[lammps-users] Is NVE always NVE in LAMMPS?

I was thinking of trying out a simulation where I can
somehow make the boundary act elastic, perhaps via
some trick with "fix spring", such that the volume of
my simulation isn't necessarily fixed. (Whether this
is a good idea is another story, but I'll leave that
for another time.)

Now as I've understood it, NVE integration simply
means that no scaling is applied to Newton's laws, so
energy (the "E" in NVE) is conserved, but the
conservation of the "N" and "V" in NVE is a byproduct
of periodic boundary conditions rather than something
enforced by the intergration algorithm itself. Is this correct?

You're correct. NVE integration assumes the N,V are constant,
but it doesn't deal with those quantities explicitly.

If you look in fix_nve.cpp, you'll see the eqs. They're
very simple.