Restart command

LAMMPS LAMMPS General Discussion
1

Dear LAMMPS Expert,

I hope this message finds you well. I have a couple of questions regarding the use of the restart command:

  1. Are force field parameters (such as Lennard-Jones parameters and bond coefficients) stored in the restart file?
  2. When running simulations on a supercomputer using the restart command, do additional configurations need to be set? I have successfully run my restart file on my personal computer, but encountered errors on the supercomputer related to unset pair coefficients, despite having already defined them.

I appreciate any insights or guidance you can provide on these matters.

Best regards,

Bahman

2

Please study the LAMMPS manual, and particularly the documentation for the restart and read_restart commands. All the answers are there.

3

Are you using the same LAMMPS version on your personal computer and on the supercomputer?

Simon

5

Where?

https://docs.lammps.org/pair_hybrid.html#mixing-shift-table-tail-correction-restart-rrespa-info

6

Yes, I used the same version of LAMMPS both to create and read the restart file. I also reviewed the documentation detailing the information included (and not included) in the restart file, confirming that force field styles and coefficients are supposed to be included. However, upon running the simulation again, I encountered the following error:

ERROR: All pair coeffs are not set (…/pair.cpp:231)

Here is the hybrid pair style configuration I used:

variable epsilon_Si equal 0.235
variable sigma_Si equal 2.23
variable Rc_Si equal “v_sigma_Si2.5"
variable epsilon_N equal 0.0017
variable sigma_N equal 3.250
variable Rc_N equal "v_sigma_N2.5”

variable epsilon_SiN equal “sqrt(v_epsilon_Siv_epsilon_N)"
variable sigma_SiN equal "0.5(v_sigma_Si+v_sigma_N)”
variable Rc_SiN equal “v_sigma_SiN*2.5”

pair_style hybrid tersoff lj/cut 8.3
pair_coeff * * tersoff Si.tersoff Si NULL
pair_coeff 2 2 lj/cut v_epsilon_N v_sigma_N v_Rc_N
pair_coeff 1 2 lj/cut v_epsilon_SiN v_sigma_SiN v_Rc_SiN
neighbor 0.3 bin

7

What you see is documented behavior. You have to improve your skills in reading the documentation and paying more attention to details.

Also, after so many posts and seeing how often we complain about people not properly quoting text, you should have learned how to do it and know that you will be singled out for not following the guidelines.

This is not correct. Please see the read_restart and pair style hybrid documentation. For the latter, I already provided the link to the doc section with restart info where it explicitly says that the pair_style command line is stored but the pair_coeff commands not.

The read_restart docs say:

force field coefficients (pair, bond, angle, etc) in some cases (see below)

and

Note that some force field styles (pair, bond, angle, etc) do not store their coefficient info in restart files. Typically these are many-body or tabulated potentials which read their parameters from separate files. In these cases you will need to re-specify the pair_coeff, bond_coeff, etc commands in your restart input script. The doc pages for individual force field styles mention if this is the case. This is also true of pair_style hybrid (bond hybrid, angle hybrid, etc) commands; they do not store coefficient info.

It cannot get more explicit than that.

1 Like
8

Thanks a lot Axel, It was my personal mistake.