2023-10-23 Meeting notes

General updates

• MT – I’m offline next Monday. Remove from calendar?

  • (General) – Yes

• LW – Comparing sage scripts and benchmarks. Differences look small/comparable, especially now that InChI has been solved.

  • MT – Did you just solve this today?

  • LW – No, I just ignored that data in the comparison and it looked good.

  • MT – Great, I’ll close that ticket.

•  

Trello

https://trello.com/b/dzvFZnv4/infrastructure?filter=label:benchmarking

• Isomorphism issue

  • Solution in #8 should work, though removing it would probably be better long-term as there’s no real benefit to InChi round-tripping

  • (rambling conversation about graph molecule standardization and representations, but not actionable in this context)

  • MT added a comment saying that this code isn’t necessary for “correct” behavior and so it’s being turned optional.

• MT – Solvation free energies update?

  • LW – No update

  • MT – Will assume this is still in purgatory.

  • .

  •  

• MT – Project-wise, how is this looking? Is progress good? Features are useful? We’d actually use it?

  • LW – I think this is going well. There’s a good cycle of user feedback that gives me confidence. In comparing this to sage benchmarking scripts this seems to overcome a lot of deficiencies that I saw there. So this very much isn’t the bottleneck for future FF releases.

  • MT – I am wondering how stuff like Rosemary benchmarks would fit in.

  • JW – My understanding is that the right way to prepare for this is to build toward the Evaluator interface. Chapin is working on biopolymer benchmarks in his own repo now, where he is making big changes. But in the future we’ll have him (and possibly MT) work together on making the biopolymer benchmarks into an Evaluator target.

  •  

  •  

  • .