I am trying to 'rerun' a simulation from dump file and rerun command by follwing input lines in sequential order in same input script:
dump test_dump all custom 1 dump.file id x y z vx vy vz
fix NPT1 MovingSystem npt temp 300 300 100 iso 1.0 1.0 100 dilate MovingSystem
run 10
unfix NPT1
rerun dump.file dump x y z vx vy vz
I get following outputs:
step etotal evdwl ecoul lx ly lz pxx pyy pzz temp
for run : 0 -1848017.1 591796.75 12511103 178.37504 182.09152 187.42603 1297.8127 1282.6458 -2498.9966 299.93886
for rerun: 0 -1848019.6 591796.28 12511183 178.37504 182.09152 187.42603 1297.8901 1283.1423 -2498.8412 299.93886
there are also inconsistent flag warnings.
So while temperature and box dimensions remain same, energy and pressure tensor values are changing.
1) If it is 'rerun' should both values be not same? Is there something wrong in my approach here?
2) Why inconsistent flag warning? from read_dump page:
"One way this can happen is if you read image flag fields from the dump file but do not also use the dump file box parameters"
but here box dimensions are same as they should be so i guess that is not the problem.
3) If they will be different, (Lyapunov instability?) will they follow original constraints or fixes?
I am trying to 'rerun' a simulation from dump file and rerun command by
follwing input lines in sequential order in same input script:
dump test_dump all custom 1 dump.file id x y z vx vy vz
fix NPT1 MovingSystem npt temp 300 300 100 iso 1.0 1.0 100 dilate
MovingSystem
run 10
unfix NPT1
rerun dump.file dump x y z vx vy vz
I get following outputs:
step etotal evdwl ecoul lx ly lz pxx pyy pzz temp
for run : 0 -1848017.1 591796.75 12511103 178.37504
182.09152 187.42603 1297.8127 1282.6458 -2498.9966 299.93886
for rerun: 0 -1848019.6 591796.28 12511183 178.37504
182.09152 187.42603 1297.8901 1283.1423 -2498.8412 299.93886
there are also inconsistent flag warnings.
So while temperature and box dimensions remain same, energy and pressure
tensor values are changing.
1) If it is 'rerun' should both values be not same? Is there something
wrong in my approach here?
you are not feeding LAMMPS *exactly* the same numbers. with default
settings, dump files write out values with about 6-7 digits precision,
internally those numbers are stored with 15-16 digits precision. the
differences you see are due to this. they are nothing to worry about, since
you'll be doing averaging over the results anyway and the statistical
convergence is typically even less. if you feel differently, the precision
for output can be increased (which will also increase the file size
accordingly).
2) Why inconsistent flag warning? from read_dump page:
"One way this can happen is if you read image flag fields from the dump
file but do not also use the dump file box parameters"
but here box dimensions are same as they should be so i guess that is not
the problem.
you are not storing image flags, so they will become inconsistent, if
molecules will move through periodic boundaries. you cannot read/restore
what isn't there.
3) If they will be different, (Lyapunov instability?) will they follow
original constraints or fixes?
lyapunov has nothing to do with this. the trajectory will be exactly what
you read in, regardless of fixes or forces. that is the whole point about
using "rerun".
axel.