1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308
//! A scalable barrier (like [`std::sync::Barrier`]) that enables multiple threads to synchronize
//! the beginning of some computation.
//!
//! This crate provides a similar interface as [`std::sync::Barrier`], but behaves much better in
//! the face of many concurrently waiting threads, and incurs a lower per-thread latency penalty
//! (see benchmarks below). The interface does differ from the standard library barrier however:
//!
//! - `Barrier` in this crate is `Clone`, and should *not* be wrapped in a `sync::Arc`.
//! - `Barrier::wait` in this crate takes a `&mut self` receiver as each thread must keep some
//! local state.
//!
//! Furthermore, when a thread blocks on `Barrier::wait`, the thread will (currently) *never* be
//! suspended, and instead spin on the barrier. For the first few spins, it will also not call
//! `sched_yield` to avoid the cost of thread sleep/wakeup. If threads are expected to reach the
//! barrier at nearly the same time, or barrier latency is critical, this is probably what you
//! want. However, if barriers are staggered and far between, then you may want to use
//! [`std::sync::Barrier`] instead, as it is better about handling long waits.
//!
//! # Examples
//!
//! ```
//! use hurdles::Barrier;
//! use std::thread;
//!
//! let mut handles = Vec::with_capacity(10);
//! let mut barrier = Barrier::new(10);
//! for _ in 0..10 {
//! let mut c = barrier.clone();
//! // The same messages will be printed together.
//! // You will NOT see any interleaving.
//! handles.push(thread::spawn(move|| {
//! println!("before wait");
//! c.wait();
//! println!("after wait");
//! }));
//! }
//! // Wait for other threads to finish.
//! for handle in handles {
//! handle.join().unwrap();
//! }
//! ```
//!
//! # Implementation
//!
//! At the time of writing, the implementation of `std::sync::Barrier` internally uses a `Mutex`,
//! which causes contention with many waiting threads, and incurs an undue performance overhead for
//! each call to `wait`.
//!
//! This crate instead implements a counter-based linear barrier as described in "3.1 Centralized
//! Barriers" in Mellor-Crummey and Scott’s paper [Algorithms for scalable synchronization on
//! shared-memory multiprocessors][1] from 1991. For a higher-level explanation, see Lars-Dominik
//! Braun's [Introduction to barrier algorithms][2].
//!
//! # Numbers
//!
//! Modern laptop with 2-core (4HT) Intel Core i7-5600U @ 2.60GHz:
//!
//! ```text
//! test tests::ours_2 ... bench: 190 ns/iter (+/- 24)
//! test tests::std_2 ... bench: 2,054 ns/iter (+/- 822)
//! test tests::ours_4 ... bench: 236 ns/iter (+/- 2)
//! test tests::std_4 ... bench: 11,913 ns/iter (+/- 60)
//! ```
//!
//! Dell server with 2x 10-core (20HT) Intel Xeon E5-2660 v3 @ 2.60GHz across two NUMA nodes:
//!
//! ```text
//! test tests::ours_4 ... bench: 689 ns/iter (+/- 9)
//! test tests::std_4 ... bench: 4,762 ns/iter (+/- 151)
//! test tests::ours_8 ... bench: 1,380 ns/iter (+/- 13)
//! test tests::std_8 ... bench: 17,545 ns/iter (+/- 288)
//! test tests::ours_16 ... bench: 2,970 ns/iter (+/- 33)
//! test tests::std_16 ... bench: 38,215 ns/iter (+/- 469)
//! test tests::ours_32 ... bench: 3,838 ns/iter (+/- 129)
//! test tests::std_32 ... bench: 94,266 ns/iter (+/- 12,243)
//! ```
//!
//! [1]: https://dl.acm.org/citation.cfm?doid=103727.103729
//! [2]: https://6xq.net/barrier-intro/
//! [`std::sync::Barrier`]: https://doc.rust-lang.org/std/sync/struct.Barrier.html
#![deny(missing_docs)]
#![cfg_attr(feature = "nightly", feature(test))]
#[cfg(feature = "nightly")]
extern crate test;
extern crate parking_lot_core;
use std::sync::{atomic, Arc};
struct BarrierInner {
gsense: atomic::AtomicBool,
count: atomic::AtomicUsize,
max: usize,
}
/// A barrier which enables multiple threads to synchronize the beginning of some computation.
pub struct Barrier {
inner: Arc<BarrierInner>,
lsense: bool,
used: bool,
}
/// A `BarrierWaitResult` is returned by [`wait`] when all threads in the [`Barrier`]
/// have rendezvoused.
///
/// # Examples
///
/// ```
/// use hurdles::Barrier;
///
/// let mut barrier = Barrier::new(1);
/// let barrier_wait_result = barrier.wait();
/// ```
///
/// [`wait`]: struct.Barrier.html#method.wait
/// [`Barrier`]: struct.Barrier.html
pub struct BarrierWaitResult(bool);
impl Barrier {
/// Creates a new barrier that can block a given number of threads.
///
/// A barrier will block `n-1` threads which call [`wait`] and then wake up all threads at once
/// when the `n`th thread calls [`wait`].
