The ATLAS Trigger System in Run-2 Rhys Owen 1 University of - - PowerPoint PPT Presentation

▶
the atlas trigger system in run 2
SMART_READER_LITE
LIVE PREVIEW

The ATLAS Trigger System in Run-2 Rhys Owen 1 University of - - PowerPoint PPT Presentation

Feb 05, 2024 351 likes •834 views

The ATLAS Trigger System in Run-2 Rhys Owen 1 University of Birmingham 1 14 Febuary 2018 Particle Physics Seminar Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 1 / 29 Introduction In Run-2 of the LHC

slide-1
SLIDE 1

The ATLAS Trigger System in Run-2

Rhys Owen1

University of Birmingham1 14 Febuary 2018 Particle Physics Seminar

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 1 / 29

slide-2
SLIDE 2

Introduction

In Run-2 of the LHC increased centre-of-mass energies and instantaneous luminosity have lead to increases in the trigger rate but this is constrained by hardware requirements. The easiest solution to reduce the rate again would be to increase the energy thresholds used by the trigger, however this would severely curtail the ATLAS physics programme. This required significant upgrades at Level-1 and optimisations in the HLT to maintain signal efficiency while reducing the rate of events.

EgammaTriggerPublicResults

Entries / 2 GeV

10 20 30 40 50 60

3

10 ×

ν e → W ATLAS

  • 1

13 TeV, 81 pb

Data

  • Syst. Unc.

⊕ MC Stat. ν e → W Multijet

  • e

+

e → Z ν τ → W Other backgrounds

[GeV]

e T

p 20 30 40 50 60 70 80 90 100 Data / Pred. 0.8 0.9 1 1.1 1.2

  • Phys. Lett. B 759 (2016) 601

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 2 / 29

slide-3
SLIDE 3

The ATLAS Detector

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 3 / 29

slide-4
SLIDE 4

The ATLAS Detector: Sub-detectors

General purpose detector at the LHC. Several detector technologies and components used to detect and identify final state particles. Can be roughly split into layers, tracking, calorimetry and muon spectrometry. Responsibility of the trigger and data acquisition system to select and record “interesting” events at a reduced rate to disk. Due to detector design different information available to trigger system as the trigger decision progresses.

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 4 / 29

slide-5
SLIDE 5

Run 2 Conditions

LHC bunches filled with protons collide at 40 MHz Providing an instantaneous luminosity which peaked at 20.6 × 1033 cm−2 s−1 This leads to a large number of p-p interactions which could all produce a signature of interest.

Day in 2017 01/05 02/06 05/07 07/08 08/09 11/10 12/11 15/12 ]

  • 1

s

  • 2

cm

33

Peak Luminosity per Fill [10 5 10 15 20 25

= 13 TeV s

ATLAS Online Luminosity

LHC Stable Beams

  • 1

s

  • 2

cm

33

10 × Peak Lumi: 20.6

initial calibration

Mean Number of Interactions per Crossing 10 20 30 40 50 60 70 80 /0.1]

  • 1

Recorded Luminosity [pb 50 100 150 200 250 300 350 Online, 13 TeV ATLAS

  • 1

Ldt=86.5 fb

∫

> = 13.4 µ 2015: < > = 25.1 µ 2016: < > = 38.1 µ 2017: < > = 32.0 µ Total: <

initial 2017 calibration

https://twiki.cern.ch/twiki/bin/view/AtlasPublic/LuminosityPublicResultsRun2 Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 5 / 29

slide-6
SLIDE 6

The ATLAS Detector: Trigger / DAQ

Level-1 Level-1 Accept Level-1 Muon Endcap sector logic Barrel sector logic Level-1 Calo CP (e,γ,τ) CMX JEP (jet, E) CMX Central Trigger MUCTPI L1Topo CTP CTPCORE CTPOUT Preprocessor nMCM Detector Read-Out ROD FE ROD FE ROD FE

...

DataFlow Read-Out System (ROS) Data Collection Network Data Storage Muon detectors Calorimeter detectors High Level Trigger (HLT) Processors RoI Event Data Fast TracKer (FTK) TileCal Accept Pixel/SCT Tier-0

≈ 1.5 kHz 100 kHz 40 MHz Level-1 - reduced granularity information at full rate HLT - full granularity information at reduced rate

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 6 / 29

slide-7
SLIDE 7

Level-1 Trigger

Level-1 - reduced granularity information at full rate Hardware based trigger Primarily derived from calorimeter and muon systems Provides a rate reduction from 40 MHz to 100 kHz limited by the maximum readout rate of the front end electronics. Also provides Regions Of Interest (ROIs) as the starting point for software algorithms. Significant hardware and firmware updates in Run-2

Level-1 Level-1 Accept Level-1 Muon Endcap sector logic Barrel sector logic Level-1 Calo CP (e,γ,τ) CMX JEP (jet, E) CMX Central Trigger MUCTPI L1Topo CTP CTPCORE CTPOUT Preprocessor nMCM Muon detectors Calorimeter detectors RoI TileCal

EventDisplayRun2Physics Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 7 / 29

slide-8
SLIDE 8

Level-1 Trigger: Updates

The largest update was the inclusion of Topological triggering with the L1Topo module. Other systems need to provide L1Topo with information This is done with Trigger OBjects (TOBs) which represent the potential physics objects which have been detected. Similar to the ROIs which are sent to the HLT.

