Drazen Urch
7bd1550195
* Checkpoint * cpu cycle ffi * Rename * mixnode feature * Bundle libcpucycles
89 lines
4.3 KiB
HTML
89 lines
4.3 KiB
HTML
<html>
|
|
<head>
|
|
<meta http-equiv="content-type" content="text/html; charset=utf-8">
|
|
<meta name="viewport" content="width=device-width, initial-scale=1">
|
|
<style type="text/css">
|
|
html{overflow-y:scroll}
|
|
body{font-family:sans-serif}
|
|
p,ul,ol,blockquote,pre{font-size:0.9em;line-height:1.6em}
|
|
li p{font-size:1.0em}
|
|
blockquote p{font-size:1.0em}
|
|
tt{font-size:1.2em}
|
|
code{font-size:1.2em}
|
|
h1{font-size:1.5em}
|
|
h2{font-size:1.3em}
|
|
h3{font-size:1.0em}
|
|
h1 a{text-decoration:none}
|
|
table{border-collapse:collapse}
|
|
th,td{border:1px solid black}
|
|
table a{text-decoration:none}
|
|
table tr{font-size:0.9em;line-height:1.6em}
|
|
.links a:hover{text-decoration:underline}
|
|
.links a:active{text-decoration:underline}
|
|
.links img{width:200px;padding-left:1em}
|
|
.links td{border:0px;padding-top:0.5em;padding-bottom:0.5em}
|
|
.headline{padding:0;font-weight:bold;font-size:1.5em;vertical-align:top;padding-bottom:0.5em;color:#125d0d}
|
|
.navt{display:inline-block;box-sizing:border-box;-moz-box-sizing:border-box;-webkit-box-sizing:border-box;
|
|
min-width:14%;margin:0;padding:0;padding-left:0.5em;padding-right:0.5em;vertical-align:center;
|
|
font-weight:bold;font-size:1.1em;text-align:center;border:1px solid black}
|
|
.here{border-bottom:0px;background-color:#ffffff}
|
|
.away{background-color:#125d0d;}
|
|
.away a{text-decoration:none;display:block;color:#ffffff}
|
|
.away a:hover,.away a:active{text-decoration:underline}
|
|
.main{margin:0;padding-top:0em;padding-bottom:1%;clear:both}
|
|
</style>
|
|
<title>
|
|
Intro</title>
|
|
</head>
|
|
<body>
|
|
<div class=headline>
|
|
libcpucycles
|
|
</div>
|
|
<div class=nav>
|
|
<div class="navt here">Intro
|
|
</div><div class="navt away"><a href=download.html>Download</a>
|
|
</div><div class="navt away"><a href=install.html>Install</a>
|
|
</div><div class="navt away"><a href=api.html>API</a>
|
|
</div><div class="navt away"><a href=counters.html>Counters</a>
|
|
</div><div class="navt away"><a href=selection.html>Selection</a>
|
|
</div><div class="navt away"><a href=security.html>Security</a>
|
|
</div></div>
|
|
<div class=main>
|
|
<p>libcpucycles is a public-domain microlibrary for counting CPU cycles.
|
|
Cycle counts are not as detailed as
|
|
<a href="https://gamozolabs.github.io/metrology/2019/08/19/sushi_roll.html">Falk diagrams</a>
|
|
but are the most precise timers available to typical software; they are
|
|
central tools used in understanding and improving software performance.</p>
|
|
<p>The libcpucycles <a href="api.html">API</a> is simple: include <code><cpucycles.h></code>, call
|
|
<code>cpucycles()</code> to receive a <code>long long</code> whenever desired, and link with
|
|
<code>-lcpucycles</code>.</p>
|
|
<p><a href="counters.html">Internally</a>, libcpucycles understands machine-level
|
|
cycle counters for amd64 (both PMC and TSC), arm32, arm64 (both PMC and
|
|
VCT), mips64, ppc32, ppc64, riscv32, riscv64, s390x, sparc64, and x86.
|
|
libcpucycles also understands four OS-level mechanisms, which give
|
|
varying levels of accuracy: <code>mach_absolute_time</code>, <code>perf_event</code>,
|
|
<code>CLOCK_MONOTONIC</code>, and, as a fallback, microsecond-resolution
|
|
<code>gettimeofday</code>.</p>
|
|
<p>When the program first calls <code>cpucycles()</code>, libcpucycles automatically
|
|
benchmarks the available mechanisms and <a href="selection.html">selects</a> the
|
|
mechanism that does the best job. Subsequent <code>cpucycles()</code> calls are
|
|
thread-safe and very fast. An accompanying <code>cpucycles-info</code> program
|
|
prints a summary of cycle-counter accuracy.</p>
|
|
<p>For comparison, there is a simple-sounding <code>__rdtsc()</code> API provided by
|
|
compilers, but this works only on Intel/AMD CPUs and is generally noisier
|
|
than PMC. There is a <code>__builtin_readcyclecounter()</code> that works on more
|
|
CPUs, but this works only with <code>clang</code> and has the same noise problems.
|
|
Both of these mechanisms put the burden on the caller to figure out what
|
|
can be done on other CPUs. Various packages include their own more
|
|
portable abstraction layers for counting cycles (see, e.g., FFTW's
|
|
<a href="https://github.com/FFTW/fftw3/blob/master/kernel/cycle.h"><code>cycle.h</code></a>,
|
|
used to automatically select from among multiple implementations
|
|
provided by FFTW), but this creates per-package effort to keep up with
|
|
the latest cycle counters. The goal of libcpucycles is to provide
|
|
state-of-the-art cycle counting centrally for all packages to use.</p><hr><font size=1><b>Version:</b>
|
|
This is version 2023.01.06 of the "Intro" web page.
|
|
</font>
|
|
</div>
|
|
</body>
|
|
</html>
|