///
/// # Examples
///
/// ```
/// use hurdles::Barrier;
/// let mut barrier = Barrier::new(10);
/// ```
///
/// [`wait`]: struct.Barrier.html#method.wait
pub fn new(n: usize) -> Self {
Barrier {
used: false,
lsense: true,
inner: Arc::new(BarrierInner {
gsense: atomic::AtomicBool::new(true),
count: atomic::AtomicUsize::new(n),
max: n,
}),
}
}
/// Blocks the current thread until all threads have rendezvoused here.
///
/// Barriers are re-usable after all threads have rendezvoused once, and can be used
/// continuously.
///
/// A single (arbitrary) thread will receive a [`BarrierWaitResult`] that returns `true` from
/// [`is_leader`] when returning from this function, and all other threads will receive a
/// result that will return `false` from [`is_leader`].
///
/// # Examples
///
/// ```
/// use hurdles::Barrier;
/// use std::thread;
///
/// let mut handles = Vec::with_capacity(10);
/// let mut barrier = Barrier::new(10);
/// for _ in 0..10 {
/// let mut c = barrier.clone();
/// // The same messages will be printed together.
/// // You will NOT see any interleaving.
/// handles.push(thread::spawn(move|| {
/// println!("before wait");
/// c.wait();
/// println!("after wait");
/// }));
/// }
/// // Wait for other threads to finish.
/// for handle in handles {
/// handle.join().unwrap();
/// }
/// ```
///
/// [`BarrierWaitResult`]: struct.BarrierWaitResult.html
/// [`is_leader`]: struct.BarrierWaitResult.html#method.is_leader
pub fn wait(&mut self) -> BarrierWaitResult {
self.used = true;
self.lsense = !self.lsense;
if self.inner.count.fetch_sub(1, atomic::Ordering::SeqCst) == 1 {
// we're the last to reach the barrier -- release all
self.inner
.count
.store(self.inner.max, atomic::Ordering::SeqCst);
self.inner
.gsense
.store(self.lsense, atomic::Ordering::SeqCst);
BarrierWaitResult(true)
} else {
// wait for everyone to reach the barrier
let mut wait = parking_lot_core::SpinWait::new();
while self.inner.gsense.load(atomic::Ordering::SeqCst) != self.lsense {
// XXX: in theory we could go even further and park the thread eventually
wait.spin();
}
BarrierWaitResult(false)
}
}
}
impl Clone for Barrier {
fn clone(&self) -> Self {
assert!(!self.used);
Barrier {
used: false,
lsense: self.lsense,
inner: self.inner.clone(),
}
}
}
impl BarrierWaitResult {
/// Returns whether this thread from [`wait`] is the "leader thread".
///
/// Only one thread will have `true` returned from their result, all other
/// threads will have `false` returned.
///
/// [`wait`]: struct.Barrier.html#method.wait
///
/// # Examples
///
/// ```
/// use hurdles::Barrier;
///
/// let mut barrier = Barrier::new(1);
/// let barrier_wait_result = barrier.wait();
/// assert_eq!(barrier_wait_result.is_leader(), true);
/// ```
pub fn is_leader(&self) -> bool {
self.0
}
}
#[cfg(test)]
mod tests {
use super::Barrier;
use std::sync::mpsc::{channel, TryRecvError};
use std::thread;
#[cfg(feature = "nightly")]
use test::Bencher;
#[cfg(feature = "nightly")]
const BENCH_THREADS: usize = 4;
#[cfg(feature = "nightly")]
#[cfg_attr(feature = "nightly", bench)]
fn ours(b: &mut Bencher) {
let mut barrier = Barrier::new(BENCH_THREADS);
for _ in 0..(BENCH_THREADS - 1) {
let mut barrier = barrier.clone();
thread::spawn(move || loop {
barrier.wait();
});
}
b.iter(move || { barrier.wait(); })
}
#[cfg(feature = "nightly")]
#[cfg_attr(feature = "nightly", bench)]
fn std(b: &mut Bencher) {
use std::sync::{self, Arc};
let barrier = Arc::new(sync::Barrier::new(BENCH_THREADS));
for _ in 0..(BENCH_THREADS - 1) {
let barrier = barrier.clone();
thread::spawn(move || loop {
barrier.wait();
});
}
b.iter(move || { barrier.wait(); })
}
#[test]
fn test_barrier() {
const N: usize = 10;
let mut barrier = Barrier::new(N);
let (tx, rx) = channel();
for _ in 0..N - 1 {
let mut c = barrier.clone();
let tx = tx.clone();
thread::spawn(move || { tx.send(c.wait().is_leader()).unwrap(); });
}
// At this point, all spawned threads should be blocked,
// so we shouldn't get anything from the port
assert!(match rx.try_recv() {
Err(TryRecvError::Empty) => true,
_ => false,
});
let mut leader_found = barrier.wait().is_leader();
// Now, the barrier is cleared and we should get data.
for _ in 0..N - 1 {
if rx.recv().unwrap() {
assert!(!leader_found);
leader_found = true;
}
}
assert!(leader_found);
}
}