[QEE] Collection MR for further line optimisations 2025
Changes
MR | Status | BW Changes | Stream | Short description |
---|---|---|---|---|
!4151 (merged) + !4278 (merged) + !4140 (cf37b3ad) | Merged | negligible reduction | Full | Remove some unwanted low-mass events from Z->ll lines by adding a composite mass cut |
!4216 (merged) | Merged | tbd overall change, significant reduction in relevant lines | Full | PID added to DisplacedDiMuon{, _ss}
|
!4192 (merged) | Merged | est. 20% reduction of QEE Full-stream Sprucing | Full | add an isolation cut to Hlt2QEE_SingleHighPtMuon |
!4126 (merged) | Merged | + ~10 MB/s at Exclusive Sprucing | Full | loosen prescales Hlt2QEE_IncDiJet{15,20,25,30}GeVFull
|
!4310 (merged) | Merged | - ~20 MB/s at (HLT2/Sprucing) Turbo | Turbo | Replace persistreco with rawbanks for DrellYan lines |
Postponed to next MR
!4129 | Draft, two Reviews | tbd | Turbo | overhaul inclusive HNL line to a TTrack-vertex |
---|---|---|---|---|
!4094 | Draft | tbd | Turbo | Reworking diphoton for 2025 |
BW Impact
As of 11-Feb, the changes to Turbo/Full stream from this MR evaluated via most recent BW Test are found to be:
- HLT2_Turbo: est. 470 -> 360 MB/s (-23%)
- HLT2_Full: est. 560 -> 550 MB/s (-2%)
- Spruce_Turbo: 231 -> 211 MB/s (-9%)
- Spruce_Full: 151 -> 123 MB/s (-19%)
See [!4140 (comment 9051820)] for more information
ToDo
-
New ci-test + BW evaluation -
Update desc. with new numbers -
Approval from RTAWP3 and DPAWP1 coord. for bw changes
Edited by Alessandro Bertolin