---

HSF Weekly Meeting #160, 28 February, 2019

Present/Contributors: Graeme Stewart, André Sailer, Marko Petric, Tommaso Boccali, Danilo Piparo, Pere Mato, Stefan Roiser, Andrea Valassi, Eduardo Rodrigues, David Lange, Agnieszka Dziurda, Serhan Mete, Andrea Rizzi, Caterina Doglioni, Ed Moyse, Jim Amundson, Liz Sexton-Kennedy, Mark Hodgkinson, Daniel Elvira, Heather Gray, Markus Klute, Gloria Corti

News, general matters

• LHCC WLCG Meeting:
  • Graeme presented slides from software projects and HSF.
  • Question asked about Geant vectorisation R&D:
    • Emphasised again the intent to deliver improvements into production Geant4.
  • Discussion about accelerator integration and framework from CMS presentation:
    • Pointed to JLab workshop as a place to discuss this.
  • Support from referees for common software work.
• AccelNet:FRESCO support letter
  • See agenda for project description and draft letter.
  • To clarify, the HSF in the project application is described as:
    • HSF: The HEP Software Foundation (HSF) encourages cooperation and common development of software in HEP. The HSF hosts a number of software-focused working groups and projects where developers are collaborating. These include groups on physics generator models, detector simulation techniques, event reconstruction techniques and approaches to software packaging.
  • i.e. HSF as a forum to build networks and locate community experts.
  • André can you mention CLIC?
    • Yes, now it’s there.
  • Meeting approved the letter of support.

Google Summer of Code 2019

• We are accepted as an organisation again.
• There is a meeting for all mentors tomorrow at 14h CERN time.
  • Will cover how to interact with students and the timeline.
• Proposal for Advisory Committee:
  • https://docs.google.com /document/d/17ifO0i4O5JsBVKkcx3j75Z_iW2bXFnCbqFB2YirlJQw/edit?usp=sharing
  • Need help to select projects against slots (assuming slots < projects).
    • Should also assess given the potential students.
    • 3 people proposed - better if these are not mentors.
    • Criteria: novelty, student rank, impact, etc.
• Please review and comment on the proposal in the next week, aim to approve in next week’s meeting.
  • If you would be willing to serve as on the AC please let HSF Coordination know.

HSF/WLCG/OSG Workshop News and Planning

• Indico registration is still open.
  • 221 people registered.
• There is a Google Sheet linked from the main page to help coordinate car sharing from Dulles (you take your own responsibility for this!).

Session Planning (see block timetable)

• Reminder: David Lange, Michel Jouvin and Graeme Stewart organising for HSF.
• General Sessions:
  • Plenary sessions all fixed now.
• HSF Sessions:
  • Rooms:
    • Will do a sound check with Amber and LOC next week to verify Vidyo connection is acceptable.
  • For planning other HSF Parallel Sessions, see WG activity reports below.
• We will contact Amber about JLab entry requirements (some worries that people may need CVs due to changes in DOE rules).

Activity and Working Group Updates

Data Analysis

• Draft agenda for JLAB discussed, here a draft (speakers being contacted).
  • 1st session: Where are we now, where are we going ?
    • 20’ talk on analysis needs and how they are served today (Paul)
    • 20’ talk on what’s missing today (Andrea)
    • 20’ talk on evolution of HEP computing implications for analysis (data distribution, access to resources, hardware platforms, …) (TBD)
    • 30’ discussion
  • 2nd session: Technology
    • 20’ talk on the future techs (Danilo)
    • 20’ talk on Python and frameworks for declarative analyses(Jim/Nick)
      • Check for overlaps with PyHEP session talk on Coffea!
    • 20’ talk on ROOT Declarative framework: RDF (Stefan)
    • 30’ discussion
• Timetable should go into Indico, even if not 100% settled.

Detector Simulation

• Agenda for the HSF meeting is finalised – both talks and speakers.
  • Overall philosophy: tried to assign a balance of speakers from experiments. Expect each speaker to summarise one topic focusing both on their own work and also the perspective from other experiments
• Next topical meeting on 6 March: https://indico.cern.ch/event/782508/
  • Need to add agenda entries for that meeting.

Reconstruction and Software Triggers

• Agenda for HSF meeting finalised (talks and speakers).
  • Session 1: real-time analysis, talks from LHCb, ATLAS, CMS, ALICE.
  • Session 2: ML/accelerators for reconstruction, talks on LHCb tracking, GPU for IceCube, ML in FPGA, JLab reconstruction challenges.
  • Plan to have a chat/meeting with all speakers to prepare the discussion session.
    • We (WG coords) will look at the talks and prepare 2-3 questions in advance so that we can have a discussion at the workshop.
      • General direction of questions:
        • Is there room for collaboration and where?
        • Are upgrades & detectors real-time-analysis ready?
• Topical meetings: would like to have a shared calendar with other groups (e.g., IRIS-HEP) to pick dates for our own without conflicts.
  • It’s already hard enough for us to find time where we’re all available :).
  • Should we have an HSF + others google calendar?
    • This is possible with the current HSF calendar.
    • Graeme will send instructions on how to edit and add entries.

Software Tools

• JLab session well defined.

PyHEP

• JLab 1.5h session is finalised. Outside speaker from Anaconda.

Training

• PR in progress for listing training events on the HSF website.

Event Generators

• Will have regular open meetings every two-weeks, first is this afternoon at 16h00 after this meeting (but on a different Vidyo room). 28th Feb 16h00.

Licensing

• Discussion on Gaudi Plugin Service / Gaudi License with CERN KT.
  • Re. code ©CERN, responsibility for re-licensing open source code is given to the CERN group that develops the code. In this case EP-LBC.
    • Thus is should be fine to relicense from GPLv3 -> Apache2.
  • With that change the intention is to ©Gaudi to CERN, for the benefit of the ATLAS and LHCb collaborations; and to apply an Apache2 licence.
• CERN have a webpage that recommends GPLv3:
  • https://legal.web.cern.ch/licensing/software-terms-use
  • We think this is anachronistic - Graeme will follow up with KT.
• Still no response from FastJet authors.

Packaging

• Held a meeting on 27 February.
  • NixCon (Chris Burr attended).
  • Spack update (HEP people now have merge rights for our packages; we also have our own Slack channel).

CWP

Roadmap

• We got final proofs from CSBS.
  • Eduardo and Michel noticed quite a few errors, which have been pointed out to Springer and will be corrected.

---