I want to simulate the thermal degradation of a polymer and have several problems:
I wrote an input script (please find it in the attached), using pair-style reax/c which requires atom-style charge.
Do I have to assign 0 for neutral atoms or do I have to assign the Z-number of the atom?,
from the manual I would say 0...
When I run my script without the fix deposit, I lose my atoms after one or two timesteps (although having periodic boundary conditions) but get no error.
Where does that comes from?, is it somehow possible to visualize the bond formation of reax/c?
When I run my script with the fix deposit I get the error message: Too many ghost atoms what should not occur by using 78 atoms in total and running lammps in serial...
The “Too many ghost atoms” is an issue with combining fix deposit and pair_style reaxc from USER-REAXC. The version of pair_style reaxc in the KOKKOS package solves this problem. Regarding your other questions, I suggest you start out by studying and experimenting with the LAMMPS ReaxFF examples provided in examples/reax/*.
thanks for your reply! I now built LAMMPS with the KOKKOS package and changed pair style to reax/c/kk, but when I run lmp_serial … I get Error: Unknown Pair-Style.
When I made LAMMPS I used make yes-kokkos and it installed successfully, should I have installed other packages, too?
I rebuilt lammps and now reax/c/kk is listed, thank you for that.
When I now use lmp_serial -sf kk … it throws the error:
terminate called after throwing an instance of ‘std::runtime_error’
what(): Constructing View and initializing data with uninitialized execution space
Traceback functionality not available