Differences

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

--- doku:lammps [2014/10/29 13:58] – sh
+++ doku:lammps [2014/11/04 09:36] – sh
@@ Line 200: / Line 200: @@
 </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
+</code>
-NOTE2: To get information about packages that will be used
+NOTE_2: To get information about included packages
 <code>
 make package-status
@@ Line 242: / Line 249: @@
 === For example /opt/sw/lammps/examples/kokkos/ on VSC-1 ===
-. Change into some temporary directory and prepare the submit script for SGE (fermi queue) using 1 GPU or 2 GPUs per node.
+. 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
@@ Line 261: / Line 268: @@
 # single GPU
 /opt/sw/lammps/lmp_fermi_vsc1 -k on t 6  -sf kk -in ./in.kokkos
-mv ./log.lammps ./log.lammps_fermi.1_gpu.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.2_gpu.kokkos
+mv ./log.lammps ./log.lammps_fermi.2gpu.kokkos
+</code>
+. Change into the fermi group and submit it to the appropriate queue.
+<code>
+sg fermi
+qsub.py ./sge_lammps_fermi_kokkos.scrpt
 </code>
+. Compare the results - ./log.lammps_fermi.[1,2]gpu.kokkos - to the reference log files in /opt/sw/lammps/examples/kokkos.