1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
|
<!DOCTYPE html>
<html class="writer-html5" lang="en">
<head>
<meta charset="utf-8" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
<title>11.8. Processor and memory affinity — Open MPI 5.0.8 documentation</title>
<link rel="stylesheet" type="text/css" href="../_static/pygments.css" />
<link rel="stylesheet" type="text/css" href="../_static/css/theme.css" />
<!--[if lt IE 9]>
<script src="../_static/js/html5shiv.min.js"></script>
<![endif]-->
<script data-url_root="../" id="documentation_options" src="../_static/documentation_options.js"></script>
<script src="../_static/jquery.js"></script>
<script src="../_static/underscore.js"></script>
<script src="../_static/_sphinx_javascript_frameworks_compat.js"></script>
<script src="../_static/doctools.js"></script>
<script src="../_static/sphinx_highlight.js"></script>
<script src="../_static/js/theme.js"></script>
<link rel="index" title="Index" href="../genindex.html" />
<link rel="search" title="Search" href="../search.html" />
<link rel="next" title="11.9. MPI-IO tuning options" href="mpi-io/index.html" />
<link rel="prev" title="11.7.4. Static cluster configurations" href="large-clusters/static-cluster-config.html" />
</head>
<body class="wy-body-for-nav">
<div class="wy-grid-for-nav">
<nav data-toggle="wy-nav-shift" class="wy-nav-side">
<div class="wy-side-scroll">
<div class="wy-side-nav-search" >
<a href="../index.html" class="icon icon-home">
Open MPI
</a>
<div role="search">
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
<input type="text" name="q" placeholder="Search docs" aria-label="Search docs" />
<input type="hidden" name="check_keywords" value="yes" />
<input type="hidden" name="area" value="default" />
</form>
</div>
</div><div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="Navigation menu">
<ul class="current">
<li class="toctree-l1"><a class="reference internal" href="../quickstart.html">1. Quick start</a></li>
<li class="toctree-l1"><a class="reference internal" href="../getting-help.html">2. Getting help</a></li>
<li class="toctree-l1"><a class="reference internal" href="../release-notes/index.html">3. Release notes</a></li>
<li class="toctree-l1"><a class="reference internal" href="../installing-open-mpi/index.html">4. Building and installing Open MPI</a></li>
<li class="toctree-l1"><a class="reference internal" href="../features/index.html">5. Open MPI-specific features</a></li>
<li class="toctree-l1"><a class="reference internal" href="../validate.html">6. Validating your installation</a></li>
<li class="toctree-l1"><a class="reference internal" href="../version-numbering.html">7. Version numbers and compatibility</a></li>
<li class="toctree-l1"><a class="reference internal" href="../mca.html">8. The Modular Component Architecture (MCA)</a></li>
<li class="toctree-l1"><a class="reference internal" href="../building-apps/index.html">9. Building MPI applications</a></li>
<li class="toctree-l1"><a class="reference internal" href="../launching-apps/index.html">10. Launching MPI applications</a></li>
<li class="toctree-l1 current"><a class="reference internal" href="index.html">11. Run-time operation and tuning MPI applications</a><ul class="current">
<li class="toctree-l2"><a class="reference internal" href="environment-var.html">11.1. Environment variables set for MPI applications</a></li>
<li class="toctree-l2"><a class="reference internal" href="networking/index.html">11.2. Networking support</a></li>
<li class="toctree-l2"><a class="reference internal" href="multithreaded.html">11.3. Running multi-threaded MPI applications</a></li>
<li class="toctree-l2"><a class="reference internal" href="dynamic-loading.html">11.4. Dynamically loading <code class="docutils literal notranslate"><span class="pre">libmpi</span></code> at runtime</a></li>
<li class="toctree-l2"><a class="reference internal" href="fork-system-popen.html">11.5. Calling fork(), system(), or popen() in MPI processes</a></li>
<li class="toctree-l2"><a class="reference internal" href="fault-tolerance/index.html">11.6. Fault tolerance</a></li>
<li class="toctree-l2"><a class="reference internal" href="large-clusters/index.html">11.7. Large Clusters</a></li>
<li class="toctree-l2 current"><a class="current reference internal" href="#">11.8. Processor and memory affinity</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#processor-affinity">11.8.1. Processor affinity</a></li>
<li class="toctree-l3"><a class="reference internal" href="#memory-affinity">11.8.2. Memory affinity</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="mpi-io/index.html">11.9. MPI-IO tuning options</a></li>
<li class="toctree-l2"><a class="reference internal" href="coll-tuned.html">11.10. Tuning Collectives</a></li>
<li class="toctree-l2"><a class="reference internal" href="benchmarking.html">11.11. Benchmarking Open MPI applications</a></li>
<li class="toctree-l2"><a class="reference internal" href="heterogeneity.html">11.12. Building heterogeneous MPI applications</a></li>
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="../app-debug/index.html">12. Debugging Open MPI Parallel Applications</a></li>
<li class="toctree-l1"><a class="reference internal" href="../developers/index.html">13. Developer’s guide</a></li>
<li class="toctree-l1"><a class="reference internal" href="../contributing.html">14. Contributing to Open MPI</a></li>
<li class="toctree-l1"><a class="reference internal" href="../license/index.html">15. License</a></li>
<li class="toctree-l1"><a class="reference internal" href="../history.html">16. History of Open MPI</a></li>
<li class="toctree-l1"><a class="reference internal" href="../man-openmpi/index.html">17. Open MPI manual pages</a></li>
