Hello,
I am studying canonical thermal equilibration of CNTs. I have two builds of LAMMPS on the computer cluster I use. I have been (unknowingly) running my simulations in a version from 2008(not sure of the month). After realizing there is a build from Oct. 2010 on the system, I began running simulations in that version. To my surprise, the different builds return different thermal_style outputs with identical input files. I output total energy, potential energy, kinetic energy, and temperature. The KE and temp data is mostly the same, but the TE and PE are very different. The ONLY change I made in the input file was in the ‘fix nvt’ command. The 2008 command is: “fix NVT all nvt 300 300 0.1 drag 0.2” and the 2010 version requires the ‘temp’ keyword to be placed before you specify the start/stop temperatures and damping parameter, i.e., “fix NVT all nvt temp 300 300 0.1 drag 0.2”. Below is my input file for the 2010 version.
#NVT ensemble: CNT(10,0)
#Variable Definitions
units metal
boundary p p p
atom_style atomic
read_data coordinate.10_0_10
pair_style tersoff
pair_coeff * * SiC.tersoff C
newton on
neighbor 2.0 bin
neigh_modify every 3 delay 3
timestep 0.0005
thermo 1
#Canonical Ens, 100000 steps
fix NVT all nvt temp 300 300 0.01 drag 0.2 #temp keyword omitted to run in 2008
thermo_style custom step etotal pe ke temp
dump DUMP all atom 10 dump.10_0_1
log log.10_0_10_NVT
run 100000
Thanks for any insight as to why this is occuring.
Andy
Hello,
I am studying canonical thermal equilibration of CNTs. I have two builds of
LAMMPS on the computer cluster I use. I have been (unknowingly) running my
simulations in a version from 2008(not sure of the month). After realizing
there is a build from Oct. 2010 on the system, I began running simulations
in that version. To my surprise, the different builds return different
thermal_style outputs with identical input files. I output total energy,
potential energy, kinetic energy, and temperature. The KE and temp data is
mostly the same, but the TE and PE are very different. The ONLY change I
very different at the fist step or over time?
made in the input file was in the 'fix nvt' command. The 2008 command is:
"fix NVT all nvt 300 300 0.1 drag 0.2" and the 2010 version requires the
'temp' keyword to be placed before you specify the start/stop temperatures
and damping parameter, i.e., "fix NVT all nvt temp 300 300 0.1 drag 0.2".
Below is my input file for the 2010 version.
#NVT ensemble: CNT(10,0)
#Variable Definitions
units metal
boundary p p p
atom_style atomic
read_data coordinate.10_0_10
pair_style tersoff
pair_coeff * * SiC.tersoff C
newton on
neighbor 2.0 bin
neigh_modify every 3 delay 3
timestep 0.0005
thermo 1
#Canonical Ens, 100000 steps
fix NVT all nvt temp 300 300 0.01 drag 0.2 #temp keyword omitted to
run in 2008
thermo_style custom step etotal pe ke temp
dump DUMP all atom 10 dump.10_0_1
log log.10_0_10_NVT
run 100000
Thanks for any insight as to why this is occuring.
the nose-hoover thermostats were replaced by a new
implementation that provides nose-hoover chain thermostats.
diverging trajectories are to be expected, but the averages
for each property should remain the same.
axel.
More generally, you should download the current version.
No one, especially me, is going to look at possible issues
in versions from long ago.
Steve