[lammps-users] "FENE bond too long" warnings after the "Total wall time" line in log file

LAMMPS LAMMPS Mailing List Mirror
1

Dear Lammps community,

I simulate a long bead-on-spring (FENE+WCA) chain (2000 monomers) in a crowded (WCA beads) radially confinement. I use

fix F_nve all nve/limit 0.01
fix F_langevin all langevin 1.0 1.0 10.0 ${randseed}

for free energy minimization with dt=0.01 for 1e6 steps. As the initial conformation of the polymer, I placed the monomers on a helix in the input data file. In this way, the chain is small enough and the data file can be used as an input or a large set of input parameters. During the equilibration, I expect to get some “FENE bond too long” warning due to this unrealistic initial condition.

I run more than 150 long simulation with the same data file but different set of parameter (except # of monomers). All of these simulations are finished successfully without any dangerous neighbour build lists.

However, I have seen several lines of “FENE bond too long” warnings at ends of several the log files (less than 20 logs), bellow the “Wall time” line. I read the previous discussions on the “FENE bog too long” warning and realized that Lammps print the warnings after the thermo outputs. Regarding this in mind, I have some questions and one suggestion:
Are these warnings about some timesteps in the equilibration phase of my simulation?

At the end of each simulation, I use the write_data command to dump a data file. I get the following warning about this command: “WARNING: Not all mixed pair coeffs generated from mixing. Use write_data with ‘pair ij’ option to store all pair coeffs. (src/write_data.cpp:312)”. Since the “FENE bond too long” warnings are printed after this command and its warning, is it possible that they are related to the write_data command?

My suggestion is this: it would be great to have the tilmestep at which the warning or error is reported.

Please see the attached, log (the warnings are at the end of this file), data, and input files.
I use Lammps version 29 Oct 2020.

Thank you in advance for your help.
Kind regards,
Amir

input.lmp (4.61 KB)

output.txt (555 KB)

chain.2000.data (199 KB)

2

Dear Lammps community,

[…]

However, I have seen several lines of “FENE bond too long” warnings at ends of several the log files (less than 20 logs), bellow the “Wall time” line. I read the previous discussions on the “FENE bog too long” warning and realized that Lammps print the warnings after the thermo outputs. Regarding this in mind, I have some questions and one suggestion:
Are these warnings about some timesteps in the equilibration phase of my simulation?

as can be gathered from the source code (hint, hint: the warning message prints the name of the source file and the line in that source file where you would need to look)the timestep number is the first number printed, the next two at the atom ids and the fourth number is the distance between them.

At the end of each simulation, I use the write_data command to dump a data file. I get the following warning about this command: “WARNING: Not all mixed pair coeffs generated from mixing. Use write_data with ‘pair ij’ option to store all pair coeffs. (src/write_data.cpp:312)”. Since the “FENE bond too long” warnings are printed after this command and its warning, is it possible that they are related to the write_data command?

no. they must be buffered by your MPI library and are probably from a processor different from MPI rank 0 which does normally all output.
if anything, this is a problem of the MPI library. output from “remote” processes is normally block buffered to avoid negative impact to parallel performance.

My suggestion is this: it would be great to have the tilmestep at which the warning or error is reported.

that is already the case.

axel.

3

Oops! I forgot to look at the source code for more information. I presumed that the warning is all the information I have/need and the the first number is the bond number.

Thanks for your support.
Amir