Skip to end of metadata
Go to start of metadata

Attendees:

Canonical Meeting Template (as a reminder):

1. Review of meeting agenda
2. Addressing issues

  • Description/Status of issue
  • Review of action items from previous meetings if any
  • Recommended actions and discussion of same
  • Close issue or identify next action items

3. Status updates from SMT members

  • Status update on significant activities
  • Review of prior open issues if any (and not addressed earlier in agenda)
  • Identification of new issues if any (and not addressed earlier in agenda)

Agenda/Notes

1. Review of Meeting Agenda: Any comments/additions? 

 

2. XSEDE2 Reporting (Leslie Froeschl

Previously:

  • Previous discussion archived: SMT Meeting - April 9, 2020
  • (4/23) RY4 annual reporting will begin on May 1. 
      • Tuesday 5/5/2020: L3 text due to PMs; PMs give L3 text to L2s

      • Thursday 5/7/2020: All L2 text (including L3 text) due to PMs

    • Supplement doc
      • Holly working on prelim budget (should have it tomorrow)
        • Can repurpose unspent funds for year 6 (e.g., participant support funds)
        • Need to project funds we can't spend due to COVID-19 crisis
      • Should have draft to John early next week for his review.
      • All subaward partners so far have indicated interest in participating an additional year. 
  • (5/7) All text due to your PM TODAY for the annual report. Please get your inputs in today to keep us on schedule!
  • (5/21) Supplement request was submitted on Tuesday. Annual report was submitted late yesterday (on time). Thank you to all for your efforts! We won't need to worry about reporting again until August.
  • (6/18) Enjoy the break from reporting. We'll be back at it with IPR12 beginning Aug 1.
    • Panel report should be received next week. 
  • (7/16) Thanks to everyone for your work on the response to the panel report. John submitted our response yesterday. Recommendations have been created based on the items mentioned in the report. Appropriate folks were assigned those recommendations. Please be sure to go in and address them promptly.
    • The next reporting cycle begins in two short weeks on Aug. 1. Your area PM will be in touch about collecting all of the inputs for your area. L3 text will be due Tuesday, August 4. L2 text will be due Thursday, Aug. 6. Please be prompt and get your part done on time so we can keep on-track with our schedule!
    • Risk review is underway. 

Discussion:

  • IPR12 is underway and going well so far. Thank you to everyone for your efforts!
  • Reminder to all L2s to send Leslie any updates on COVID-19 related work to be included in the report by end of this week. 

Decisions:

 

3. Cancer Computer L2 Service Provider Request (Ron Payne

Previously:

  • (5/27) SPF L2 Request Letter sent to L2 for consideration.
  • (5/28) Questions posed from Bob S & Dave H:
    • (1) What is the hardware they are providing. Their website says “We are computing the cure in our 6 supercomputing clusters in Ottawa, Montreal, Kitchener, Bloomington, Minneapolis, Salt lake City, and Urbana-Champaign.” Beyond that, it’s pretty vague.

      • Donations of hardware that they repurpose in an HT condor environment/high throughput. Find sites willing to host, and site gets access to some percentage of equipment in exchange for hosting. Equipment is retired after a period of time, but should be similar to OSG. 
    • (2) Are these resources that are used solely by Cancer Computer or do they just provide spare cycles like OSG?

      • All except those used by hosting site are allocated by CC. They will delegate that to XSEDE in this case. 
      • What work host sites allow on the machines has to be in support of cancer research. 
    • (3) What fraction of the time will be allocated through XSEDE?

      • John will have to find out. Non trivial amount. 
    • (4) Do they have any special capabilities for dealing with sensitive data? That would be a big plus and offer something XSEDE doesn’t currently have. Not necessarily at the level of HIPAA compliance, but maybe ability to work with dbGaP data

      • Not at this time. It's complicated for them as an org with an intl footprint as different nations have different requirements. 
    • (5) Do we need to be concerned at all with the fact that Cancer Computer is based in Canada? Some of its hardware is based in the US, but the SP itself is Canada-based.

