invalid neighbor bin

hi, there.
i run this problem(in and data profiel attached) and it prints " ERROR on proc 0: Invalid neighor bin (…/neighbor.cpp:2093)" on the screen.
i searched on the mail list and read the mannul to find where the mistake was but failed .
it can’t can even single a step!
can u help me with that and give me some suggestion?
sincerely,
Stephen.

newmodel66.data (529 KB)

innew (952 Bytes)

hi, there.
i run this problem(in and data profiel attached) and it prints " ERROR on
proc 0: Invalid neighor bin (../neighbor.cpp:2093)" on the screen.
i searched on the mail list and read the mannul to find where the mistake
was but failed .
it can't can even single a step!

​have you looked at the energies of your zero-th step?

Memory usage per processor = 25.4611 Mbytes
Step Temp E_pair E_mol TotEng Press
       0 0 -nan 15243.118 -nan -nan

​You energy for E_pair is "-nan"​, that is usually a sign for a bad initial
structure, often indicating overlapping atoms (which may not be obvious due
to periodic boundary conditions.

if i insert the following code in your input

run 0 post no
delete_atoms overlap 0.1 all all mol yes bond yes compress yes
​run 0 post no​

​you get: ​

Deleted 129 atoms, new total = 2346
Deleted 86 bonds, new total = 2072
Deleted 43 angles, new total = 2398
Deleted 0 dihedrals, new total = 3612
Deleted 0 impropers, new total = 600

​and after that:

Step Temp E_pair E_mol TotEng Press
       0 0 32796.242 15091.641 47887.883 480609.05

when i turn off periodic boundaries, the "-nan" does not appear, so it
looks that 43 of your water molecules are within 0.1 angstrom from their
​periodic neighbors, which means, that indeed your initial geometry is
*very* bad, as it includes periodic copies of water molecules that sit
exactly on top of each other.

​axel.​