[lammps-users] pair_gromacs.error?

Hi LAMMPS developers,

There may be an error in pair_gromacs. I advise you check it. Because I push a rigid to a bilayer and use compute_group/group command to calculate the interactions of rigid and bilayer. There exists a strange resultes of energy and force.

  1. I do not know the c_fff[0] is right, because manual points that " If c_ID[N] is used, then N must be an index from 1-M where M is the length of the vector calculated by the compute." , But if we use c_fff[1]…C_fff[4], It give an error " the c_ID is not that long",.

2.The energy is decreased but all the forces of x/y/z is positive when pushing two group closely along z-axis. I mean if the forces is positive, the energy will increase.

  1. Especially, The result of our simple simulation using LAMMPS is counterintuitive.

compute fff bilayer group/group rigid
thermo_style custom step temp etotal press f_rigd[3] c_fff[0] c_fff[1] c_fff[2] c_fff[3]

c_fff[0] is just the scalar calculation, i.e. the energy in this case.
There is no c_fff[4], hence the error.

There isn't enough info in your email to suspect a problem
with pair gromacs. Read the doc page for compute group/group
to make sure you have the sign convention correct for force.