Removing ERROR on proc 0: Non-numeric atom coords - simulation unstable error

LAMMPS LAMMPS General Discussion
, ,
1

I am trying to use GCMC for adding CO2 molecules in the Zeolite and I did the same steps for adding water molecules.
I was able to successfully run the GCMC for adding water but in case of CO2 I am getting this problem.
WARNING: Neighbor exclusions used with KSpace solver may give inconsistent Coulombic energies (…/neighbor.cpp:654)
Setting up Verlet run …
Unit style : real
Current step : 179
Time step : 0.001
Per MPI rank memory allocation (min/avg/max) = 13.17 | 13.17 | 13.17 Mbytes
Step Temp TotEng v_nO f_fgcmc[3] f_fgcmc[4] f_fgcmc[5] f_fgcmc[6]
179 0 -157459.53 6 0 0 0 0
ERROR on proc 0: Non-numeric atom coords - simulation unstable (…/domain_omp.cpp:58)
Last command: run 100000

Here is my LAMMPS file

#----------------- Read framework ----------------------
units           real
atom_style      full
boundary p p p
#read_data       FAU_box.data
reset_timestep  00
read_restart    restartfile

#------------------ Define TIP3P water molecule ---------
molecule co2mol CO2.mol
create_atoms 0 random 3 3245 NULL mol co2mol 4585 overlap 2.0 maxtry 50

#------------------ Groups ----------------------------
group zeolite type 1 2          # framework atoms
group co2 type 3 4            # TIP3P water

#------------------ Masses -----------------------------
mass 1 15.999400  # O (framework)
mass 2 28.085501  # Si (framework)
mass 3 12.0107
mass 4 15.9994

#------------------ Charges ----------------------------
set type 1 charge -0.6000
set type 2 charge 1.2000
set type 3 charge 0.6512
set type 4 charge -0.3256

#------------------ Pair Style -------------------------
pair_style lj/cut/coul/long 12 12
pair_modify mix geometric
kspace_style pppm 1.0e-4

# Framework atom
pair_coeff 1 1 0.0585 3.06
pair_coeff 2 2 0.1625 3.96

# CO2
pair_coeff 3 3 0.05948 2.745
pair_coeff 4 4 0.17024 3.017

#------------------ Bonds ------------------------------
bond_style harmonic
bond_coeff 1 597.84 1.6100
bond_coeff 2 119.23 1.149

#------------------ Angles -----------------------------
angle_style harmonic
angle_coeff 1 596.16 109.6
angle_coeff 2 119.23 142.0
angle_coeff 3 119.23  180.00

#------------------ Dihedrals --------------------------
dihedral_style opls
dihedral_coeff 1  0.0 0.0 -352.42 0.0

#------------------ Minimization -----------------------
min_style cg
minimize 1.0e-6 1.0e-6 1000 10000

##------------------ SHAKE for water -------------------
fix shak co2 shake 1.0e-5 200 0 b 2 a 3 mol co2mol
variable n equal count(co2)
variable nO equal v_n/3

###------------------ NVT for framework -----------------
compute ctco2 co2 temp
compute_modify thermo_temp dynamic/dof yes
compute_modify ctco2 dynamic/dof yes
fix mynvt co2 nvt temp 300 300 0.1
fix_modify mynvt temp ctco2

##------------------ Timestep --------------------------
timestep 0.001

##------------------ GCMC Settings ---------------------
variable tfac equal 5.0/3.0
fix fgcmc co2 gcmc 100 100 0 0 65899 300 -0.5 0.1 mol co2mol tfac_insert ${tfac} shake shak full_energy pressure 100

#dump viz all image 250 myimage-*.ppm type type &
#    shiny 0.1 box no 0.01 view 45 45 zoom 1.8 size 1200 1500
#dump_modify viz backcolor white &
#acolor 1 yellow adiam 1 1.5 &
#acolor 2 red    adiam 2 1 &
#acolor 3 cyan   adiam 3 1 &
#acolor 4 white  adiam 4 0.5

thermo 250
thermo_style custom step temp etotal v_nO &
  f_fgcmc[3] f_fgcmc[4] f_fgcmc[5] f_fgcmc[6]

#----------------- Run -------------------------------
run 100000
#write_restart restartfile

I am using LAMMPS 2Aug2023 version.

2

Hello,

That part make little sense to me.

Shake cannot be used for linear molecules like CO2. Also your previously defined parameters don’t look like that of a rigid molecule:

#------------------ Bonds ------------------------------
bond_style harmonic
bond_coeff 1 597.84 1.6100
bond_coeff 2 119.23 1.149

#------------------ Angles -----------------------------
angle_style harmonic
angle_coeff 1 596.16 109.6
angle_coeff 2 119.23 142.0
angle_coeff 3 119.23  180.00

Best,
Simon

2 Likes
3

This version is now over two years old. Perhaps it is time for an update?

1 Like
5

Yeah but that’s the most latest version we have in our hpc. Thank you for the suggestion, I will request them to update it.

6

I removed the shake command and I am able to run my simulation.
Really appreciate it.

1 Like
7

In general, it is possible to compile LAMMPS yourself on a HPC, even without ‘root’ access.

8

Hi @Rohit_Sharma,

to add on what @simongravelle just said, LAMMPS is open-source and many people develop custom functions for their own needs. Also there are a lot of specific packages and it is nearly impossible for an HPC team to stay up-to-date to all the new changes in all the popular codes they provide.

More than updates of executables, they generally provide help and troubleshooting in setting up the compilation environment and which modules to call for execution of custom software, scripts or compiling personal codes. There is often also documentation from HPC maintenance teams on how to set up the environment yourself (which modules to call etc). This can become a very valuable skill in computational science for you to learn.