identifier "INT64_MAX" is undefined

Dear all,

I’m trying to build a CUDA version LAMMPS but get the output bellow:

**mpiicc** -O3 -openmp -DLAMMPS_MEMALIGN=64 -no-offload -xHost -fno-alias -ansi-alias -restrict -override-limits -DLAMMPS_GZIP -DLAMMPS_JPEG **-DLAMMPS_LONGLONG_TO_LONG** -I../../lib/cuda -DLMP_USER_CUDA -DLMP_USER_OMP -DMPICH_SKIP_MPICXX -DOMPI_SKIP_MPICXX=1 -I/lustre/utility/intel/impi/ -DFFT_FFTW3 -I/lustre/software/fftw3/fftw-3.3.4_impi/DOUBLE/include -I/lustre/utility/cuda/6.5//include -DUNIX -DFFT_CUFFT -DCUDA_PRECISION=1 -DX_PRECISION=2 -DV_PRECISION=2 -DCUDA_ARCH=20 -c ../neb.cpp
../neb.cpp(205): error: identifier "INT64_MAX" is undefined
if (update->laststep < 0 || update->laststep > MAXBIGINT)

compilation aborted for ../neb.cpp (code 2)

I’ve read that in such case one have to add -DLAMMPS_LONGLONG_TO_LONG but that doesn’t help.
I’m using using in Intel compiler 14.0.2 and impi, on a Intel® Xeon® E5-2670 2x 8 cores machine.
Any hints here?
Thank you.


Someone else reported this in the last day. The next patch
will fix it. You need to add #include “lmptype.h” to the top
of a handful of files b/c of an issue inside OpenMPI. The
lines were removed accidentally.


Thank you Steve,

It works.
I get now an error message and a crash when I run it but I’ll make a new post for this issue.