2022-07-19 QC meeting notes

Updates from MolSSI

Infrastructure advances

• JW – DD pinged me and I updated our PRP config, so that should be available again.

  • DD – Started up PRP workers - 45 x 16 cores

• CC – Updates on LPW’s fix for geometric issues?

  • DD – He still needs to cut a release, and we'd told him it’s not urgent.

    • https://openforcefieldgroup.slack.com/?redir=%2Farchives%2FC011Z72DFH8%2Fp1652477976388109

  • CC – Gotcha. We’ll want to check that this really solves the issue before we ask for a release.

  • DD – It may not be the best use of your time to test this out. This only affected a small part of the dataset and it’s not clear that this is the last problem that we’ll have, so we may still have the same high-level problem (torsiondrives not having all optimizations complete) even if we fix this.

  • CC – Agree

  • DD – I’ll restart this conversation on Slack to see what LPW’s position is.

  •  

Throughput status

• OpenFF Protein Capped 3-mer Backbones v1.0 - 2/54 TDs complete. 220051 from 196000 opts in last week.

  • CC + JW – This is looking good. Very cool that they’re completing.

• RNA Single Point Dataset v1.0: 4489 calculations x two QC specifications (our default, and wb97m-d3bj/def2-tzvppd)

  • default - 3479/8978 (2 * 4489 since functional and dispersion calcs were split)

  • wb97m-d3bj/def2-tzvppd - 0/4489

  • PB – if we give enough workers like spice most of the jobs will be done in a day

    • DD – Sounds good - Could you re-tag those to compute-openff-rna and I'll spin up some workers on Lilac to take them.

    • PB – Yes

User questions/issues

Science support needs