I have a job that runs fine without VTUNE, but when I follow these instructions on the wiki:

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

To use memory bandwidth analysis, execute the following command via Cobalt submission script
$ source /soft/compilers/intel/vtune_amplifier_xe/amplxe-vars.sh
$ amplxe-cl -c memory-access -knob analyze-mem-objects=true -knob sampling-interval=10 -data-limit=0 -d 60 -- mpirun -f $COBALT_NODEFILE -env KMP_AFFINITY=balanced,granularity=fine numactl -m 1 ./YourExecutable

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

I add this to my submission script and my job fails during execution.

=================================

#!/bin/bash

RANKS_PER_NODE=$1

NODES=$2

echo RANKS_PER_NODE=$RANKS_PER_NODE

echo NODES=$NODES

source /soft/compilers/intel/compilers_and_libraries/linux/mpi/bin64/mpivars.sh

source /soft/compilers/intel/vtune_amplifier_xe/amplxe-vars.sh

echo contents of "/var/run/hwloc/knl_memoryside_cache" file

cat /var/run/hwloc/knl_memoryside_cache

echo contents of Run.dat

cat Run.dat

BASE=/home/jchilders/sherpa

SHERPA=$BASE/rel-2-2-0_anl/install/bin/Sherpa

echo Starting Sherpa $SECONDS

amplxe-cl -c memory-access -knob analyze-mem-objects=true -knob sampling-interval=10 -data-limit=0 -d 60 -- mpirun -f $COBALT_NODEFILE -ppn $(($NODES*$RANKS_PER_NODE)) -env \KMP_AFFINITY=balanced,granularity=fine numactl -m 0 $SHERPA PSI_ITMIN_BY_NODE=5 PSI_ITMAX_BY_NODE=5 PSI_ITMIN=5 PSI_ITMAX=5 FINISH_OPTIMIZATION=Off MPI_SEED_MODE=1

echo Sherpa completed with exit code $? in $SECONDS seconds

=================================

Does someone see if I have misconfigured VTUNE?

J. Taylor Childers, Assistant Physicist
Argonne National Laboratory
HEP Division Bldg. 360, Rm. C-217
Office: (630) 252-4212 Mobile: (331) 302-4647
jchilders@anl.gov