<li class="toctree-l1"><a class="reference internal" href="../man-openshmem/index.html">18. OpenSHMEM manual pages</a></li>
</ul>
</div>
</div>
</nav>
<section data-toggle="wy-nav-shift" class="wy-nav-content-wrap"><nav class="wy-nav-top" aria-label="Mobile navigation menu" >
<i data-toggle="wy-nav-top" class="fa fa-bars"></i>
<a href="../index.html">Open MPI</a>
</nav>
<div class="wy-nav-content">
<div class="rst-content">
<div role="navigation" aria-label="Page navigation">
<ul class="wy-breadcrumbs">
<li><a href="../index.html" class="icon icon-home" aria-label="Home"></a></li>
<li class="breadcrumb-item"><a href="index.html"><span class="section-number">11. </span>Run-time operation and tuning MPI applications</a></li>
<li class="breadcrumb-item active"><span class="section-number">11.8. </span>Processor and memory affinity</li>
<li class="wy-breadcrumbs-aside">
<a href="../_sources/tuning-apps/affinity.rst.txt" rel="nofollow"> View page source</a>
</li>
</ul>
<hr/>
</div>
<div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
<div itemprop="articleBody">
<style>
.wy-table-responsive table td,.wy-table-responsive table th{white-space:normal}
</style><div class="section" id="processor-and-memory-affinity">
<h1><span class="section-number">11.8. </span>Processor and memory affinity<a class="headerlink" href="#processor-and-memory-affinity" title="Permalink to this heading"></a></h1>
<div class="section" id="processor-affinity">
<span id="tuning-using-paffinity-label"></span><h2><span class="section-number">11.8.1. </span>Processor affinity<a class="headerlink" href="#processor-affinity" title="Permalink to this heading"></a></h2>
<p>Open MPI supports processor affinity on a variety of systems through
process binding, in which each MPI process, along with its threads, is
“bound” to a specific subset of processing resources (cores, packages,
etc.). That is, the operating system will constrain that process to
run on only that subset.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>The operating system may allow other processes to run on the
same resources.</p>
</div>
<p>Affinity can improve performance by inhibiting excessive process
movement — for example, away from “hot” caches or NUMA memory.
Judicious bindings can improve performance by reducing resource
contention (by spreading processes apart from one another) or
improving interprocess communications (by placing processes close to
one another). Binding can also improve performance reproducibility by
eliminating variable process placement.</p>
<div class="admonition warning">
<p class="admonition-title">Warning</p>
<p>Processor affinity probably should <em>not</em> be used when a
node is over-subscribed (i.e., more processes are
launched than there are processors).</p>
<p>This can lead to a serious degradation in performance
(even more than simply oversubscribing the node). Open
MPI will usually detect this situation and automatically
disable the use of processor affinity (and display
run-time warnings to this effect).</p>
</div>
</div>
<div class="section" id="memory-affinity">
<h2><span class="section-number">11.8.2. </span>Memory affinity<a class="headerlink" href="#memory-affinity" title="Permalink to this heading"></a></h2>
<p>Memory affinity is critically important on modern servers
because most architectures exhibit Non-Uniform Memory Access (NUMA)
architectures. In a NUMA architecture, memory is physically
distributed throughout the machine even though it is virtually treated
as a single address space. That is, memory may be physically local to
one or more processors — and therefore remote to other processors.</p>
<p>Simply put: some memory will be faster to access (for a given process)
than others.</p>
<p>Open MPI supports general and specific memory affinity, meaning that
it generally tries to allocate all memory local to the processor that
asked for it. When shared memory is used for communication, Open MPI
uses memory affinity to make certain pages local to specific
processes in order to minimize memory network/bus traffic.</p>
<p>Open MPI supports memory affinity on a variety of systems, and is
controlled through the <a class="reference external" href="https://www.open-mpi.org/projects/hwloc/">Hardware Locality (hwloc)</a> library.</p>
<p>Note that memory affinity support is enabled
<em>only when processor affinity is enabled.</em> Specifically: using memory
affinity does not make sense if processor affinity is not enabled
because processes may allocate local memory and then move to a
different processor, potentially remote from the memory that it just
allocated.</p>
</div>
</div>
</div>
</div>
<footer><div class="rst-footer-buttons" role="navigation" aria-label="Footer">
<a href="large-clusters/static-cluster-config.html" class="btn btn-neutral float-left" title="11.7.4. Static cluster configurations" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left" aria-hidden="true"></span> Previous</a>
<a href="mpi-io/index.html" class="btn btn-neutral float-right" title="11.9. MPI-IO tuning options" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right" aria-hidden="true"></span></a>
</div>
<hr/>
<div role="contentinfo">
<p>© Copyright 2003-2025, The Open MPI Community.
<span class="lastupdated">Last updated on 2025-05-30 16:41:43 UTC.
</span></p>
</div>
Built with <a href="https://www.sphinx-doc.org/">Sphinx</a> using a
<a href="https://github.com/readthedocs/sphinx_rtd_theme">theme</a>
provided by <a href="https://readthedocs.org">Read the Docs</a>.
</footer>
</div>
</div>
</section>
</div>
<script>
jQuery(function () {
SphinxRtdTheme.Navigation.enable(true);
});
</script>
</body>
</html>
|
