Optimization of BdsToPpPmKpPim and BdsToPpPmPipPim for 2025 data-taking

Will add more testing result

Update(21-March) In master, we only have BdsToPpPmKpKm, and the bw of hlt2 is 5.3 MB/s

Currently, we add some tight requirement on final state, e.g. PID, track fit chi-square, and remove pv_tracks. The BW for hlt2 lines:

B->ppbarKK: 0.96 MB/s

B->ppbarKpi+c.c.: 7.4 MB/s

B->ppbarpipi: 2.6 MB/s

---

(NEW!) Update(24-March) Using 102453 events, and more tight cut on track.

B->ppbarKK: 0.5 MB/s

B->ppbarKpi+c.c.: 3.8 MB/s

B->ppbarpipi: 2.4 MB/s

with same mdf file and same events, the master version is 10.5MB/s

So clearly decrease in B->pphh.

/--------------------/ deprecated message /------------------------------/

Below are for spruce lines, we need hlt2 instead, sorry for bothering.

Update test bandwidth with snippet (https://gitlab.cern.ch/-/snippets/3293):

pppipi: 2.2 MB/s ppkpi: 4.1 MB/s ppKK: 1.7MB/s

Comments: the largest bw is below the BnoC line

 SpruceBnoC_BdsToKpKpKmKmDecision

and locate around NO.265 in the list https://lhcbpr-hlt.web.cern.ch/UpgradeRateTest/BandwidthTest_lhcb-master-mr.12834_Moore_hlt2_and_spruce_bandwidth_x86_64_v3-el9-gcc13-opt+g_2025-03-11_12:14:59_+0100/spruce__full__all_rates.html

Update(17-Mar-2025),

1. Calculate using 75.49 kHz as input Add more cuts on daughter:
2. track: F.CHI2DOF > 4, while the candidates almost gone around 3
3. require Kaon and Pion has F.ISMUON == is_muon pppipi: 2.9 MB/s ppkpi: 5.4 MB/s ppKK: 2.4 MB/s

Comparing to last MR(!3311 (merged)), we have 34.2 MB/s in total for three lines.