data analysis at cms level 1 trigger
play

Data Analysis at CMS Level-1 Trigger Zhenbin Wu (University - PowerPoint PPT Presentation

Data Analysis at CMS Level-1 Trigger Zhenbin Wu (University of Illinois at Chicago) -- On behalf of the CMS Collaboration 8/2/17 DPF 2017 1 CMS Trigger


  1. Data ¡Analysis ¡at ¡CMS ¡Level-­‑1 ¡ Trigger Zhenbin ¡Wu (University ¡of ¡Illinois ¡at ¡Chicago) -­‑-­‑ On ¡behalf ¡of ¡the ¡CMS ¡Collaboration 8/2/17 DPF ¡2017 1

  2. CMS ¡Trigger ¡System ~30MHz Detectors 100kHz Lvl-1 Frontend pipelines Readout buffers Switching network CMS ¡has ¡been ¡designed ¡with ¡a ¡2-­‑level ¡trigger ¡system 1kHz HLT Processor farms Level-­‑1 ¡trigger (L1): ¡fast ¡electronics • High ¡Level ¡Trigger ¡(HLT): ¡processor ¡farm • 8/2/17 DPF ¡2017 2

  3. Challenge ¡from ¡LHC CMS ,ntegrated LumLnosLty, SS • After ¡LS1, ¡LHC ¡has ¡achieved ¡very ¡high ¡ DDtD included fUom 2010-03-30 11:22 to 2017-07-28 01:31 87C 60 60 luminosity 7otDO ,nteJUDted LumLnoVLty ( fb − 1 ) 2010, 7 7e9, 45.0 pb − 1 2011, 7 7e9, 6.1 fb − 1 • 2016: ¡1.5*10 34 cm -­‑2 s -­‑1 → ¡PU~48 50 50 2012, 8 7e9, 23.3 fb − 1 2015, 13 7e9, 4.2 fb − 1 • 2017: ¡expecting ¡2*10 34 cm -­‑2 s -­‑1 → ¡PU~56 2016, 13 7e9, 40.8 fb − 1 40 40 2017, 13 7e9, 8.9 fb − 1 30 30 20 20 • The ¡CMS ¡trigger ¡needs ¡to ¡maintain ¡and ¡ 10 10 improve ¡performance ¡while ¡keeping ¡the ¡rate ¡ × 50 0 0 U y n O J S t v c u S D u u e c o e J 2 A 0 J A 6 1 D 1 under ¡control 1 1 1 1 1 1 1 1 DDte (87C) CMS PeDk LuPLnosLty Per DDy, SS DDtD included fUom 2010-03-30 11:22 to 2017-07-28 01:31 87C 3eDk DelLveUed LuPLnoVLty ( Hz / nb ) 2010, 7 7e9, PDx. 203.8 Hz /µ b 2017 ¡has ¡ • 2011, 7 7e9, PDx. 4.0 Hz / nb 2012, 8 7e9, PDx. 7.7 Hz / nb 20 20 exceeded ¡2016 ¡ 2015, 13 7e9, PDx. 5.1 Hz / nb 2016, 13 7e9, PDx. 15.3 Hz / nb 2017, 13 7e9, PDx. 16.9 Hz / nb peak ¡luminosity 15 15 10 10 • Expecting ¡2017 ¡ peak ¡luminosity ¡ 5 5 at ¡20Hz/nb × 10 0 0 n n n n n n n D D D D D D D J J J J J J J 1 1 1 1 1 1 1 DDte (87C) 8/2/17 DPF ¡2017 3

  4. Level-­‑1 ¡Trigger The ¡CMS ¡Level-­‑1 ¡trigger ¡system ¡has ¡been ¡fully ¡Phase ¡1 ¡upgraded ¡in ¡2016 8/2/17 DPF ¡2017 4

  5. The ¡Level-­‑1 ¡Menu • The ¡bandwidth ¡allocated ¡per ¡trigger ¡ object ¡type ¡in ¡2016 ¡L1Menu for ¡ 10 34 cm -­‑2 s -­‑1 ¡ luminosity • Note : ¡fractions ¡are ¡inclusive ¡-­‑> ¡no ¡ attempt ¡to ¡correct ¡for ¡overlaps ¡ • Trigger ¡strategy ¡for ¡2017: ¡data ¡ analysis ¡at ¡Level-­‑1 ¡trigger • Object ¡matching ¡with ¡offline ¡ reconstruction ¡ • More ¡sophisticated ¡trigger ¡ algorithm ¡matching ¡to ¡offline ¡ data ¡analysis A. ¡Tapper, ¡ICHEP2016 8/2/17 DPF ¡2017 5

