ERROR on proc 0: Out of range atoms - cannot compute MSM (../msm.cpp:1447)

Dear LAMMPS users,

I have been recently struggling with getting an error when I try to equilibrate a water box. I have used both SPCE and TIP3P model for water and created the structure using solvate tool of vmd. I equilibrated the system under NVT ensemble with 1 fs time step and 100.0 fs for damping factor. When I use periodic boundary conditions with pppm for kspace_style in all three directions, the simulation successfully proceeds and thermodynamic properties of the system converge to the desired values on average. However, when I use non-periodic boundary conditions in all three directions with msm kspace style, I get the following error during simulation:

ERROR on proc 0: Out of range atoms - cannot compute MSM (…/msm.cpp:1447)

Could anyone please give me any advice regarding how to deal with this problem. I can provide my data files if they’re needed. I appreciate your attention.

Regards,

Farshad

Dear LAMMPS users,

I have been recently struggling with getting an error when I try to
equilibrate a water box. I have used both SPCE and TIP3P model for water and
created the structure using solvate tool of vmd. I equilibrated the system
under NVT ensemble with 1 fs time step and 100.0 fs for damping factor. When
I use periodic boundary conditions with pppm for kspace_style in all three
directions, the simulation successfully proceeds and thermodynamic
properties of the system converge to the desired values on average.
However, when I use non-periodic boundary conditions in all three directions
with msm kspace style, I get the following error during simulation:

ERROR on proc 0: Out of range atoms - cannot compute MSM (../msm.cpp:1447)

Could anyone please give me any advice regarding how to deal with this
problem. I can provide my data files if they're needed. I appreciate your

well, have you checked whether the claim of the error message - atoms
that are (too far) outside the box - is true?

axel.

Stan may have some additional ideas.

Steve

Thanks you all for your comments.

I used fixed boundary conditions in all three directions. I am doing a test to see whether the problem is due to intercommunication between processors or not. Because, yesterday I ran my simulation on one and two processors. The case on one processor was kept running, but the one on two processors ended up giving the following error after 13700 steps:

ERROR on proc 0: Out of range atoms - cannot compute MSM (…/msm.cpp:1447)

The thermodynamic out put looks like this:

Step PotEng TotEng Press Temp Volume

12900 -8590.9513 -8111.681 10.819455 114.27523 41940.113
13000 -8582.8456 -8109.359 135.16285 112.89619 41940.113
13100 -8576.5002 -8090.496 -57.232757 115.88083 41940.113
13200 -8565.9387 -8069.8907 -131.20414 118.27563 41940.113
13300 -8550.0363 -8052.6932 409.2915 118.58443 41940.113
13400 -8556.0719 -8052.9895 233.23153 119.95288 41940.113
13500 -8563.083 -8071.384 -330.60615 117.23868 41940.113
13600 -8554.9835 -8085.6313 -77.486196 111.9104 41940.113
13700 -8553.1265 -8075.8104 68.420099 113.80928 41940.113

ERROR on proc 0: Out of range atoms - cannot compute MSM (…/msm.cpp:1447)

My input file is as follows. I have also attached the water box datafile.

units real
dimension 3
boundary f f f
atom_style full

read_data lmp_waterbox.dat
neighbor 3.0 bin
neigh_modify delay 0 every 1 check yes

Assign Charges

group hy type 1 # H_water
group ox type 2 # O_water
set group ox charge -0.8476

set group hy charge 0.4238

Flexible SPC/E Potential Parameters

Zhang et al., Fluid Phase Equilibria, 262 (2007) 210-216

pair_style lj/cut/coul/msm 8.5
pair_coeff 1 1 0.0 0.0 # H-H
pair_coeff 1 2 0.0 0.0 # H-O
pair_coeff 2 2 0.1553 3.166 # O-O

bond_style harmonic
bond_coeff 1 7.669 1.0
angle_style harmonic
angle_coeff 1 1.829 109.47
kspace_style msm 1.0e-5

