<div dir="ltr"><div><br></div><div>I have an issue with the "libint" library and please see the attached file. <br></div><div><br></div><div>Best,</div><div>Esam<br></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Tue, Jun 22, 2021 at 4:22 PM Esam Elkandro <<a href="mailto:elka...@gmail.com">elka...@gmail.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr"><div>Thank you so much Jeff.<br></div><div><br></div><div>I follow your suggestion about "
conda install -c conda-forge cp2k=8.2.0=py38_openmpi_1" and the installation was successful with "cp2k.psmp". <br></div><div>After installation, I got this message "
mpiexec --mca opal_cuda_support 1" what does this mean?<br></div><div><br></div><div><b>Here is my job file: </b><br></div><div></div><div><br></div><div>#!/bin/bash<br>#SBATCH -J MPI<br>#SBATCH --nodes=2<br>#SBATCH --ntasks=64<br>#SBATCH --time=0-24:00:00<br>#SBATCH --export=all <br>#SBATCH --out=Foundry-%j.out <br><br><br>module load openmpi/4.0.3/intel/2020.0<br>module load anaconda/2020.7<br>source activate cp2k<br>####mpiexec --mca opal_cuda_support -np 128 cp2k.psmp -i NVE_md.inp -o md4.out<br></div><div>
mpiexec -np 128 cp2k.psmp -i NVE_md.inp -o md4.out <br></div><div><br></div><div><br></div><div>##### After installation ####</div><div><br></div><div>
"Preparing transaction: done<br>Verifying transaction: done<br>Executing transaction: |<br>For Linux 64, Open MPI is built with CUDA awareness but this support is disabled by default.<br>To enable it, please set the environment variable OMPI_MCA_opal_cuda_support=true before<br>launching your MPI processes. Equivalently, you can set the MCA parameter in the command line:<br>mpiexec --mca opal_cuda_support 1 ...<br><br>In addition, the UCX support is also built but disabled by default.<br>To enable it, first install UCX (conda install -c conda-forge ucx). Then, set the environment<br>variables OMPI_MCA_pml="ucx" OMPI_MCA_osc="ucx" before launching your MPI processes.<br>Equivalently, you can set the MCA parameters in the command line:<br>mpiexec --mca pml ucx --mca osc ucx ...<br>Note that you might also need to set UCX_MEMTYPE_CACHE=n for CUDA awareness via UCX.<br>Please consult UCX's documentation for detail."
</div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Tue, Jun 22, 2021 at 3:24 PM Jeff Terstriep <<a href="mailto:jeff.t...@gmail.com" target="_blank">jeff.t...@gmail.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">You only want to use the 'conda install -c conda-forge cp2k' command. The second conda install using the label installs the much older cp2k 6.1.<div><br></div><div>That said, cp2k installs the single-node version. If you want cp2k.psmp using MPI, install with 'conda install -c conda-forge cp2k=8.2.0=py38_openmpi_1'. Search <a href="http://anaconda.org" target="_blank">anaconda.org</a> for cp2k and look under the files tab to see the latest version and build data.</div><div><br></div><div>Jeff<br><br></div><div class="gmail_quote"><div dir="auto" class="gmail_attr">On Tuesday, June 15, 2021 at 10:52:34 AM UTC-5 <a href="mailto:el...@gmail.com" target="_blank">el...@gmail.com</a> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><br><div>Dear cp2k community,</div><div><br></div><div>I installed the cp2k under our three clusters and every time I don't get executable file "cp2k.psmp" to use it. Even though, I run my calculations by using "cp2k.popt", however, these calculations is very slow and I don't see any difference if I use multi node comparing single node. How to install the executable file "psmp"? <br></div><div><br></div><div>Hint: Our cluster has more than 50 nodes and each node has 64 CPU. <br></div><div>Please check my installation steps in two different clusters: <br></div><div><br></div><div>#### 1st way ########<br></div><div>conda create --name cp2k<br>conda activate cp2k<br>conda install -c conda-forge cp2k<br>conda install -c conda-forge/label/cf202003 cp2k<br>conda update -n base -c defaults conda<br></div><div><br></div><div>
#### 2nd way ######## <br></div><div><br></div><div>source /opt/intel/bin/ifortvars.sh intel64<br> <br>wget <a href="https://github.com/cp2k/cp2k/releases/download/v5.1.0/cp2k-5.1.tar.bz2" rel="nofollow" target="_blank">https://github.com/cp2k/cp2k/releases/download/v5.1.0/cp2k-5.1.tar.bz2</a><br>tar xvf cp2k-5.1.tar.bz2<br>cd cp2k-5.1/tools/toolchain/<br>./install_cp2k_toolchain.sh --install-all --with-mkl=system</div><div><br>cd ~<br>cp /root/cp2k-5.1/tools/toolchain/install/arch/* cp2k-5.1/arch/<br>source /root/cp2k-5.1/tools/toolchain/install/setup<br> <br>cd cp2k-5.1<br>make -j 28 ARCH=local VERSION="sopt sdbg ssmp popt pdbg "<br><br></div><div><br></div><div>Best,</div><div>Esam<br></div></blockquote></div>
<p></p>
-- <br>
You received this message because you are subscribed to a topic in the Google Groups "cp2k" group.<br>
To unsubscribe from this topic, visit <a href="https://groups.google.com/d/topic/cp2k/t3Yp00Pt3yg/unsubscribe" target="_blank">https://groups.google.com/d/topic/cp2k/t3Yp00Pt3yg/unsubscribe</a>.<br>
To unsubscribe from this group and all its topics, send an email to <a href="mailto:cp...@googlegroups.com" target="_blank">cp...@googlegroups.com</a>.<br>
To view this discussion on the web visit <a href="https://groups.google.com/d/msgid/cp2k/3609ffa3-80f1-45db-a0dc-721bd8d1e3a8n%40googlegroups.com?utm_medium=email&utm_source=footer" target="_blank">https://groups.google.com/d/msgid/cp2k/3609ffa3-80f1-45db-a0dc-721bd8d1e3a8n%40googlegroups.com</a>.<br>
</blockquote></div>
</blockquote></div>