2022-05-17 Meeting notes

Updates from MolSSI

• BP: No server issues this week. Added many-body-expansion feature, qcf-fractal-next is almost near to release. Had working sessions with DD to change things in qcsubmit. Will be busy this week with admin stuff.

  • DD: Another session to take a look at few more improvements in qcsubmit.

  • BP: We can coordinate that.

Throughput status

• OpenFF Protein Capped 1-mer Sidechains v1.3 - 45/61 TDs

  • CC: Lee-Ping’s answer on gradients being large say that even if we start from random starting points we may not able to converge, let’s move this to end-of-life. Still an open question is whether we can use the partial torsion scans.

• SPICE PubChem Set 2 Single Points Dataset v1.2: 19 left (almost all are consistent errors)

• SPICE PubChem Set 3 Single Points Dataset v1.2: 31 left (almost all are consistent errors)

• SPICE PubChem Set 4 Single Points Dataset v1.2: 70787 from 8162, 52K more remaining. We can push set 5 to compute.

User questions/issues

• PB: Can we delete old records (or change status from ‘complete’) that contain the wrong Iodine auxiliary basis?

  • BP: Yeah, we can do that in next version, changing it to ‘INVALID’

• DD: @JH are there any features in qcsubmit that we shouldn’t cut down in lieu of qcfractal objects in next release?

  • JH: That should be fine, some of the deduplication and querying features were added since we couldn’t do them qcf side before.

  • BP: Yeah, there are lot of server-side features added now which would make it much faster.

  • JH: If we can shrink json blobs that’d be another great feature.

  • DD: Can you raise an issue with the feature request?

  • JH: Sure.