DokuWiki

User Tools

Site Tools

Trace:
cluster:172

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision 		Previous revision
Next revision Both sides next revision
cluster:172 [2018/08/22 19:39]
hmeij07 [Amber] 		cluster:172 [2018/08/23 18:16]
hmeij07 [mapd]
Line 20: Line 20:
   * unmount NFS mounts while installing nvidia as root   * unmount NFS mounts while installing nvidia as root
   * install other software as regular user    * install other software as regular user 
 +  * 
 ==== Nvidia ==== ==== Nvidia ====
  
Line 86: Line 87:
   * export PATH=/usr/local/cuda/bin:$PATH   * export PATH=/usr/local/cuda/bin:$PATH
   * export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH   * export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH
 +  * export CUDA_HOME=/usr/local/cuda
  
 **Verification** **Verification**
Line 227: Line 229:
 # please be sure to verify any results against known outcomes # please be sure to verify any results against known outcomes
 export CUDA_HOME=/usr/local/cuda export CUDA_HOME=/usr/local/cuda
 +
 ./configure -cuda gnu | tee -a  amber16-install.log 2>&1  ./configure -cuda gnu | tee -a  amber16-install.log 2>&1 
 source /usr/local/amber16/amber.sh source /usr/local/amber16/amber.sh
 make install | tee -a amber16-install.log 2>&1 make install | tee -a amber16-install.log 2>&1
 Installation of pmemd.cuda complete Installation of pmemd.cuda complete
 +
 + ./configure -mpi -cuda gnu | tee -a  amber16-install.log 2>&1
 +source /usr/local/amber16/amber.sh
 +make install | tee -a amber16-install.log 2>&1
 +Installation of pmemd.cuda.MPI complete
 +
 +[hmeij@n37 amber16]$ ls -l bin/pmemd*
 +-rwxr-xr-x 1 hmeij its  3097968 Aug 22 10:12 bin/pmemd
 +lrwxrwxrwx 1 hmeij its       15 Aug 22 15:19 bin/pmemd.cuda -> \
 +pmemd.cuda_SPFP
 +-rwxr-xr-x 1 hmeij its 38851928 Aug 22 15:25 bin/pmemd.cuda_DPFP
 +-rwxr-xr-x 1 hmeij its 39436704 Aug 22 16:04 bin/pmemd.cuda_DPFP.MPI
 +lrwxrwxrwx 1 hmeij its       19 Aug 22 15:57 bin/pmemd.cuda.MPI -> \
 +pmemd.cuda_SPFP.MPI
 +-rwxr-xr-x 1 hmeij its 32950848 Aug 22 15:19 bin/pmemd.cuda_SPFP
 +-rwxr-xr-x 1 hmeij its 33531456 Aug 22 15:57 bin/pmemd.cuda_SPFP.MPI
 +-rwxr-xr-x 1 hmeij its 33405504 Aug 22 15:31 bin/pmemd.cuda_SPXP
 +-rwxr-xr-x 1 hmeij its 33990208 Aug 22 16:10 bin/pmemd.cuda_SPXP.MPI
 +-rwxr-xr-x 1 hmeij its  3647784 Aug 22 10:36 bin/pmemd.MPI
  
 </code> </code>
 +
 +**Tests**
 +
 +Although the 9.2 cuda compiled Amber passed all tests please double check your results.
 +
 +<code>
 +export DO_PARALLEL="mpirun -np 8"
 +make test >> amber16-test.log 2>&1
 +</code>
 +
 +**Finish**  
 +  * [root@n37 /]# tar -cvf /tmp/n37.chroot.ul.tar usr/local
 +  * [root@n37 /]# scp /tmp/n37.chroot.ul.tar sms_server:/var/chroots/goldimages/
 +
 +
 +==== Gromacs ====
 +
 +As root install
 +
 +  * cmake, latest version, never understand why so far ahead of distro...
 +
 +Download and extract source. Using same environment as Amber compilation.
 +
 +<code>
 +
 + cd gromacs-2018/
 + mkdir build
 + cd build
 +
 + which mpicc mpicxx
 +/share/apps/CENTOS6/openmpi/1.8.4/bin/mpicc
 +/share/apps/CENTOS6/openmpi/1.8.4/bin/mpicxx
 +
 + CC=mpicc CXX=mpicxx \
 +   /share/apps/CENTOS7/cmake/3.12.1/bin/cmake .. \
 +  -DCMAKE_INSTALL_PREFIX=/usr/local/gromacs-2018 \
 +  -DGMX_BUILD_OWN_FFTW=ON -DGMX_MPI=ON -DGMX_GPU=ON
 + CC=mpicc CXX=mpicxx make
 + CC=mpicc CXX=mpicxx make install
 +
 +</code>
 +==== Lammps ====
 +
 +As root install
 +
 +  * yum install libjpeg libjpeg-devel libjpeg-turbo libjpeg-turbo-devel 
 +  * yum install blas blas-devel lapack lapack-devel boost boost-devel
 +
 +For Lammps-22Aug18 I followed the top installation instructions at this page
 +
 +  * [[cluster:161|lammps-11Aug17]]
 +
 +The only difference in my approach was
 +
 +  * to stay with openmpi-1.8.4 (not mpich3...)
 +  * consulting the ARCH web page I choose -arch=sm_35
 +
 +Good thing we're doing this now, future versions of CUDA will not support the K20s anymore. In fact on that web site they are not mentioned, only the K40/K80 gpus. So we'll see what testing reveals.  Please double check results against previous runs. Compile as regular user and stage lmp_mpi in /usr/local/lammps-22Aug10/
 +
 +<code>
 +
 +[hmeij@n37 src]$ ll /usr/local/lammps-22Aug18/
 +total 104356
 +-rwxr-xr-x 1 hmeij its 35739800 Aug 23 08:49 lmp_mpi-double-double-with-cuda
 +-rwxr-xr-x 1 hmeij its 35555672 Aug 23 09:11 lmp_mpi-single-double-with-cuda
 +-rwxr-xr-x 1 hmeij its 35559552 Aug 23 09:53 lmp_mpi-single-single-with-cuda
 +
 +</code>
 +
 +==== mapd ====
 +
 +  * https://www.mapd.com/docs/latest/4_centos7-yum-gpu-ce-recipe.html
 +
 +<code>
 +
 +useradd -U mapd
 +
 +# mapd.repo
 +[mapd-ce-cuda]
 +name=mapd ce - cuda
 +baseurl=https://releases.mapd.com/ce/yum/stable/cuda
 +gpgcheck=1
 +gpgkey=https://releases.mapd.com/GPG-KEY-mapd
 +
 +yum  install \
 +  copy-jdk-configs java-1.8.0-openjdk-headless \
 +  javapackages-tools libxslt \
 +  lksctp-tools python-javapackages \
 +  python-lxml tzdata-java \
 +  mapd 
 +  # n37:/usr/local/src
 +
 +# User specific aliases and functions
 +export MAPD_USER=mapd
 +export MAPD_GROUP=mapd
 +export MAPD_STORAGE=/var/lib/mapd
 +export MAPD_PATH=/opt/mapd
 +# The $MAPD_STORAGE directory must be dedicated to MapD
 +
 +</code>
 +
 +
 \\ \\
 **[[cluster:0|Back]]** **[[cluster:0|Back]]**
cluster/172.txt ยท Last modified: 2020/07/15 17:52 by hmeij07

Page Tools

Except where otherwise noted, content on this wiki is licensed under the following license: CC Attribution-Share Alike 4.0 International
CC Attribution-Share Alike 4.0 International Donate Powered by PHP Valid HTML5 Valid CSS Driven by DokuWiki