Fix species makes qeq unconverged ?

Dear all,

Enclosed are two input files (with correspond output files) showing a strange behavior (at least strange for me…).

Adding (or removing) a fix species command make the QeQ unconverged (converged) and the calculation KO (OK).

The input files aim at creating bulk water from a random distribution of O and H atoms in NVT conditions.

Do you have any clue of the link between a post processing command and the charge equilibration ?

Does it mean that fix species highlights a problem in the simulation ?

I thank you in advance,



without_species_15.out (8.02 KB) (3.78 KB)

with_species_16.out (9.53 KB) (3.7 KB)

This is highly unlikely, but I will take a look when I get a chance.
Thanks, Ray


Fix species changes the re-neighboring criteria, and the fix species averaging of “2 10 1000” is not a good choice since it will then be forced to re-neighbor every 1000 steps. We have a newer version that corrects this, but for now I’d suggest you change the averaging to “2 10 20”.


Dear Ray,

I saw this warning about re-neighboring but thought Nfreq was only an output parameter, like in dump or thermo.

Thanks a lot, the calculation is going well.