software:compilation_scripts_are_here_available

The AMD Bulldozer optimized code was compiled using the following script

OPEN64
module add open64/4.5.2
module add mpi/openmpi/1.6.5/open64_4.5.2
module add acml/5.3.1/open64_4.5.2/open64_64_fma4
 
 
export NWCHEM_TOP="$(dirname $PWD)"
 
export LARGE_FILES="TRUE"
export NWCHEM_TARGET="LINUX64"
export ARMCI_NETWORK="MELLANOX"
export IB_HOME="/usr"
export IB_INCLUDE="/usr/include/infiniband"
export IB_LIB="/usr/lib64"
export MSG_COMMS="MPI"
export ENABLE_COMPONENT="yes"
export USE_MPI="y"
export USE_MPIF="y"
export USE_MPIF4="y"
 
export LIBMPI="-lmpi_f90 -lmpi_f77 -lmpi -ldl -Wl,--export-dynamic -lnsl -lutil"
export F77="mpif77"
export FC="mpif77"
export F90="mpif90"
export MPI_LOC="$MPI_HOME"
export MPI_LIB="$MPI_LOC/lib"
export MPI_INCLUDE="$MPI_LOC/include"
 
export ACMLROOT="/cluster/acml/5.3.1/open64_64_fma4"
export USE_64TO32=y
export HAS_BLAS=yes
export BLASOPT="-m64 -L$ACMLROOT/lib -lacml -lgfortran -mp"
export CPATH=$CPATH:$ACMLROOT/include
 
export LIB_DEFINES="-DDFLT_TOT_MEM=2147483648"
export NWCHEM_MODULES="all"
export CC="mpicc"
export FOPTIMIZE="-O3 -march=bdver1 -mtune=native -msse -msse2 -msse3 -mssse3 -msse4 -msse4.1 -msse4.2 -LNO:prefetch -LNO:fission=n -LNO:fusion=n -fp-accuracy -ipa -mp"
export COPTIMIZE="-O3 -march=bdver1 -mtune=native -msse -msse2 -msse3 -mssse3 -msse4 -msse4.1 -msse4.2 -LNO:prefetch  -LNO:fission=n -LNO:fusion=n -fp-accuracy -ipa -mp"
export MRCC_THEORY="TRUE"
 
 
make clean
make nwchem_config
make 64_to_32
make
GCC 4.8.1
module add gcc/4.8.1
module add mpi/openmpi/1.6.5/gcc_4.8.1
module add acml/5.3.1/gfortran_int64/gfortran64_fma4_int64
 
export NWCHEM_TOP="$(dirname $PWD)"
echo $NWCHEM_TOP
export LARGE_FILES="TRUE"
export NWCHEM_TARGET="LINUX64"
export ARMCI_NETWORK="MELLANOX"
export IB_HOME="/usr"
export IB_INCLUDE="/usr/include/infiniband"
export IB_LIB="/usr/lib64"
export MSG_COMMS="MPI"
export ENABLE_COMPONENT="yes"
export USE_MPI="y"
export USE_MPIF="y"
export USE_MPIF4="y"
 
export LIBMPI="-lmpi_f90 -lmpi_f77 -lmpi -ldl -Wl,--export-dynamic -lnsl -lutil"
export F77="mpif77"
export FC="mpif77"
export F90="mpif90"
export MPI_LOC="$MPI_HOME"
export MPI_LIB="$MPI_LOC/lib"
export MPI_INCLUDE="$MPI_LOC/include"
 
export ACMLROOT="/cluster/acml/5.3.1/gfortran64_fma4_int64"
export HAS_BLAS=yes
export BLASOPT="-m64 -L$ACMLROOT/lib -lacml -lgfortran -fopenmp"
export CPATH=$CPATH:$ACMLROOT/include
 
export LIB_DEFINES="-DDFLT_TOT_MEM=2147483648"
export NWCHEM_MODULES="all"
export CC="mpicc"
export FOPTIMIZE="-O3 -march=bdver1 -mtune=bdver1 -mprefer-avx128 -funroll-all-loops -fprefetch-loop-arrays -ftree-vectorize -mieee-fp"
export COPTIMIZE="-O3 -march=bdver1 -mtune=bdver1 -mprefer-avx128 -funroll-all-loops -fprefetch-loop-arrays -ftree-vectorize -mieee-fp"
export MRCC_THEORY="TRUE"
 
