Hi-
I have encountered an issue using a hybrid pair_style with sw. The result depends on the order in which the pair_styles are listed. I have attached a simple example script that demonstrates this. It consists of two particles that interact via an LJ potential. When the two components of the hybrid potential are listed in the order given in the attached script:
pair_style hybrid sw lj/cut 9.0
then there is no force between the two particles (see the output file ‘f’). When the order is reversed:
pair_style hybrid lj/cut 9.0 sw
then I get the correct LJ force between the two particles.
Thanks, Steven
in.lammps (423 Bytes)
Si.sw (782 Bytes)
Hi-
I have encountered an issue using a hybrid pair_style with sw. The result
depends on the order in which the pair_styles are listed. I have attached a
simple example script that demonstrates this. It consists of two particles
that interact via an LJ potential. When the two components of the hybrid
potential are listed in the order given in the attached script:
pair_style hybrid sw lj/cut 9.0
then there is no force between the two particles (see the output file 'f').
When the order is reversed:
pair_style hybrid lj/cut 9.0 sw
then I get the correct LJ force between the two particles.
which version of LAMMPS are you using? is this the 13 Dec 2016 version or newer?
than this may be related to the recent neighbor list refactoring.
i can reproduce your observation with the 9 Jan 2017, but not with the
latest stable version 17 Nov 2016.
axel.
i've recorded this on github as issue #333
https://github.com/lammps/lammps/issues/333
Oh yes, I had meant to mention that as well, sorry. I am using the 9 Jan 2017 version. I also did not get the error with an older version.
Thanks, Steve