Difference between revisions of "EPSCI Group Meeting May 3, 2021"

From epsciwiki
Jump to navigation Jump to search
 
(One intermediate revision by the same user not shown)
Line 103: Line 103:
 
* CLARA
 
* CLARA
 
** v5.0.1 release is not officially deployed as a production version in CLAS12, so nothing to report.
 
** v5.0.1 release is not officially deployed as a production version in CLAS12, so nothing to report.
* EJFAT
+
* EJFAT( Michael's notes)
** ESnet whitebox design is available for review. The design includes an option for a switch based load balancing.
+
** ESnet whitebox design is available for review. The design is based on FPGA acceleration on the compute server side and network switch-based load balancing on the DAQ side but also includes an option for DAQ side FPGA acceleration.
** ESnet suggest using hardware switches equipped with the LAB protocol, where multiple links between two switches combine to provide higher bandwidth links between them.
+
** ESnet suggest using hardware switches equipped with the LAG aggregation protocol, where multiple links between two switches combine to provide higher bandwidth links between them, however initial testing could be based on non-LAG supporting switches.
 
** PR is being prepared for the FPGA.
 
** PR is being prepared for the FPGA.
** FPGA from the ESnet will be available for tests mid July.
+
** AVNET (vendor) has requested reestablishing recently expired NDA.
 +
** Expect that FPGA from the ESnet will be available for tests no earlier than mid-July.
  
 
* Funding opportunities
 
* Funding opportunities

Latest revision as of 20:44, 3 May 2021

The meeting time is 10:00am.

Connection Info:

You can connect using BlueJeans Video conferencing (ID: 253 300 597). (Click "Expand" to the right for details -->):

Meeting URL
 https://bluejeans.com/253300597?src=join_info

Meeting ID
253 300 597

Want to dial in from a phone?

Dial one of the following numbers:
+1.888.240.2560 (US Toll Free)
(see all numbers - https://www.bluejeans.com/premium-numbers)

Enter the meeting ID and passcode followed by #

Connecting from a room system?
Dial: bjn.vc or 199.48.152.152 and enter your meeting ID & passcode

Agenda:

  1. Previous meeting
  2. Announcements
  3. Conferences and Workshops
  4. Ongoing Activities
    • EJFAT (ESnet/JLab + FPGA +Accelerated Transport)
      • Status of proposal
    • DAQ systems
      • SRO
        • SAMPA + ERSAP + JANA2 + INDRA-ASTRA = ?
      • CODA (CODA3 support, EVIO-6)
    • A.I.
      • Experimental Controls
      • AIMCEG
      • Multiple FOAs + JLab LDRD
        • Collaboration with Theory on MCGen project DE-FOA-0002493 - preproposal submitted
        • Collaboration with BNL on AI scheduling DE-FOA-0002482
        • Collaboration with INDRA-ASTRA - ?
        • Collaboration with Sergey F. on AI + FPGA - deferred awaiting roadmap
        • Surrogate Models proposal (LDRD) - preproposal submitted
        • Amplitude Analysis Inverse Problem (LDRD) - preproposal submitted
    • Offsite Computing
      • OSG
      • NERSC, PSC, IU
  5. AOT



Minutes:

Attendees: Nathan B., Kishan R., Vardan G., Thomas B., Mike G., Torri J., Carl T.

  • Announcements
    • Fortnight paper: There was a suggestion that EPSCI group every focus group (AI, SRO, FPGA based data transfer, etc.) will come up with a paper that would benefit their research.
    • Summer internship program: 5 high school students will be working with our group during the June-July time frame.
  • Workshop and conferences
    • 3 accepted papers at VCHEP2021
    • TriDAS group will provide a presenter for the talk on TriDAS+JANA2 paper.
  • CI + SPACK
    • Summer intern (Jillian) will work on the JLAB common environment.
  • EIC
    • Graham informed on a new working group formation at the EIC, including David Lawrence, Christiano, and Dmitry Romanov.
    • Thomas reported that EIC will be able to submit jobs on OSG through newly created EIC sub VO on the JLAB OSG dedicated resources.
    • Brayan is going to increase fair share for the EIC VO, effectively treating EIC as a the 5th experimental Hall.
  • JANA2
    • Nathan continues working on JANA2 GlueX port, and reported a sizable progress achieved last week.
  • CLARA
    • v5.0.1 release is not officially deployed as a production version in CLAS12, so nothing to report.
  • EJFAT( Michael's notes)
    • ESnet whitebox design is available for review. The design is based on FPGA acceleration on the compute server side and network switch-based load balancing on the DAQ side but also includes an option for DAQ side FPGA acceleration.
    • ESnet suggest using hardware switches equipped with the LAG aggregation protocol, where multiple links between two switches combine to provide higher bandwidth links between them, however initial testing could be based on non-LAG supporting switches.
    • PR is being prepared for the FPGA.
    • AVNET (vendor) has requested reestablishing recently expired NDA.
    • Expect that FPGA from the ESnet will be available for tests no earlier than mid-July.
  • Funding opportunities
    • Graham is leading multiple efforts to establish collaborations and projects with real funding opportunities, including an official collaboration with ESnet, ASCR funded 3 phase Data Reduction Pipeline project, and collaboration with BNL on AI model based workflows.
  • SRO
    • Ed sent Vardan detailed instructions how to operate the SAMPA setup.
    • Vardan was able to start the readout and record data from 5 front-end cards in 10 separate files (2 stream per card).
    • Carl and Vardan are continuing to work on VTP and SAMPA setups in the indra-lab.
  • CODA
    • Carl is finalizing the EVIO-6 release
    • Vardan is working on JCEdit to implement FPGA based ROC, a new CODA component type.
  • AI
    • AI Experimental Controls
      • Torri is leading discussions at the AI WG weekly meetings related to defining critical design requirements and conditions for the project.
    • AICEG
      • Kishan continues to work with Nabuo on the AICEG project. He is evaluating Nabuo’s proposal to use a Lambda layer in the model that will help to incorporate user defined augmented features.
    • Hydra
      • Kishan he is trying to use layer-wise relevance propagation method to identify important sections of the histogram image.
      • Two summer interns will work on Hydra project this summer.
  • OSG
    • Thomas reported that the new feature to bundle similar jobs in the mc-wrapper is chocking CONDOR when the number of bundled jobs in the queue is higher than certain amount (>2000 or so). An obvious solution to this problem, according to Thomas, would be to introduce a limit on the number of jobs in the queue.
    • Two summer students/interns will work on mc-wrapper this summer.