Level-1 Level-1 Accept Level-1 Muon Endcap sector logic Barrel sector logic Level-1 Calo CP (e,γ,τ) CMX JEP (jet, E) CMX Central Trigger MUCTPI L1Topo CTP CTPCORE CTPOUT Preprocessor nMCM Muon detectors Calorimeter detectors RoI TileCal

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 8 / 29

slide-9
SLIDE 9

Level-1 Trigger: L1Calo

The Level-1 calorimeter trigger. Analogue sum of calorimeter cells provided by both electromagnetic and hadronic calorimeter. Fast digitisation performed to produce “trigger towers” (typically 0.1 × 0.1 in ∆η × ∆φ) Separate sub-systems then search for clusters compatible with electromagnetic, tau and hadronic jet like energy deposits Cables carrying analogue signals from calorimeters.

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 9 / 29

slide-10
SLIDE 10

Level-1 Trigger: L1Calo

The Level-1 calorimeter trigger. Analogue sum of calorimeter cells provided by both electromagnetic and hadronic calorimeter. Fast digitisation performed to produce “trigger towers” (typically 0.1 × 0.1 in ∆η × ∆φ) Separate sub-systems then search for clusters compatible with electromagnetic, tau and hadronic jet like energy deposits The electromagnetic algorithm is based on windows such as this, where the sums of towers around a local maximum are calculated.

Vertical sums

! !

Horizontal sums

! ! ! !

Electromagnetic isolation ring Hadronic inner core and isolation ring Electromagnetic calorimeter Hadronic calorimeter Trigger towers ("# × "$ = 0.1 × 0.1) Local maximum/ Region-of-interest

JINST 3 (2008) P03001 Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 9 / 29

slide-11
SLIDE 11

Level-1 Trigger: L1Calo - Run 2 Upgrades

Digitisation

◮ nMCM - new Multi Chip Module,

updated digitisation and dynamic baseline subtraction.

Processing

◮ CPM - Cluster Processor Module,

updated algorithm to allow ET-dependent isolation

Architecture

◮ CMX - Common Merger

eXtended, merge Trigger OBjects (TOBs) instead of threshold counts and forward to the Level-1 topological system.

]

  • 1

s

  • 2

cm

30

Instantaneous luminosity / bunch [10 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Average L1_XE50 rate / bunch [Hz] 0.5 1 1.5 2 2.5 ATLAS Operations

= 13 TeV s 2015 Data, 50 ns pp Collision Data without pedestal correction with pedestal correction

L1CaloTriggerPublicResults Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 10 / 29

slide-12
SLIDE 12

Level-1 Trigger: L1Calo - Run 2 Upgrades

Digitisation

◮ nMCM - new Multi Chip Module,

updated digitisation and dynamic baseline subtraction.

Processing

◮ CPM - Cluster Processor Module,

updated algorithm to allow ET-dependent isolation

Architecture

◮ CMX - Common Merger

eXtended, merge Trigger OBjects (TOBs) instead of threshold counts and forward to the Level-1 topological system.

[GeV]

T

E 10 20 30 40 50 60 70 80 90 100 Efficiency 0.2 0.4 0.6 0.8 1 1.2 1.4

L1_EM20VH L1_EM20VHI

ATLAS Preliminary

  • 1

=13 TeV, 60.3 pb s Data 2016, EgammaTriggerPublicResults Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 10 / 29

slide-13
SLIDE 13

Level-1 Trigger: L1Muon

The Level-1 muon trigger is based

  • n dedicated triggering chambers

RPCs (TGCs) found in the barrel (endcap)

2 4 6 8 10 12 14 m 16 2 4 6 8 10 12 m

Large (odd numbered) sectors

BIL BML BOL EEL EIL CSC 1 2 3 4 5 6 EIL4 1 2 3 4 5 6 1 2 3 4 5 6 TGCs 1 2 3 4 5 1 2

3

End-cap magnet RPCs y 1 2

TGCs

EEL 2

End-cap toroid z

η=2.4 η=1.3 η=1.0

TGC-FI

η=1.9

TileCal

  • Eur. Phys. J. C 77 (2017) 317

Green: Active, Red: Ready for 2018 data taking.

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 11 / 29

slide-14
SLIDE 14

Level-1 Trigger: L1Muon - Run 2 Upgrades

Algorithm

◮ Additional logic requiring a

coincidence between the inner TGC layers (TGC-FI) or the TileCal and the outer layers. Reducing the trigger rate by up to 10% for the unprescaled muon trigger.

