Choose timezone
Your profile timezone:
Topical ROOT Team Meeting: Preparation for the RUW2025
ROOT Team Meeting
Zoom Meeting ID
63815034196
Host
Danilo Piparo
Alternative host
Philippe Canal
Useful links
Join via phone
Zoom URL
Shift handover: Danilo -> Bertrand
- suffering from issues with the new mac vm's exchange between Niels and Olivier, Stephan fixed on friday by enabling full builds (no incrementals, caching, downloads etc)
- Olivier reports an issue on one of the mac15 VMs with cmake, checked the .profile file and turns out the path to cmake is missing, complete mystery, Stephan thinks it could be connected to the problem above, Olivier put back the path to cmake in the file and it seems to work
- Giulio (ALICE) brings uo the issue "If I compile with -Dcuda=off -Dcudnn=OFF on linux or even on macOS, I get a bunch of symbols for NVPTX in libCling which I would not expect", Lukas investigated but doesn't have an answer yet, he asks for help
Liaison report
- CMS:
- Vincenzo will be presenting at the CMS software days, giving a general status report of the project and answering specific questions reported by CMS core software
- Bertrand: are the slides presented accessible to the team? --> No
- Stephan: it would be good if we can centralize the slides, to see the latest decks that were presented and easily craft presentations in the future by mixing and matching
- Vincenzo: we can add them to the internal documentation
- ALICE
- Giacomo and Vincenzo had a talk with Giulio about their use of computation graphs in RDF, it was very enlightening and yielded in actionable items to tackle what we thought was 1 feature request turned out to be 2 or at least 3, still to make a concrete plan with a timeline
- Vincenzo to provide an example for Giulio and see if it is actually feasible
Meeting reports
- IO meeting report
- discussion on the posters for the users workshop, initially 2 RNTuple posters suggested by Danilo, decided to create 1 poster with a general overview of RNTuple, would also require less work
- Key4HEP
- Stephan went as an obserber: everyone complaining ONNX so unstable, etc, Stephan mentions SOFIE in ROOT to them, they didn't seem receptive, Stephan thinks if we go with a working example in their context they will be very happy about it
- Stephan: it would be good if everyone involved in rbatchgenerator, inference etc go for a coffee and brainstorm a strategy to talk about this (Vincenzo and Martin for now)
- Spot Meeting
- Qual task of students: currently running reconstruction jobs outputting into ttree with lzma, the task is to compare to ttree with zstandard and rntuple with lzma and zstandard
- results: RNTuple performs generally better than ttree
- they observed ttree with zstandard performs similarily to rntuple with lzma (not unexpected)
- there is a codimd with instructions to reproduce this job for our own benchmarks if needed
Users Workshop
- Stephan: i think it's a good idea to have the topical sessions split into 2 days to give people time to think about some points and bring them up again in the next session
- Stephan: we don't have parallel sessions, does it really make sense to have summaries? Can we convert it to an "open-mic feedback session for each track"? Bertrand thinks it's a good idea
- Florine: we could use tools like mentimeter to make it interractive
- Vincenzo: we are not sure if we do summaries or one general closing remarks session, but we can do this as well, we will discuss it
- Stephan: is lunch included on friday? --> No
- Bertrand: can the timeline be adjusted in case we get extra registration?
- Vincenzo: we can accept extra registrations, but not talks will be allocated, maybe posters at best
- Stephan: makes up discover the compact style to show the overview of the week agenda on indico
- Stephan: we should add the template for the posters to the internal documentation
- Stephan: doubtful that 3 hours is enough
- Bertrand: the audience is very different from the summer students course, we need to adjust
- Vincenzo: 3 hours is what we already have today for the basics training, we will shave off the extremely basic stuff, and inject some of the advanced stuff,
- Stephan: we can't make everybody happy if we put everyone in the same room (a large spectrum of expertise with ROOT) and have one person present at the projector, i suggest to split the room into a beginners table and an experts table, and make it an asychronous hands-on session with instructors walking around individuals running the notebooks at their own pace
- Vincenzo: i think it would be a bit chaotic to try and organize them in corners, logistically complicated given we are 60 people and we have one room
- Vincenzo: I will ask Pedro if they have foldable tables int he conference room or if it's just chairs
- Vincenzo: we will reach out to the participants about the training session, request them to bring laptops and share the summer students course video to prepare
- Stephan: we are getting 30 or 40 accounts for people with no CERN accounts to use SWAN
- Florine: from my experience there is always someone running the notebook at the front, you can choose to follow
- Stephan: then we shouldn't mention hands-on, the alternative is to do 1 hour tutorials then the rest is hands-on
- Lukas: Fons is retiring, did we consider getting him a farewell speech?
- Florine we get him a t-shirt
There are minutes attached to this event.
Show them.