  6. L1 ¡e/ɣ in ¡2017 Dynamic ¡clustering ¡improved ¡energy ¡containment ¡ • and ¡resolution Cluster ¡shape ¡veto: • • Discriminate ¡e/ɣ from ¡jets ¡using ¡cluster ¡ shape ¡and ¡EM ¡energy ¡fraction Cluster ¡energy ¡calibrated ¡with ¡LUT • Isolation ¡re-­‑optimized ¡for ¡2017 ¡scenarios • High ¡p T isolation ¡WP ¡for ¡single ¡e/ɣ trigger • • Low ¡p T isolation ¡WP ¡for ¡cross ¡trigger CMS Preliminary -1 -1 221 pb (2016) + 53 pb (2017) (13 TeV) a.u. 0.08 γ L1 e/ candidates 2017 data 0.07 2016 data 0.06 0.05 Updated ¡calibration ¡ 0.04 improved ¡energy ¡resolution, ¡ 0.03 comparing ¡to ¡2016 0.02 0.01 0 − − − − − 0.5 0.4 0.3 0.2 0.1 0 0.1 0.2 0.3 0.4 0.5 L1 offline offline (E - E )/E T 8/2/17 DPF ¡2017 T 6 T

  7. L1 ¡e/ɣ in ¡2017 -1 -1 -1 -1 CMS Preliminary CMS Preliminary 221 pb (2016) + 53 pb (2017) (13 TeV) 221 pb (2016) + 53 pb (2017) (13 TeV) Efficiency Efficiency 1 1 L1 Single EG Threshold: 38 GeV + isolation 0.8 0.8 0.6 0.6 L1EG(X) OR L1EGIso(X-2) OR L1EGIsoEr(X-4) <PU> = 24 0.4 0.4 2017 Data Z -> ee 2017 Data, X=40 in GeV unit e p >35 GeV 2016 Data T 0.2 0.2 2016 Data, X=42 in GeV unit <PU> = 24 0 0 20 40 60 80 100 10 20 30 40 50 60 70 80 90 100 n E [GeV] vtx T • New ¡isolation ¡WP ¡improved ¡e/ɣ • Isolation ¡and ¡H/E ¡retuned ¡for ¡2017 ¡ efficiency ¡with ¡respect ¡to ¡2016 condition • Two ¡curves ¡at ¡same ¡trigger ¡rate • Better ¡turn ¡on ¡at ¡same ¡trigger ¡rate 8/2/17 DPF ¡2017 7

  8. L1 ¡Muon ¡in ¡2017 L1 ¡muon ¡p T > ¡25 ¡GeV Improved ¡muon ¡efficiency ¡w.r.t 2016: RPC ¡added ¡to ¡both ¡barrel ¡and ¡end-­‑cap ¡track ¡finder • • Retuned ¡the ¡muon ¡p T assignment 8/2/17 DPF ¡2017 8

  9. Level-­‑1 ¡Global ¡Trigger • It ¡is ¡implemented ¡in ¡ μTCA technology ¡with ¡ large ¡FPGAs • It ¡receives ¡and ¡ synchronizes ¡inputs, ¡ and ¡issues ¡the ¡L1 ¡ trigger ¡decisions • High ¡flexibility ¡in ¡trigger ¡ menu ¡design • It ¡supports ¡analysis-­‑like ¡ conditions, ¡such ¡as ¡ invariant ¡mass, ¡ transverse ¡mass ¡ 8/2/17 DPF ¡2017 9

