Crates.io | queuecheck |
lib.rs | queuecheck |
version | 0.1.1 |
source | src |
created_at | 2017-03-25 02:04:25.592314 |
updated_at | 2018-08-17 06:45:09.971604 |
description | A thread-safe queue testing and benchmarking library. |
homepage | |
repository | https://github.com/KyleMayes/queuecheck |
max_upload_size | |
id | 9142 |
size | 29,068 |
A thread-safe queue testing and benchmarking library.
Supported on the stable, beta, and nightly Rust channels.
Released under the Apache License 2.0.
The below tests the unbounded MPMC queue from the standard library by producing 100,000 items using two producer threads which are then consumed by one consumer thread.
use std::sync::mpsc::{self, Receiver, Sender};
let (producer, consumer) = mpsc::channel();
queuecheck_test!(
// enqueue/dequeue operation pairs
100_000,
// producer threads
vec![producer.clone(), producer],
// consumer threads
vec![consumer],
// produce operation
|p: &Sender<String>, i: String| p.send(i).unwrap(),
// consume operation
|c: &Receiver<String>| c.try_recv().ok()
);
The below benchmarks the latency of the unbounded MPMC queue from the standard library by producing 100,000 items using two producer threads which are then consumed by one consumer thread.
use std::sync::mpsc::{self, Receiver, Sender};
let (producer, consumer) = mpsc::channel();
let latency = queuecheck_bench_latency!(
// warmup and measurement enqueue/dequeue operation pairs
(1_000, 100_000),
// producer threads
vec![producer.clone(), producer],
// consumer threads
vec![consumer],
// produce operation
|p: &Sender<i32>, i: i32| p.send(i).unwrap(),
// consume operation
|c: &Receiver<i32>| c.try_recv().ok()
);
latency.report("mpmc", &[50.0, 70.0, 90.0, 95.0, 99.09]);
The below benchmarks the throughput of the unbounded MPMC queue from the standard library by producing 100,000 items using two producer threads which are then consumed by one consumer thread.
use std::sync::mpsc::{self, Receiver, Sender};
let (producer, consumer) = mpsc::channel();
let ops = queuecheck_bench_throughput!(
// warmup and measurement enqueue/dequeue operation pairs
(1_000, 100_000),
// producer threads
vec![producer.clone(), producer],
// consumer threads
vec![consumer],
// produce operation
|p: &Sender<i32>, i: i32| p.send(i).unwrap(),
// consume operation
|c: &Receiver<i32>| c.try_recv().ok()
);
println!("{:.3} operation/second", ops);