• Stars
    star
    32
  • Rank 773,496 (Top 16 %)
  • Language
    Elixir
  • License
    MIT License
  • Created about 9 years ago
  • Updated about 7 years ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

DefMemo - Ryuk's little puppy! Bring apples.

DefMemo

A memoization macro (defmemo) for Elixir.

Build Status

Adapted from : (Gustavo Brunoro) https://gist.github.com/brunoro/6159378

I found Gustavo's Gist when looking at memoization and elixir and fixed it to work with version 1.0.x. Since then I've fixed a few of the problems with the original implementation:

  • will correctly memoize the results of functions with identical signatures but in different modules.

  • will work with 'when' guard clauses in function definitions. (That was fun!)

  • Added lots of lovely tests.

Usage

Add defmemo to your mix.exs file:

{:defmemo, "~> 0.1.0"}

And run:

mix deps.get

Before using a defmemo'd function (it's fine to define them), start_link must be called. e.g.

DefMemo.start_link

or you can add :defmemo into the applications section of your mix.exs:

[applications: [:logger, :defmemo]]

Example

defmodule FibMemo do
  import DefMemo

  defmemo fibs(0), do: 0
  defmemo fibs(1), do: 1
  defmemo fibs(n), do: fibs(n - 1) + fibs(n - 2)

  def fib_10 do
    fibs(10)
  end
end

Performance

As you would expect for something like fibs, memoization provides dramatic performance improvements:

UNMEMOIZED VS MEMOIZED
***********************
fib (unmemoized)
function -> {result, running time(μs)}
==================================
fibs(30) -> {832040, 31089}
fibs(30) -> {832040, 31833}

FibMemo (memoized)
==================================
fibs(30) -> {832040, 79}
fibs(30) -> {832040, 3}
fibs(50) -> {12586269025, 103}
fibs(50) -> {12586269025, 3}

Note that these have also improved from version 0.1 to 0.1.1. The above numbers are on the low end of the spectrum with access ranging from 2 to 15 μs for me.

TODO

  • Add test for supervisor crashing.

  • Look at injecting the type of result table used.

  • Better documentation.

  • More tests (alwaaaays with the testing!)

  • Test with some biger data (e.g. for something like web crawling)

  • ~~Supervisor ~~

  • Redis Based ResultTable - I've been playing with this - obviously there are limitations on type and it's slower than gen server but there are of course circumstances where it could be useful but for the most part its not a good "fit".