Coverage

◮ Additional RPC chambers made

  • perational in the bottom of the

spectrometer increase coverage by 3.6%.

Architecture

◮ An additional module

MUCTPI2TOPO was introduced to transmit muon TOBs to the Level-1 topological system.

2 4 6 8 10 12 14 m 16 2 4 6 8 10 12 m

Large (odd numbered) sectors

BIL BML BOL EEL EIL CSC 1 2 3 4 5 6 EIL4 1 2 3 4 5 6 1 2 3 4 5 6 TGCs 1 2 3 4 5 1 2

3

End-cap magnet RPCs y 1 2

TGCs

EEL 2

End-cap toroid z

η=2.4 η=1.3 η=1.0

TGC-FI

η=1.9

TileCal

  • Eur. Phys. J. C 77 (2017) 317

Green: Active, Red: Ready for 2018 data

  • taking. Arrow indicates path of background

beam particle.

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 12 / 29

slide-15
SLIDE 15

Level-1 Trigger: L1Muon - Run 2 Upgrades

Algorithm

◮ Additional logic requiring a

coincidence between the inner TGC layers (TGC-FI) or the TileCal and the outer layers. Reducing the trigger rate by up to 10% for the unprescaled muon trigger.

Coverage

◮ Additional RPC chambers made

  • perational in the bottom of the

spectrometer increase coverage by 3.6%.

Architecture

◮ An additional module

MUCTPI2TOPO was introduced to transmit muon TOBs to the Level-1 topological system.

L1_MU15

η

3 − 2 − 1 − 1 2 3 Number of triggers [nb-1] 20 40 60 80 100 120 ATLAS =13TeV s

  • 1

L dt = 11.1 pb

∫

L1_MU15 w/o FI coincidence,

  • 1

L dt = 20.6 pb

∫

L1_MU15 w/ FI coincidence, L1_MU15

η

3 − 2 − 1 − 1 2 3 Rate reduction 0.2 − 0.2 0.4 0.6 0.8 1 ATLAS =13TeV s L1_MU15 rate reduction

  • Eur. Phys. J. C 77 (2017) 317

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 12 / 29

slide-16
SLIDE 16

Level-1 Trigger: L1Muon - Run 2 Upgrades

Algorithm

◮ Additional logic requiring a

coincidence between the inner TGC layers (TGC-FI) or the TileCal and the outer layers. Reducing the trigger rate by up to 10% for the unprescaled muon trigger.

Coverage

◮ Additional RPC chambers made

  • perational in the bottom of the

spectrometer increase coverage by 3.6%.

Architecture

◮ An additional module

MUCTPI2TOPO was introduced to transmit muon TOBs to the Level-1 topological system.

φ

  • ffline muon

2 − 2

Efficiency

0.5 1

  • 1

L1_MU20, Data 2015, 3.2 fb

  • 1

L1_MU20, Data 2016, 127 pb

ATLAS Preliminary

=13 TeV s=13 TeV s | < 1.05

µ

η > 25 GeV, |

µ T

, p µ µ → Z | < 1.05

µ

η > 25 GeV, |

µ T

, p µ µ → Z

MuonTriggerPublicResults Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 12 / 29

slide-17
SLIDE 17

Level-1 Trigger: L1Topo

Receives TOBs from both L1Calo and L1Muon systems

◮ Muon TOBs represent reduced

granularity in η/φ and have three energy thresholds.

◮ Calo TOBs retain the L1Calo

granularity and contain isolation information.

Topological combinations of trigger

  • bjects add discrimination allowing

low thresholds to be maintained.

  • Eur. Phys. J. C 77 (2017) 317

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 13 / 29

slide-18
SLIDE 18

Level-1 Trigger: L1Topo

An example are 2τ triggers. As used for the H → ττ analysis. The di-tau system is expected to be boosted and therefore have a small ∆R separation. Adding a requirement ∆R < 2.9 at Level-1 leads to a significant reduction in rates.

]

  • 1

s

  • 2

cm

33

  • Inst. luminosity [10

9 9.5 10 10.5 11 11.5 12 Rate [kHz] 10

2

10

53.5 22.9 6.7 5.9 3.8

= 13 TeV s Data 2016, ATLAS Trigger Operation

>12 GeV without isolation cut

T

2

τ

>20 GeV, p

T

1

τ

L1: p >12 GeV

T

2

τ

>20 GeV, p

T

1

τ

L1: p >25 GeV

T jet

>12 GeV, p

T

2

τ

>20 GeV, p

T

1

τ

L1: p )<2.9

2

τ ,

1

τ R( ∆ >12 GeV,

T

2

τ

>20 GeV, p

T

1

τ

L1Topo: p >25 GeV

T jet

)<2.9, p

2

τ ,

1

τ R( ∆ >12 GeV,

T

2

τ

>20 GeV, p

T

1

τ

L1Topo: p

ATLAS-CONF-2017-061 Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 14 / 29

slide-19
SLIDE 19

Level-1 Trigger: L1Topo

An example are 2τ triggers. As used for the H → ττ analysis. The di-tau system is expected to be boosted and therefore have a small ∆R separation. Adding a requirement ∆R < 2.9 at Level-1 leads to a significant reduction in rates.

