Deploying to gh-pages from @ ea39c45 🚀

Author: luraess

Diff for: software_install/index.html

@@ -373,11 +373,24 @@ <h3 id=julia_mpi_gpu_on_piz_daint ><a href="#julia_mpi_gpu_on_piz_daint" class=h
 
 </ol>
 <pre><code class="sh hljs">srun -n4 julia --project &lt;my_mpi_script.jl&gt;</code></pre>
+<p>If you do not want to use an interactive session you can use the <code>sbatch</code> command to launch an MPI job remotely on daint. Example of a <code>sbatch_mpi_daint.sh</code> you can launch &#40;without need of an allocation&#41; as <a href="https://github.com/eth-vaw-glaciology/course-101-0250-00/blob/main/scripts/l8_scripts/l8_sbatch_mpi_daint.sh"><code>sbatch sbatch_mpi_daint.sh</code></a>:</p>
+<pre><code class="sh hljs"><span class=hljs-meta >#!/bin/bash -l</span>
+<span class=hljs-comment >#SBATCH --job-name=&quot;diff2D&quot;</span>
+<span class=hljs-comment >#SBATCH --output=diff2D.%j.o</span>
+<span class=hljs-comment >#SBATCH --error=diff2D.%j.e</span>
+<span class=hljs-comment >#SBATCH --time=00:05:00</span>
+<span class=hljs-comment >#SBATCH --nodes=4</span>
+<span class=hljs-comment >#SBATCH --ntasks-per-node=1</span>
+<span class=hljs-comment >#SBATCH --partition=normal</span>
+<span class=hljs-comment >#SBATCH --constraint=gpu</span>
+<span class=hljs-comment >#SBATCH --account class04</span>
+
+srun -n4 bash -c <span class=hljs-string >&#x27;julia --project &lt;my_julia_mpi_gpu_script.jl&gt;&#x27;</span></code></pre>
+<div class=note ><div class=title >💡 Note</div>
+<div class=messg >The scripts above can be found in the <a href="https://github.com/eth-vaw-glaciology/course-101-0250-00/blob/main/scripts/l8_scripts/">scripts</a> folder.</div></div>
 <h4 id=cuda-aware_mpi_on_piz_daint ><a href="#cuda-aware_mpi_on_piz_daint" class=header-anchor >CUDA-aware MPI on Piz Daint</a></h4>
 <div class=warning ><div class=title >⚠️ Warning&#33;</div>
 <div class=messg >There is currently an issue on the Daint software stack with CuDA-aware MPI. For now, make sure <strong>not to run</strong> with CUDA-aware MPI, i.e., having both <code>MPICH_RDMA_ENABLED_CUDA</code> and <code>IGG_CUDAAWARE_MPI</code> set to 0.</div></div>
-<div class=note ><div class=title >💡 Note</div>
-<div class=messg >The scripts above can be found in the <a href="https://github.com/eth-vaw-glaciology/course-101-0250-00/blob/main/scripts/l8_scripts/">scripts</a> folder.</div></div>
 <div class=page-foot >
   <div class=copyright >
     <a href="https://github.com/eth-vaw-glaciology/course-101-0250-00/"><b>Edit this page on <img class=github-logo  src="https://unpkg.com/ionicons@5.1.2/dist/svg/logo-github.svg"></b></a><br>