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
doku:lammps [2014/10/29 12:32] – shdoku:lammps [2014/11/04 09:46] (current) – sh
Line 174: Line 174:
 cp ./lmp_vsc1   /opt/sw/lammps/lmp_vsc1 cp ./lmp_vsc1   /opt/sw/lammps/lmp_vsc1
 </code> </code>
 +
 +NOTE: To get information about included packages
 +<code>
 +make package-status 
 +</code>
 +
  
 ===== LAMMPS on GPUs ====== ===== LAMMPS on GPUs ======
Line 197: Line 203:
 make clean-vsc1 make clean-vsc1
 make vsc1 make vsc1
-cp ./lmp_vsc1   /opt/sw/lammps/lmp_fermi_vsc1+cp ./lmp_vsc1   /opt/sw/lammps/lmp_fermi[kepler]_vsc1
 </code>  </code> 
      
-NOTE: for the fermi nodes use ''CUDA_ARCH = -arch=sm_20'', and for the kepler nodes use ''CUDA_ARCH = -arch=sm_35''+NOTE: for the fermi nodes use ''CUDA_ARCH = -arch=sm_20'', and for the kepler nodes use ''CUDA_ARCH = -arch=sm_35''; so switching from building for 'fermi' to building for 'kepler' requires re-compilation of package ''gpu'', ie 
 +<code> 
 +      cd .../wherever/it/is/lammps-28Jun14/lib/gpu 
 +      cp Makefile.fermi Makefile.kepler 
 +      vi Makefile.kepler   ( set CUDA_ARCH = -arch=sm_35 ) 
 +      make -f Makefile.kepler clean 
 +      make -f Makefile.kepler 
 +      cd ../../src 
 +      ... continue with the above block to build lmp_vsc1 
 +</code>
        
-NOTE2: To get information about packages that will be used+ 
 + 
 + 
 +===== Testing LAMMPS ====== 
 + 
 +==== MPI-Parallel ==== 
 + 
 +=== For example /opt/sw/lammps/examples/peptide/ on VSC-1 === 
 + 
 +1. Change into some temporary directory and prepare the submit script for SGE. 
 + 
 +vi ./sge_lammps_peptide.scrpt
 <code> <code>
-make package-status +#$ -N lammps_peptide 
 +#$ -S /bin/bash 
 +#$ -cwd 
 +#$ -pe mpich 8 
 +#$ -V 
 +  
 +export LD_LIBRARY_PATH=/opt/sw/openkim-api/1.6.3/lib/libkim-api-v1.so:$LD_LIBRARY_PATH 
 +export LD_LIBRARY_PATH=/opt/intel/mkl/lib/intel64:$LD_LIBRARY_PATH 
 +export LD_LIBRARY_PATH=/opt/intel/composer_xe_2013_sp1.3.174/compiler/lib/intel64:$LD_LIBRARY_PATH 
 +export I_MPI_FABRICS=shm:dapl  
 + 
 +mpirun -machinefile $TMPDIR/machines -np $NSLOTS /opt/sw/lammps/lmp_vsc1  -in  ./in.peptide 
 +</code> 
 + 
 +2. Submit it and compare the results - log.lammps - to the reference log files in /opt/sw/lammps/examples/peptide.  
 +<code> 
 +qsub ./sge_lammps_peptide.scrpt 
 +</code> 
 + 
 + 
 +==== GPU fermi ==== 
 + 
 +=== For example /opt/sw/lammps/examples/kokkos/ on VSC-1 === 
 + 
 +1. Change into some temporary directory and prepare the submit script for SGE (fermi queue). Two consecutive short test runs will employ 1 GPU and 2 GPUs per node. 
 + 
 +vi ./sge_lammps_fermi_kokkos.scrpt 
 +<code> 
 +#$ -N lammps_kokkos 
 +#$ -S /bin/bash 
 +#$ -cwd 
 +#$ -pe smp 12 
 +#$ -V 
 +#$ -q fermi  
 +#$ -P fermi 
 + 
 +export LD_LIBRARY_PATH=/opt/sw/openkim-api/1.6.3/lib/libkim-api-v1.so:$LD_LIBRARY_PATH 
 +export LD_LIBRARY_PATH=/opt/intel/mkl/lib/intel64:$LD_LIBRARY_PATH 
 +export LD_LIBRARY_PATH=/opt/intel/composer_xe_2013_sp1.3.174/compiler/lib/intel64:$LD_LIBRARY_PATH 
 +export I_MPI_FABRICS=shm:dapl 
 + 
 +# single GPU 
 +/opt/sw/lammps/lmp_fermi_vsc1 -k on t 6  -sf kk -in ./in.kokkos 
 +mv ./log.lammps ./log.lammps_fermi.1gpu.kokkos 
 + 
 +# 2 x GPU 
 +mpirun -np 2 /opt/sw/lammps/lmp_fermi_vsc1 -k on t 6 -sf kk -in ./in.kokkos 
 +mv ./log.lammps ./log.lammps_fermi.2gpu.kokkos 
 +</code> 
 + 
 + 
 +2. Change into the fermi group and submit it to the appropriate queue. 
 +<code> 
 +sg fermi 
 +qsub.py ./sge_lammps_fermi_kokkos.scrpt 
 +</code> 
 + 
 +3. Compare the results - ./log.lammps_fermi.[1,2]gpu.kokkos - to the reference log files in /opt/sw/lammps/examples/kokkos.  
 + 
 + 
 +==== GPU kepler ==== 
 + 
 +=== For example /opt/sw/lammps/examples/kokkos/ on VSC-1 === 
 + 
 +1. Change into some temporary directory and prepare the submit script for SGE (kepler queue). The test is analogous to the one for fermi, hence two consecutive short runs employing 1 GPU or 2 GPUs per node. 
 + 
 +vi ./sge_lammps_kepler_kokkos.scrpt 
 +<code> 
 +#$ -N lammps_kokkos 
 +#$ -S /bin/bash 
 +#$ -cwd 
 +#$ -pe smp 12 
 +#$ -V 
 +#$ -q kepler  
 +#$ -P fermi 
 + 
 +export LD_LIBRARY_PATH=/opt/sw/openkim-api/1.6.3/lib/libkim-api-v1.so:$LD_LIBRARY_PATH 
 +export LD_LIBRARY_PATH=/opt/intel/mkl/lib/intel64:$LD_LIBRARY_PATH 
 +export LD_LIBRARY_PATH=/opt/intel/composer_xe_2013_sp1.3.174/compiler/lib/intel64:$LD_LIBRARY_PATH 
 +export I_MPI_FABRICS=shm:dapl 
 + 
 +# single GPU 
 +/opt/sw/lammps/lmp_kepler_vsc1 -k on t 6  -sf kk -in ./in.kokkos 
 +mv ./log.lammps ./log.lammps_kepler.1gpu.kokkos 
 + 
 +# 2 x GPU 
 +mpirun -np 2 /opt/sw/lammps/lmp_kepler_vsc1 -k on t 6 -sf kk -in ./in.kokkos 
 +mv ./log.lammps ./log.lammps_kepler.2gpu.kokkos 
 +</code> 
 + 
 + 
 +2. Change into the fermi group and submit it to the appropriate --- kepler --- queue. 
 +<code> 
 +sg fermi 
 +qsub.py ./sge_lammps_kepler_kokkos.scrpt
 </code> </code>
  
 +3. Compare the results - ./log.lammps_kepler.[1,2]gpu.kokkos - to the reference log files in /opt/sw/lammps/examples/kokkos. 
  • doku/lammps.1414585973.txt.gz
  • Last modified: 2014/10/29 12:32
  • by sh