) τ , τ Roffline( ∆ 0.5 1 1.5 2 2.5 3 Efficiency 0.2 0.4 0.6 0.8 1

ATLAS Preliminary = 13 TeV s Data 2016, L1Topo Commissioning

)<2.9

2

τ ,

1

τ R( ∆ >12 GeV,

T

2

τ

>20 GeV, p

T

1

τ

: p L1Topo

ATLAS-CONF-2017-061 Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 14 / 29

slide-20
SLIDE 20

HLT

The higher level trigger runs

  • ffline-like algorithms

Final trigger decision is an OR of many independent trigger chains. Each chain is defined as a series of algorithms with the ability to abort execution part way though to save CPU.

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 15 / 29

slide-21
SLIDE 21

HLT

ROS

Patch panel ATLAS cavern Patch panel Surface 150 m

TPU SFO

RoIB

x98 4 x 10 Gbps 1 Gbps x6 x40 x50 8 x 10 Gbps x2 2x 10 Gbps 8 x 10 Gbps

Router cluster

CERN Permanent Storage

sw-data-tpu-xx

sw-data-core-01 sw-data-core-02

sw-data-edge-01

2x 10 Gbps 2 x 10 Gbps 0 -20 % 20 -40 % 40 -60 % 60 -80 % 80 -100 %

Link utjlizatjon

2 x 10 Gbps

Detector Readout Level – 1 Regions

  • f Interest

https://twiki.cern.ch/twiki/bin/view/AtlasPublic/ApprovedPlotsDAQ

In Run-1 HLT consisted of two levels, the first one with faster algorithms and mostly regional reconstruction, and the second one with full event reconstruction with higher precision. Updated in Run-2 to be an integrated system to save resources and simplify processing.

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 16 / 29

slide-22
SLIDE 22

HLT: Electrons and Photons - Algorithm

Trigger reconstruction of electrons and photons share a similar chain of algorithms. Both seeded by L1Calo EM regions

  • f interest.

Calorimeter clustering is performed using higher granularity calorimeter cells (typically 0.025 × 0.025 in ∆η × ∆φ) Precision tracks extrapolated to the second layer of the EM calorimeter. Electrons use a likelihood based identification using calorimeter, tracking and transition radiation information. Photon identification based on calorimeter variables only.

Level-1 EM RoI Calorimeter Clustering Within 0.4x0.4 (ΔηxΔφ) Fast Tracking Within 0.2 (Δη) Offline-like cluster calibration Precision Tracking Within 0.05x0.05 (ΔηxΔφ) Particle Identification

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 17 / 29

slide-23
SLIDE 23

HLT: Electrons and Photons - Performance

[GeV]

T

Offline isolated photon E

2

10

3

10 Trigger Efficiency 0.2 0.4 0.6 0.8 1 1.2 1.4

HLT_g25_medium_L1EM20VH HLT_g35_medium_L1EM20VH HLT_g140_ HLT_g200_loose

ATLAS Preliminary

= 13 TeV s Data 2017,

tight

EgammaTriggerPublicResults

[GeV]

T

Offline isolated electron E 20 40 60 80 100 120 140 Trigger Efficiency 0.2 0.4 0.6 0.8 1 1.2 1.4

HLT_e26_lhtight_nod0_ivarloose Data ee MC → Z

ATLAS Preliminary

  • 1

=13 TeV, 15.4 fb s Data 2017, EgammaTriggerPublicResults

The electromagnetic triggers performed well during 2017. The single unprescaled electron threshold was maintained at 26 GeV with a loose track based isolation. The single unprescaled photon threshold was 140 GeV.

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 18 / 29

slide-24
SLIDE 24

HLT: Muons - Algorithm

Muon reconstruction proceeds in two stages. A first “fast” reconstruction is performed on each Level-1 muon candidate with the pT assigned by a lookup table based on MDT measurements. These tracks are then extrapolated to the inner detector to create combined muons. The second “Precision” pass produces a more accurate fit of the track at the cost of processing speed.

