Dear Users,
I am restarting the trajectory with 3Feb16 version. The following error pops out.
Dear Users,
I am restarting the trajectory with 3Feb16 version. The following error pops out.
read_restart rA.rest
orthogonal box = (-54.3192 -54.3192 -54.3192) to (54.3192 54.3192 54.3192)
2 by 3 by 4 MPI processor grid
27572 atoms
ERROR: Did not assign all restart atoms correctly (…/read_restart.cpp:513)
Although, It runs and restarts properly within old version 17Jan14
I have checked with different input files within the same version and they are restarting fine.
I do not have any idea, why it is not working with the particular files.
Restart files are not guaranteed to be portable between different versions of lammps, especially not, if they are two years apart.
Axel
What you should be able to do is use the old version to
read the old restart file. Then convert the info to
a data file (either via the older restart2data command
or via the newer -r command line switch). Then you
can use the data file as input to the current version of LAMMPS.
Steve
Thanks for the mail.
I am starting the calculations in new version and restarting them in new version.
This I have compared against old version.
It works fine in old version but not in new version.
The error pops out when I am using runstyle respa. Surprisingly, it vanish with runstyle verlet.
Not sure the reason though.
Thanks
Karteek
Thanks for the mail.
I am starting the calculations in new version and restarting them in new
version.This I have compared against old version.
It works fine in old version but not in new version.
that is not an argument.
The error pops out when I am using runstyle respa. Surprisingly, it vanish
with runstyle verlet.Not sure the reason though.
what does it matter? it is not *supposed* to work. you've been told so
now by two people. what is the point of insisting it should work? and
on what grounds? how about actually following the advice of the *lead
developer* of the software you are using rather than embarrassing
yourself in public?
if you had done some research in the mailing list archives on this
topic, you should have seen, that this is a regularly discussed issue
and there are plenty of explanations in previous discussion on why
LAMMPS behaves the way it does and why using data files is the only
approach that works reliably (and not just by chance and with the
danger of uncontrolled hidden data corruption).
axel.