ACE grid computes with validation examples #5
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary
This PR introduces ACE descriptor grid computes, similar to the way descriptor grid computes were introduced for SNAP. It allows one to evaluate ACE descriptors at points on a grid defined in LAMMPS. The ACE descriptor values are calculated based on the grid and atomic neighbors within user-specified cutoffs, with ACE descriptors defined up to some user-specified body order. This is analogous to the implementation for the grid implementation of SNAP descriptors.
It also updates the relevant
pair_pace
and CMake dependencies to a more recent version of thelammps-user-pace
library from ICAMS. Examples were included that validate the ACE grid computes in a similar way that the SNAP grid computes were validated. Log files have been provided for both serial and parallel tests of the grid computes.Related Issue(s)
N/A
Author(s)
James Goff (Sandia National Laboratory)
Email: [email protected]
Licensing
By submitting this pull request, I agree, that my contribution will be included in LAMMPS and redistributed under either the GNU General Public License version 2 (GPL v2) or the GNU Lesser General Public License version 2.1 (LGPL v2.1).
Backward Compatibility
N/A
Implementation Notes
The ACE per-atom computes were added to the MALA branch of LAMMPS, and may now be used as described in the LAMMPS development branch docs. For example, a user must provide a '.yace' file that contains the generalized Clebsch-Gordan coefficients and the ACE descriptor hyperparameters.
General syntax:
compute <computename> <group> pace <coeff_and_parameter_file.yace> bikflag dgradflag
Example syntax:
compute pace all pace coupling_coefficients.yace 1 0
The local and global grid computes are added for ACE descriptors similarly and may be used, respectively, with:
compute <computename> <group> pace/grid/local grid <ngrida> <ngridb> <ngridc> <coeff_and_parameter_file.yace>
compute mygridlocal all pace/grid/local grid 4 4 4 coupling_coefficients.yace
Further examples on how to use them (and expected outputs) may be found with the SNAP grid computes in
examples/snap
.Post Submission Checklist
Further Information, Files, and Links
As with other ACE computes and ACE potentials, the
coupling_coefficient.yace
files that contain the generalized Clebsch-Gordan coefficients and the ACE descriptor hyperparameters must be generated outside of LAMMPS. Examples on how to generate these for atom-centered ACE may be found in FitSNAP.The ACE grid computes are set up such that grid points always adopt
type 1
in LAMMPS andmu0 = 0
in the ACE basis (which correspond to the first element type). Depending on the desired use cases for the ACE grid compute, it may be desirable for the user to prune anymu0 != 0
ACE descriptors from the<coeff_and_parameter_file.yace>
file. (Doing so will reduce memory overhead for multielement ACE grid computes by preventing allocation for atom-centered ACE descriptor data.)