compute pair/local command


compute ID group-ID pair/local value1 value2 ... keyword args ...
  • ID, group-ID are documented in compute command

  • pair/local = style name of this compute command

  • one or more values may be appended

  • value = dist or dx or dy or dz or eng or force or fx or fy or fz or p1 or p2 or …

    dist = pairwise distance
    dx,dy,dz = components of pairwise distance
    eng = pairwise energy
    force = pairwise force
    fx,fy,fz = components of pairwise force
    p1, p2, ... = pair style specific quantities for allowed N values
  • zero or more keyword/arg pairs may be appended

  • keyword = cutoff

    cutoff arg = type or radius


compute 1 all pair/local eng
compute 1 all pair/local dist eng force
compute 1 all pair/local dist eng fx fy fz
compute 1 all pair/local dist fx fy fz p1 p2 p3


Define a computation that calculates properties of individual pairwise interactions. The number of datums generated, aggregated across all processors, equals the number of pairwise interactions in the system.

The local data stored by this command is generated by looping over the pairwise neighbor list. Info about an individual pairwise interaction will only be included if both atoms in the pair are in the specified compute group, and if the current pairwise distance is less than the force cutoff distance for that interaction, as defined by the pair_style and pair_coeff commands.

The value dist is the distance between the pair of atoms. The values dx, dy, and dz are the \((x,y,z)\) components of the distance between the pair of atoms. This value is always the distance from the atom of higher to the one with the lower atom ID.

The value eng is the interaction energy for the pair of atoms.

The value force is the force acting between the pair of atoms, which is positive for a repulsive force and negative for an attractive force. The values fx, fy, and fz are the \((x,y,z)\) components of force on atom I. For pair styles that apply non-central forces, such as granular pair styles, these values only include the \((x,y,z)\) components of the normal force component.

A pair style may define additional pairwise quantities which can be accessed as p1 to pN, where \(N\) is defined by the pair style. Most pair styles do not define any additional quantities, so \(N = 0\). An example of ones that do are the granular pair styles which calculate the tangential force between two particles and return its components and magnitude acting on atom \(I\) for \(N \in \{1,2,3,4\}\). See individual pair styles for details.

When using pN with pair style hybrid, the output will be the Nth quantity from the sub-style that computes the pairwise interaction (based on atom types). If that sub-style does not define a pN, the output will be 0.0. The maximum allowed \(N\) is the maximum number of quantities provided by any sub-style.

When using pN with pair style hybrid/overlay the quantities from all sub-styles that provide them are concatenated together into one long list. For example, if there are 3 sub-styles and 2 of them have additional output (with 3 and 4 quantities, respectively), then 7 values (p1 up to p7) are defined. The values p1 to p3 refer to quantities defined by the first of the two sub-styles. Values p4 to p7 refer to quantities from the second of the two sub-styles. If the referenced pN is not computed for the specific pairwise interaction (based on atom types), then the output will be 0.0.

The value dist, dx, dy and dz will be in distance units. The value eng will be in energy units. The values force, fx, fy, and fz will be in force units. The values pN will be in whatever units the pair style defines.

The optional cutoff keyword determines how the force cutoff distance for an interaction is determined. For the default setting of type, the pairwise cutoff defined by the pair_style command for the types of the two atoms is used. For the radius setting, the sum of the radii of the two particles is used as a cutoff. For example, this is appropriate for granular particles which only interact when they are overlapping, as computed by granular pair styles. Note that if a granular model defines atom types such that all particles of a specific type are monodisperse (same diameter), then the two settings are effectively identical.

Note that as atoms migrate from processor to processor, there will be no consistent ordering of the entries within the local vector or array from one timestep to the next. The only consistency that is guaranteed is that the ordering on a particular timestep will be the same for local vectors or arrays generated by other compute commands. For example, pair output from the compute property/local command can be combined with data from this command and output by the dump local command in a consistent way.

Here is an example of how to do this:

compute 1 all property/local patom1 patom2
compute 2 all pair/local dist eng force
dump 1 all local 1000 tmp.dump index c_1[1] c_1[2] c_2[1] c_2[2] c_2[3]


For pairs, if two atoms I,J are involved in 1–2, 1–3, and 1–4 interactions within the molecular topology, their pairwise interaction may be turned off, and thus they may not appear in the neighbor list, and will not be part of the local data created by this command. More specifically, this will be true of I,J pairs with a weighting factor of 0.0; pairs with a non-zero weighting factor are included. The weighting factors for 1–2, 1–3, and 1–4 pairwise interactions are set by the special_bonds command. An exception is if long-range Coulombics are being computed via the kspace_style command, then atom pairs with weighting factors of zero are still included in the neighbor list, so that a portion of the long-range interaction contribution can be computed in the pair style. Hence in that case, those atom pairs will be part of the local data created by this command.

Output info

This compute calculates a local vector or local array depending on the number of keywords. The length of the vector or number of rows in the array is the number of pairs. If a single keyword is specified, a local vector is produced. If two or more keywords are specified, a local array is produced where the number of columns = the number of keywords. The vector or array can be accessed by any command that uses local values from a compute as input. See the Howto output page for an overview of LAMMPS output options.

The output for dist will be in distance units. The output for eng will be in energy units. The output for force, fx, fy, and fz will be in force units. The output for pN will be in whatever units the pair style defines.




The keyword default is cutoff = type.