Reduce bandwidth by adding ECal-EOP and HCal-EOP cut for electorn high pt line and Zee line
FYI @hyin @lugrazet, We want to add ECal-EOP and HCal-EOP cut rather than prescale for reducing bandwidth on electron high pt line. Changed pt > 15GeV as 10 GeV and add ECal-EOP and HCal-EOP cut for Zee line. Although the singlehighptelectron line BW has some rising, the total effect can larger reduce the BW ~20MB/s such as WEJet
, WZToEEE
and so on.
Local Estimation of BW
__With Changes:__
Ran over 100,000 events.
Full-stream-spruce filesize = 22M
BW = 1 MHz * 22 MB / 100,000 = `220 MB/s`
__Without Changes:__
Ran over 100,000 events.
Full-stream-spruce filesize = 24M
BW = 1 MHz * 24 MB / 100,000 = `240 MB/s`
From the local BW test, we can reduce ~20MB/s.
Diffing the output for relevant lines I see that between qee_upgrade_24
to this branch there's a large inrease in SingleHighPtElectron events being selected at sprucing yet you report a net negative due to the large reductions elsewhere.
< LAZY_AND: SpruceQEE_DiElectronSameSignHighMassDecisionWithOutput #=592 Sum=12 Eff=|( 2.027027 +- 0.579191)%|
< LAZY_AND: SpruceQEE_SingleHighPtElectronDecisionWithOutput #=592 Sum=10 Eff=|( 1.689189 +- 0.529638)%|
< LAZY_AND: SpruceQEE_SingleHighPtElectronIsoDecisionWithOutput #=592 Sum=2 Eff=|(0.3378378 +- 0.238484)%|
< LAZY_AND: SpruceQEE_SingleHighPtElectronPrescaleDecisionWithOutput #=592 Sum=6 Eff=|( 1.013514 +- 0.411663)%|
< LAZY_AND: SpruceQEE_SingleVHighPtElectronDecisionWithOutput #=592 Sum=4 Eff=|(0.6756757 +- 0.336695)%|
< LAZY_AND: SpruceQEE_ZToEEDecisionWithOutput #=592 Sum=18 Eff=|( 3.040541 +- 0.705683)%|
> LAZY_AND: SpruceQEE_DiElectronSameSignHighMassDecisionWithOutput #=507 Sum=0 Eff=|( 0.000000 +- 0.00000 )%|
> LAZY_AND: SpruceQEE_SingleHighPtElectronDecisionWithOutput #=507 Sum=25 Eff=|( 4.930966 +- 0.961572)%|
> LAZY_AND: SpruceQEE_SingleHighPtElectronIsoDecisionWithOutput #=507 Sum=1 Eff=|(0.1972387 +- 0.197044)%|
> LAZY_AND: SpruceQEE_SingleHighPtElectronPrescaleDecisionWithOutput #=507 Sum=2 Eff=|(0.3944773 +- 0.278387)%|
> LAZY_AND: SpruceQEE_SingleVHighPtElectronDecisionWithOutput #=507 Sum=4 Eff=|(0.7889546 +- 0.392918)%|
> LAZY_AND: SpruceQEE_ZToEEDecisionWithOutput #=507 Sum=0 Eff=|( 0.000000 +- 0.00000 )%|
Also these lines slightly reduce, assumedly due to their dependency on Hlt2_SingleHighPtElectronFull
.
< LAZY_AND: SpruceQEE_WEJetDecisionWithOutput #=592 Sum=6 Eff=|( 1.013514 +- 0.411663)%|
< LAZY_AND: SpruceQEE_WEJetJetDecisionWithOutput #=592 Sum=11 Eff=|( 1.858108 +- 0.555011)%|
< LAZY_AND: SpruceQEE_WGammaToEPhotonDecisionWithOutput #=592 Sum=10 Eff=|( 1.689189 +- 0.529638)%|
< LAZY_AND: SpruceQEE_WZToEEEDecisionWithOutput #=592 Sum=2 Eff=|(0.3378378 +- 0.238484)%|
> LAZY_AND: SpruceQEE_WEJetDecisionWithOutput #=507 Sum=2 Eff=|(0.3944773 +- 0.278387)%|
> LAZY_AND: SpruceQEE_WEJetJetDecisionWithOutput #=507 Sum=10 Eff=|( 1.972387 +- 0.617542)%|
> LAZY_AND: SpruceQEE_WGammaToEPhotonDecisionWithOutput #=507 Sum=5 Eff=|(0.9861933 +- 0.438859)%|
> LAZY_AND: SpruceQEE_WZToEEEDecisionWithOutput #=507 Sum=1 Eff=|(0.1972387 +- 0.197044)%|
Edited by Jiuzhao Li