LAMMPS WWW Site - LAMMPS Documentation - LAMMPS Mailing List Archives
Re: [lammps-users] Report bug on Thermo
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [lammps-users] Report bug on Thermo

From: Axel Kohlmeyer <akohlmey@...24...>
Date: Wed, 02 May 2018 21:12:09 +0000


You are reminding me of an idea to change how output is handled, that I tried to pitch a few years back. I think it is time to revisit this idea and discuss it with Steve. 


Dr. Axel Kohlmeyer  akohlmey@...24...
College of Science & Technology, Temple University, Philadelphia PA, USA
International Centre for Theoretical Physics, Trieste. Italy.

On Wed, May 2, 2018, 13:46 Robert Hoy <robertscotthoy@...24...> wrote:
Haha!  It would, however, be nice to be able to specify that the thermo output has be written to disk at MOST every N steps.  On some machines, runs will go for a very long time without writing anything to log.lammps.   On several occasions when runs on our cluster have crashed or timed-out,  I've had to restart from an earlier restart file than I otherwise would have used, simply to avoid having gaps in the log file.  This could be fixed by adding an option requiring thermo to flush whenever a restart file is written (as opposed to flushing every line).


On Tue, May 1, 2018 at 11:42 AM, Stefan Paquay <stefanpaquay@...24...> wrote:
I don't think this is an issue. In fact, it's a feature. You don't want your simulation to have to wait for frequent writes to disk, do you?
If you want to have it print immediately, there is an option to make the thermo output flush after every write called "flush". Check the docs here:
However, unlike with toilets, there is a good reason to not flush in general.

On Mon, Apr 30, 2018 at 2:48 PM, Lu Wang <mm229196@...7278...> wrote:
The data appear in the log file during simulation, but with a few delays, say, 100 timesteps.

How can I remove this issue?
Check out the vibrant tech community on one of the world's most
engaging tech sites,!
lammps-users mailing list