Carl LercheandGitHub ed5a94eb2d executor: rewrite the work-stealing thread pool (#1657)
This patch is a ground up rewrite of the existing work-stealing thread
pool. The goal is to reduce overhead while simplifying code when
possible.

At a high level, the following architectural changes were made:

- The local run queues were switched for bounded circle buffer queues.
- Reduce cross-thread synchronization.
- Refactor task constructs to use a single allocation and always include
  a join handle (#887).
- Simplify logic around putting workers to sleep and waking them up.

**Local run queues**

Move away from crossbeam's implementation of the Chase-Lev deque. This
implementation included unnecessary overhead as it supported
capabilities that are not needed for the work-stealing thread pool.
Instead, a fixed size circle buffer is used for the local queue. When
the local queue is full, half of the tasks contained in it are moved to
the global run queue.

**Reduce cross-thread synchronization**

This is done via many small improvements. Primarily, an upper bound is
placed on the number of concurrent stealers. Limiting the number of
stealers results in lower contention. Secondly, the rate at which
workers are notified and woken up is throttled. This also reduces
contention by preventing many threads from racing to steal work.

**Refactor task structure**

Now that Tokio is able to target a rust version that supports
`std::alloc` as well as `std::task`, the pool is able to optimize how
the task structure is laid out. Now, a single allocation per task is
required and a join handle is always provided enabling the spawner to
retrieve the result of the task (#887).

**Simplifying logic**

When possible, complexity is reduced in the implementation. This is done
by using locks and other simpler constructs in cold paths. The set of
sleeping workers is now represented as a `Mutex<VecDeque<usize>>`.
Instead of optimizing access to this structure, we reduce the amount the
pool must access this structure.

Secondly, we have (temporarily) removed `threadpool::blocking`. This
capability will come back later, but the original implementation was way
more complicated than necessary.

**Results**

The thread pool benchmarks have improved significantly:

Old thread pool:

```
test chained_spawn ... bench:   2,019,796 ns/iter (+/- 302,168)
test ping_pong     ... bench:   1,279,948 ns/iter (+/- 154,365)
test spawn_many    ... bench:  10,283,608 ns/iter (+/- 1,284,275)
test yield_many    ... bench:  21,450,748 ns/iter (+/- 1,201,337)
```

New thread pool:

```
test chained_spawn ... bench:     147,943 ns/iter (+/- 6,673)
test ping_pong     ... bench:     537,744 ns/iter (+/- 20,928)
test spawn_many    ... bench:   7,454,898 ns/iter (+/- 283,449)
test yield_many    ... bench:  16,771,113 ns/iter (+/- 733,424)
```

Real-world benchmarks improve significantly as well. This is testing the hyper hello
world server using: `wrk -t1 -c50 -d10`:

Old scheduler:

```
Running 10s test @ http://127.0.0.1:3000
  1 threads and 50 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   371.53us   99.05us   1.97ms   60.53%
    Req/Sec   114.61k     8.45k  133.85k    67.00%
  1139307 requests in 10.00s, 95.61MB read
Requests/sec: 113923.19
Transfer/sec:      9.56MB
```

New scheduler:

```
Running 10s test @ http://127.0.0.1:3000
  1 threads and 50 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   275.05us   69.81us   1.09ms   73.57%
    Req/Sec   153.17k    10.68k  171.51k    71.00%
  1522671 requests in 10.00s, 127.79MB read
Requests/sec: 152258.70
Transfer/sec:     12.78MB
```
2019-10-19 11:09:40 -07:00
2019-09-19 12:11:21 -07:00
2019-01-06 23:25:55 -08:00
2019-09-17 10:54:22 -04:00

Tokio

NOTE: Tokio's master is currently undergoing heavy development. This branch and the alpha releases will see API breaking changes and there are currently significant performance regressions that still need to be fixed before the final release. Use the v0.1.x branch for stable releases.