      • A lot of resources are at US institutions. IN, IL, UT & others. Could provide some level of support to Canadian researchers, but see this similar to OSG. Need to be careful we're not giving away resources across international boundaries. John will bring this up with Bob to see if he can provide guidance. 
      • We wouldn't change allocations process–would still need to be US researcher. 
  • (6/18) Approved pending answers to the following questions:
    • Amount of time that will be allocated through XSEDE 
    • magnitude of hardware that will be available. 
    • NSF approval given that CC is based in Canada 
  • (7/16) Need to get note off to Bob to explain relationship between CC & XSEDE. Concern about whether foreign entity is drawing substantive/inappropriate resources from XSEDE. CC wants visibility. Secondary benefit would be peer review for things placed on that resource. They have been held up in deployments. 100 nodes deployed at 3 locations in Canada. 8-10 nodes at UT, WI, IN. 300 deployed at IL. John doesn't expect this will be a problem, but we need to document what is happening carefully. 
    • Ron Payne to follow up with Roy about amount of time to be allocated.
    • HT Condor based. Victor will follow up for kickoff meeting once approved. 

Discussion:

  • (8/13) Discussion with Bob C. has been positive. Cancer Computer has indicated that, "we have 24 64-core servers (1,536 cores), which we could dedicate to this specifically, and replace some of the Suns that had issues with power supplies. We could have this ready for a fall shipment." John's input is, "If we go with 1000 cores, this translates to ~8M core-hours/year or ~8M SUs/year in XSEDE-speak. Given that the proposals that will be appropriate for this resource will be limited to cancer-related work, I think this is a reasonable number, particularly to start.  We can then monitor this for a while and adjust as necessary."
  • Pivoting focus of org to "computing for the cure." May have more PIs able to make use of the resource. Timing should be over the next few weeks. 
    • Will this org be responsible for providing documentation or will hosting orgs provide this? John believes they have documentation on their website but will check. Similar to OSG. HTCondor based. Should be straight-forward. Some cancer researchers won't be familiar with HTCondor. 
  • Expect more resources to be made available over time.  

Decisions:

  • Ron will finalize approval/acceptance letter and copy John/Victor/Ruth. Victor can then take over getting them integrated. 


4. Update on terminology task force (Linda Akli ) 

Previously:

  • (6/18) Added to agenda.

    • Received multiple complaints about offensive language used in XSEDE bootcamp/parallel MPI. Linda talked to John Urbanic, Susan Mehringer, and Ken Hackworth. Have done a good job promoting code of conduct. They use wording/terminology that refers to "master slave." 2 people sent emails about how offensive this was. That terminology is not controlled by us. There are alternative terms out there, but Open MPI has not done so. They are working on this, however. Don't want to guide people to use non-offensive terms if that creates problems coding. 
    • Recommendation: slide for trainers. Code of conduct acknowledges that there are offensive, but this is not controlled by XSEDE but by outside entities. That will help trainers as they are uncomfortable as they don't endorse the terminology. Allows participants to know XSEDE does not condone. Talked about how me might influence developers of Open MP, MPI (friends on steering committee), GIT, spark, tensorflow
    • John: Language is used to describe programming model. Language used in official standard for MPI. MPI standards body has recognized this is a problem and is working to rectify it. Expectation of vote on change to standard at SC20. Proposed language is "manager worker." Didn't realize Open MP has similar issue. Linda to ask training team to dig deeper to find a contact with Open MP. John Townswill send communication to Open MP organization (CEO, chair of language committee) to address this issue of use of "master" in their documentation. 
    • Linda could not find anything on ACM site regarding this, which was a surprise. Linda Akli to reach out to John West at TACC (ACM diversity chair) to ask about this. 
    • Linda Akli to forward complaints to John. 
    • John: Appoint task force to review all online content/documentation including wiki to look for such offensive terms.  CEE will take the lead in pulling this together. Maytal DahanLinda Akli
  • (7/16) Susan Lindsey & Susan Mehringer co-leading. All areas of XSEDE except Ops & XCI have representation on committee. Great participation so far. 
      • Lizanne to assign an eval person to participate
    • Get list of terminology
    • Workplan, schedule set up.
    • Web pages, training materials in the mix. 
    • Tools: do we need PIF for ECSS staff to provide some AI for the process. 
    • Potential guidelines going out to make sure we don't collect new materials that might be considered offensive. 
    • If anyone can participate in hunting for offensive terms, or if your org has any lists, please reach out to the team. 
    • Google folder set up for documents etc. 
    • Believe we may be in the lead on doing this work in the community. ACM has addressed some parts with facial recognition, AI, but not this issue in particular. ACM diversity/inclusivity council was going to look into this as well. Will be difficult for them due to extensive ACM library. 
    • NCSA has done some work generating a list of terms & alternative language. John will share this list. 
    • Need to include why a term is offensive and what an alternate could be. Context is very important. 

