finds.dev

devonestes/fast-elixir

Writing Fast Elixir -- Collect Common Elixir idioms.

View on GitHubHomepage ↗

★ 1,289Elixirupdated Nov 2023

// the take

fast-elixir is a benchmark-backed catalog of Elixir idioms where more than one semantically equivalent way to write something exists, showing which is actually faster using benchee. It's for intermediate-to-advanced Elixir developers who want real numbers instead of folklore about things like ets vs GenServer or list concatenation vs cons.

Every idiom ships with actual benchee output (ips, deviation, memory usage) rather than just an assertion, and the runnable .exs source for each example is linked so you can reproduce or extend it. The selection covers genuinely non-obvious BEAM tradeoffs — ets/persistent_term/GenServer state access, IO lists vs string interpolation, `++` vs cons+flatten — that documentation rarely quantifies. It's inspired by and structured like fast-ruby, so the format is proven and easy to skim.

Last push was November 2023 and the benchmarks themselves are pinned to Elixir 1.11-rc / OTP 23, so several comparisons may no longer hold on current BEAM versions with newer JIT and binary optimizations. There's no CI that re-runs the benchmarks, so a reader has no way to know if a given result is still accurate versus quietly stale. It's a static reference, not a library, so there's nothing to actually adopt — the real risk is treating a three-year-old micro-benchmark as current truth for a production decision.

We read the README, the directory tree and a key source file in October 2026. Nobody pays to be listed. How we judge →

search the same ground:

// at a glance

Stars
1,289
Forks
43
Language
Elixir
Last commit
8 Nov 2023

// nearby

If fast-elixir isn't quite it.

All repos →