Processing time per RoI [ms] 20 40 60 80 100 120 Normalised entries

4 −

10

3 −

10

2 −

10

1 −

10

MS-only <time>: 8.2 ms Combined <time>: 6.2 ms

ATLAS = 13 TeV s Processing time per RoI [ms] 500 1000 1500 2000 2500 3000 Normalised entries

3 −

10

2 −

10

1 −

10 1

Precision Muon <time>: 239.5 ms

ATLAS = 13 TeV s

  • Eur. Phys. J. C 77 (2017) 317

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 19 / 29

slide-25
SLIDE 25

HLT: Muons - Performance

Barrel

[GeV]

T

  • ffline muon p

20 40 60 80 100

Efficiency

0.5 1

L1 MU20 HLT mu26_ivarmedium or mu60 HLT mu26_ivarmedium or mu60 with respect to L1

ATLAS Preliminary

  • 1

=13 TeV, Data 2017, 15 fb s

  • 1

=13 TeV, Data 2017, 15 fb s µ µ → Z | < 1.05

µ

η | µ µ → Z | < 1.05

µ

η | MuonTriggerPublicResults

End-cap

[GeV]

T

  • ffline muon p

20 40 60 80 100

Efficiency

0.5 1

L1 MU20 HLT mu26_ivarmedium or mu60 HLT mu26_ivarmedium or mu60 with respect to L1

ATLAS Preliminary

  • 1

=13 TeV, Data 2017, 15 fb s

  • 1

=13 TeV, Data 2017, 15 fb s µ µ → Z | < 2.4

µ

η 1.05 < | µ µ → Z | < 2.4

µ

η 1.05 < | MuonTriggerPublicResults

HLT Muon reconstruction is ≈ 100% with respect to the Level-1 trigger. Single unprescaled muon threshold set at 26 GeV

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 20 / 29

slide-26
SLIDE 26

HLT: Tau Leptons - Algorithm

The algorithm starts from the Level-1 TAU ROI. Two-stage fast tracking

◮ First a leading PT track is

identified within ∆R < 0.1 of the cluster centre.

◮ Further tracks are then identified

∆R < 0.4 from the leading track but originating within a fixed window along the beam pipe.

Tracks are counted as Core ∆R < 0.2 or Wide 0.2 < ∆R < 0.4 Particle identification is provided by a boosted decision tree similar to that used offline

Level-1 TAU RoI Calorimeter Clustering Within 0.2 (ΔR=√Δη²xΔφ²) Two Stage Fast Tracking 1 – 3 Core Tracks Precision Tracking Particle Identification 1 or Fewer Wide Tracks

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 21 / 29

slide-27
SLIDE 27

HLT: Tau Leptons - Algorithm

The algorithm starts from the Level-1 TAU ROI. Two-stage fast tracking

◮ First a leading PT track is

identified within ∆R < 0.1 of the cluster centre.

◮ Further tracks are then identified

∆R < 0.4 from the leading track but originating within a fixed window along the beam pipe.

Tracks are counted as Core ∆R < 0.2 or Wide 0.2 < ∆R < 0.4 Particle identification is provided by a boosted decision tree similar to that used offline

10

2015 √s =

  • 10

2015 √s =

  • One-stage tracking RoI

Two-stage tracking: 1st stage RoI Two-stage tracking: 2nd stage RoI Plan view beam line

  • Eur. Phys. J. C 77 (2017) 317

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 21 / 29

slide-28
SLIDE 28

HLT: Tau Leptons - Algorithm

The algorithm starts from the Level-1 TAU ROI. Two-stage fast tracking

◮ First a leading PT track is

identified within ∆R < 0.1 of the cluster centre.

◮ Further tracks are then identified

∆R < 0.4 from the leading track but originating within a fixed window along the beam pipe.

Tracks are counted as Core ∆R < 0.2 or Wide 0.2 < ∆R < 0.4 Particle identification is provided by a boosted decision tree similar to that used offline

[GeV]

T

Offline track p 1 2 3 4 5 67 10 20 30

2

10 Efficiency 0.9 0.92 0.94 0.96 0.98 1 1.02

Fast Track Finder (Stage 1) Fast Track Finder (Stage 2) Precision Tracking (Stage 2) Fast Track Finder (Stage 1) Fast Track Finder (Stage 2) Precision Tracking (Stage 2) Fast Track Finder (Stage 1) Fast Track Finder (Stage 2) Precision Tracking (Stage 2)

ATLAS

Data 2015 √s = 13 TeV > 1 GeV

T

  • ffline track p

25 GeV Tau Trigger

_

  • Eur. Phys. J. C 77 (2017) 317

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 21 / 29

slide-29
SLIDE 29

HLT: Tau Leptons - Algorithm

The algorithm starts from the Level-1 TAU ROI. Two-stage fast tracking

◮ First a leading PT track is

identified within ∆R < 0.1 of the cluster centre.

◮ Further tracks are then identified

∆R < 0.4 from the leading track but originating within a fixed window along the beam pipe.

Tracks are counted as Core ∆R < 0.2 or Wide 0.2 < ∆R < 0.4 Particle identification is provided by a boosted decision tree similar to that used offline

Processing time per RoI [ms] 50 100 150 200 250 Normalised entries

  • 5

10

  • 4

10

  • 3

10

  • 2

10

  • 1

10 1 ATLAS Operations

Data 2015 √s = 13 TeV Tau trigger: Fast Track Finder _ _ Two-stage: 1st stage mean = 23.1 ± 0.11 ms Single-stage: mean = 66.2 ± 0.34 ms ___ Two-stage: 2nd stage mean = 21.4 ± 0.09 ms . . .

