[Pw_forum] error loading shared libraries on parallel execution

sreekar guddeti colonel.sreekar at gmail.com
Sun Sep 6 01:20:45 CEST 2009


i know this issuehas been addressed and documented in troubleshooting
section of the users guide.
but i giveup in despair trying for a whole day to figure this problem
i run my jobs on rocks cluster by using SGE's facility of submitting batch
jobs
http://www.rocksclusters.org/roll-documentation/sge/5.2/submitting-batch-jobs.html

what i 'apparently' observe(or doubtfully infer) is that i can successfully
run a single parallel job, but on submitting a second job i get the error
____________________________________________
/home3/colonel/espresso-4.0.5/bin/pw.x: error while loading shared
libraries: libgfortran.so.3: cannot open shared object file: No such
file or directory
____________________________________________


i find out the path for the library and added to the LD_LIBRARY_PATH by
writing
_______________________________________
#set the library path to include gfortran libraries
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/lib64:/usr/lib


$ECHO
$ECHO "$LD_LIBRARY_PATH"
$ECHO
_______________________________________

in the file which acts as script for qsub ($qsub -pe orte 4 dosroutine.qsub)
which is

dosroutine.qsub
-----------------------------------------------------------------

#!/bin/bash
#
#$ -cwd
#$ -j y
#$ -S /bin/bash
#

#extract the info about no of processors involved from command line
arguments of 'qsub'
PROCESSORS=$NSLOTS

#heuristically assign the no of processors per pool NPR
NPR=4
#as a result no of pools are give by
NPK=`expr $PROCESSORS / $NPR`

#!/bin/bash
#
#
#Script for performing a dos calculation on a parallel processor
WORKINGDIR=`pwd`
ECHO="echo"

#set the library path to include gfortran libraries
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/lib64:/usr/lib


$ECHO
$ECHO "$LD_LIBRARY_PATH"
$ECHO

# set the needed environment variables

PREFIX=`cd /home3/colonel/espresso-4.0.5 ; pwd`
$ECHO $PREFIX
BIN_DIR=$PREFIX/bin
PSEUDO_DIR=$PREFIX/pseudo
TMP_DIR=$HOME/tmp
PARA_PREFIX="/opt/openmpi/bin/mpirun -np $PROCESSORS"
PARA_POSTFIX="-npool $NPK"

# required executables and pseudopotentials
$ECHO
$ECHO "  executables directory: $BIN_DIR"
$ECHO "  pseudo directory:      $PSEUDO_DIR"
$ECHO "  temporary directory:   $TMP_DIR"

#create results directory
for DIR in "$TMP_DIR" "$WORKINGDIR/results" ; do
    if test ! -d $DIR ; then
        mkdir $DIR
    fi
done
cd $WORKINGDIR/results


# variables to represent programs
PW_COMMAND="$PARA_PREFIX $BIN_DIR/pw.x $PARA_POSTFIX"
DOS_COMMAND="$PARA_PREFIX $BIN_DIR/dos.x $PARA_POSTFIX"
PROJWFC_COMMAND="$PARA_PREFIX $BIN_DIR/projwfc.x $PARA_POSTFIX"


# DOS calculation for 0Ni0
cat > 0ni0.dos.in << EOF
 &control
    calculation='nscf'
    restart_mode='from_scratch',
    prefix='0ni0',
    pseudo_dir = '$PSEUDO_DIR/',
    outdir='$TMP_DIR/'
 /
 &system
    ibrav=2, celldm(1) =6.48, nat=1, ntyp=1,
    nspin = 2,  starting_magnetization(1)=0.7,
    ecutwfc = 24.0, ecutrho = 288.0, nbnd=8,
    occupations='tetrahedra'
 /
 &electrons
    conv_thr = 1.0e-10
    mixing_beta = 0.7
 /
ATOMIC_SPECIES
 Ni 58.69 NiUS.RRKJ3.UPF
ATOMIC_POSITIONS
 Ni 0.0 0.0 0.0
K_POINTS {automatic}        !special points generated by tetrahedra method
 12 12 12 0 0 0
EOF

$ECHO "  running DOS calculation for 0Ni0 ...\c"
$PW_COMMAND < 0ni0.dos.in > 0ni0.dos.out
$ECHO
$ECHO "   done"

-------------------------------------------------------------

the output i get is

************************************************
:/usr/lib64:/usr/lib

/home3/colonel/espresso-4.0.5

  executables directory: /home3/colonel/espresso-4.0.5/bin
  pseudo directory:     /home3/colonel/espresso-4.0.5/pseudo
  temporary directory:   /home3/colonel/tmp
  running DOS calculation for 0Ni0 ...\c
/home3/colonel/espresso-4.0.5/bin/pw.x: error while loading shared
libraries: libgfortran.so.3: cannot open shared object file: No such
file or directory
/home3/colonel/espresso-4.0.5/bin/pw.x: error while loading shared
libraries: libgfortran.so.3: cannot open shared object file: No such
file or directory
/home3/colonel/espresso-4.0.5/bin/pw.x: error while loading shared
libraries: libgfortran.so.3: cannot open shared object file: No such
file or directory
/home3/colonel/espresso-4.0.5/bin/pw.x: error while loading shared
libraries: libgfortran.so.3: cannot open shared object file: No such
file or directory

   done
*************************************************
plz suggest solutions which donot require root permissions , as i dont have
thanks in advance
-- 
Sreekar Guddeti
Department of Physics
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.democritos.it/pipermail/pw_forum/attachments/20090906/2e064769/attachment-0001.htm 


More information about the Pw_forum mailing list