Discussion:

  • The group met Monday. Found some industry lists with terminology of concern. If you run across any concerning words, please inform a member of the task force.
  • John to send Linda a list from NCSA.

Decisions:


5. Interconnections Diagrams (Tim Boerner )

Previously:

  • (8/13) Added to agenda.

Discussion:

  • Reference:  REC-249 - Getting issue details... STATUS
  • connection between scientific impact, workforce development, and contributions made by the XSEDE 2.0 project
  • Interconnections diagrams could help address this. Need to get diagrams to a place that they work together. 
  • Will pull together L2s to discuss.
  • Would be useful as part of big picture/XSEDE overview presentations. Would also be helpful as we respond to future solicitations. 

Decisions:



6. All-Staff Meeting Prep (Ron Payne  )

Previously:

Discussion:

Decisions:



Around the call up

Follow this basic outline:

  • Status update on significant activities
  • Review of prior open issues if any (and not addressed earlier in agenda)
  • Identification of new issues if any (and not addressed earlier in agenda)
  • Identification of recent improvements implemented

Updates from SMT members:

  • PI (John/Ron) - 
    • Extension approved for another 12 mos of funding. 
    • Still pressing NSF on when we can expect to see a solicitation
  • Tim
    • no update
  • CEE (Kelly) - 
    • finishing all-staff slides
    • PEARC20: John moderated panel re. new resources. This was valuable from training perspective.     
    • Upping sponsorship for TAPIA. Will have 8-9 registrations available. Will have opportunities to interview people for jobs. Want to open virtual booth and make maximum use of it. Excellent source of people with CS & computational backgrounds.

  • ECSS (Bob S/Phil/Sergiu) - 
    • Preparing for new systems coming online. 
    • ECSS all-hands meeting will be held Aug. 28
    • ECSS symposia have had an increase in attendance. Possibly due to increased engagement with orgs that align with XSEDE. Will be asking people about this to better understand where people are coming from. 

  • XCI (Dave L/Craig) - 
    • Good discussions about community software repository

    • Texas A&M got an MRI (FASTER), so that will be another new SP. 12 new SPs coming in .

    • Johns Hopkins federation letter is still moving slowly. Waiting to hear from them re. questions/concerns. 
        

  • Ops (Greg/Victor) -   
    •  No update
       
  • RAS (Dave H/Ken/Nathan) - 
    • The new Fields of Science list has been rolled out into production.

    • Prepping for XRAC meeting. 

    • Testing for Expanse & Bridges2

  • Program Office (Ron/Leslie) - 
    • Reminder: Enter Publications in XUP and documents in IDEALS
    • Reminder: Reference collaborations in tracking worksheet: https://docs.google.com/spreadsheets/d/19YRAPeAForh5I9Xvhj8nKjROe3-JN7a1caLYbNamPaw
    • Reminder: Address recommendations (Jira) 
    • XSEDE all-staff meeting is tomorrow (8/14) at 3pm ET. 
    • September Quarterly Meeting will be remote (Sept. 1-2). Sessions can be requested at any time leading up to the meeting. XSEDE Quarterly Meeting - September 1-2, 2020 (Remote via Zoom) 
    • All supporting documents have been provided for the PY10 funding amendments. This and the PY10 spend plans provided by all partners will set the stage for the PY11 budget discussions that will begin in the second half of September 2020. More to come from BusOps.
  • SP Forum (Ruth) -  
    • Had presentations from each of the 5 new systems. 

    • Next week's meeting will feature Open On Demand. 

  • UAC (Emre) -  
    • No update
       
  • NSF (Bob C) - No update


Next Meeting:     SMT - August 27, 2020


  • No labels