Skip to content

RD: Add isolation to hlt2 lines (b --> xll and baryonic)

Paloma Laguarta Gonzalez requested to merge palaguar_b2dll_hlt2 into master

The lines that were modified are B+ —> pi+ e+e-, for the mesonic lines and Lb —> p+ pi- e+e- , Lb —> p+ pi- mu+mu- for the baryonic lines. In both cases, what was done was add isolation information using the rd_isolation tool. In order to test the impact of these modifications, bandwidth tests were run following the guidelines provided here https://indico.cern.ch/event/1365216/contributions/5790687/attachments/2792768/4870747/HLT2BWRTADPAGeneral_RossHunter_020224.pdf. Initially, these tests were run with the minbias sample, but the low occurrence of these decays was requiring to use a high number of events of the sample resulting in a big consumption of computation resources and still very low statistics. Therefore, we decided to run on signal MC samples of B- —> K- Jpsi for the mesonic lines and Lb —> p+K- Jpsi for the baryonic lines. After the tests, we look at the increase in event size (which is computed by checking the size of the output file of the test and dividing by the number of selected events by the line). Taking into account this increase and using the latest BW tests we can estimate the actual BW values that result from adding the isolation to the mentioned lines.

  • MESONIC:

Before isolation info added:

Event size = 72Mb / 593 output events = 0.1214 Mb

After isolation info added:

Event size = 91Mb / 593 output events = 0.1534 Mb

Edit: The line BuToPimMuMu was modified to add isolation in !2978 (merged) and already merged, therefore no addition is done for this line in this MR.

Increase of ~26% --> From latest BW test: BW = 4Mb/s --> BW_with_isolation = 5Mb/s

  • BARYONIC:

Before isolation info added:

Event size = 23Mb / 181 output events = 0.1271 Mb

After isolation info added:

Event size = 31Mb / 181 output events = 0.1713 Mb

Increase of ~35% --> From latest BW test: BW = 2Mb/s --> BW_with_isolation = 2.7Mb/s

Note that the tests were run using signal MC, and since min bias events are typically less busy, we can consider these values as upper limits of the BWs.

FYI @cmarinbe @lcalefic

Edited by Paloma Laguarta Gonzalez

Merge request reports