LIGGGHTS: possible conflict with standard lammps?

Submitted by phsieh2005 on Fri, 02/26/2010 - 13:41

Hi, Christoph,

Thanks for the help with running the LIGGGHTS examples.

After I ran the movingMeshGran and meshGran examples successfully, I tried to run the pour example in the lammps/examples folder (I included the granular package when I compiled lammps/LIGGGHTS). This example failed to run. I then installed the official lammps on another computer with the granular package included. The pour example ran without any problem. In the LIGGGHTS-beta version, is there a conflict between LIGGHTS and lammps?

Below are the error messages:
hsieh@tarrytown:~/cfdem/LIGGGHTS-beta/lammps-24Jan10/examples/pour> mpirun -np 4 ./lmp_mpich2 < in.pour
LAMMPS (15 Jan 2010)
Created orthogonal box = (-10 -10 -0.5) to (10 10 16)
2 by 2 by 1 processor grid
ERROR: Illegal pair_style command
hsieh@tarrytown:~/cfdem/LIGGGHTS-beta/lammps-24Jan10/examples/pour>
------------------------------------------------------------------------------------------------------
Below listed the packages installed:
hsieh@tarrytown:~/cfdem/LIGGGHTS-beta/lammps-24Jan10/src> make package-status
Installed YES: package ASPHERE
Installed YES: package CLASS2
Installed YES: package COLLOID
Installed NO: package DIPOLE
Installed NO: package DSMC
Installed NO: package GPU
Installed YES: package GRANULAR
Installed YES: package KSPACE
Installed YES: package MANYBODY
Installed NO: package MEAM
Installed YES: package MOLECULE
Installed NO: package OPT
Installed NO: package PERI
Installed NO: package POEMS
Installed NO: package PRD
Installed NO: package REAX
Installed NO: package XTC

Installed NO: package USER-ACKLAND
Installed NO: package USER-ATC
Installed NO: package USER-CD-EAM
Installed NO: package USER-CG-CMM
Installed NO: package USER-EWALDN
Installed NO: package USER-IMD
Installed NO: package USER-SMD
hsieh@tarrytown:~/cfdem/LIGGGHTS-beta/lammps-24Jan10/src>
-----------------------------------------------------------------------------------------
Pei

Forums:

ckloss's picture

ckloss | Fri, 02/26/2010 - 15:06

Hello Pei,

In principle, LIGGGHTS should be able do everything LAMMPS can do, with improvements for granular systems.

It is correct that the "pour" example from LAMMPS will not work with LIGGGHTS.

The reason is that in LIGGGHTS, the granular pair style formulation is different from LAMMPS, for the following reasons:
+ To allow different particle/wall materials within one simulation
+ To derive stiffness / damp coefficients from models coming from literature
+ To add macroscopic cohesion

All of these features are not not possible with standard LAMMPS. Please find the doc_update dir for a beta version of the LIGGGHTS doc. As you can see, the pair style gran command has change to allow the new features.

IMPORTANT NOTE: Please do NOT make a package update of the granular package, as the LIGGGHTS files of the beta version are only in the src dir.
If you have already done, so, please overwrite all files in the src dir with the files in the src dir from the LIGGGHTS tarball

htm (not verified) | Fri, 10/22/2010 - 04:27

Christoph, Pei
how would I have to modify the pour 2d lammmps example to make it work with liggghts? I am new at both and interested in using liggghts because of improved granular stuff
thanks!
Horacio

ckloss's picture

ckloss | Fri, 10/22/2010 - 07:49

If you want to get started, there are LIGGGHTS examples that come with the tarball - so unless you want to study this very example, there is no need to modify the pour 2d lammmps example. But if you want to do that - the LIGGGHTS example "newContactModels" contains 2 input scripts - one that works with LIGGGHTS and one that works with LAMMPS. Here you can see the difference.

Christoph

htm (not verified) | Fri, 10/29/2010 - 23:38

Hi again
Thanks for your reply, what I would like to test is a 2d simulation, that is why I was asking previously how would I modify the 2dpour example in LAMMPS to make it works with LIGGGHTS. Or put another way, how would I enforce 2d on the simulation on the pour example that comes with LIGGGHTS? I tried using the "enforce 2D" or something like that that comes with the LAMMPS 2dpour example but then I have the issue with the "particletemplate/sphere" argument in fix and can't find the appropriate one to generate the pour. Any suggestion would be useful :)
Thanks for your help!
H

ckloss's picture

ckloss | Sat, 10/30/2010 - 00:00

Horacio,

what enforce 2d does is, as the doc says: Zero out the z-dimension velocity and force on each atom in the group
If you insert the particles 3d, as with fix pour/dev, they will still be 3d with enforce2d. What you could do is insert particles on a lattice in 2d.

If you ask me for my personal opinion, 2d simulations are not that useful, and I personally will not do any code development for 2d simulations

Christoph

htm (not verified) | Sat, 10/30/2010 - 16:03

Christoph
Thanks for your answer and your comment about 2d simulations, you are probably right about not being that useful. Since I am interested in a quasi-2d problem I was reducing the problem maybe to much. I am now able to set the problem in 3d closer to the experimental setup (which is 3d even when they try to have a flow as close to 2d as possible) and take into account the effects of walls thanks to LIGGGHTS!
Again thanks for your help and comments on this issue.
I'll probably read you again on another post!
Horacio