2020-12-18 QCA Submission Meeting notes

Date

Dec 18, 2020

New advancements
- STANDARDS-based versioning #137
- Heavy basis sets!
- Month basis sets!
New submissions
- PB: Genentech OptimizationDataset (ready for review)
- Dataset consolidation for Hyesu benchmark sets
Upcoming infrastructure improvements
- Dataset index on qca-dataset submission #147
- Stale jobs fix on QCFractal server/managers
Upcoming science support
- Enforced c1 symmetry in psi4 is almost ready
Larger advances
- Automated FF coverage gap identification, torsion prioritization, submission generation
- Benchmarking (dashboard, etc.)

Item	Presenter	Notes

Item	Presenter	Notes
MolSSI compute	Ben	Do y’all need more compute from MolSSI? DD: not at this time
STANDARDSv3	David	DD: can lay out implementation components in January due to holiday
Heavy and month basis sets	Josh	DD: what are these seasonal basis sets? BP: monthly datasets started with “aug” for “augmented”, and then folks started reducing some of the components of that “aug” set and called them “jul”, “jun”, JH: not going to try and support all variations, because it gets really messy real fast; taking the approach of adding support when it’s asked for
Torsiondrive benchmark consolidation	Josh	JH: PR is up! Ready for review. DD: Added Hyesu and myself as reviewer
Dataset index	Josh	Ready to go! DD: Merged
Grid settings	Josh	SB’s dataset has one result that failed to converge; probably needs a finer grid spec DD: should we re-roll with double the grid density for SCF? BP: not sure if it would actually help with an SCF convergence error thinking about ways we could introduce more of the STDOUT into the provenance DD: in discussions with partners, they’re interested in keeping file output from a calculation from the manager basically, telling the manager not to clean up after itself BP: may be an option in QCEngine related to this JH: will play with this one result and see if changing grid settings yields any different result