pair_style pace command

Accelerator Variants: pace/kk


pair_style pace ... keyword values ...
  • one or more keyword/value pairs may be appended

    keyword = product or recursive or chunksize
      product = use product algorithm for basis functions
      recursive = use recursive algorithm for basis functions
      chunksize value = number of atoms in each pass


pair_style pace
pair_style pace product chunksize 2048
pair_coeff * * Cu-PBE-core-rep.ace Cu


Pair style pace computes interactions using the Atomic Cluster Expansion (ACE), which is a general expansion of the atomic energy in multi-body basis functions. (Drautz). The pace pair style provides an efficient implementation that is described in this paper (Lysogorskiy).

In ACE, the total energy is decomposed into a sum over atomic energies. The energy of atom i is expressed as a linear or non-linear function of one or more density functions. By projecting the density onto a local atomic base, the lowest order contributions to the energy can be expressed as a set of scalar polynomials in basis function contributions summed over neighbor atoms.

Only a single pair_coeff command is used with the pace style which specifies an ACE coefficient file followed by N additional arguments specifying the mapping of ACE elements to LAMMPS atom types, where N is the number of LAMMPS atom types:

  • ACE coefficient file

  • N element names = mapping of ACE elements to atom types

Only a single pair_coeff command is used with the pace style which specifies an ACE file that fully defines the potential. Note that unlike for other potentials, cutoffs are not set in the pair_style or pair_coeff command; they are specified in the ACE file.

The pair_style pace command may be followed by the optional keyword product or recursive, which determines which of two algorithms is used for the calculation of basis functions and derivatives. The default is recursive.

The keyword chunksize is only applicable when using the pair style pace with the KOKKOS package on GPUs and is ignored otherwise. This keyword controls the number of atoms in each pass used to compute the atomic cluster expansion and is used to avoid running out of memory. For example if there are 8192 atoms in the simulation and the chunksize is set to 4096, the ACE calculation will be broken up into two passes (running on a single GPU).

See the pair_coeff page for alternate ways to specify the path for the ACE coefficient file.

Mixing, shift, table, tail correction, restart, rRESPA info

For atom type pairs I,J and I != J, where types I and J correspond to two different element types, mixing is performed by LAMMPS with user-specifiable parameters as described above. You never need to specify a pair_coeff command with I != J arguments for this style.

This pair style does not support the pair_modify shift, table, and tail options.

This pair style does not write its information to binary restart files, since it is stored in potential files. Thus, you need to re-specify the pair_style and pair_coeff commands in an input script that reads a restart file.

This pair style can only be used via the pair keyword of the run_style respa command. It does not support the inner, middle, outer keywords.

Styles with a gpu, intel, kk, omp, or opt suffix are functionally the same as the corresponding style without the suffix. They have been optimized to run faster, depending on your available hardware, as discussed on the Accelerator packages page. The accelerated styles take the same arguments and should produce the same results, except for round-off and precision issues.

These accelerated styles are part of the GPU, INTEL, KOKKOS, OPENMP, and OPT packages, respectively. They are only enabled if LAMMPS was built with those packages. See the Build package page for more info.

You can specify the accelerated styles explicitly in your input script by including their suffix, or you can use the -suffix command-line switch when you invoke LAMMPS, or you can use the suffix command in your input script.

See the Accelerator packages page for more instructions on how to use the accelerated styles effectively.


This pair style is part of the ML-PACE package. It is only enabled if LAMMPS was built with that package. See the Build package page for more info.


recursive, chunksize = 4096

(Drautz) Drautz, Phys Rev B, 99, 014104 (2019).

(Lysogorskiy) Lysogorskiy, van der Oord, Bochkarev, Menon, Rinaldi, Hammerschmidt, Mrovec, Thompson, Csanyi, Ortner, Drautz, TBD (2021).