| <!DOCTYPE html><html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width, initial-scale=1.0"><meta name="generator" content="rustdoc"><meta name="description" content="Performance characteristics."><meta name="keywords" content="rust, rustlang, rust-lang, performance"><title>arc_swap::docs::performance - Rust</title><link rel="preload" as="font" type="font/woff2" crossorigin href="../../../SourceSerif4-Regular.ttf.woff2"><link rel="preload" as="font" type="font/woff2" crossorigin href="../../../FiraSans-Regular.woff2"><link rel="preload" as="font" type="font/woff2" crossorigin href="../../../FiraSans-Medium.woff2"><link rel="preload" as="font" type="font/woff2" crossorigin href="../../../SourceCodePro-Regular.ttf.woff2"><link rel="preload" as="font" type="font/woff2" crossorigin href="../../../SourceSerif4-Bold.ttf.woff2"><link rel="preload" as="font" type="font/woff2" crossorigin href="../../../SourceCodePro-Semibold.ttf.woff2"><link rel="stylesheet" href="../../../normalize.css"><link rel="stylesheet" href="../../../rustdoc.css" id="mainThemeStyle"><link rel="stylesheet" href="../../../ayu.css" disabled><link rel="stylesheet" href="../../../dark.css" disabled><link rel="stylesheet" href="../../../light.css" id="themeStyle"><script id="default-settings" ></script><script src="../../../storage.js"></script><script defer src="../../../main.js"></script><noscript><link rel="stylesheet" href="../../../noscript.css"></noscript><link rel="alternate icon" type="image/png" href="../../../favicon-16x16.png"><link rel="alternate icon" type="image/png" href="../../../favicon-32x32.png"><link rel="icon" type="image/svg+xml" href="../../../favicon.svg"></head><body class="rustdoc mod"><!--[if lte IE 11]><div class="warning">This old browser is unsupported and will most likely display funky things.</div><![endif]--><nav class="mobile-topbar"><button class="sidebar-menu-toggle">☰</button><a class="sidebar-logo" href="../../../arc_swap/index.html"><div class="logo-container"><img class="rust-logo" src="../../../rust-logo.svg" alt="logo"></div></a><h2></h2></nav><nav class="sidebar"><a class="sidebar-logo" href="../../../arc_swap/index.html"><div class="logo-container"><img class="rust-logo" src="../../../rust-logo.svg" alt="logo"></div></a><h2 class="location"><a href="#">Module performance</a></h2><div class="sidebar-elems"></div></nav><main><div class="width-limiter"><nav class="sub"><form class="search-form"><div class="search-container"><span></span><input class="search-input" name="search" autocomplete="off" spellcheck="false" placeholder="Click or press ‘S’ to search, ‘?’ for more options…" type="search"><div id="help-button" title="help" tabindex="-1"><a href="../../../help.html">?</a></div><div id="settings-menu" tabindex="-1"><a href="../../../settings.html" title="settings"><img width="22" height="22" alt="Change settings" src="../../../wheel.svg"></a></div></div></form></nav><section id="main-content" class="content"><div class="main-heading"><h1 class="fqn">Module <a href="../../index.html">arc_swap</a>::<wbr><a href="../index.html">docs</a>::<wbr><a class="mod" href="#">performance</a><button id="copy-path" onclick="copy_path(this)" title="Copy item path to clipboard"><img src="../../../clipboard.svg" width="19" height="18" alt="Copy item path"></button></h1><span class="out-of-band"><a class="srclink" href="../../../src/arc_swap/docs/performance.rs.html#1-87">source</a> · <a id="toggle-all-docs" href="javascript:void(0)" title="collapse all docs">[<span class="inner">−</span>]</a></span></div><details class="rustdoc-toggle top-doc" open><summary class="hideme"><span>Expand description</span></summary><div class="docblock"><p>Performance characteristics.</p> |
| <p>There are several performance advantages of <a href="../../type.ArcSwap.html"><code>ArcSwap</code></a> over <a href="std::sync::RwLock"><code>RwLock</code></a>.</p> |
| <h3 id="lock-free-readers"><a href="#lock-free-readers">Lock-free readers</a></h3> |
| <p>All the read operations are always <a href="https://en.wikipedia.org/wiki/Non-blocking_algorithm#Lock-freedom">lock-free</a>. Most of the time, they are actually |
| <a href="https://en.wikipedia.org/wiki/Non-blocking_algorithm#Wait-freedom">wait-free</a>. They are <a href="https://en.wikipedia.org/wiki/Non-blocking_algorithm#Lock-freedom">lock-free</a> from time to time, with at least <code>usize::MAX / 4</code> accesses |
| that are <a href="https://en.wikipedia.org/wiki/Non-blocking_algorithm#Wait-freedom">wait-free</a> in between.</p> |
| <p>Writers are <a href="https://en.wikipedia.org/wiki/Non-blocking_algorithm#Lock-freedom">lock-free</a>.</p> |
| <p>Whenever the documentation talks about <em>contention</em> in the context of <a href="../../type.ArcSwap.html"><code>ArcSwap</code></a>, it talks |
| about contention on the CPU level ‒ multiple cores having to deal with accessing the same cache |
| line. This slows things down (compared to each one accessing its own cache line), but an |
| eventual progress is still guaranteed and the cost is significantly lower than parking threads |
| as with mutex-style contention.</p> |
| <h3 id="speeds"><a href="#speeds">Speeds</a></h3> |
| <p>The base line speed of read operations is similar to using an <em>uncontended</em> <a href="std::sync::Mutex"><code>Mutex</code></a>. |
