<span id="install"></span><span id="index-0"></span><h1>Installation<a class="headerlink" href="#installation" title="Permalink to this headline">¶</a></h1>
<div class="section" id="supported-platforms">
<h2>Supported platforms<a class="headerlink" href="#supported-platforms" title="Permalink to this headline">¶</a></h2>
<p id="index-1">CVX is supported on 32-bit and 64-bit versions of Linux, Mac OSX, and Windows. For 32-bit
platforms, MATLAB version 7.5 (R2007b) or later is required; for 64-bit platforms, MATLAB
version 7.8 (R2009a) or later is required. There are some important platform-specific
cautions, however:</p>
<ul class="simple">
<li>Gurobi support requires Matlab 7.7 (R2008b) or later.</li>
<li>32-bit Linux: the Gurobi solver is not available for this platform, as Gurobi is phasing
out support for 32-bit Linux altogether.</li>
<li>Older versions of Mac OS X (e.g. 10.5) ship with Java 1.5. The standard version of
CVX works properly on this platform, but CVX Professional support requires Java 1.6.
To restore this support, upgrade your operating system or Java installation.</li>
</ul>
<p>As of version 2.0, support for versions 7.4 (R2007a) or older has been discontinued.
If you need to use CVX with these older versions of Matlab, please use CVX 1.22 or
earlier, which will remain available indefinitely on the CVX Research web site. However,
this version is no longer supported, and will not receive bug fixes or improvements.
We strongly encourage you to update your Matlab installation to the latest version
possible.</p>
<div class="admonition note" id="index-2">
<p class="first admonition-title">Note</p>
<p class="last">If you wish to use CVX with Gurobi or MOSEK, they must be installed and accessible
from MATLAB <em>before</em> running <tt class="docutils literal"><span class="pre">cvx_setup</span></tt>. See <a class="reference internal" href="#extsolv"><em>below</em></a> for more details.</p>
</div>
<ol class="arabic">
<li><p class="first">Retrieve the latest version of CVX from <a class="reference external" href="http://cvxr.com/cvx/download">the web site</a>.
You can download the package as either a <tt class="docutils literal"><span class="pre">.zip</span></tt> file or a <tt class="docutils literal"><span class="pre">.tar.gz</span></tt> file.</p>
</li>
<li><p class="first">Unpack the file anywhere you like; a directory called <tt class="docutils literal"><span class="pre">cvx</span></tt> will be
created. There are two important exceptions:</p>
<ul class="simple">
<li><em>Do not</em> place CVX in Matlab’s own <tt class="docutils literal"><span class="pre">toolbox</span></tt> directory, Octave’s built-in
scripts directory.</li>
<li><em>Do not</em> unpack a new version of CVX on top of an old one. We recommend moving the
old version out of the way, but do not delete it until you are sure the new
version is working as you expect.</li>
</ul>
</li>
<li><p class="first">Start Matlab or Octave. <em>Do not add CVX to your path by hand.</em></p>
</li>
<li><p class="first">Change directories to the top of the CVX distribution, and run the <tt class="docutils literal"><span class="pre">cvx_setup</span></tt>
command. For example, if you installed CVX into <tt class="docutils literal"><span class="pre">C\personal\cvx</span></tt> on
<p>The <tt class="docutils literal"><span class="pre">cvx_setup</span></tt> function performs a variety of tasks to verify that your
installation is correct, sets your Matlab/Octave search path so it can find all of the CVX
program files, and runs a simple test problem to verify the installation.</p>
</li>
<li><p class="first">In some cases—usually on Linux—the <tt class="docutils literal"><span class="pre">cvx_setup</span></tt> command may instruct you to
create or modify a <tt class="docutils literal"><span class="pre">startup.m</span></tt> file that allows you to use CVX without having
to type <tt class="docutils literal"><span class="pre">cvx_setup</span></tt> every time you re-start Matlab.</p>
<span id="licinstall"></span><span id="index-3"></span><h2>Installing a CVX Professional license<a class="headerlink" href="#installing-a-cvx-professional-license" title="Permalink to this headline">¶</a></h2>
<p>If you acquire a license key for CVX Professional, the only change required to the above
steps is to include the name of the license file as an input to the <tt class="docutils literal"><span class="pre">cvx_setup</span></tt> command.
For example, if you saved your license file to <tt class="docutils literal"><span class="pre">~/licenses/cvx_license.mat</span></tt> on a Mac,
<span id="extsolv"></span><span id="index-4"></span><h2>Solvers included with CVX<a class="headerlink" href="#solvers-included-with-cvx" title="Permalink to this headline">¶</a></h2>
<p>All versions of CVX include copies of the solvers
in the directories <tt class="file docutils literal"><span class="pre">cvx/sedumi</span></tt> and <tt class="file docutils literal"><span class="pre">cvx/sdpt3</span></tt>, respectively. When you
run <cite>cvx_setup</cite>, CVX will automatically add these solvers to its solver list.</p>
<p>If you have downloaded a CVX Professional Solver Bundle, then the solvers