thermo 100
thermo_style custom step pe etotal press temp vol
thermo_modify norm no
dump mydcd all dcd 500 nvt.dcd
dump_modify mydcd unwrap no

timestep 1.0

velocity all create 50.0 482748 dist gaussian
fix water_bond all shake 1e-5 20 0 b 1 a 1
fix 1 all nvt temp 50.0 298.0 100.0
run 50000
write_restart restart.nvt_50_298.*
unfix 1
fix 2 all nvt temp 298.0 298.0 100.0
run 50000
write_restart restart.nvt_298.*

Stan may have some additional ideas.

Steve

lmp_waterbox.dat (157 KB)

I ran the simulation on single processor and still getting the same error. I also replaced the fixed boundary with shrink-wrapped boundary. Nothing changed.

Thanks you all for your comments.

I used fixed boundary conditions in all three directions. I am doing a test to see whether the problem is due to intercommunication between processors or not. Because, yesterday I ran my simulation on one and two processors. The case on one processor was kept running, but the one on two processors ended up giving the following error after 13700 steps:

ERROR on proc 0: Out of range atoms - cannot compute MSM (…/msm.cpp:1447)

The thermodynamic out put looks like this:

Step PotEng TotEng Press Temp Volume

12900 -8590.9513 -8111.681 10.819455 114.27523 41940.113
13000 -8582.8456 -8109.359 135.16285 112.89619 41940.113
13100 -8576.5002 -8090.496 -57.232757 115.88083 41940.113
13200 -8565.9387 -8069.8907 -131.20414 118.27563 41940.113
13300 -8550.0363 -8052.6932 409.2915 118.58443 41940.113
13400 -8556.0719 -8052.9895 233.23153 119.95288 41940.113
13500 -8563.083 -8071.384 -330.60615 117.23868 41940.113
13600 -8554.9835 -8085.6313 -77.486196 111.9104 41940.113
13700 -8553.1265 -8075.8104 68.420099 113.80928 41940.113

ERROR on proc 0: Out of range atoms - cannot compute MSM (…/msm.cpp:1447)

My input file is as follows. I have also attached the water box datafile.

units real
dimension 3
boundary f f f
atom_style full

read_data lmp_waterbox.dat
neighbor 3.0 bin
neigh_modify delay 0 every 1 check yes

Assign Charges

group hy type 1 # H_water
group ox type 2 # O_water
set group ox charge -0.8476

set group hy charge 0.4238

Flexible SPC/E Potential Parameters

Zhang et al., Fluid Phase Equilibria, 262 (2007) 210-216

pair_style lj/cut/coul/msm 8.5
pair_coeff 1 1 0.0 0.0 # H-H
pair_coeff 1 2 0.0 0.0 # H-O
pair_coeff 2 2 0.1553 3.166 # O-O

bond_style harmonic
bond_coeff 1 7.669 1.0
angle_style harmonic
angle_coeff 1 1.829 109.47
kspace_style msm 1.0e-5

thermo 100
thermo_style custom step pe etotal press temp vol
thermo_modify norm no
dump mydcd all dcd 500 nvt.dcd
dump_modify mydcd unwrap no

timestep 1.0

velocity all create 50.0 482748 dist gaussian
fix water_bond all shake 1e-5 20 0 b 1 a 1
fix 1 all nvt temp 50.0 298.0 100.0
run 50000
write_restart restart.nvt_50_298.*
unfix 1
fix 2 all nvt temp 298.0 298.0 100.0
run 50000
write_restart restart.nvt_298.*

Stan may have some additional ideas.

Steve

I ran the simulation on single processor and still getting the same error. I
also replaced the fixed boundary with shrink-wrapped boundary. Nothing
changed.

careful with shrink wrap. your box seems to have lots of vacuum and
with "s" boundaries you loose atoms quickly. you need to use "m"
instead.

axel.