| However, <a href="../../struct.ArcSwapAny.html#method.load"><code>load</code></a> suffers no contention from any other read operations and only slight |
| ones during updates. The <a href="../../struct.ArcSwapAny.html#method.load_full"><code>load_full</code></a> operation is additionally contended only on |
| the reference count of the <a href="std::sync::Arc"><code>Arc</code></a> inside ‒ so, in general, while <a href="std::sync::Mutex"><code>Mutex</code></a> rapidly |
| loses its performance when being in active use by multiple threads at once and |
| <a href="std::sync::RwLock"><code>RwLock</code></a> is slow to start with, <a href="../../type.ArcSwap.html"><code>ArcSwap</code></a> mostly keeps its performance even when read by |
| many threads in parallel.</p> |
| <p>Write operations are considered expensive. A write operation is more expensive than access to |
| an <em>uncontended</em> <a href="std::sync::Mutex"><code>Mutex</code></a> and on some architectures even slower than uncontended |
| <a href="std::sync::RwLock"><code>RwLock</code></a>. However, it is faster than either under contention.</p> |
| <p>There are some (very unscientific) <a href="https://github.com/vorner/arc-swap/tree/master/benchmarks">benchmarks</a> within the source code of the library, and the |
| <a href="../../strategy/type.DefaultStrategy.html" title="crate::DefaultStrategy"><code>DefaultStrategy</code></a> has some numbers measured on my computer.</p> |
| <p>The exact numbers are highly dependant on the machine used (both absolute numbers and relative |
| between different data structures). Not only architectures have a huge impact (eg. x86 vs ARM), |
| but even AMD vs. Intel or two different Intel processors. Therefore, if what matters is more |
| the speed than the wait-free guarantees, you’re advised to do your own measurements.</p> |
| <p>Further speed improvements may be gained by the use of the <a href="../../cache/struct.Cache.html"><code>Cache</code></a>.</p> |
| <h3 id="consistency"><a href="#consistency">Consistency</a></h3> |
| <p>The combination of <a href="https://en.wikipedia.org/wiki/Non-blocking_algorithm#Wait-freedom">wait-free</a> guarantees of readers and no contention between concurrent |
| <a href="../../struct.ArcSwapAny.html#method.load"><code>load</code></a>s provides <em>consistent</em> performance characteristics of the synchronization mechanism. |
| This might be important for soft-realtime applications (the CPU-level contention caused by a |
| recent update/write operation might be problematic for some hard-realtime cases, though).</p> |
| <h3 id="choosing-the-right-reading-operation"><a href="#choosing-the-right-reading-operation">Choosing the right reading operation</a></h3> |
| <p>There are several load operations available. While the general go-to one should be |
| <a href="../../struct.ArcSwapAny.html#method.load"><code>load</code></a>, there may be situations in which the others are a better match.</p> |
| <p>The <a href="../../struct.ArcSwapAny.html#method.load"><code>load</code></a> usually only borrows the instance from the shared <a href="../../type.ArcSwap.html"><code>ArcSwap</code></a>. This makes |
| it faster, because different threads don’t contend on the reference count. There are two |
| situations when this borrow isn’t possible. If the content gets changed, all existing |
| <a href="../../struct.Guard.html"><code>Guard</code></a>s are promoted to contain an owned instance. The promotion is done by the |
| writer, but the readers still need to decrement the reference counts of the old instance when |
| they no longer use it, contending on the count.</p> |
| <p>The other situation derives from internal implementation. The number of borrows each thread can |
| have at each time (across all <a href="../../struct.Guard.html"><code>Guard</code></a>s) is limited. If this limit is exceeded, an owned |
| instance is created instead.</p> |
| <p>Therefore, if you intend to hold onto the loaded value for extended time span, you may prefer |
| <a href="../../struct.ArcSwapAny.html#method.load_full"><code>load_full</code></a>. It loads the pointer instance (<a href="std::sync::Arc"><code>Arc</code></a>) without borrowing, which is |
| slower (because of the possible contention on the reference count), but doesn’t consume one of |
| the borrow slots, which will make it more likely for following <a href="../../struct.ArcSwapAny.html#method.load"><code>load</code></a>s to have a slot |
| available. Similarly, if some API needs an owned <code>Arc</code>, <a href="../../struct.ArcSwapAny.html#method.load_full"><code>load_full</code></a> is more convenient and |
| potentially faster then first <a href="../../struct.ArcSwapAny.html#method.load"><code>load</code></a>ing and then cloning that <a href="std::sync::Arc"><code>Arc</code></a>.</p> |
| <p>Additionally, it is possible to use a <a href="../../cache/struct.Cache.html"><code>Cache</code></a> to get further speed improvement at the |
| cost of less comfortable API and possibly keeping the older values alive for longer than |
| necessary.</p> |
| </div></details></section></div></main><div id="rustdoc-vars" data-root-path="../../../" data-current-crate="arc_swap" data-themes="ayu,dark,light" data-resource-suffix="" data-rustdoc-version="1.66.0-nightly (5c8bff74b 2022-10-21)" ></div></body></html> |
