[lammps-users] about PPPM

Dear,

When I run the following code(in blue) with LAMMPS, system give me a error notice(in red)
“Out of range atoms - cannot compute PPPM”, manual says that One or more atoms are attempting to map their charge to a PPPM grid point that is not owned by a processor. This is usually because an atom has moved to far in a single timestep.

But I don’t know how solve it with kspace_modify command. my simulation box is 250 ppp.
How can I deal with some question like this? I hope somebody give me some adivce. Thanks in advance.

Wende Tian

LAMMPS (5 Oct 2007)

read_data data.dna
Scanning data file …
1 = max bonds/atom
1 = max angles/atom
Reading data file …
orthogonal box = (-123.311 -123.311 -123.311) to (123.311 123.311 123.311)
1 by 1 by 2 processor grid
240 atoms
179 bonds
178 angles
Finding 1-2 1-3 1-4 neighbors …
2 = max # of 1-2 neighbors
2 = max # of special neighbors

restart 5000 restart.%.dna
reset_timestep 0
run 4PPPM initialization …
G vector = nan
grid = 48 48 60
RMS precision = nan
brick FFT buffer size/proc = 85833 69120 15606
Setting up run …
Memory usage per processor = 96.3655 Mbytes
Step Temp PotEng KinEng TotEng E_coul E_pair E_bond
0 23.980464 nan 35.820819 nan nan nan 12.25924
ERROR: Out of range atoms - cannot compute PPPM

echo screen
dimension 3
boundary p p p
units lj

neighbor 0.3 bin
neigh_modify every 1 delay 5 check yes

atom_style full
bond_style harmonic
special_bonds 0.0 1.0 1.0
angle_style harmonic

Soft peotential push-off

read_data data.dna

group polymer type 1
group ion type 2

velocity all create 2.0 1112345
bond_coeff 1 3968.5 1.2246
angle_coeff 1 70 180

pair_style soft 2.5
pair_coeff * * 1.0 30.0 1.12246

fix 1 all nve
fix 2 all langevin 2.0 1.5 1000 699443

thermo 500

dump 1 all atom 100 dump1.project

run 10000

#Main run
#The cutoff distances are changeable here to different affinities between atoms

timestep 0.001

pair_style hybrid/overlay lj/expand 1.12246 coul/long 10.0
pair_coeff 1 1 lj/expand 1.0 1.12246 0.0 1.2246
pair_coeff 1 2 lj/expand 1.0 0.98 0.0 0.98
pair_coeff 2 2 lj/expand 1.0 0.6735 0.0 0.6735
pair_coeff * * coul/long
dielectric 80.0
pair_modify shift yes

kspace_style pppm 0.0001
kspace_modify mesh 24 24 30 order 6

fix 5 all langevin 1.5 1.2 1000 699447

thermo 100
thermo_style custom step temp pe ke etotal ecoul epair ebond

dump 2 polymer atom 1000 dump.polymer.project
dump 3 ion atom 1000 dump.ion.project
dump 4 all atom 1000 dump.project

restart 5000 restart.%.dna
reset_timestep 0
run 50000

Your system seems to contain badly overlapped atoms. Try minimizing the system first and then run MD.

Arnab

The problem is you are getting a G-vector of nan by the
PPPM setup. The question is why. Please run
your script with the latest version of LAMMPS (22 Jan 08
release) - there was an upgrade to the PPPM setup between
5 Oct 07 and now.

If you stilll have a problem, please post your input
script and Paul will take a look at it.

Steve