Hi,
I compiled nwchem6.6 on OpenSuse with MKL mpich and gcc/gfortran the following settings:
export NWCHEM_TOP=/programme/nwchem-6.6
export NWCHEM_TARGET=LINUX64
export NWCHEM_MODULES="all"
export NWCHEM_LONG_PATHS=Y
export USE_NOFSCHECK=Y
export FC=gfortran
export CC=gcc
export USE_MPI=y
export USE_MPIF=y
export LIBMPI="-lmpich -lfmpich -L/programme/mpich-3.2-intel64-bin/lib64"
export HAS_BLAS=yes
export MPI_LOC=/programme/mpich-3.2-intel64-bin
export MPI_LIB=/programme/mpich-3.2-intel64-bin/lib64
export MPI_INCLUDE=$MPI_LOC/include
export MKLROOT=/opt/intel/mkl/
export MKLLIB=/opt/intel/mkl/lib/intel64
export MKLINC=/opt/intel/mkl/include
export BLASOPT="-L$MKLLIB -lmkl_gf_ilp64 -lmkl_core -lmkl_sequential -lpthread -lm"
export LAPACK_LIB="-L$MKLLIB -lmkl_gf_ilp64 -lmkl_core -lmkl_sequential -lpthread -lm"
export BLAS_LIB="-L$MKLLIB -lmkl_gf_ilp64 -lmkl_core -lmkl_sequential -lpthread -lm"
but I get the following error when I try to do some ccsd(t) calculations:
ccsd_t: MA error dbl
I recompiled without the Intel-MKL but got the same result.
So I think it might be a problem with the compiler?
The version on the machine is:
gcc version 6.2.1 20160830 [gcc-6-branch revision 239856] (SUSE Linux)
Any idea how to fix the problem?
Thank You
|