[lammps-users] ERROR: Non-numeric pressure - simulation unstable

Dear Lammps users,

I want to relax the wurtzite structure by SW polymorphic potentiel , but when I try to do simulation with Lammps, I got an error.
I have attached the input data and the code for relaxation.
LAMMPS (29 Sep 2021 - Update 1)

relaxation for lattice constant

P1 Initialization basic condition

units metal
#package omp 0 neigh yes # Nthreads=0 for default value set by OMP_NUM_THREADS environment
dimension 3
#how MPI tasks are mapped to the simulation box
processors * * *
boundary p p p
atom_style atomic
#atom_modify

P2 setup box and create atoms and groupen atoms

read struct.data

read_data perfectwz.data
Reading data file …
orthogonal box = (-1.1275740 0.0000000 0.0000000) to (13.530880 10.416080 8.7716640)
5 by 4 by 2 MPI processor grid
reading atoms …
420 atoms
read_data CPU = 0.009 seconds

neighbor 2.0 bin
neigh_modify delay 0 every 1 check yes

P3 Potential config: Pair style for atomic interactions

pair_style polymorphic
pair_coeff * * GaN_sw.poly N Ga
Reading polymorphic potential file GaN_sw.poly with DATE: 2015-12-09
#pair_style sw
#pair_coeff * * GaN.sw N Ga

P4 Define and initialize atomic configuration

fix F1 all box/relax aniso 0 vmax 0.01 fixedpoint 0 0 0

P4 Fixes for time integration and constraints

thermo 1
thermo_style custom step press etotal vol lx ly lz

compute atmE all pe/atom
dump D1 all custom 100 rlx_struct.dump.* id type x y z fx fy fz c_atmE

P5 Run or minimize

min_style cg
minimize 1.0e-15 1.0e-15 1000 10000
Neighbor list info …
update every 1 steps, delay 0 steps, check yes
max neighbors/atom: 2000, page size: 100000
master list distance cutoff = 5.36
ghost atom cutoff = 5.36
binsize = 2.68, bins = 6 4 4
1 neighbor lists, perpetual/occasional/extra = 1 0 0
(1) pair polymorphic, perpetual
attributes: full, newton on
pair build: full/bin/atomonly
stencil: full/bin/3d
bin: standard
Per MPI rank memory allocation (min/avg/max) = 5.611 | 5.617 | 5.621 Mbytes
Step Press TotEng Volume Lx Ly Lz
0 -nan -nan 1339.2895 14.658454 10.41608 8.771664
ERROR: Non-numeric pressure - simulation unstable (…/fix_box_relax.cpp:732)
Last command: minimize 1.0e-15 1.0e-15 1000 10000

in.LattCont (1.15 KB)

perfectwz.data (13 KB)

Most likely your initial geometry is incorrect leading to overlapping atom positions.
Have you visualized it and also checked whether there are no overlaps through periodic boundaries?
You can also do a simple check in your input. If you add a command like:

delete_atoms overlap 0.1 all all

There should be no atoms deleted. If there are, your data file has a bad geometry.