Crate romu

source ·
Expand description

A pseudo random number generator using the Romu algorithm.

This pseudo random number generator (PRNG) is not intended for cryptographic purposes. This crate only implements the 64-bit “RomuTrio” generator, since it’s the recommended generator by the original author.

§Non-linear random number generator

Romu is a non-linear random number generator. That means that the period is probabilistic and is based on the seed. The bigger the needed period is, the higher the chance it is that the actual period is “too small”.

Following formula is given by the author:

ⓘ
    P(|cycle contains x<= 2^k|) = 2^k-s+7
        k is size of random numbers needed + 1.
        s is the state size.

Example chances for getting a “too small” period:

  • When 2^62 * 64 bit numbers are needed (32 EiB) -> 2^-122 chance
  • When 2^39 * 64 bit numbers are needed (4 TiB) -> 2^-146 chance
  • When 2^36 * 64 bit numbers are needed (512 GiB) -> 2^-149 chance

You can read more about the theory behind Romu in the official paper and it’s unique selling points on the official website of the original author.

§Seeding

When the user calls the new() or default() functions of a generator, the implementation tries to use the best available randomness source to seed the generator (in the following order):

  1. The crate getrandom to seed from a high quality randomness source of the operating system. The feature getrandom must be activated for this.
  2. Use the functionality of the standard library to create a low quality randomness seed (using the current time, the thread ID and a memory address). The feature std must be activated for this.
  3. Use a memory address as a very low randomness seed. If Address Space Layout Randomization (ASLR) is supported by the operating system, this should be a pretty “random” value.

It is highly recommended using the no_std compatible getrandom feature to get high quality randomness seeds.

The user can always create / update a generator with a user provided seed value.

If the tls feature is used, the user should call the seed() function to seed the TLS before creating the first random numbers, since the TLS instance is instantiated with a fixed value.

§SIMD

The crate currently provides three generators that tries to use auto vectorization to speed up the generation of large amounts of random numbers.

  • Rng128 - This should be used when the processor has access to 128-bit SIMD (SSE2 / NEON).
  • Rng256 - This should be used when the processor has access to 256-bit SIMD (AVX2).
  • Rng512 - This should be used when the processor has access to 512-bit SIMD (AVX512).

The nightly only feature unstable_simd uses the core::simd create to implement the SIMD. Users should test the available generators for their workload and verify if they can accelerate them using the SIMD functionality.

§Features

The crate is no_std compatible.

  • std - If getrandom is not used or returns an error, the generator will use the thread name and the current instance time to create a seed value. Enabled by default.
  • tls - Create static functions to use a thread local version of the generator. Enabled by default.
  • getrandom - Uses the getrandom crate to create a seed of high randomness. Enabled by default.
  • unstable_tls - Uses the unstable thread_local feature of Rust nightly. Improves the call times to the thread local functions greatly.
  • unstable_simd - Uses the unstable core::simd crate of Rust nightly to provide special SIMD versions of the generator which can be used to create large amount of random data fast.

Structs§

  • Rng
    Implements RomuTrio.
  • Rng128
    Implements RomuTrio with 128-bit width.
  • Rng256
    Implements RomuTrio with 256-bit width.
  • Rng512
    Implements RomuTrio with 512-bit width.

Enums§

  • SeedSource
    Defines which source the seed was created from.

Functions§

  • booltls
    Generates a random bool value.
  • f32tls
    Generates a random f32 value in range (0..1).
  • f64tls
    Generates a random f64 value in range (0..1).
  • fill_bytestls
    Fills a mutable [u8] slice with random values.
  • i8tls
    Generates a random i8 value.
  • i16tls
    Generates a random i16 value.
  • i32tls
    Generates a random i32 value.
  • i64tls
    Generates a random i64 value.
  • isizetls
    Generates a random isize value.
  • mixtls
    Mixes the state, which should improve the quality of the random numbers.
  • mod_u8tls
    Generates a random u8 value in range (0..n).
  • mod_u16tls
    Generates a random u16 value in range (0..n).
  • mod_u32tls
    Generates a random u32 value in range (0..n).
  • mod_u64tls
    Generates a random u64 value in range (0..n).
  • mod_usizetls
    Generates a random usize value in range (0..n).
  • range_i8tls
    Generates a random i8 value in the given range.
  • range_i16tls
    Generates a random i16 value in the given range.
  • range_i32tls
    Generates a random i32 value in the given range.
  • range_i64tls
    Generates a random i64 value in the given range.
  • range_isizetls
    Generates a random isize value in the given range.
  • range_u8tls
    Generates a random u8 value in the given range.
  • range_u16tls
    Generates a random u16 value in the given range.
  • range_u32tls
    Generates a random u32 value in the given range.
  • range_u64tls
    Generates a random u64 value in the given range.
  • range_usizetls
    Generates a random usize value in the given range.
  • seedtls
    Seeds the thread local instance from the best available randomness source.
  • seed_sourcetls
    Shows which source was used to acquire the seed for the thread local instance.
  • seed_with_64bittls
    Seeds the thread local instance with the given 64-bit seed.
  • seed_with_192bittls
    Seeds the thread local instance with the given 192-bit seed.
  • shuffletls
    Randomly shuffles a slice.
  • u8tls
    Generates a random u8 value.
  • u16tls
    Generates a random u16 value.
  • u32tls
    Generates a random u32 value.
  • u64tls
    Generates a random u64 value.
  • usizetls
    Generates a random usize value.