[lammps-users] Query on array neighbor


I am trying to use negative velocity ramping method to reduce the
volume of the system. There are two molecules which are separated by
finite distance. When I am trying squeeze them together; I am getting
following errors.

ERROR on proc 0: Failed to allocate -575198064 bytes for array neighbor:binhead

if I change the option of inner and outer distance in my input script

pair_style lj/charmm/coul/charmm 20.00 20.10

the bytes value changes.

It seems like once the chain is coming closer to the another chain, it
starts feeling the presence of the other chain but neightbor array is
not getting changed.

Could anybody has any suggestion to solve the problem?


The alloc error could be due to the simulation box being
very large. The neighbor routines have to fill the box
with bins, so if it is huge they run out of memory. A large
box could happen with shrink-wrap boundary conditions
(non-periodic) if you blew the molecules apart from
each other b/c you pushed them too close. Check what the
coords of the atoms/molecules are immediately before
the crash ...



I'm trying to simulate surface premelting using lammps. When I heat a

Created box = (-81.6 -408 -81.6) to (81.6 102 81.6)

to the bulk melting point of the material there should be surface premelting
(large disorder of surface atoms in comparison to the rest of the lattice).
However, I find that the rmsd on the y-axis is very high (larger than the
lattice constant for a very small temperature increase). I'm calculating the
rmsd by taking the square root of the msd obtained with fix msd. I notice
that these large oscillations in the y-axis (which are always larger than
the ones in the x- and z-axis) decrease when I decrease the size of the box
on the y-axis. I'm using shrink wrapping on the y-axis for the boundary

I was wondering why this is happening?

Jaime Sanchez

Hello everyone,

When using the boundary condition s I notice that there are large
oscillations of the lattice on the axis perpendicular to the shrink wrapped
boundary (this also happens for the boundary condition f). These
oscillations look like the lattice on that axis behaves as a spring in a
compression/tension cycle. Why are these oscillations happening? When
getting the msd (or rmsd) info from the axis perpendicular to the boundary,
say the y-axis, these oscillations are evident in that the atoms oscillate
in that axis to distances larger than the lattice constant (even for low
temperatures). Also, it seems that these oscillations are increased as the
size of the domain increases on that axis.

I'm setting fix nve and ramping the temperature of the domain using fix
temp/rescale up to the melting point. Any suggestions?

Jaime Sanchez

I would try to visualize your system and see if you
can tell what is happening in the y-direction and if
it looks odd.