Skip to content

Fix to PIDe filtering problem

Problem

The old version of this code was doing the PIDe inversion by running filter_anti_pid_e. This functionality has been moved to this this builder.

Consequences

The filtering function was filtering by PID from [-1,+1] as shown below:

image

the intended cut should be in [-1, 1.5]. This is expected to increase the bandwidth.

Rates

The bandwidth W.R.T the full RD lines is attached. The table concerning the current module is below:

rate (before) rate (after) difference difference (%)
Hlt2RD_BuToKpEE_MVA_misid 149.686 161.026 11.3398 7.57576
Hlt2RD_BuToKpEE_MVA_cal 106.595 102.059 -4.53594 -4.25532
Hlt2RD_B0ToKpPimEE_MVA_misid 49.8953 58.9672 9.07188 18.1818
Hlt2RD_LbTopKEE_MVA_misid 45.3594 54.4313 9.07188 20
Hlt2RD_BuToKpEE_SameSign_MVA 54.4313 45.3594 -9.07188 -16.6667
Hlt2RD_BsToPhiEE_MVA_misid 40.8234 38.5555 -2.26797 -5.55556
Hlt2RD_BuToKpEE_MVA 47.6274 36.2875 -11.3398 -23.8095
Hlt2RD_B0ToKpPimEE_MVA_cal 34.0195 29.4836 -4.53594 -13.3333
Hlt2RD_BuToKpMuE_MVA 22.6797 22.6797 0 0
Hlt2RD_B0ToKpPimEE_MVA 15.8758 18.1438 2.26797 14.2857
Hlt2RD_BsToPhiEE_MVA 11.3398 13.6078 2.26797 20
Hlt2RD_LbTopKMuE_MVA 11.3398 11.3398 0 0
Hlt2RD_LbTopKEE_MVA 9.07188 11.3398 2.26797 25
Hlt2RD_LbTopKEE_SameSign_MVA 9.07188 9.07188 0 0
Hlt2RD_BuToKpMuMu_SameSign_MVA 6.80391 6.80391 0 0
Hlt2RD_B0ToKpPimEE_SameSign_MVA 6.80391 6.80391 0 0
Hlt2RD_B0ToKpPimMuE_MVA 4.53594 4.53594 0 0
Hlt2RD_LbTopKMuMu_SameSign_MVA 4.53594 4.53594 0 0
Hlt2RD_LbTopKMuMu_MVA 4.53594 4.53594 0 0
Hlt2RD_BuToKpMuMu_MVA 4.53594 4.53594 0 0
Hlt2RD_B0ToKpPimMuMu_SameSign_MVA 2.26797 2.26797 0 0
Hlt2RD_B0ToKpPimMuMu_MVA 2.26797 2.26797 0 0
Hlt2RD_BsToPhiEE_SameSign_MVA 4.53594 2.26797 -2.26797 -50
Hlt2RD_BsToPhiMuMu_SameSign_MVA 0 0 0 nan
Hlt2RD_BsToPhiMuMu_MVA 0 0 0 nan
Hlt2RD_BsToPhiMuE_MVA 0 0 0 nan

Bandwidth

The bandwiths are:

Before: 595 Mb => 235 Mb After : 604 Mb => 239 Mb

It ran over 568000 events with a rate of 1.3 MHz. The increase in bandwidth is 9 Mb. However the last MR showed a decrease of -29 Mb. Thus the net decrease should be 20Mb with both MRs combined. The compression was done with zstd -3

Edited by Angel Campoverde

Merge request reports

Loading