  10. The ¡Vector ¡Boson ¡Fusion ¡Process A ¡Higgs ¡boson ¡can ¡be ¡produced ¡from ¡the ¡ • fusion ¡of ¡the ¡two ¡vector ¡bosons The ¡Higgs ¡boson ¡decay ¡products ¡are ¡often ¡ • located ¡in ¡the ¡central ¡of ¡the ¡detector The ¡outgoing ¡quarks ¡producing ¡high-­‑p T jets ¡ • in ¡the ¡forward ¡region ¡of ¡the ¡detector, ¡ forming ¡large ¡angular ¡separation ¡and ¡large ¡ invariant ¡mass The ¡VBF ¡topology ¡allows ¡selecting ¡region ¡of ¡phase ¡space ¡with ¡excellent ¡signal ¡to ¡ background ¡ratio, ¡is ¡one ¡of ¡the ¡most ¡sensitive ¡categories ¡for ¡the ¡SM ¡H→ 𝜐𝜐 analysis 8/2/17 DPF ¡2017 10

  11. The ¡VBF ¡Trigger ¡Strategy • At ¡least ¡one ¡jet ¡with ¡E T > ¡X • At ¡least ¡two ¡jets ¡with ¡E T > ¡Y • In ¡the ¡collection ¡of ¡jets ¡with ¡E T > ¡Y, ¡at ¡least ¡a ¡pair ¡with ¡m jj > ¡Z -1 CMS CMS Preliminary 6.3 pb (13 TeV) Preliminary (13 TeV) Rate reduction factor Efficiency 2016 Data 0.82 2016 Simulation 1 → τ τ ZeroBias <PU> = 55 VBF H 0.81 0.8 0.8 lead jet at least 1 online jet with E > X GeV T sub jet 0.79 at least 2 online jets with E > 35 GeV 0.6 T and m > 620 GeV jj 0.78 0.4 0.77 0.76 0.2 lead jet at least 1 online (offline) jet with E > X (X') GeV 0.75 T sub jet at least 2 online (offline) jets with E > 35 (45) GeV T and m > 620 (720) GeV jj 0.74 0 30 40 50 60 70 80 90 100 110 120 130 40 60 80 100 120 lead jet lead jet E online [GeV] E offline [GeV] T T At ¡X=90GeV, ¡a ¡~95% ¡rate ¡reduction ¡for ¡only ¡~5% ¡efficiency ¡loss ¡is ¡achieved ¡ 8/2/17 DPF ¡2017 11

  12. The ¡VBF ¡H→ 𝜐𝜐 • A ¡dedicated ¡hadronic ¡ 𝜐𝜐 algorithm ¡designed ¡for ¡SM ¡H→ 𝜐𝜐 analysis • It ¡selects ¡events ¡with ¡a ¡pair ¡of ¡ 𝜐 with ¡p T >32GeV ¡with ¡a ¡ 14kHz ¡allocated ¡trigger ¡bandwidth ¡at ¡Level-­‑1 CMS CMS Preliminary (13 TeV) Preliminary (13 TeV) Events/2.5 GeV Events/40 GeV Only VBF selection Only VBF selection 400 1000 Only DiTau selection Only DiTau selection DiTau AND VBF DiTau AND VBF 350 With ¡VBF ¡trigger, ¡ 800 2016 Simulation 2016 Simulation 300 ~60% ¡gain ¡in ¡ → τ τ → τ τ VBF H VBF H acceptance ¡for ¡an ¡ 250 600 acceptable ¡ 200 400 increase ¡in ¡rate. 150 100 200 50 0 0 20 30 40 50 60 70 80 90 100 110 120 400 600 800 1000 1200 1400 1600 1800 2000 τ sub offline m [GeV] offline p [GeV] jj T 8/2/17 DPF ¡2017 12

  13. Summary • The ¡CMS ¡level-­‑1 ¡trigger ¡system ¡is ¡fully ¡commissioned ¡in ¡2016 • New ¡calibration ¡and ¡tuning ¡improved ¡object ¡performance ¡in ¡ 2017 ¡data ¡taking – While ¡keeping ¡both ¡rate ¡and ¡pileup ¡dependence ¡under ¡ control • Further ¡exploit ¡new ¡functionalities ¡from ¡L1 ¡global ¡trigger, ¡which ¡ support ¡new ¡trigger ¡algorithms – Bringing ¡analysis-­‑like ¡condition ¡down ¡to ¡CMS ¡Level-­‑1 ¡trigger ¡ system 8/2/17 DPF ¡2017 13

  14. BACKUP 8/2/17 DPF ¡2017 14

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend


More recommend