Hi everyone,
I'm trying to load a restart file but it keeps error-exiting when I fix my mesh surfaces. I have the original input file that created the restart state, I've checked the log that was output for this case, and as far as I can tell I am specifying the exact same fix properties with the restarted input file.
Normally with these kinds of restart file errors I have the work-around that I copy the fix command from the original input, unfix it, then make the fix with a different name. But this time I'm stuck because to the best of my knowledge, I _am_ copying the fix from the original input but I'm still getting the error.
I thought I could somehow open the restart file to discover if there's information about what it wants in my restarted fix. I looked at LAMMPS' restart2data tool but it won't work (saying something about different formats). Is there any other way to access the restart file to help me figure out where I'm going wrong?
ckloss | Tue, 05/14/2013 - 12:58
Hi jtvanlew,
Hi jtvanlew,
did you use the same LIGGGHTS version to create and to read the restart file?
If yes, can you post a simple case that shows the issue?
If no, there is no guaranteed restart file consistency along with version updates of LIGGGHTS (we're not Microsoft...)
Christoph
jtvanlew | Mon, 05/20/2013 - 23:13
fixed, sort of
haha yeah - I am not expecting Microsoft level of legacy support :) They were the same version.
I had done a chain of restart files like:
in.pack
in.press (load pack restart data)
in.evap (load press restart data)
in.Heat (load evap restart data)
in.HeatRestart (the first heat didn't carry through long enough to thermal steady-state, so load heat restart and keep going)
and it was on that last heat restart where my walls wouldn't get fixed properly. I doubt I could make a simple case to recreate this.
However, I was able to get around it. Typically, so long as there wasn't a change in the wall behavior (i.e. at first they have no temperature, then I later specify one) I could load the exact same fix command and have no issue. Following advice from the forum, if I _did_ have property changes, I would call the original, unfix it, and fix a new one with the new properties. I wasn't able to do that for the last restart file.
So, instead, I wouldn't ever try to call the original CAD fix at all. I could call a new fix to the same CAD file, Then, so long as I called the same fix for wall/gran/hertz/history mesh -- but referencing only the new wall fixes -- it would load with no problem.
I don't know if this is an elegant way to take care of my problem -- but so long as it keeps working for me now, I am not going to complain.