6 releases (breaking)
new 0.5.0 | Nov 24, 2024 |
---|---|
0.4.0 | Nov 22, 2024 |
0.3.0 | Nov 20, 2024 |
0.2.0 | Nov 19, 2024 |
0.1.1 | Nov 16, 2024 |
#321 in Algorithms
615 downloads per month
105KB
3K
SLoC
Cheap and dirty RNGs
use voxell_rng::prelude::*;
use voxell_rng::time_seeded::TimeSeededXorShift32;
// seeds using os entropy
let mut rng = TimeSeededXorShift32::generate().unwrap();
rng.next_u32();
Welcome to the land of unreproducible builds
Use this crate if you need simple random number generators for your project
and you don't want to depend on a big library like rand
.
You can seed your RNGs using the system time voxell_rng::time_seeded
or runtime entropy voxell_rng::runtime_seeded
.
There are 5 RNGs available:
SplitMix64
: a 64-bit RNG with 64-bit output used for seeding other RNGsXorShift32
: a 32-bit Xorshift RNG with 32-bit outputXorShift128
: a 128-bit Xorshift RNG with 64-bit output (Recommended)XoRoShiRo128
: a 128-bit XoRoShiRo RNG with 64-bit outputPcg8
throughPcg128
: the PCG family of RNGs
All RNGs implement BranchRng
which is a simple trait that provides a branch_rng
method
for creating a new divergent RNG from the current one. The resulting RNG will have a different
state and will produce different random numbers without needing to specify a new seed.
Examples
- Create a new RNG using a constant seed and use it:
use voxell_rng::rng::XorShift32;
// create the rng
let mut rng = XorShift32::new(0xcafebabe as u64);
// generate a new number
rng.next_f32();
- Seed your RNGs using the system time
use voxell_rng::time_seeded::TimeSeededXorShift32;
let mut rng = TimeSeededXorShift32::generate().unwrap();
rng.next_f32();
- Seed your RNGs using runtime entropy
use voxell_rng::rng::XorShift32;
// Default implementation for non-PCG RNG's use OS entropy
let mut rng = XorShift32::default();
rng.next_f32();
- Create new RNGs from a master RNG for divergent thread local RNGs:
use voxell_rng::prelude::*;
use voxell_rng::rng::XoRoShiRo128;
let mut master_rng = XoRoShiRo128::new(0xabad1dea as u64);
let thread_handles = (0..16)
.map(|_| {
let rng = master_rng.branch_rng();
std::thread::spawn(move || {
let mut thread_local_rng = rng;
for _ in 0..1000 {
thread_local_rng.next_u64();
}
})
})
.collect::<Vec<_>>();
Dependencies
~86–255KB