Skip to content

feat: add ring planarity benchmark#681

Open
lwalew wants to merge 1 commit into
ddmms:mainfrom
lwalew:feat/add-ring-planarity
Open

feat: add ring planarity benchmark#681
lwalew wants to merge 1 commit into
ddmms:mainfrom
lwalew:feat/add-ring-planarity

Conversation

@lwalew

@lwalew lwalew commented Jul 8, 2026

Copy link
Copy Markdown

Pre-review checklist for PR author

PR author must check the checkboxes below when creating the PR.

Summary

Migrates the ring planarity benchmark from the MLIP Audit suite. For each of a set of small organic molecules with aromatic rings, an NVT molecular dynamics simulation is run at 300 K starting from a QM-optimised reference geometry (selected from QM9), and the deviation of the ring atoms from their best-fit plane is measured over the trajectory. Aromatic rings are planar, so a lower average deviation is better. The mean planarity deviation per model is reported in the metrics table, and the distribution of sampled deviations is shown as a histogram.

This is the first MLIP Audit migration to include the shared mlipaudit wiring (the mlipaudit optional extra and git source in pyproject.toml, the mlip_audit framework badge, and ml_peg/calcs/utils/mlipaudit.py), which overlaps with the bond length distribution PR (#666) and Leon's PRs (#660, #644); whichever lands first introduces this wiring and the others can drop it.

Linked issue

Resolves #680

Progress

  • Calculations
  • Analysis
  • Application
  • Documentation

The benchmark input data still needs to be uploaded to the ML-PEG S3 bucket at inputs/molecular_dynamics/ring_planarity/ring_planarity.zip (layout ring_planarity/ring_planarity_data.json) before it can run end-to-end.

Testing

Not yet run end-to-end. Style checks (pre-commit) pass locally.

New decorators/callbacks

None required.

@ElliottKasoar ElliottKasoar added the new benchmark Proposals and suggestions for new benchmarks label Jul 8, 2026
@lwalew lwalew force-pushed the feat/add-ring-planarity branch from 7f4a5dc to 0b74b47 Compare July 8, 2026 11:22
Computational cost
------------------

High: one MD simulation per molecule, each 1,000,000 steps. Faster inference can be achieved

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

gpu estimate?

@joehart2001

Copy link
Copy Markdown
Collaborator

Thanks for the pr! same as #658 could you send the data + have a look at the element filtering?

docs: https://ddmms.github.io/ml-peg/developer_guide/filter.html

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

new benchmark Proposals and suggestions for new benchmarks

Projects

None yet

Development

Successfully merging this pull request may close these issues.

New benchmark Ring Planarity

3 participants