Choose timezone
Your profile timezone:
EP R&D Software Working Group Meeting
→
Europe/Zurich
32/1-A24 (CERN)
EP R&D Software Working Group Meeting
Zoom Meeting ID
62893115044
Host
Graeme A Stewart
Alternative host
Andre Sailer
Useful links
Join via phone
Zoom URL
Software R&D Working Meeting Minutes
2025-02-19
Room: André, Giacomo, Aaron, Joana, Pyiush, Peter, Swathi, Mateusz, Juan, Aurora, Anna, Juan
Remote: Joshua, Leonhard, Florine, Lukas,
Apologies:
News
Please consider presenting on one of the EP R&D Seminars.
Key4hep Status report - Swathi Sasikumar
- Inheriting two decades of software research in the context of iLCSoft for linear colliders. With new types of lepton colliders on the horizon, proposal to share components to reduce maintenance and develop a common turnkey software stack: Key4hep.
- EDM4hep: common data model definition, growing along stakeholders needs. Notably includes associations between objects in the data model, e.g. particles and tracks
- Podio: I/O in TTree and RNTuple. The produced files can be used outside of EDM4hep, but a limitation is that the relations between objects have to be resolved manually.
- PandoraPFA: particle flow algorithm developed to study particle flow calorimetry. Limitation is that it was aimed at high granularity CALICE calorimeters. The goal is to enable use of PandoraPFA across different calorimeter geometries.
Questions
Anna: You mentioned collecting all photons into one?
Swathi: may happen that photons are reconstructed as two different particles. Maybe due to photon conversions.
Anna: you wouldn't be able to reconstruct it if it was an actual physical event, wouldn't be able to connect which particle produces the photon.
Swathi: this was possible in this case because it's a one photon event. The statistics for this case is not too much. In my plot, the small tail is due to the layer height was given wrongly. Along the electrode and not the radius.
Anna: The XML with the pandora settings is in the repo?
Swathi: In key4hep, in CLD_o4_v05.
Juan: Are there any changes done in Pandora here?
Swathi: Mostly the reconstruction parameters. I tuned the calibration of photon.
Anna: These new calibration constants are in the CLD config?
Swathi: Not yet, we can upload them. The PR comes with the option file, the plugin, the driver file.
Peter: I don't know how would this translate from a CALICE calorimeter to liquid argon?
Swathi: In terms of physics are going to be similar.
Peter: I wasn't sure if we need some retuning.
Swathi: One could redo the same plot not with a particle gun but with a complete event, see what it looks like. The distribution will not be ideal, but in my work I focused more on the shoulder of the distribution that was due to the wrong configuration of the layer height.
Vincenzo: Slide 11, mention the data flow with vectors of collections, are these collections all of the same type or do you use a variant as the element type of the vector?
Juan: yes, all of the same type
Vincenzo: Slide 10. For the new gaudi algorithms, how do you ensure the user-defined classes with operator() do not carry internal state?
Juan: The operator() is const, so the class cannot have mutable internal state, although users can always force it if they want, but it's now clearer that should not be done.
Mateusz: Also, new gaudi algorithms are thread-safe by design and the "boilerplate" to start a new algorithm is much clearer than before on how it should be used.
There are minutes attached to this event.
Show them.
