an N-core computer. "

How we program multicores - Joe Armstrong

i.e. your expectations should cap out at

Once you replace core with computer the network overhead will start to take its toll.

And I want to measure that.