A runtime for writing reliable, asynchronous, and slim applications with the Rust programming language. It is:

  • Fast: Tokio's zero-cost abstractions give you bare-metal performance.

  • Reliable: Tokio leverages Rust's ownership, type system, and concurrency model to reduce bugs and ensure thread safety.

  • Scalable: Tokio has a minimal footprint, and handles backpressure and cancellation naturally.

Crates.io MIT licensed Build Status Gitter chat

Website | Guides | API Docs | Chat

Overview

Tokio is an event-driven, non-blocking I/O platform for writing asynchronous applications with the Rust programming language. At a high level, it provides a few major components:

  • A multithreaded, work-stealing based task scheduler.
  • A reactor backed by the operating system's event queue (epoll, kqueue, IOCP, etc...).
  • Asynchronous TCP and UDP sockets.

These components provide the runtime components necessary for building an asynchronous application.

Example

A basic TCP echo server with Tokio:

use tokio::net::TcpListener;
use tokio::prelude::*;
use std::net::SocketAddr;

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    let addr = "127.0.0.1:8080".parse::<SocketAddr>()?;
    let mut listener = TcpListener::bind(&addr).await?;

    loop {
        let (mut socket, _) = listener.accept().await?;

        tokio::spawn(async move {
            let mut buf = [0; 1024];

            // In a loop, read data from the socket and write the data back.
            loop {
                let n = match socket.read(&mut buf).await {
                    // socket closed
                    Ok(n) if n == 0 => return,
                    Ok(n) => n,
                    Err(e) => {
                        println!("failed to read from socket; err = {:?}", e);
                        return;
                    }
                };

                // Write the data back
                if let Err(e) = socket.write_all(&buf[0..n]).await {
                    println!("failed to write to socket; err = {:?}", e);
                    return;
                }
            }
        });
    }
}

More examples can be found here. Note that the master branch is currently being updated to use async / await. The examples are not fully ported. Examples for stable Tokio can be found here.

Getting Help

First, see if the answer to your question can be found in the Guides or the API documentation. If the answer is not there, there is an active community in the Tokio Gitter channel. We would be happy to try to answer your question. Last, if that doesn't work, try opening an issue with the question.

Contributing

🎈 Thanks for your help improving the project! We are so happy to have you! We have a contributing guide to help you get involved in the Tokio project.

Project layout

The tokio crate, found at the root, is primarily intended for use by application developers. Library authors should depend on the sub crates, which have greater guarantees of stability.

The crates included as part of Tokio are:

  • tokio-executor: Task executors and related utilities. Includes a single-threaded executor and a multi-threaded, work-stealing, executor.

  • tokio-fs: Filesystem (and standard in / out) APIs.

  • tokio-codec: Utilities for encoding and decoding protocol frames.

  • tokio-io: Asynchronous I/O related traits and utilities.

  • tokio-macros: Macros for usage with Tokio.

  • tokio-net: Event loop that drives I/O resources as well as TCP, UDP, and unix domain socket apis.

  • tokio-timer: Time related APIs.

In addition to the crates in this repository, the Tokio project also maintains several other libraries, including:

  • tracing (formerly tokio-trace): A framework for application-level tracing and async-aware diagnostics.

  • mio: A low-level, cross-platform abstraction over OS I/O APIs that powers tokio.

  • bytes: Utilities for working with bytes, including efficient byte buffers.

Supported Rust Versions

Tokio is built against the latest stable, nightly, and beta Rust releases. The minimum version supported is the stable release from three months before the current stable release version. For example, if the latest stable Rust is 1.29, the minimum version supported is 1.26. The current Tokio version is not guaranteed to build on Rust versions earlier than the minimum supported version.

License

This project is licensed under the MIT license.

Contribution

Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in Tokio by you, shall be licensed as MIT, without any additional terms or conditions.

S
Description
A runtime for writing reliable asynchronous applications with Rust. Provides I/O, networking, scheduling, timers, ...
Readme MIT
66 MiB
Languages
Rust 100%