ERROR: One or more rigid bodies are a single particle in water tip4p simulation

LAMMPS LAMMPS Mailing List Mirror
1

I am trying to simulate TIP4P water by adding 4th massless site (1e-10 mass) and simulating using rigid/nvt/small. My system consist of water, some organic molecules and some non moving atoms.

whenever I am trying to run simulation i get

ERROR: One or more rigid bodies are a single particle (../fix_rigid_small.cpp:1613)

however when i delete all other molecules except water one, it runs smoothly. Simulation also works under NVT if i delete water molecules. Hence other molecules are also fine.

But all system together gives above error. What are the possible reasons?

Rigid molecules are defined in seperate file as follows:
variable bodies atomfile water_groups.txt
fix 1 water rigid/nvt/small custom v_bodies 300 300 100
run 10

2

Have you tried to use pair_style lj/cut/tip4p/cut? It will add the 4th massless site for you.

http://lammps.sandia.gov/doc/Section_howto.html#howto-8



6. How-to discussions — LAMMPS documentation
lammps.sandia.gov
The example input scripts included in the LAMMPS distribution and highlighted in Section 7 also show how to setup and run various kinds of simulations.

Simon

3

The error message is pretty clear. Your variable v_bodies
must be assigning a single atom to a particular rigid body,
which is not allowed. I suggest you dump the variable to a file (dump custom)
and examine what it defines.

Steve

4

Thank you for reply, I did what you told using following commands

variable bodies atomfile water_groups.txt
dump 1 all custom 1 dump.txt v_bodies
# delete_atoms group all_but_water
run 0

There is ony one differnce between running it as just rigid bodies (water) and with water+ other molecules, there is a 0 group in dump files which has all the other atoms.
I checked the body file using simple python script, it does not contain any single atom defined as group.

5

Thank you for reply, I did what you told using following commands

variable bodies atomfile water_groups.txt
dump 1 all custom 1 dump.txt v_bodies
# delete_atoms group all_but_water
run 0

There is ony one differnce between running it as just rigid bodies (water)
and with water+ other molecules, there is a 0 group in dump files which has
all the other atoms.
I checked the body file using simple python script, it does not contain
any single atom defined as group.

​perhaps your python script has a bug?
at any rate, without a simple/complete test example to reproduce​ the issue
and the confirmation, that this issue persists with the latest development
version of LAMMPS, there is not likely much more that we can do here.

axel.

6

Its just 13 lines of code, not much scope of bug! besides it works fine with just water. In any case I understand. Bit late here so will work tomorrow in more organized fashion to see if i can find the trigger, because very recently I was able to run similar file without fuss. In the mean if anyone wants to give a go, or perhaps confirm that problem is only at my end, I am attaching tar file with lammps script and data file i am using (its bit laarge, 26 MB hence the one drive link,[https://1drv.ms/u/s!Ave1F6LmCN-jgwJjsW5lxiJM_7d3] and sorry for the messy formatting!).

I ran it using Lammps 11 may 2018 version (git repo) usin lmp_serial binary.________________________________________

7

yes. but it is not accelerated right? no support for gpu and xeon phi. thats why i went for making tip4p model my self. the method to do that was discussed by axel last week only in mailing list