_

  • Eur. Phys. J. C 77 (2017) 317

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 21 / 29

slide-30
SLIDE 30

HLT: Tau Leptons - Algorithm

The algorithm starts from the Level-1 TAU ROI. Two-stage fast tracking

◮ First a leading PT track is

identified within ∆R < 0.1 of the cluster centre.

◮ Further tracks are then identified

∆R < 0.4 from the leading track but originating within a fixed window along the beam pipe.

Tracks are counted as Core ∆R < 0.2 or Wide 0.2 < ∆R < 0.4 Particle identification is provided by a boosted decision tree similar to that used offline

Level-1 TAU RoI Calorimeter Clustering Within 0.2 (ΔR=√Δη²xΔφ²) Two Stage Fast Tracking 1 – 3 Core Tracks Precision Tracking Particle Identification 1 or Fewer Wide Tracks

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 21 / 29

slide-31
SLIDE 31

HLT: Tau Leptons

Efficiency

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6

τ τ → MC Z MC stat. error Data 2016 Data syst. error Data stat. + sys. error

ATLAS Preliminary

  • 1

= 13 TeV, 33 fb s , 1-prong

had

τ

µ

τ → Z HLT tau25 medium trigger

[GeV]

T

Offline tau p

30 40 50 60 70 100 200 300 Data/exp. 0.5 1 1.5 ATLAS-CONF-2017-061

Efficiency

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6

τ τ → MC Z MC stat. error Data 2016 Data syst. error Data stat. + sys. error

ATLAS Preliminary

  • 1

= 13 TeV, 33 fb s , 3-prong

had

τ

µ

τ → Z HLT tau25 medium trigger

[GeV]

T

Offline tau p

30 40 50 60 70 100 200 300 Data/exp. 0.5 1 1.5 ATLAS-CONF-2017-061

Single Tau threshold set at 160 GeV Use of two level tracking essential to identify candidates against increasing hadronic backgrounds.

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 22 / 29

slide-32
SLIDE 32

HLT: Missing Transverse Momentum

The increased number of hadronic interactions makes these triggers sensitive to the increase in instantaneous luminosity. The improvements to the Level-1 digitisation mean it is possible to keep the threshold relatively low (50 GeV) This is needed for a typical analysis selection of 200 GeV Several algorithms are run in parallel but due to the overlap between the resource intensive parts (clustering) this does not add much overhead. The algorithm pufit is used extensively to reduce the rate from pile up contributions.

(Z) [GeV]

T

p

50 100 150 200 250 300

Efficiency

0.2 0.4 0.6 0.8 1

  • 1

= 13 TeV, 15.4 fb s Data 2017, µ µ → Z

ATLAS Preliminary

L1_XE50 HLT_xe110_pufit_L1XE50

MissingEtTriggerPublicResults Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 23 / 29

slide-33
SLIDE 33

HLT: Missing Transverse Momentum

The increased number of hadronic interactions makes these triggers sensitive to the increase in instantaneous luminosity. The improvements to the Level-1 digitisation mean it is possible to keep the threshold relatively low (50 GeV) This is needed for a typical analysis selection of 200 GeV Several algorithms are run in parallel but due to the overlap between the resource intensive parts (clustering) this does not add much overhead. The algorithm pufit is used extensively to reduce the rate from pile up contributions.

> µ < 10 15 20 25 30 35 40 45 50 55 Trigger cross section [nb] 10 20 30 40 50 60 ATLAS Trigger Operations

HLT_xe110_mht_L1XE50 HLT_xe110_pufit_L1XE50

= 13 TeV s Data 2016 / 2017,

MissingEtTriggerPublicResults

pufit Details: Eur. Phys. J. C 77 (2017) 317

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 23 / 29

slide-34
SLIDE 34

HLT: Hadronic Jets

Jet triggers cover single and multi-jet topologies Jets are constructed using the anti-kT algorithm operating on calorimeter clusters. Radius parameters 0.4 and 1.0 are used. Some chains also include tracking information in order to improve the resolution subject to resource constraints.

380 400 420 440 460 480 500 520 540 [GeV]

T

p Leading offline jet 0.2 0.4 0.6 0.8 1 1.2 Per-event trigger efficiency

> 450 GeV

T

p HLT, 2016 calibration 2017 calib., calorimeter only 2017 calib., with tracks

ATLAS Preliminary

1 −

= 13 TeV, 21.9 fb s Data 2017, |<2.8 η 1 jet with | ≥ Offline selection:

JetTriggerPublicResults Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 24 / 29

slide-35
SLIDE 35

HLT: b-Jets

Several analyses rely on “b-jets” where the jet is initiated by the decay of a B hadron indicating a bottom quark in the final state. For example H → b¯ b The trigger uses the MV2 algorithm which uses inputs from the impact parameter, displaced vertexing and jet structure algorithms in a configuration close to the offline configuration. Two stage fast tracking is again employed to aid in the finding of the primary vertex avoiding the performance cost of having to perform tracking over the whole detector.