make clean
make nwchem_config
make

The INTEL AVX optimized, MKL linked code was compiled using the following script

module add intel/composer/2013_4_183
module add intel/mkl/2013_4_183
module add mpi/intelmpi/4.0.3.008
 
export NWCHEM_TOP="$(dirname $PWD)"
export LARGE_FILES="TRUE"
export NWCHEM_TARGET="LINUX64"
export ENABLE_COMPONENT="yes"
export ARMCI_NETWORK="OPENIB"
export MSG_COMMS="MPI"
export LIB_DEFINES="-DDFLT_TOT_MEM=2147483648"
export NWCHEM_MODULES="all"
 
export USE_MPI="y"
export USE_MPIF="y"
export USE_MPIF4="y"
 
export LIBMPI="-lmpigf -lmpigi -lmpi_ilp64 -lmpi"
export FC="ifort"
export MPI_LOC="$MPI_HOME"
export MPI_LIB="$MPI_LOC/lib64"
export MPI_INCLUDE="$MPI_LOC/include64"
 
export BLASOPT="-L/cluster/intel/composer_xe_2013.4.183/mkl/lib/intel64 -lmkl_scalapack_ilp64 -lmkl_intel_ilp64 -lmkl_sequential -lmkl_core -lmkl_blacs_intelmpi_ilp64 -lpthread -lm"
export BLAS_SIZE=8
 
export FOPTIMIZE="-O3 -mAVX -xSSE2,SSE3,SSSE3,SSE4.1,SSE4.2,AVX -no-prec-div -funroll-loops -unroll-aggressive"
export COPTIMIZE="-O3 -mAVX -xSSE2,SSE3,SSSE3,SSE4.1,SSE4.2,AVX -no-prec-div -funroll-loops"
 
make clean
make nwchem_config
make

The INTEL AVX optimized, ACML linked code was compiled using the following script

module add intel/composer/2013_4_183
module add acml/5.3.1/intel/ifortran_fma4_int64
module add mpi/intelmpi/4.0.3.008
 
 
export NWCHEM_TOP="$(dirname $PWD)"
echo $NWCHEM_TOP
export LARGE_FILES="TRUE"
export NWCHEM_TARGET="LINUX64"
export ARMCI_NETWORK="OPENIB"
export MSG_COMMS="MPI"
export ENABLE_COMPONENT="yes"
export USE_MPI="y"
export USE_MPIF="y"
export USE_MPIF4="y"
 
export LIBMPI="-lmpigf -lmpigi -lmpi_ilp64 -lmpi"
export F77="mpif77"
export FC="ifort"
export F90="mpif90"
export MPI_LOC="$MPI_HOME"
export MPI_LIB="$MPI_LOC/lib64"
export MPI_INCLUDE="$MPI_LOC/include64"
 
export ACMLROOT="/cluster/acml/5.3.1/ifort64_fma4_int64"
export BLASOPT="-m64 -L$ACMLROOT/lib -lacml"
export CPATH=$CPATH:$ACMLROOT/include
export HAS_BLAS=yes
 
export LIB_DEFINES="-DDFLT_TOT_MEM=2147483648"
export NWCHEM_MODULES="all"
export CC="mpicc"
export FOPTIMIZE="-O3 -mAVX -xSSE2,SSE3,SSSE3,SSE4.1,SSE4.2,AVX -no-prec-div -funroll-loops -multiple-processes=8 -unroll-aggressive"
export COPTIMIZE="-O3 -mAVX -xSSE2,SSE3,SSSE3,SSE4.1,SSE4.2,AVX -no-prec-div -funroll-loops"
export MRCC_THEORY="TRUE"
 
make clean
make nwchem_config
make
  • software/compilation_scripts_are_here_available.txt
  • Last modified: 2013/08/01 15:08
  • by noskov