lammps crashes

Hi,

I have found a strange issue with lammps (with reax). The code runs fine when the timestep=0.1 but exits with a segmentation fault when I put timestep=0.5.
The backtrace from core file is:

Program terminated with signal 11, Segmentation fault.
#0 0x0815c37e in LAMMPS_NS::Neighbor::bin_atoms() ()
(gdb) backtrace
#0 0x0815c37e in LAMMPS_NS::Neighbor::bin_atoms() ()
#1 0x08160be7 in LAMMPS_NS::Neighbor::half_bin_no_newton(LAMMPS_NS::NeighList*) ()
#2 0x08158d10 in LAMMPS_NS::Neighbor::build() ()
#3 0x0820e527 in LAMMPS_NS::Verlet::run(int) ()
#4 0x081f385d in LAMMPS_NS::Run::command(int, char**) ()
#5 0x0813a83c in LAMMPS_NS::Input::execute_command() ()
#6 0x0813b253 in LAMMPS_NS::Input::file() ()
#7 0x081414e0 in main ()

I tried to debug with:
gdb lmp_ubuntu
run < input
During startup program terminated with signal SIGKILL, Killed.

The program runs fine on example files in reax directory.

My output when timestep=0.1:

Step Atoms Temp Press Enthalpy TotEng PotEng KinEng E_pair eb ev et
20 928 0 24286.79 -118100.4 -122804.58 -122804.58 0 -122804.58 -164555.17 2238.2132 -135.18425
30 928 0.089326043 23984.756 -118159.62 -122805.3 -122805.55 0.2468267 -122805.55 -164556.7 2237.8782 -135.31901
40 928 0.15971699 23664.939 -118222.36 -122806.11 -122806.55 0.44133173 -122806.55 -164558.29 2237.5146 -135.47286
50 928 0.23630163 23573.485 -118240.71 -122806.83 -122807.49 0.65295125 -122807.49 -164559.9 2237.1329 -135.56733
60 928 0.32675176 23358.931 -118283.04 -122807.49 -122808.39 0.90288403 -122808.39 -164561.11 2236.7075 -135.72528
70 928 0.39833944 23228.934 -118308.83 -122808.23 -122809.33 1.1006959 -122809.33 -164562.56 2236.3185 -135.91092
80 928 0.44264537 23224.5 -118310.23 -122808.95 -122810.17 1.2231225 -122810.17 -164564.49 2235.9576 -136.05399
90 928 0.55570451 23193.396 -118316.75 -122809.48 -122811.01 1.5355288 -122811.01 -164565.32 2235.5864 -136.25497
100 928 0.59408565 22885.514 -118377.03 -122810.22 -122811.86 1.641584 -122811.86 -164565.75 2235.2976 -136.45841

My output when timestep=0.5:

Step Atoms Temp Press Enthalpy TotEng PotEng KinEng E_pair eb ev et
20 928 0 24286.79 -118100.4 -122804.58 -122804.58 0 -122804.58 -164555.17 2238.2132 -135.18425
30 928 8.3039536e+09 2.2786772e+09 3.8242598e+10 2.2945557e+10 -18432.434 2.2945575e+10 -18432.434 -37228.326 2689.7337 347.49761

The temperature rises significantly (10^9) and other energy terms too.

Thanks,

Vivek

Hi,
I have found a strange issue with lammps (with reax). The code runs fine
when the timestep=0.1 but exits with a segmentation fault when I put
timestep=0.5.

please explain where there is a strangeness.

if you use too long a time step for the physics of
your system, you should see _exactly_ this behavior,
i.e. kinetic energy exploding.

axel.

on top of my previous remarks, there is one issue with your
system. do you expect to have this high a pressure?

some atoms have to be _very_ close together and that
will most certainly limit your ability to use a large time step.

axel.

as you can figure out, i am learning to perform these calculations… so, i do not know a lot of things currently.
i need to get some feel for “typical numbers”.

Thanks for your remarks !

as you can figure out, i am learning to perform these calculations.... so, i
do not know a lot of things currently.
i need to get some feel for "typical numbers".

in that case, it would be _really_ helpful if you tell
people what kind of a system/setup under what
intended conditions you are after.

i've been simply guessing that the use of reax as a force field
makes it a bit unlikely that you are looking at high-pressure
structural transformations of solids. :wink:

axel