1 of 9

EPSCI

Group Meeting�

September 1, 2026

1

2 of 9

DAQ: Streaming Readout + CODA

2

Accomplishments

  1. Streaming CODA
    • CLAS12 streaming DAQ simulation (in progress)
    • EIC streaming testbed in Hall-D (done)
    • Virtual crate project. Design requirements (back burner)
    • EVIO AI AutoResearch optimizations (Dmitry Romanov)
    • EVIO bug fixes
  2. PET
    • E2SAR integration into the pipeline
      • Running on Perlmutter (done)
      • Checking observability with ERSAP (in progress)
  3. ERSAP
    • Observability support implementation for HAIDIS (done)
    • ERSAP extended paper (in progress)
    • CHEP26 paper: "Event Identification in Streaming DAQ using ERSAP" (in progress)

EPSCI Group Meeting Sep 1 ,2026

3 of 9

EIC Computing and Software

3

Reconstruction

  1. ePIC Workshop on Streaming Reconstruction

​

  • Yet another round of work on the timeframe splitting PR
    • Implemented the “JOmniUnfolderGenerator”
    • https://github.com/eic/EICrecon/pull/2824

​

​

Prod

  • Working on automations and better standardizations of requests

EPSCI Group Meeting Sep 1 ,2026

4 of 9

4

LD2606 and AmSC Data Services - Data Movement

Aug/19 – Sep/1 Summary

  1. How trustful AI reasoning is in an automatic agentic workflow for research paper writing?
    • Pros – Sharp findings in the first run; …
    • Cons – Over-claiming; contradictory details; buzz advanced words; …
    • “Supervise” AI if the delivery can not be tested
  2. LDRD
    • Coding: Wait for Raiqa to merge PR and test on the Scicomp testbed
    • Paper: maybe postpone to something later?
      • TODO: Fix all the testbed nodes to run experiments!!!
  3. SC workshop notification will come out this Saturday/(next week?).
  4. Sep/30 full-day ALCF seminar – “Service-Enabled Science: Working with Jobs, Data, AI, and Agents”, https://www.alcf.anl.gov/events/service-enabled-science-jobs-data-ai-agents

“Unsupervised”

“Supervised”

EPSCI Group Meeting Sep 1 ,2026

5 of 9

Genesis Mission / American Science Cloud

5

Accomplishments

  1. one …
  2. two …
  3. three …

BPSW

  • Trying to finish off knowledge base contribution
  • Gearing up for the hackathon

EPSCI Group Meeting Sep 1 ,2026

6 of 9

Coupling Experiment to Accelerator Controls

6

TwinD

  1. Benchmarking based on standard code
    • Oh the horrific debugging…..(regex)
  2. Optimized visualizations

EPSCI Group Meeting Sep 1 ,2026

7 of 9

JANA2

7

JANA2

​

  1. Bugfix release: https://github.com/JeffersonLab/JANA2/releases/tag/v2026.03.01�
    • Parent event cleared prematurely. Introduced in v2026.03.00. Issue #515, PR #516.
    • Pool depletion and hang in JUnfoldArrow due to state machine bug. Introduced in v2026.03.00. PR #514.
    • JEventSourceGenerator::SetLevel() was set to private. PR #513.
    • janacontrol links against libzmq target instead of inconsistently defined CMake variable. PR #512.

LD2613: Distributed workflow steering�

  • Development of FY27 proposal (Deadline deferred until 21 Sept)
  • Nathan: Separating Shell from Conductor so that Shell can run locally while Conductor runs in VM
  • Jon: Researched MCP server integrations with C++

AmSC-SUF

EPSCI Group Meeting Sep 1 ,2026

8 of 9

Hydra

8

Accomplishments

  1. one …
  2. two …
  3. three …

Development

  • Just about through the backlog of PRs
    • Siamese refactor
  • Ready to tag and deploy new version
  • Polishing hydra_train.py
  • Refactoring a new script made for HPO after Thomas’s review

Other

  • Trying to obtain broader data use permissions for Hydra development

EPSCI Group Meeting Sep 1 ,2026

9 of 9

Miscellaneous

9

Accomplishments

  1. one …
  2. two …
  3. three …

MCWrapper

  • Autorenewal fixed!!!!!!
    • Other outages and gremlins
  • Started a large campaign

EPSCI Group Meeting Sep 1 ,2026