HSF Weekly Meeting #147, 11 October, 2018
Present/Contributors: Graeme Stewart, Serhan Mate, Marilena Bandieramonte, Dario Menasce, Sudhir Malik, Davide Costanzo, Andrea Valassi, Jim Amundson, Pere Mato, Witek Pokorski, Eduardo Rodrigues, Tommaso Boccali, Simone Campana, David Lange, Ed Moyse, Pete Elmer, Robert Langenberg, Mark Neubauer, Caterina Doglioni, Liz Sexton-Kennedy, Daniel Elvira
News, general matters
- Maria Girone contacted us for an HSF speaker for the CMS ECOM2X (committee looking at software and computing evolution for Run3 and Run4). Eduardo agreed to do that.
HSF/WLCG/OSG workshop
- OSG have now confirmed their participation in the 3-way workshop at JLab 18-22 March next year.
- Next step is to try and identify topics we want to make sure are
covered, and prepare a draft agenda.
- Meeting next week for WLCG/HSF people.
- Please send any early thoughts to Graeme/Michel/David (organising for HSF).
New Working Groups
- Draft mandate document.
- Some comments made by Caterina and Frank (thanks!). Now all
converted into concrete suggestions.
- Each proposed change was discussed and refined, then we agreed on it.
- Concrete deliverables were raised, how much can we push these?
- Can’t put too many constraints on conveners for things that it is not in their power to deliver. Main deliverable is a functioning community.
- Dario - make this bottom up.
- Meeting agreed on the mandate now.
- Would like to now call for further nominations.
- One or two weeks?
- Two weeks. Graeme will send the call.
- One or two weeks?
Generators Software Computing Workshop
- Indico agenda.
- Organisation progressing well, quite a number of speakers confirmed, 25 people registered.
- Next meeting for the organisers immediately after this one…
Activity updates
Packaging
- Next meeting tentatively next week (get in touch with Graeme or Ben for any topics to present).
Software Forum
- Successful meeting
yesterday with a
presentation on optimisation work done in CERN IT.
- Some discussion over estimated gains from using vectorised instructions.
- Minutes in preparation.
- Open slot in two weeks time, 24 October.
- Any volunteers?
- Then further topics will go into 2019.
PyHEP
- We want to converge and identify a representative from the Intensity Frontier to join the efforts. If you know of a great candidate (yourself or …), please shout ASAP.
CWP
-
Publication status for Individual WG Papers
- Event/Data Processing Frameworks
- Abstract added and few mistakes fixed in this PR; references and final author list still need done (should not take too long).
- Event/Data Processing Frameworks
AOB
- HSF Logo in vector format - reminder of action on Benedikt.
- There was a discussion with Claire Boudarios (ATLAS) about outreach
activities and the HSF’s involvement there.
- Graeme - no activity in this area as far as I am aware.
- Is anyone interested in this topic?
- Dario and Pete are - Graeme will pass on contact details.
- http://hepsoftware.org/ has been down for a week - tracked as a GitHub issue.
Training Discussion
- Our goal is to start a concrete activity.
- IRIS-HEP has a small training component.
- FIRST-HEP is totally dedicated to training.
- Try and strengthen training in the whole community.
- Tier 1 - build out of StarterKit, and Software Carpentries.
- Tier 2 - this is HEP specific, but could organise more “boot camps” and BOFs.
-
Tier 3 - already an established network of schools.
- Dario - we didn’t yet have time to work on more specific plans
- What about University engagement? Traditionally this is quite
weak in Europe.
- Our model adapts to an inhomogeneous intake already, and seems a much more tractable problem to start with graduate students.
- Liz - most physics computing teaching done by physicists already.
- If funding is US based, how can we make sure that there are
similar initiatives for Europe?
- CERN based training will help (but students still need to get here and be funded…).
- Also doing regional events at which students will be present.
- What about University engagement? Traditionally this is quite
weak in Europe.
- Liz - if something at FNAL should we include neutrino people?
- Yes.
- Pulling off something practical (Tier 1 training and
survey, detailed on the final slide) is more important.
- Meeting agreed to that.
- CMS: FNAL is natural.
- ATLAS: CERN? Could be a good place to bootstrap? Graeme,
Davide to follow up.
- Key point to build up this cohort of trainers, who then have their own regional connections.
- Sudhir - want to tie this to a concrete CMS event, maybe beginning of next year.
- Training survey - there was a draft floating around, need to
re-circulate it.
- Graeme thinks ATLAS did do something in this area (perhaps tangentially), he will follow up.
- Promote training as part of people’s mental landscape