Skip to main content

core/internal/PRNG

Collection of pseudo-random number generators

The algorithms deliver deterministic statistical randomness, not cryptographic randomness.

Algorithm 1: 128-bit Seiran PRNG See: https://github.com/andanteyk/prng-seiran

Algorithm 2: SFC64 and SFC32 (Chris Doty-Humphrey’s Small Fast Chaotic PRNG) See: https://numpy.org/doc/stable/reference/random/bit_generators/sfc64.html

Copyright: 2023 MR Research AG Main author: react0r-com Contributors: Timo Hanke (timohanke)

Class SFC64​

class SFC64(p : Nat64, q : Nat64, r : Nat64)

Constructs an SFC 64-bit generator. The recommended constructor arguments are: 24, 11, 3.

Example:

import PRNG "mo:core/internal/PRNG";

let rng = PRNG.SFC64(24, 11, 3);

For convenience, the function SFC64a() returns a generator constructed with the recommended parameter set (24, 11, 3).

Function init​

func init(seed : Nat64)

Initializes the PRNG state with a particular seed

Example:



### Function `initPre`
``` motoko no-repl
func initPre()

Initializes the PRNG state with a hardcoded seed. No argument is required.

Example:

Function init3​

func init3(seed1 : Nat64, seed2 : Nat64, seed3 : Nat64)

Initializes the PRNG state with three state variables

Example:

Function next​

func next() : Nat64

Returns one output and advances the PRNG's state

Example:

Function sfc64a​

func sfc64a() : SFC64

SFC64a is the same as numpy. See: sfc64_next()