| 5:14:32 PM PDT - Thu, Aug 23rd 2012 |
|
Hi,
I compiled nwchem 6.1.1 with intel 11.1 and mvapich2. Everything works just fine (passed the tests and runs smoothly) except that one additional core is always used on one of the allocated hosts during the parallel execution. For example, If I request 24 cores on 4 hosts (6 for each), I always find that one of the hosts will use a 7th core if allowed to do so, which will certainly affects the other running jobs. Any idea why this is the case? Is there any means to prevent the use of additional core than requested? Thanks!
best wishes
JT
P.S. here is the setup used for the compilation.
export LARGE_FILES=TRUE
export USE_NOFSCHECK=TRUE
export TCGRSH=/usr/bin/ssh
export USE_MPI=y
export USE_MPIF=y
export USE_MPIF4=y
export IB_HOME=/opt
export IB_INCLUDE=/opt/include/infiniband
export IB_LIB=$IB_HOME/lib64
export IB_LIB_NAME="-L/usr/lib64 -lrdmacm -libumad -libverbs -lpthread -lrt"
export ARMCI_NETWORK=OPENIB
export MSG_COMMS=MPI
export NWCHEM_TOP=/usr/apps/NWCHEM/nwchem-6.1.1-src
export NWCHEM_TARGET=LINUX64
export MPI_LOC=/usr/apps/mpi/intel/mvapich2-1.7
export MPI_LIB=$MPI_LOC/lib
export MPI_INCLUDE=$MPI_LOC/include
export LIBMPI="-lmpichf90 -lmpich -lopa -lmpl -lpthread -lrdmacm -libverbs -libumad -ldl -lrt"
export NWCHEM_MODULES="all python"
export HAS_BLAS=y
export MKL64=/usr/apps/intel-11.1/fortran/mkl/lib/em64t
export BLASOPT="-L${MKL64} -Wl,--start-group -lmkl_intel_ilp64 -lmkl_sequential -lmkl_core -Wl,--end-group"
export PYTHONHOME=/nas02/apps/python-2.6.5
export PYTHONVERSION=2.6
|