Number of tracks 10 20 30 40 50 60 70 80 Vertex finding efficiency 0.2 0.4 0.6 0.8 1 1.2

T T T T T T

> 55 GeV > 55 GeV > 55 GeV > 110 GeV > 110 GeV > 110 GeV > 260 GeV > 260 GeV > 260 GeV

T T T

Jet trigger E Jet trigger E Jet trigger E

ATLAS

Data 2015 √s = 13 TeV

  • ffline track pT > 1 GeV

b-jet trigger vertex tracking _

  • Eur. Phys. J. C 77 (2017) 317

Full algorithm details: Eur. Phys. J. C 77 (2017) 317

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 25 / 29

slide-36
SLIDE 36

HLT: b-Jets

Several analyses rely on “b-jets” where the jet is initiated by the decay of a B hadron indicating a bottom quark in the final state. For example H → b¯ b The trigger uses the MV2 algorithm which uses inputs from the impact parameter, displaced vertexing and jet structure algorithms in a configuration close to the offline configuration. Two stage fast tracking is again employed to aid in the finding of the primary vertex avoiding the performance cost of having to perform tracking over the whole detector.

10 20 30 40 50 60 70 Mean Number of Interactions per Crossing 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 Efficiency w.r.t. Offline b-tagging

Online 77% Online 70% Online 60% Online 50% Online 40%

ATLAS Preliminary

= 13 TeV s Data 2017 Offline b-tagging efficiency 70% BJetTriggerPublicResults

Full algorithm details: Eur. Phys. J. C 77 (2017) 317

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 25 / 29

slide-37
SLIDE 37

HLT: B-Physics

Several di-muon triggers are defined for selecting J/ψ, B and Υ(nS) states. These rely on a low di-muon threshold at Level-1 and the relevant invariant mass selection in the HLT. Even small increases in the threshold for either leg can have a large effect on the efficiency.

) [GeV]

−

µ

+

µ m( 3 4 5 6 7 8 9 10 11 12 Entries / 10 MeV

3

10

4

10

5

10

6

10

7

10

ψ J/ B S) n ( ϒ

) > 4 GeV

2

µ (

T

p ) > 4 GeV,

1

µ (

T

p ) > 4 GeV

2

µ (

T

p ) > 6 GeV,

1

µ (

T

p ) > 6 GeV

2

µ (

T

p ) > 6 GeV,

1

µ (

T

p

) > 20 GeV

1

µ (

T

p Single muon trigger: ) > 4 GeV

2

µ (

T

p ) > 4 GeV,

1

µ (

T

p Supporting dimuon trigger:

ATLAS Preliminary

= 13 TeV s

  • 1

Ldt = 3.2 fb

∫

BPhysicsTriggerPublicResults Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 26 / 29

slide-38
SLIDE 38

HLT: B-Physics

Several di-muon triggers are defined for selecting J/ψ, B and Υ(nS) states. These rely on a low di-muon threshold at Level-1 and the relevant invariant mass selection in the HLT. Even small increases in the threshold for either leg can have a large effect on the efficiency. A good example of where L1Topo can help alleviate a Level-1 bottleneck.

Luminosity block [~60s] 100 120 140 160 180 200 220 240 Rate [Hz] 1000 2000 3000 4000 5000 6000

> 6 GeV

µ T

L1: 2 x p [0.2,1.5] ∈

µ µ

R ∆ [2, 9] GeV, ∈

µ µ

> 6 GeV, m

µ T

L1Topo: 2 x p

ATLAS Trigger Operations = 13 TeV s Data 2017, Run taken on Jun 17, 2017 TriggerOperationPublicResults Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 26 / 29

slide-39
SLIDE 39

Trigger Level Analysis

There are not only rate restrictions at Level-1. It is also important to consider the rate to disk from the HLT and the available resources for prompt reconstruction. Jets for example have a Level-1 threshold O(100GeV ) but a HLT threshold O(400GeV ). One solution being considered is to perform the analysis selection online in the trigger and vastly decrease the data volume by only saving the selected objects. The plot shows a search for Di-Jet resonances using this technique.

Events 1

2

10

4

10

6

10

8

10

[TeV]

jj

m

500 600 700 800 900 1000

TLA/Offline (j110) 0.5 1 1.5

ATLAS Preliminary

  • 1

s=13 TeV, 3.4 fb |y*| < 0.6 TLA jets Offline jets selected by any single-jet trigger Offline jets selected by j110 single-jet trigger

ATLAS-CONF-2016-030 Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 27 / 29

slide-40
SLIDE 40

Trigger Level Analysis II

TLA represents a high HLT rate but tiny bandwidth user.

TriggerOperationPublicResults Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 28 / 29

slide-41
SLIDE 41

Trigger Level Analysis II

TLA represents a high HLT rate but tiny bandwidth user.

