Choose timezone
Your profile timezone:
Powered by Indico
v3.3.5
https://cern.zoom.us/j/64087077370?pwd=bX5hTzCoYEwMnTsUaT4MX0JPxwnn8l.1
Important notes for everybody:
Figure out how much time each of your subexercises should take
Label stuff as bonus material if you think it’s too much
TEST TEST TEST. Send to students/friends/whoever. You can ask summer students (or bachelors/masters students in general) to test at least the basic setup of your exercises since they should have relatively clean areas
Please fill out this google form (https://forms.gle/hbmk9Tdu6vNjnQBK8), once per exercise, by September 11, so we have all the important info for each exercise.
We have made a central gitlab https://gitlab.cern.ch/cmsdas-hamburg-2025 group for you to create repos here, if you would like (not mandatory)
We have made a central AFS area on lxplus (/afs/cern.ch/cms/Tutorials/CMSDASinHamburg2025) where you can put input files. Try not to fill it up with TBs of data - we have also asked for central EOS space in the meantime.
Please check the 2024 pre-exercises (https://cmsdas-cern-2024.docs.cern.ch/) to see if we should add anything there, or if all of this is actually needed for your exercises. If an existing pre-exercise is not needed, we will keep it there as bonus, and simply not make it mandatory.
We will have a final 1-2 hr meeting with facilitators in early October, once the registration closes and we have assigned participants to each exercise. Won’t be as long as these other meetings, but just to touch base one more time before the DAS and answer any last minute questions
Ying+Patrick for both exercises, as Alberto might not be in Hamburg for DAS
Pre-exercises:
Check if participants get cernbox area and how much space
Jets (Andreas + Dennis)
Try to emphasize Run 3. Will switch to Run 3 samples
Still need to figure out how much space is needed per participant
MET is not in list of exercises, but most people need it. Will add a bit
PU reweighting in CAT exercise? Need to follow up with Aliya, Daniel, etc. Included in long exercise (also)?
Follow up with flavor tagging exercise (Phillip and Uttiya)
Use SWAN. Which version? Same for everyone? Matteo: no, for muons we need something specific
Add logging into SWAN to pre-exercises: Matthias S will do. Maybe already part of Jupiter pre-exercise?
Should know if need particular library (particularly ROOT library or interface to CMSSW)
Combine limit setting (Kuan, Suman)
Based on combine exercise from terascale school
Still need to write script to write the data card, histogram input, and visualize the outcome
Pre-exercise to install combine (also needed for Beatriz/Jan’s exercise, instructions wil be the same): Beatriz will prepare
Probably needs 2 more weeks to finish 🙁
Reuse something from previous school? Check DAS from June DAS
Combine unfolding (Jan and Beatriz)
Only need ~1 MB storage (only histograms)
Not fully done with exercise but pretty far
Need to test that the timing works. Maybe a little bit long
At least half of the time (1 hour) on the Combine exercise (1 hour for the intro lecture and mini exercises)
Machine learning (Jorn and Matthias)
Have 5 exercises ready (just need to add to gitlab), 3 are optional
Already tested 🙂
Input samples accessible through SWAN, so probably they dont need to download it per participant
If have to download: 300-400 kB
How many GPU resources do we ask for? Need 1 node per participant. Ask O&C/ ticket with CERN IT.
If resources are still swamped, can run on CPUs and/or locally
Need part of pre-exercises: CERN account with 2FA. Already implicitly part of accessing lxplus as of today, but we can write it explicitly too
Tracking (Jan, Brunella, Marco)
Updated exercises to Run 3 🙂
Will update the CMSSW release to match the Combine one
Computing space needed is 200 MB. Participants don’t need to download inputs
Preparing slides, mostly a copy of last year (15-20 min)
Additional info includes PU: link to PU reweighting from Andreas/Dennis on jets - Jan will check
E/gamma (Alberto, Ying)
3 exercises, almost ready and on gitlab
3rd exercise requires some disk space and cernbox. How much disk space? Maybe 2-3 GB
Additional facilitator: they will ask Patrick to help
Muons (Matteo)
Recommend to take a look at beforehand: Awkward Arrays
5 exercises plus one bonus, but 5 seem relatively short
Flavor tagging (Uttiya, Philipp)
4 exercises in notebooks, mostly ready
Based on previous DAS school
Input data is in BTV central EOS space
Have the breakdown of how much time per exercise :)
Dennis asked them to include a bit about top/W tagging - just to mention it? They will mention in their intro slides
Taus (Andrea, Alexei)
2 notebooks based on 2024 DAS (40 min for 1st exercise, 1 hour for second)
Need to update a bit for the 2nd notebook to nanoAOD v15
Central storage space needed for the input tuples (< 1GB) - could even put in the repo
Make central CERN EOS space for the school? We ask Matteo how they did it last year at CERN
Generators (Dominic)
Exercises mostly ready, just updating CMSSW versions
Still need to make slides
Will send to Michel to test (need to add him to indico and our slides)
How much output space do the participants need? Will check. Maybe 1 GB
REANA (Pallabi)
Probably can complete everything in 2 hours
Based on NPS exercises
Optional pre-exercise: go through linked slides from Pallabi’s slide 3
Exercise on AXOL1TL on Run 3 data
1GB
Will move docker image part to the beginning, as this is important (no longer keep as a bonus)
Will make on lxplus machines, Pallabi is sure htat works
Total of 4 exercises
Applying central correction (Daniel)
Not done, was focusing on the long exercise
Have the correctionlib…
Can make a short exercise during the week of September 15, will take a few days, but will be starting from 0
Can start from code samples in POG areas, maybe 2-3 of them
Need to follow up with Aliya and Freya about their exercises