6 releases

0.2.0 Feb 19, 2023
0.1.3 Jan 21, 2023
0.0.0-pre-release0.0 Jan 16, 2023

#988 in Concurrency

MIT license

88KB
412 lines

This crates provides a sequentially locking Ring Buffer. It allows for a fast and non-writer-blocking SPMC-queue, where all consumers read all messages.

Usage

There are two ways of consuming from the queue. If threads share a ReadGuard through a shared reference, they will steal queue items from one anothers such that no two threads will read the same message. When a ReadGuard is cloned, the new ReadGuard's reading progress will no longer affect the other one. If two threads each use a separate ReadGuard, they will be able to read the same messages.

use sling::*;

let buffer = RingBuffer::<_, 256>::new();

let mut writer = buffer.try_lock().unwrap();
let mut reader = buffer.reader();

std::thread::scope(|s| {
    let reader = &reader;
    for t in 0..8 {
        s.spawn(move || {
            for _ in 0..100 {
                if let Some(val) = reader.pop_front() {
                    println!("t: {}, val: {:?}", t, val);
                };
            }
        });
    }

    for i in 0..100 {
        writer.push_back([i, i, i]);
    }
});

Important!

It is also important to keep in mind, that slow readers will be overrun by the writer if they do not consume messages quickly enough. This can happen quite frequently if the buffer size is not large enough. It is advisable to test applications on a case-by-case basis and find a buffer size that is optimal to your use-case.

Benchmarks

Ping speeds compared to Crossbeam-Channel's mpmc channel and lockfree's spmc channel.

violin-plot-1

violin-plot-2 Disclaimer: These benchmarks were performed on a personal computer and may not be prepresentative of relative performance on other computers.

Specifications of the benchmarking system:

Architecture:            x86_64
  CPU op-mode(s):        32-bit, 64-bit
  Address sizes:         39 bits physical, 48 bits virtual
  Byte Order:            Little Endian
CPU(s):                  16
  On-line CPU(s) list:   0-15
Vendor ID:               GenuineIntel
  Model name:            Intel(R) Core(TM) i7-10700K CPU @ 3.80GHz
    CPU family:          6
    Model:               165
    Thread(s) per core:  2
    Core(s) per socket:  8
    Socket(s):           1
    Stepping:            5
    CPU(s) scaling MHz:  68%
    CPU max MHz:         5100.0000
    CPU min MHz:         800.0000
    BogoMIPS:            7602.45

Virtualization features:
  Virtualization:        VT-x
Caches (sum of all):
  L1d:                   256 KiB (8 instances)
  L1i:                   256 KiB (8 instances)
  L2:                    2 MiB (8 instances)
  L3:                    16 MiB (1 instance)
NUMA:
  NUMA node(s):          1
  NUMA node0 CPU(s):     0-15

Flamegraph

flamegraph

Dependencies

~0–26MB
~330K SLoC