run command


run N keyword values ...
  • N = # of timesteps

  • zero or more keyword/value pairs may be appended

  • keyword = upto or start or stop or pre or post or every

    upto value = none
    start value = N1
      N1 = timestep at which 1st run started
    stop value = N2
      N2 = timestep at which last run will end
    pre value = no or yes
    post value = no or yes
    every values = M c1 c2 ...
      M = break the run into M-timestep segments and invoke one or more commands between each segment
      c1,c2,...,cN = one or more LAMMPS commands, each enclosed in quotes
      c1 = NULL means no command will be invoked


run 10000
run 1000000 upto
run 100 start 0 stop 1000
run 1000 pre no post yes
run 100000 start 0 stop 1000000 every 1000 "print 'Protein Rg = $r'"
run 100000 every 1000 NULL


Run or continue dynamics for a specified number of timesteps.

When the run style is respa, N refers to outer loop (largest) timesteps.

A value of N = 0 is acceptable; only the thermodynamics of the system are computed and printed without taking a timestep.

The upto keyword means to perform a run starting at the current timestep up to the specified timestep. E.g. if the current timestep is 10,000 and “run 100000 upto” is used, then an additional 90,000 timesteps will be run. This can be useful for very long runs on a machine that allocates chunks of time and terminate your job when time is exceeded. If you need to restart your script multiple times (reading in the last restart file), you can keep restarting your script with the same run command until the simulation finally completes.

The start or stop keywords can be used if multiple runs are being performed and you want a fix command that changes some value over time (e.g. temperature) to make the change across the entire set of runs and not just a single run. See the doc page for individual fixes to see which ones can be used with the start/stop keywords.

For example, consider this fix followed by 10 run commands:

fix         1 all nvt 200.0 300.0 1.0
run         1000 start 0 stop 10000
run         1000 start 0 stop 10000
run         1000 start 0 stop 10000

The NVT fix ramps the target temperature from 200.0 to 300.0 during a run. If the run commands did not have the start/stop keywords (just “run 1000”), then the temperature would ramp from 200.0 to 300.0 during the 1000 steps of each run. With the start/stop keywords, the ramping takes place over the 10000 steps of all runs together.

The pre and post keywords can be used to streamline the setup, clean-up, and associated output to the screen that happens before and after a run. This can be useful if you wish to do many short runs in succession (e.g. LAMMPS is being called as a library which is doing other computations between successive short LAMMPS runs).

By default (pre and post = yes), LAMMPS creates neighbor lists, computes forces, and imposes fix constraints before every run. And after every run it gathers and prints timings statistics. If a run is just a continuation of a previous run (i.e. no settings are changed), the initial computation is not necessary; the old neighbor list is still valid as are the forces. So if pre is specified as “no” then the initial setup is skipped, except for printing thermodynamic info. Note that if pre is set to “no” for the very 1st run LAMMPS performs, then it is overridden, since the initial setup computations must be done.


If your input script changes the system between 2 runs, then the initial setup must be performed to insure the change is recognized by all parts of the code that are affected. Examples are adding a fix or dump or compute, changing a neighbor list parameter, or writing restart file which can migrate atoms between processors. LAMMPS has no easy way to check if this has happened, but it is an error to use the pre no option in this case.

If post is specified as “no”, the full timing summary is skipped; only a one-line summary timing is printed.

The every keyword provides a means of breaking a LAMMPS run into a series of shorter runs. Optionally, one or more LAMMPS commands (c1, c2, ..., cN) will be executed in between the short runs. If used, the every keyword must be the last keyword, since it has a variable number of arguments. Each of the trailing arguments is a single LAMMPS command, and each command should be enclosed in quotes, so that the entire command will be treated as a single argument. This will also prevent any variables in the command from being evaluated until it is executed multiple times during the run. Note that if a command itself needs one of its arguments quoted (e.g. the print command), then you can use a combination of single and double quotes, as in the example above or below.

The every keyword is a means to avoid listing a long series of runs and interleaving commands in your input script. For example, a print command could be invoked or a fix could be redefined, e.g. to reset a thermostat temperature. Or this could be useful for invoking a command you have added to LAMMPS that wraps some other code (e.g. as a library) to perform a computation periodically during a long LAMMPS run. See this section of the documentation for info about how to add new commands to LAMMPS. See this section of the documentation for ideas about how to couple LAMMPS to other codes.

With the every option, N total steps are simulated, in shorter runs of M steps each. After each M-length run, the specified commands are invoked. If only a single command is specified as NULL, then no command is invoked. Thus these lines:

variable q equal x[100]
run 6000 every 2000 "print 'Coord = $q'"

are the equivalent of:

variable q equal x[100]
run 2000
print "Coord = $q"
run 2000
print "Coord = $q"
run 2000
print "Coord = $q"

which does 3 runs of 2000 steps and prints the x-coordinate of a particular atom between runs. Note that the variable “$q” will be evaluated afresh each time the print command is executed.

Note that by using the line continuation character “&”, the run every command can be spread across many lines, though it is still a single command:

run 100000 every 1000 &
  "print 'Minimum value = $a'" &
  "print 'Maximum value = $b'" &
  "print 'Temp = $c'" &
  "print 'Press = $d'"

If the pre and post options are set to “no” when used with the every keyword, then the 1st run will do the full setup and the last run will print the full timing summary, but these operations will be skipped for intermediate runs.


You might wish to specify a command that exits the run by jumping out of the loop, e.g.

variable t equal temp
run 10000 every 100 "if '$t < 300.0' then 'jump SELF afterrun'"

However, this will not work. The run command simply executes each command one at a time each time it pauses, then continues the run.

Instead, you should use the fix halt command, which has additional options for how to exit the run.


When not using the upto keyword, the number of specified timesteps N must fit in a signed 32-bit integer, so you are limited to slightly more than 2 billion steps (2^31) in a single run. When using upto, N can be larger than a signed 32-bit integer, however the difference between N and the current timestep must still be no larger than 2^31 steps.

However, with or without the upto keyword, you can perform successive runs to run a simulation for any number of steps (ok, up to 2^63 total steps). I.e. the timestep counter within LAMMPS is a 64-bit signed integer.


The option defaults are start = the current timestep, stop = current timestep + N, pre = yes, and post = yes.