when should finish() method of a fix be called?

LAMMPS LAMMPS Mailing List Mirror
1

Dear All,

I have a very naive question, but
I noticed that a finish() method of a
(ATC_TransferHardy) fix is not called.

Methods like pre_init_integrate(),
post_init_integrate(), pre_final_integrate()
are called as expected.

Should not finish() be called at the end of each run command?

Kind regards,
Denis Davydov

2

Dear All,

I have a very naive question, but
I noticed that a finish() method of a
(ATC_TransferHardy) fix is not called.

Methods like pre_init_integrate(),
post_init_integrate(), pre_final_integrate()
are called as expected.

Should not finish() be called at the end of each run command?

there is no finish() method defined in fix.h why should it be called then?

axel.

3

Dear Axel,

Dear All,

I have a very naive question, but
I noticed that a finish() method of a
(ATC_TransferHardy) fix is not called.

Methods like pre_init_integrate(),
post_init_integrate(), pre_final_integrate()
are called as expected.

Should not finish() be called at the end of each run command?

there is no finish() method defined in fix.h why should it be called then?

I guess the right name for a method to be called after each run is complete is
post_run() ?

p/s/ did finish() method ever exist in LAMMPS code?
Just wonder why in ATC fix it is documented as
" // called after the end of a "run" "

Kind regards,
Denis.

4

Hi Denis, when we started we fully implemented the fix functionality inherited from fix.h. However, not all are used, which is as it should be. The whole point of the class structure of Lammps fixes is that only the needed functionality should be added.
Jeremy

5

Dear Jeremy,

Thank you very much for making clear this.

Kind regards,
Denis.