[lammps-users] Error to Float point exception for stretching simulations

Dear lammps-users,

I’m trying to simulate the stretching for diamond nanowires, first I minimize the structure, and then temperature relaxation to 300K, but when stretched along Z, there is an error for floating point exception and the following is the in file. Another question is why the Lx and Ly is also increase instead of decreasing?

#This is the log. File#

LAMMPS (9 Jan 2009)

units metal

boundary s s s

atom_style atomic

dimension 3

#Atom definition

read_data data.CH

orthogonal box = (-1.09707 -1.09707 13.1701) to (11.6799 11.6799 398.112)

1 by 2 by 32 processor grid

13700 atoms

pair_style airebo 3.0 0 0

pair_coeff * * CH.airebo C H

neighbor 2.0 nsq

neigh_modify delay 0 every 10 check yes page 100000

region 2 block INF INF INF INF INF 48.7669 units box

group left region 2

1246 atoms in group left

region 3 block INF INF INF INF 362.4 INF units box

group right region 3

1287 atoms in group right

group middle subtract all left right

11167 atoms in group middle

velocity left set 0.0 0.0 0.0 units box

fix 4 left setforce NULL NULL 0.0

velocity right set 0.0 0.0 0.0 units box

fix 5 right setforce NULL NULL 0.0

minimize 1.0e-6 1.0e-6 1000 100000

Memory usage per processor = 2.72927 Mbytes

Step Temp E_pair E_mol TotEng Press Volume

0 0 -60837.918 0 -60837.918 88008.548 62842.152

252 0 -73509.805 0 -73509.805 -65338.965 85839.623

Loop time of 424.325 on 64 procs for 252 steps with 13700 atoms

Minimization stats:

Stopping criterion = energy tolerance

Energy initial, next-to-last, final =

-60837.9182901 -73509.750813 -73509.8045783

Force two-norm initial, final = 706.347 4.66326

Force max component initial, final = 15.017 1.72312

Final line search alpha, max atom move = 0.0050467 0.00869608

Iterations, force evaluations = 252 524

velocity middle create 600.0 4928459 dist gaussian units box

fix 1 all nvt 300 300 100 drag 0.2

compute myTemp middle temp

thermo_style custom step temp etotal

thermo_modify lost warn flush yes

thermo 500

timestep 0.0005

run 50000

Memory usage per processor = 2.02649 Mbytes

Step Temp TotEng

252 489.0576 -72643.813

500 265.6557 -72639.757

1000 271.89835 -72639.552

50000 288.23934 -72609.464

50252 291.30806 -72609.205

unfix 1

fix 1 all nvt 300 300 100

run 50000

Memory usage per processor = 2.50336 Mbytes

Step Temp TotEng

50252 291.30806 -72609.205

50500 292.35759 -72608.989

100000 300.14503 -72573.819

100252 300.12216 -72573.53

unfix 1

unfix 5

fix 1 all nve

velocity right set 0.0 0.0 0.0003 units box

velocity middle ramp vz 0.0 0.0003 z 3.5667 103.4343 sum yes units box

compute s middle stress/atom

compute strea middle reduce sum c_s[1]

variable stressa equal c_strea/vol

compute streb middle reduce sum c_s[2]

variable stressb equal c_streb/vol

compute strec middle reduce sum c_s[3]

variable stressc equal c_strec/vol

thermo_style custom step c_myTemp etotal pe press v_stressa v_stressb v_stressc vol lx ly lz

thermo_modify lost warn flush yes

run 500000

Memory usage per processor = 2.59091 Mbytes

Step myTemp stressc Volume Lx Ly Lz

100252 326.51092 0.048937861 2.6429595e+11 7292.3672 7908.463 4582.7893

101000 362.50276 0.03799822 2.707081e+11 7318.2076 7952.266 4651.635

102000 448.37422 0.017477795 2.7940256e+11 7352.7889 8010.8067 4743.5342

srun: error: req727: task52: Floating point exception

srun: Terminating job

Please anybody have idea of such problems? Or there are some wrong in my in.file? I would be very thankful.

Sincerely yours,

Jing

Dear lammps-users,
    I'm trying to simulate the stretching for diamond nanowires, first
I minimize the structure, and then temperature relaxation to 300K, but
when stretched along Z, there is an error for floating point exception
and the following is the in file. Another question is why the Lx and
Ly is also increase instead of decreasing?

a) there seems to be a difficulty to integrate the equations of
motion and that would make the calculation crash. this happens
when your timestep is too large or you stretch too fast or both.

b) the increasing volume is an indication of atoms having evaporated
from your nanowire. they will move on forever because of using
shrinkwrap boundary conditions.

in both cases, you first have to look at your trajectory in a
visualization program to see what is going on.

axel

Dear axel,

Thank you for you advices. I have seen the trajectory of atoms form VMD. Yes, the atoms indeed have evaporated and the system has collapsed, but I have the structure optimized, perhaps I will exchange the shrink-wrap boundary conditions for others, so I will have a try for change the boundary conditions、timestep and the velocity of stretching, thank you very much!

在2009-11-30,“Axel Kohlmeyer” <akohlmey@…24…> 写道: