2023-10-24 alchemiscale Working Group meeting notes

Participants

  • @Jeffrey Wagner

  • @Irfan Alibay

  • Ian Kenney

  • @David Dotson

  • Meghan Osato

  • @David W.H. Swenson

  • @James Eastwood

  • @Mike Henry

  • @Richard Gowers

  • Jenke Scheen

Goals

  • alchemiscale.org user group

    • user questions / issues / feature requests

    • compute resources status

    • current stack versions:

      • alchemiscale: 0.2.1

      • gufe: 0.9.4

      • openfe: 0.13.0

      • perses: protocol-neqcyc

      • openmmforcefields: 0.12.0

  • openmmforcefields 0.12.0 deployed to alchemiscale.org under build v0.2.1-2

    • should allow for custom OFFXML in FF fields for Protocol Settings

  • IP : Protein-ligand benchmarks working group update

  • IP : feflow development:

  • alchemiscale development : sprint ends today: next sprint spans 10/25 - 11/6

    • architecture overview : PL Benchmarks on FAH - Architecture v6.drawio

    • coordination board : alchemiscale : Phase 3 - Folding@Home, new features, optimizations, targeted refactors

      • call for volunteers for available issues

    • alchemiscale 0.3.0 milestone:

    • alchemiscale-fah: 0.1.0 milestone

    • review Completed cards

    • updates on In Review, In Progress, and Available cards

Discussion topics

Notes

Notes

  • alchemiscale.org user group

    • user questions / issues / feature requests

      • (Lots of jobs in flight, OpenFE

    • compute resources status

      • ~200 compute services running on NRP, compute running on lilac. A few jobs getting rerun but the queue is dwindling. So room for more submissions.

      • MH – On lilac, do you filter by host? Because the GPUs are heterogenous, so I wonder if some jobs are landing on 1080s and others on newer cards.

        • DD – Good Q… looks like we’re submitting to all GPU nodes excepting the LT16 nodes.

    • current stack versions:

      • alchemiscale: 0.2.1

      • gufe: 0.9.4

      • openfe: 0.13.0

        • MH – 0.14 release pending. Troubleshooting some little issues from jobs run on Lilac, not sure what root cause is so I’ll get back to you on this.

      • perses: protocol-neqcyc

      • openmmforcefields: 0.12.0

  • openmmforcefields 0.12.0 deployed to alchemiscale.org under build v0.2.1-2

    • should allow for custom OFFXML in FF fields for Protocol Settings

    • IA – Looking forward to this, but we haven’t tested internally on our end.

    • DD – MO, this may be of interest to you as well.

  • IP : Protein-ligand benchmarks working group update

    • JW – Any blockers that OFF can help with?

    • IA – Big thing we need is testing - This is also jumbled up with something we’re debugging

    • JS – Is there a meeting for PLB work?

      • IA – Not regularly

      • JS – I don’t have the bandwidth for debugging, but I could look into results and stuff (that’s something I already do a lot of)

      • IA – I’ll ping IP to see if we can meet up next week.

      • JW – Could be helpful for MO to also be involved

        • IA – MO is already involved

      • JW – Ok, let me know if there’s any trouble with permissions/authority on making changes/releases. I can approve anything from the OpenFF side.

    •  

  • IP : feflow development:

    • DD – Is there interest in relative hybrid topology stuff getting in here?

    • IA – The consensus with me and RG is that eventually it will get it, but we don’t have a timeline for this yet. We cover this at the OFE/Perses sync

    • MH – Right, we used to cover this at our regular meetings but due to schedule conflicts we’ve fallen a bit behind. We’ve untangled schedules and those will resume in a few weeks.

  • alchemiscale development : sprint ends today: next sprint spans 10/25 - 11/6

    • architecture overview : PL Benchmarks on FAH - Architecture v6.drawio

    • coordination board : alchemiscale : Phase 3 - Folding@Home, new features, optimizations, targeted refactors

      • call for volunteers for available issues

    • alchemiscale 0.3.0 milestone:

    • alchemiscale-fah: 0.1.0 milestone

      • DD – alchemiscale-fah 1 – Having an ongoing discussion with Sukrit and JC about data models wrt F@H compatibility. This is ongoing. A little challenging because I don’t actually have access to F@H, so I’m just building to documentation.

      • DD – Alchemiscale fah 3 – Noneqcyc protocol

        • JS – Who wrote the code to do noneqcyc on f@h during moonshot? JC? Doyou have access to the code?

        • DD – Either F@H devs or JC. I have access to that code but it’s not the bottleneck - Most of the work is in having an interface for the execution model (we will be dynamically generating runs, some other differences).

        • JW – Any way we can get you better access?

        • DD – No, I have access, there’s just a process to getting running on work servers that involves some testing. So I need to get better prepared before I jump into the testing phase.

        • JS – Is there a way to run a F@H worker on your workstation and have only that pull your project?

        • DD – That’s possible, but that could still gunk up the central server. So I need to do some more local testing before I strain my good graces with the F@H admins

    • review Completed cards

    • updates on In Review, In Progress, and Available cards

      •  

JW – Governance discussion next week - Could be a good time to review milestones, involvement, contributions, approvership.

Action items

Decisions