package core

  1. Overview
  2. Docs
Legend:
Library
Module
Module type
Parameter
Class
Class type

Memoization of OCaml functions of a single argument.

The default caching policy is to remember everything for the lifetime of the returned closure, but cache_size_bound allows one to specify an upper bound on cache size. Whenever a cache entry must be forgotten in order to obey this bound, we pick the least-recently-used one. The functions raise exceptions if cache_size_bound is negative or zero.

As you can tell from the type, the function that is memoized is the function of the first argument. To memoize a function with multiple arguments, pack them up in a tuple. See ../test/src/memo_argument.mlt for some examples.

This module does not detect or prevent infinite loops (e.g., due to a recursive call that repeats an argument).

The implementation is not thread-safe.

type ('a, 'b) fn = 'a -> 'b

A type definition to indicate that the expected use outputs a function

val general : ?hashable:'a Core__.Hashtbl.Hashable.t -> ?cache_size_bound:Base.Int.t -> ('a -> 'b) -> ('a, 'b) fn

Returns a memoized version of a function with a single argument.

Of course, if the supplied function is recursive, only the outer calls are memoized; recursive calls are not. For that, see recursive below.

val recursive : hashable:'a Core__.Hashtbl.Hashable.t -> ?cache_size_bound:Base.Int.t -> (('a -> 'b) -> 'a -> 'b) -> ('a, 'b) fn

recursive is like general but can be used to memoize recursive functions in such a way that the recursive calls are memoized as well.

As a concrete example, consider the following definition of the Fibonacci function.

let rec fib x = if x < 2 then x else fib (x - 1) + fib (x - 2) 

We can create a memoized version of this by first creating a non-recursive version of fib called fib_nonrecursive, where the recursive knot has been untied.

let fib_nonrecursive fib x = if x < 2 then x else fib (x - 1) + fib (x - 2) 

Here, rather than recursively calling itself, the function calls a function provided to it as an argument.

We can now use recursive to retie the recursive knot, injecting memoization at that point.

let fib = Memo.recursive ~hashable:Int.hashable fib_nonrecursive 

Note that calling recursive ~hashable f_nonrecursive does the partial application, f_nonrecursive f, at the time it is called, so that any side-effects or expensive computations that happen at the partial application stage happen just once, not once per evaluation.

recursive does not detect or prevent infinite loops, e.g., Memo.recursive ~hashable (fun f x -> f x) will just run until it overflows the stack.

Finally, note that recursive keeps memory around between invocations of the produced function, which may not be what you want if you're only trying to optimize recursive calls. You can achieve this with recursive by eta-expanding.

let fib x = Memo.recursive ~hashable:Int.hashable fib_nonrecursive x 

Note that the above would be a mistake when using general, since it would completely obviate the point of the call, but makes sense for recursive, since it would still optimize recursive calls within one outer invocation.

val unit : (Base.Unit.t -> 'a) -> (Base.Unit.t, 'a) fn

efficient special case for argument type unit

val of_comparable : (module Comparable.S_plain with type t = 'a) -> ('a -> 'b) -> ('a, 'b) fn

Use a comparable instead of hashable type

OCaml

Innovation. Community. Security.