Question about SWM4NDP water simulation

[luohancfd]

Dear Developers,

I'm interested in your great work. I'm currently trying to use drude oscillator to simulate water under strong electric field. However, I met a problem with the equilibrium of swm4-ndp water. Your example runs pretty well on my machine. But my input file with 256 water fails. I try to mimic the work done by G. Lamoureux, A.D. MacKerell, Jr. and B. Roux A simple polarizable model of water based on classical Drude oscillators. So I create a 19.7165A cubic box and fill with 256 water. The bond and lj coefficient are the same. But I use fix shake to hold Oxygen-Hydrogen and Oxygen-M. After drude/transfrom, I use fix npt for atoms and fix nvt for drude. However, I'll always get error like non-numeric box dimensions- simulation unstable or neighbot list overflow. When I use my data file and your input file, the condition is the same. I wonder if you have any idea about the source of error. I upload my input file and data file to google drive. Thanks for your kind help.https://goo.gl/gEdQVN

[alaindequidt]

Le 07/07/2015 22:30, Han Luo a écrit :

Dear Developers,

I'm interested in your great work. I'm currently trying to use drude
oscillator to simulate water under strong electric field. However, I
met a problem with the equilibrium of swm4-ndp water. Your example
runs pretty well on my machine. But my input file with 256 water
fails. I try to mimic the work done by G. Lamoureux, A.D. MacKerell,
Jr. and B. Roux A simple polarizable model of water based on classical
Drude oscillators. So I create a 19.7165A cubic box and fill with 256
water. The bond and lj coefficient are the same. But I use fix shake
to hold Oxygen-Hydrogen and Oxygen-M. After drude/transfrom, I use fix
npt for atoms and fix nvt for drude. However, I'll always get error
like non-numeric box dimensions- simulation unstable or neighbot list
overflow. When I use my data file and your input file, the condition
is the same. I wonder if you have any idea about the source of error.
I upload my input file and data file to google drive. Thanks for your
kind help.https://goo.gl/gEdQVN https:/%0A%20/goo.gl/gEdQVN

—
Reply to this email directly or view it on GitHub
#1.

Hello,
Thanks for your interest.
I see a possible explanation for your problem:
It seems that your fix shake only constrains the OM distance: the M
"virtual" site is free to rotate around the O, whereas it shouldn't.
You could perhaps add a HOM angle (or HM bonds) and shake it also, but I
am not sure it would be efficient because:

1. the mass of M is so tiny (so its dynamics is much more sensitive)
2. the out-of-plane motion of the M site (out of the HOH plane) will not
   be efficiently constrained because there is no shaked dihedral
   That's why we advocate to use fix rigid for swm4ndp instead of shake (as
   in the examples in the repository).
   Best regards,
   Alain

Alain DEQUIDT
+33 (0) 4 73 40 71 94

Maître de Conférences
Université Blaise Pascal Clermont II, ICCF
Chimie 7
24 avenue des landais
63171 Aubiere Cedex