2022-04-05 QC Meeting notes

Updates from Molssi

• BP – Posted quick start instructions, posted on QCA slack channel. Added some new features for managing calculation IDs and datasets, added better ways to handle optimizations within one torsiondrive. But largely I think the features that I’ll add will come from your feedback.

• PB – DD set up a copy of the qca-dataset-submission repo to test how everything works. Seems to be going well.

• BP – The server will now automatically restart failed jobs. This can be configured based on the type of error, but most of our errors now are “unknown error”

•  

Infrastructure needs/advances

• DD has spun up a mirror of qca-dataset-submission repo to test compatiblity and features of qcfractal-next release,

• DD’s update: Now working through validation issues, which should just be incompatibilities between current QCSubmit and next QCPortal. Will make necessary changes to QCSubmit in next: https://github.com/openforcefield/openff-qcsubmit/tree/next. PRs of few test submissions are up, more testing this week.

Throughput status

• OpenFF Protein Capped 1-mer Sidechains v1.2: 39/46 TD

  • CC submitted a revised version with modified initial structures for 5

• SPICE PubChem Set 2 Single Points Dataset v1.2: From 84398 to 108311 calcs, slow progress this week due to scarce availability of PRP pods (reported by DD last week)

User questions/issues

Science support needs