Generate geometrically-spaced dump sequences for LAMMPS simulations of glassy systems.
This tool creates the timing files needed for LAMMPS loops and visualizes the resulting dump schedule.
Required Parameters
timestep (dt) [LJ]?
The simulation timestep in LJ (Lennard-Jones) time units. This is the fundamental time resolution of your molecular dynamics simulation. Typical values range from 0.001 to 0.005 LJ.
tau_equil [LJ]?
The equilibration time - the duration of the long sequence during which the system equilibrates. The long sequence will target ending near this time. Typically 10-100x tau_alpha. You can use expressions like 100*tau_alpha or 20*tau_alpha.
tau_alpha [LJ]?
The alpha relaxation time - the characteristic timescale for structural relaxation in glassy systems. This is typically measured from the decay of self-intermediate scattering functions or mean-squared displacement. Used to set the geometric ratio and short sequence length.
total_time_long_sequence [LJ]?
Target end time for the long sequence. Defaults to tau_equil. The optimal number of terms (n) is auto-selected to get the final dump as close to this value as possible. You can use expressions like 0.5*tau_equil or 20*tau_alpha.
total_time_short_sequence[LJ]?
Total duration of all short sequences combined. Defaults to 100*tau_alpha. This determines how long after equilibration you continue recording. Divided by the short sequence length to get the number of repeats.
long_sequence_filename
?
Filename pattern for long-sequence trajectory files. The default traj<n>.dcd matches the files opened by 2_time_aging.py. Use <n> as the file-number placeholder.
short_sequence_filename
?
Filename pattern for short sequence trajectory files. Use <n> as placeholder for the file number. Examples: short<n>.dcd, prod_<n>.dcd, traj_short_<n>.dcd
[+] Advanced Settings (not recommended to change)
dumps_per_short_sequence?
Target number of dumps per short sequence. Due to rounding (floor function), the actual number may be slightly less. Default of 27 typically yields ~25 actual dumps. Higher values give finer time resolution but more data.
length_of_short_sequence[LJ]?
Duration of each individual short sequence. Leave blank to use 0.90*tau_alpha, or 0.90*tau_rot when tau_rot is supplied. An explicitly entered value always takes precedence.
tau_rot [LJ, optional]?
Optional and unused by default. Enter the rotational relaxation time as a number or an expression such as 0.25*tau_alpha to add it to the resolution plot. When this field is present and length_of_short_sequence is blank, that length defaults to 0.90*tau_rot. Once defined, tau_rot can also be used in other schedule expressions such as 0.9*tau_rot.
Generated Commands
Run these commands with geo-rev11.py to generate the LAMMPS loop files, or download them directly below.
Total number of particles/atoms in your simulation. Used to estimate file sizes based on DCD trajectory format.
Note: This estimates the size of trajectory files (.dcd) that LAMMPS will generate during your simulation - NOT the download buttons below.
DCD format: 12 bytes per particle per frame (3 floats x 4 bytes for x, y, z)
~100 bytes header overhead per single-frame DCD file
Formula: size = frames * ((12 * particles) + 100)
Sequence
Dumps
Timestep Range
Time Range (LJ)
Est. Trajectory Size
Long (Equilibration)
--
--
--
--
Short (Production)
--
--
--
--
Total
--
--
Analysis Script Parameters
--
Analysis Command Examples
Generate a schedule to fill the cyclic analysis command examples.
2-time cyclic command
--
S4 cyclic command
--
Download LAMMPS Loop Files
These downloads contain small text files (~KB) for use in LAMMPS loops: dump_times.txt (when to dump),
run_times.txt (how long to run between dumps), and file_names.txt (output trajectory names).
Visualization
Preview the dump schedule that will be used in your LAMMPS simulation.
Dump Times
On log scale, geometrically spaced points appear linearly spaced.
Run Times (Intervals)
Resolution at tau_alpha
Available time separations (lags) that the short-sequence schedule provides.
Red: tau_alpha. Purple, when tau_rot is entered: rotational relaxation time.
Filled regions bracket each target using its nearest available lag on either side. Dashed outlines mark the widest unsampled interval from 0.8 to 1.2 times that relaxation time; smaller local worst gaps mean denser useful information around the target.
--
Compare storage-aware schedules
This comparison does not change the generated schedule. It searches repeated, identical short sequences whose combined duration stays within 1% of the requested production length. The storage allowance is relative to the current schedule's estimated short-production DCD size; an aging run's long sequence would still be regenerated normally.
Candidate cycle lengths range from 0.5 times the smaller relaxation time to 1.25 times the larger one, with sampling-density goals from 5 through 80 frames per short sequence. Each reported option includes its own Resolution at tau_alpha plot and a direct schedule ZIP; aging exports regenerate the matching long sequence for that option.
Generate a schedule with tau_rot to compare storage-aware alternatives.
LAMMPS Loop
Changelog
July 28, 2026 — unified analysis commands
Cyclic examples now use 2_time.py and s4_tb_fft_ww.py. Supplying -c selects cyclic mode in 2_time.py; supplying -G selects it in s4_tb_fft_ww.py. Existing schedules and DCD files do not need to be regenerated; update the analysis command names.
Generated analysis commands now include run-level parallelism and both filename prefixes. Each example uses -P <n_runs>, and the two-time command now receives the short-file prefix through -n. Request approximately one analysis worker per independent run, and replace <n_runs> with the same value for -r and -P.
The analysis presentation and restart-tolerant publish steps were synchronized. Linear and cyclic analysis are now explained as modes of the unified scripts, and temporary restart/DCD files are published with /bin/mv -f. Newly downloaded preempt-tolerant templates safely replace a stale destination only at the final publish step.
July 20, 2026 — rotational resolution and storage-aware schedules
tau_rot is now an optional Advanced Setting. Time expressions can use it, including values such as 0.9*tau_rot. When supplied, a blank short-sequence length defaults to 0.90*tau_rot; when omitted, the existing 0.90*tau_alpha default is unchanged.
The resolution plot can now show both tau_alpha and tau_rot. It reports the widest local lag gap from 0.8 to 1.2 times each relaxation time so sparse coverage is visible even when a lag lands exactly on the target.
An optional comparison balances local resolution against storage. It searches schedules within 1% of the requested production length, applies a percentage storage allowance, plots each reported option, and provides a direct schedule ZIP. The comparison appears only when tau_rot is supplied. Downloading an option does not replace the active schedule; aging downloads regenerate that option's matching long sequence.
July 2026 — denser long schedules and restart-ready exports
Long trajectories now default to traj<n>.dcd. This matches 2_time_aging.py; short production files remain short<n>.dcd. Existing simulation files are not renamed. Continue an existing run with its original schedule and names.
The final long runtime now matches the final short-sequence runtime. The long region is denser than the former tau_alpha/2 endpoint. Line counts, seamline, and the first short-file index can change. Regenerate and download all three schedule files together.
Long+short exports now include Legacy and Preempt-tolerant LAMMPS templates. Legacy is for uninterrupted runs; the preempt-tolerant template documents the required LAMMPS and Slurm restart integration.
Stitched ZIPs now contain two .in files. The three schedule-text formats are unchanged, although regenerated schedules can have different contents. No Python analysis script needs to be changed.