I want to shear a 2d system of colloidal particles which are rigid
molecules. Each molecule consist of 3 types of atoms. To shear the system
I give the top two layers and bottom two layers of particles the same
constant velocity along +ve x-direction and -ve x-direction respectively,
set all other forces on those particles to 0 by fix setforce and integrate
them by rigid-nve. For the middle particles I integrate them by rigid/nve
and thermostat them by fix langevin. I have used boundary conditions p f
p. To distinguish between top, middle, and bottom layers of particles I
need total 9 types of atoms and I have defined all the interactions
properly in a separate file. I can share the input file if you like.
The problem is when I run the simulation, it exits immediately saying:
MPI_ABORT invoked on rank 0 in communicator MPI_COMM_WORLD with errorcode 1.
and in the log.lammps file it shows:
AMMPS (7 Dec 2010)
# Simple rigid body triskelions
dimension 2
units lj
atom_style molecular
atom_modify map array sort 0 0.0
boundary p f p
bond_style harmonic
read_data fort.50
2 = max bonds/atom
orthogonal box = (-170.203 -148.509 -0.5) to (170.203 148.35 0.5)
1 by 1 by 1 processor grid
2688 atoms
128 bonds
4 = max # of 1-2 neighbors
12 = max # of 1-3 neighbors
There are harmonic bonds between top particles and between bottom
particles. I have defined the atom positions as well as bonds in the file
fort.50. I don't understand why this problem is arising. Can anyone give
any suggestion regarding this?
I also want to know whether it is physical to shear a colloidal system in
this way?
Any comment in this regard is highly appreciated.
Thanks
I want to shear a 2d system of colloidal particles which are rigid
molecules. Each molecule consist of 3 types of atoms. To shear the system
I give the top two layers and bottom two layers of particles the same
constant velocity along +ve x-direction and -ve x-direction respectively,
set all other forces on those particles to 0 by fix setforce and integrate
them by rigid-nve. For the middle particles I integrate them by rigid/nve
and thermostat them by fix langevin. I have used boundary conditions p f
p. To distinguish between top, middle, and bottom layers of particles I
need total 9 types of atoms and I have defined all the interactions
properly in a separate file. I can share the input file if you like.
The problem is when I run the simulation, it exits immediately saying:
MPI_ABORT invoked on rank 0 in communicator MPI_COMM_WORLD with errorcode 1.
and in the log.lammps file it shows:
AMMPS (7 Dec 2010)
# Simple rigid body triskelions
Please try running with a current version of lammps. If there is a
bug, it may have been fixed already.
Thanks for your reply.
I have tried to simulate the system with LAMMPS (9 Nov 2011) but with same
results. I will install the most recent version and try the same. But, can
you give me a hint to the reason behind this problem?
Thanks for your reply.
I have tried to simulate the system with LAMMPS (9 Nov 2011) but with same
results. I will install the most recent version and try the same. But, can
you give me a hint to the reason behind this problem?
if i knew, i would have said so. it looks like some
misdirected MPI communication, but your guess
is as good as mine.
With no useful error message, it's hard to
say what is happening. The concept
of your model sounds fine. Assuming
this is a LAMMPS error, and not some
MPI issue, then I suggest you start
with a simple system (no shear, NVE,
no rigid fix) and "turn on" your features
one at a time ...