20241104
participants: chenguang，kaili，hongbo，reda，junquan，yingqi，manqi，mingshui，xuai，shanzhen，chen ye， jiarong， xiaotian， peng zhang， Shiyi, Hengne, and online Da Xu, Yang Zhang, Fangyi Guo, Ligang Xia, Xiang Chen, Ying Yang, Hao Zhu
- general information: recommendations from IDRC review
- select fewer channels, aimed at demonstrating that the reference detector reaches adequate performance for physics. Include some simple topology (e.g. Z→mumu). Encompass H, Z, W and top physics.
- Foresee in the TDR results and figures about performance on basic objects (leptons, photons, jets) as a function of energy and polar angle
- The performance of crystal ECAL on boson mass resolution and Jet Origin ID should be simulated in a consistent way.
- ……
- Two task forces have already started their regular meetings: Jet/Clusters on Wednesday 2pm, Tracking/PID on Friday 2pm
- Chenguang presents summary of Tracking/PID studies
- tracking efficiency in 24.10.0 looks OK now
- current PID eff (for kaon, <90% ) seems significantly lower than requirement (95%, also from Manqi’s NIMA 2023 PID paper)
- need to find out the underline differences
- suggest to investigate more at lower level distributions, e.g. dedx resolution as function of cos theta … check Manqi’s paper
- the comparison should be apple-to-apple
- Kaili presents updates on
- dimuon lower tail (13% of events with m(mumu)<120 GeV for 125 GeV Higgs) understood - due to FSR
- need to investigate further how to better recover FSR for muons
- matching eff of dimuon ~ 70%, perhaps due to reading tracks from PFO
- need to check if PFO collection contains all tracks
- suggest also to double check with reading tracks from CompleteTracks collection directly
- need lepton ID which should be developed soon
- for recoil mass distribution, one should check if beam energy spread is included (~100 MeV)
- CPU/memory usage
- 24.10.0 memory issues with muon chamber
- once muon chamber removed -> 100 Hbb events in one hour, and with 4 GB
- ZvvHbb event size: 15 MB/event, mainly sim (14MB)
- Initialization requires ~10 mins per job.
- CPU farm, only 200 cpu jobs running simultaneously (likely due to limited nodes with 4 GB memory)
- Later in software group meeting, Weidong points out that CEPC farm has > 1600 cores in HTCondor.
- to be sorted out with computing center
- duplicated track issue
- happens at 0.7% for Z->mumu events. 2 PFOs with very similar PxPyPzE.
- more often in lower energy region
- check MC truth with Geant4 sim (including muon decay information)
- Should find out a solution at analysis level
- to investigate together with Tracking experts (Chengdong), should send them the exact events that have duplicated tracks
- Round table
- Reda working on photon performance with particle gun MC
- Shiyi - smuon
- Xiaotian shows plots: - dN/dx distributions of 12GeV 72 degree (costheta~0.3) samples, which looks significantly different from PID paper
- to be followed up with TPC team, and to understand how it was done in the paper
- Peng Zhang - will put more time on cepc jet performance
- Yang Zhang show update on tracks
- Ligang Xia plan to take some concrete and urgent tasks (e.g. differential tracking performance, PID related) with his student
- Ye Chen - plan on muon study
There are minutes attached to this event.
Show them.
