Struct SeededRng 

pub struct SeededRng { /* private fields */ }

A seeded xoshiro256++ generator with a standard normal sampler.

It serializes as its 256-bit state, written as four hexadecimal strings ("0x…", because JSON readers such as JavaScript’s lose integers above 2⁵³), and the spare normal deviate, so a run can be checkpointed and resumed bit for bit. The all-zero state is rejected: xoshiro would stay at zero forever.

Implementations

impl SeededRng

pub fn seed_from_u64(seed: u64) -> Self

Seeds the generator: the four state words are the first four SplitMix64 outputs starting from seed. SplitMix64 never gives four zero words in a row, so every seed is valid.

pub fn for_stream(seed: u64, keys: &[u64]) -> Self

A generator for one stream of a seeded run, keyed by keys: a Monte Carlo sample’s index and the variable it draws, say. The stream depends only on seed and keys, so a sample draws the same numbers whatever else the run draws, however many samples it has and however they are shared among threads: the counter-based idea of J. K. Salmon, M. A. Moraes, R. O. Dror and D. E. Shaw, “Parallel random numbers: as easy as 1, 2, 3”, Proc. SC11 (2011), https://doi.org/10.1145/2063384.2063405.

The key folds into one word with SplitMix64’s output function mix, a bijection on 64-bit words: h₀ = seed, hᵢ = mix(hᵢ₋₁ ⊕ mix(kᵢ + γ)) with γ SplitMix64’s increment, and the generator is SeededRng::seed_from_u64(hₙ). Two different keys give the same hₙ only by a collision of 64-bit hashes, about one chance in 2⁶⁴ for a pair of streams; with no keys the stream is seed_from_u64(seed)’s.

pub fn next_u64(&mut self) -> u64

The next 64 random bits (the xoshiro256++ output function and state transition).

pub fn uniform(&mut self) -> f64

A uniform deviate on [0, 1): the top 53 bits of SeededRng::next_u64 times 2⁻⁵³, so every value is a multiple of 2⁻⁵³.

pub fn standard_normal(&mut self) -> f64

A standard normal deviate (mean 0, variance 1) by the Marsaglia–Bray polar method.

Each accepted pair gives two deviates; the second is returned by the next call. On average a pair costs 4/π ≈ 1.27 attempts, two uniforms each.

Trait Implementations

impl Clone for SeededRng

fn clone(&self) -> SeededRng

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for SeededRng

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for SeededRng

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for SeededRng

fn eq(&self, other: &SeededRng) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for SeededRng

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for SeededRng