installation.html

<div class="code-box">
    
    <center><h1 id="installation">Installation</h1></center>
    
    <p>When you have Python 3.10 or higher installed (CPython, 64-bit), </p>
        <codebox>
            <p>python -m pip install pyoomph</p>
        </codebox>
    
    <p>should install the recent release. 
    For maximum performance and system-specific information, please refer to the sections below. </p>
  
    <p>If you cannot manage to install it, refer to our <a href="https://pyoomph.readthedocs.io/">tutorial</a>. If this cannot help, you can ask for help (c.diddens@utwente.nl)</p>

    <center><h2 id="on-windows">On Windows</h2></center>

        <p>For maximum performance, also install <a href="https://docs.microsoft.com/visualstudio/msbuild/msbuild">Microsoft Build Tools</a>, available for download <a href="https://aka.ms/vs/16/release/vs_buildtools.exe">here</a>. </p>

        <p>Verify whether everything runs fine by </p>

        <codebox>
            <p>python -m pyoomph check all</p>
        </codebox>

    <center><h2 id="on-linux">On Linux</h2></center>
    
    <p>If you have installed via pip (see above), just make sure that you have the <code>gcc</code> compiler installed and check via</p>
        <codebox>
            <p>python -m pyoomph check all</p>
        </codebox>
    
    <center><h2 id="on-mac">On Mac</h2></center>

        <p><strong>If you have a recent Mac with an Apple silicon (arm64) chip</strong>, mind the following:
        Python cannot install the fast <code>MKL Pardiso solver</code> on arm64 natively and will use <code>Accelerate</code> by default, which becomes unacceptably slow for larger problems. There are two workarounds, either install pyoomph in a <code>Rosetta 2 terminal</code> to use MKL Pardiso or install <code>PETSc/SLEPc</code> and use <code>MUMPS</code> as fast linear solver instead.</p>
        
        <ul><li><p>For the first option, you must run the installation command above in a <code>Rosetta 2 terminal</code>, see <a href="https://www.courier.com/blog/tips-and-tricks-to-setup-your-apple-m1-for-development/">here</a> how to set it up (note that recent systems must be handled as described <a href="https://developer.apple.com/forums/thread/718666">here</a>). Also, please make sure to not upgrade your <code>mkl</code> package, i.e. enforce it to version 2021.4.0.</p>
        
             <codebox>
                <p>python3 -m pip install mkl==2021.4.0</p>
            </codebox>
            
<p> See <a href="https://pyoomph.readthedocs.io/en/latest/tutorial/installation/source/mac.html">here</a> for further details. Also, you might have to use a less recent version of python, see <a href="https://github.com/pyoomph/pyoomph/blob/main/Mac_arm64_with_Pardiso.md">here</a> for details.</p> </li>
        
        <li><p>For the second option (PETSc with MUMPS), please refer to <a href="https://pyoomph.readthedocs.io/en/latest/tutorial/installation/petscslepc.html">the installation instructions</a>.</p></li>       
</ul>
        <p><strong>In any case</strong>, make sure to have the <code>XCode</code> developer tools, e.g., by installing them via</p>

            <codebox>
                <p>xcode-select --install</p>
            </codebox>
	        <p>(in Rosetta2 for the first option) and test pyoomph via</p>

            <codebox>
                <p>python -m pyoomph check all</p>
            </codebox>


    <center><h2 id="compilation-from-source">Compilation from source</h2></center>

       Compilation from source on Linux or Mac systems is described <a href="https://github.com/pyoomph/pyoomph/blob/main/INSTALL.md#compilation-from-source">in our github repository</a> and <a href="https://pyoomph.readthedocs.io/en/latest/tutorial/installation/compile.html">our tutorial</a>.<br>
       Just clone our code
        <codebox>
            <p>git clone <a href="https://github.com/pyoomph/pyoomph.git">https://github.com/pyoomph/pyoomph.git</a></p>
        </codebox>            
       and follow the steps in the file <code>INSTALL.md</code> therein.
            
</div>