ERROR on proc 0: Rigid body atoms 19 20 missing on proc 0 at step 1

LAMMPS LAMMPS General Discussion
1

Dear all,

I have LAMMPS (29 Aug 2024 - Update 1) version installed with CUDA 12.2. I am using the TIP4P water molecule model to simulate the surface tension of liquid films. I designed a simulation domain with a size of LxLyLz=100Å30Å30Å. When we use the “implicit.txt” and “TIP3P-Ewald.txt” in the attachment to implicitly construct the tip4p water molecules, the simulation works fine.

However, the reference we want to reproduce uses Ewald sums to deal with the electrostatic interactions. According to the LAMMPS Manual: The pppm/tip4p style is identical to the pppm style except that it adds a charge at the massless fourth site in each TIP4P water molecule. It should be used with pair styles with a tip4p/long in their style name.

We think that we cannot use pair_style lj/cut/tip4p/long and kspace ewald at the same time. So we refer to the manual and try to use the explicit template of TIP4P. The input file is “implicit.txt” and “TIP4P.txt" in the attachment.

But in the calculation, the simulation has an error in the first step. ERROR on proc 0: Rigid body atoms 19 20 missing on proc 0 at step 1. Based on the previous experience of bond loss, I adjusted the number of molecules appropriately. But until I set only one water molecule in the simulation domain, the error ERROR: Lost atoms: original 4 current 0 (src/thermo.cpp:494) will appear.

I tried to reduce the time step or change neigh_modify. But it did not solve the problem. I would like to ask if any scholars have encountered similar problems and successfully solved them?

Many thanks in advance for any guidance on this issue

Deyang

explicit.txt (3.1 KB)
implicit.txt (3.2 KB)
TIP3P-Ewald.txt (612 Bytes)
TIP4P.txt (383 Bytes)

2

Ewald summation and PPPM are equivalent. The main issue with the Ewald kspace style is its bad scaling with system size (O(N^(2/3)) vs. O(N log(N)). You could easily confirm that by setting up a TIP3P water geometry and compare the forces between the different kspace styles while everything else is exactly the same.

3

I understand. Maybe there is something wrong with my reproduction process, which leads to different results from the reference. I will continue to try. Thank you very much for your reply and help!