stabilizer_qec 2.0.0

Quantum error-correction simulation and decoding: Stim-format circuits and error models, a bit-parallel sampler, sparse-blossom, correlated and belief matching, BP+OSD, window decoders, bivariate bicycle codes, lattice surgery, Clifford+T circuits with coherent noise, and sampling and BP on the GPU.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

//! Work over shots cut into one contiguous range per thread. Used by the Python bindings; the
//! WebAssembly build has one thread and does not call it.
#![cfg_attr(not(feature = "python"), allow(dead_code))]

use std::ops::Range;

/// `threads = 0` is every core; never more threads than units of work.
pub fn resolve_threads(threads: usize, units: usize) -> usize {
    let threads = if threads == 0 { std::thread::available_parallelism().map_or(1, |n| n.get()) } else { threads };
    threads.clamp(1, units.max(1))
}

/// `work` over `0..n`, one contiguous range per thread, the results in order.
pub fn parallel<T: Send>(n: usize, threads: usize, work: impl Fn(Range<usize>) -> Vec<T> + Sync) -> Vec<T> {
    let threads = resolve_threads(threads, n);
    let chunk = n.div_ceil(threads).max(1);
    std::thread::scope(|scope| {
        let work = &work;
        let handles: Vec<_> =
            (0..threads).map(|t| scope.spawn(move || work((t * chunk).min(n)..((t + 1) * chunk).min(n)))).collect();
        handles.into_iter().flat_map(|h| h.join().expect("a worker thread panicked")).collect()
    })
}