HLT2 configuration for light ion runs
Fully based on hlt2 2024 configuration for PbPb -> same hlt2 reconstruction config + hlt2 minbias lines with persist reco
Differencies:
- Streams:
ionraw,hlt2calib,nobias,lumi. - For fixed target config, minimum bias strategy is chosen too (no
ionstream as in 2024 PbPb run)
For more information, summary table with expected conditions for each sample to be taken during the special run (as of 23/06/2025):
Event sizes consider persistReco + RawBanks. They were computed with private simulation and verified with official samples with expected-2025 conditions:
- /MC/Dev/pO-Beam6800GeV-3400GeV-expected-2025-MagUp-Fix1-EPOS/Sim10f/30000000/DIGI
- /MC/Dev/OO-Beam2680GeV-expected-2025-MagUp-Fix1-EPOS/Sim10f/30000000/DIGI
More details regarding the estimations can be found in this codim: https://codimd.web.cern.ch/JwWZqclQRlyWvxLAffu8_g#
For more information, the long track reconstruction performance (efficiency and ghost rate) as function of the SciFi Hits for each collider mode systems (OO, NeNe and pO):
hlt2_OO_full_BestLong_Forward_eff_vs_FTHits_gprob0.65_evts2000.pdf
hlt2_pO_full_BestLong_Forward_gr_vs_FTHits_gprob0.65_evts2000.pdf
hlt2_OO_full_BestLong_Forward_eff_vs_FTHits_gprob0.65_evts2000.pdf
hlt2_OO_full_BestLong_Forward_gr_vs_FTHits_gprob0.65_evts2000.pdf
hlt2_NeNe_full_BestLong_Forward_eff_vs_FTHits_gprob0.65_evts2000.pdf
hlt2_NeNe_full_BestLong_Forward_gr_vs_FTHits_gprob0.65_evts2000.pdf
Black line is using the baseline pp 2025 reconstruction with gprob < 0.65. Blue line is using fast reco hlt2 reconstruction with gprob < 0.65. And green line is baseline pp 2025 reconstruction with gprob < 0.65 but using gprob tuning using OO simulation (more info in !5051 (closed)). Since the improvement is not that significant, using the pp ghost prob tuning is considered
Goes with lhcb-datapkg/PRConfig!574 (merged) for the qmt test