Mixed up atom-ids in fix store/state!

Submitted by PaulWinkler on Tue, 08/27/2013 - 12:03

Hi,

to differ moving particles from fixed, I am using fix store/state to store a displace/atom compute as per atom vector on different time steps. So I can make a dump after the last run with all displacements over several time steps and calculate the magnitude of all displacements.

...
compute disp all displace/atom
...
run...
...
fix disp1 all store/state 0 c_disp[1] c_disp[2] c_disp[3] c_disp[4]
...
run...
...
fix disp2 all store/state 0 c_disp[1] c_disp[2] c_disp[3] c_disp[4]
...

Some atoms are fixed with the fix freeze, in the normal dump file this atoms have no displacements. That's all ok, but when I use the values from the fix store/state, it seems that at least one vector mixed up the atom-ids. Atoms which are fixed (displacement = 0 in dump file) get an displacement and mobile not. The stored data seems to be mixed up afterwards or atom-ids changed...
How is the store/state values are stored? By atom-id? In LAMMPS you can store the atom-id to store/state, that wouldn't make sense at all then... Thought atom-ids are time consistence?

Best wishes,

Paul