Why the unilization ratio of CPU is more than 100% in the running process of nwchem?


Click here for full thread
Clicked A Few Times
Hi, Edoapra
Thank you for your kind helps!
I had set environmental variables ( OMP_NUM_THREADS=1 MKL_NUM_THREADS=1)
in PBS running script.
However, the running cpu cores are still more than the setting CPU cores!
PBS script is presented as follows:
   
  1. /bin/bash
  2. PBS -N nwchem
  3. PBS -l nodes=1:ppn=16
  4. PBS -q new
export MKL_NUM_THREADS=1
export OMP_NUM_THREADS=1

project_name=C6H6.nw
cd $PBS_O_WORKDIR
export LD_LIBRARY_PATH=/public/program/intel/12.1/composer_xe_2011_sp1.11.339/mkl/lib/intel64:/public/program/mpi/openmpi/1.6.2/lib$LD_LIBRARY_PATH
ulimit -s unlimited
NSLOTS=`cat ${PBS_NODEFILE} | wc -l`
NWCHEM_PATH=/public/program/nwchem-6.1-s/bin/LINUX64
  1. OPENMPI_PATH=/public/program/mpi/mpich2-1.5/bin
OPENMPI_PATH=/public/program/mpi/openmpi/1.6.5-intel2013/bin
OUTDIR=/tmp/$USER/$PBS_JOBID.$PBS_JOBNAME

mkdir -p $OUTDIR
cp -rf ${PBS_O_WORKDIR}/* ${OUTDIR}/
cd $OUTDIR

time ${OPENMPI_PATH}/mpirun --mca btl tcp,self --bind-to-core -hostfile $PBS_NODEFILE -np $NSLOTS ${NWCHEM_PATH}/nwchem ${project_name} > ${project_name}.out

cp -rf ${OUTDIR}/* ${PBS_O_WORKDIR}/
rm -rf $OUTDIR





Thank you again!