TriggerOperationPublicResults Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 28 / 29

slide-42
SLIDE 42

Summary

The LHC is performing well and delivering instantaneous luminosities above its design value. The increased number of interactions per bunch crossing add pressure to the trigger system due to the increased event complexity. Several notable improvements to the trigger system during the first LHC long shutdown provide good tools to mitigate these challenges.

◮ Level-1 improved calorimeter isolation and the introduction of topological

triggering can avoid a bottle neck at the front end readout and help the HLT by providing better seeds.

◮ The single stage HLT allows for chains with a flexible set of algorithms which

can share outputs reducing any unnecessary duplication of calculations.

Given the anticipated running conditions in 2018 the trigger will also be able to perform well for the rest of Run-2 before the next round of planned updates.

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 29 / 29

slide-43
SLIDE 43

Backup

Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 30 / 29

slide-44
SLIDE 44

Trigger Rates from 2015 @ 5 × 1033cm−2s−1

Trigger Typical offline selection Trigger Selection Level-1 Rate HLT Rate Level-1 [GeV] HLT [GeV] [kHz] [Hz] L = 5 × 1033 cm−2s−1 Single leptons Single iso µ, pT > 21GeV 15 20 7 130 Single e, pT > 25GeV 20 24 18 139 Single µ, pT > 42GeV 20 40 5 33 Single τ, pT > 90GeV 60 80 2 41 Two leptons Two µ’s, each pT > 11GeV 2 × 10 2 × 10 0.8 19 Two µ’s, pT > 19, 10GeV 15 18, 8 7 18 Two loose e’s, each pT > 15GeV 2 × 10 2 × 12 10 5 One e & one µ, pT > 10, 26GeV 20 (µ) 7, 24 5 1 One loose e & one µ, pT > 19, 15GeV 15, 10 17, 14 0.4 2 Two τ’s, pT > 40, 30GeV 20, 12 35, 25 2 22 One τ, one µ, pT > 30, 15GeV 12, 10 (+jets) 25, 14 0.5 10 One τ, one e, pT > 30, 19GeV 12, 15 (+jets) 25, 17 1 3.9 Three leptons Three loose e’s, pT > 19, 11, 11GeV 15, 2 × 7 17, 2 × 9 3 < 0.1 Three µ’s, each pT > 8GeV 3 × 6 3 × 6 < 0.1 4 Three µ’s, pT > 19, 2 × 6GeV 15 18, 2 × 4 7 2 Two µ’s & one e, pT > 2 × 11, 14GeV 2 × 10 (µ’s) 2 × 10, 12 0.8 0.2 Two loose e’s & one µ, 2 × 8, 10 2 × 12, 10 0.3 < 0.1 pT > 2 × 11, 11GeV One photon One γ, pT > 125GeV 22 120 8 20 Two photons Two loose γ’s, pT > 40, 30GeV 2 × 15 35, 25 1.5 12 Two tight γ’s, pT > 25, 25GeV 2 × 15 2 × 20 1.5 7 Single jet Jet (R = 0.4), pT > 400GeV 100 360 0.9 18 Jet (R = 1.0), pT > 400GeV 100 360 0.9 23 Emiss

T

Emiss

T

> 180GeV 50 70 0.7 55 Multi-jets Four jets, each pT > 95GeV 3 × 40 4 × 85 0.3 20 Five jets, each pT > 70GeV 4 × 20 5 × 60 0.4 15 Six jets, each pT > 55GeV 4 × 15 6 × 45 1.0 12 b−jets One loose b, pT > 235GeV 100 225 0.9 35 Two medium b’s, pT > 160, 60GeV 100 150, 50 0.9 9 One b & three jets, each pT > 75GeV 3 × 25 4 × 65 0.9 11 Two b & two jets, each pT > 45GeV 3 × 25 4 × 35 0.9 9 B−physics Two µ’s, pT > 6, 4GeV 6, 4 6, 4 8 52 plus dedicated J/ψ-physics selection Total 70 1400

ATL-DAQ-PUB-2016-001 Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 31 / 29

slide-45
SLIDE 45

Trigger Rates from 2016 @ 1.2 × 1034cm−2s−1

Trigger Typical offline selection Trigger Selection

Level-1 Peak HLT Peak

Level-1 (GeV) HLT (GeV)

Rate (kHz) Rate (Hz) L = 1.2 × 1034 cm−2s−1

Single leptons Single isolated µ, pT > 27 GeV 20 26 (i) 13 133 Single isolated tight e, pT > 27 GeV 22 (i) 26 (i) 20 133 Single µ, pT > 52 GeV 20 50 13 48 Single e, pT > 61 GeV 22 (i) 60 20 13 Single τ, pT > 170 GeV 60 160 5 15 Two leptons Two µ’s, each pT > 15 GeV 2 × 10 2 × 14 1.5 21 Two µ’s, pT > 23, 9 GeV 20 22, 8 13 30 Two loose e’s, each pT > 18 GeV 2 × 15 2 × 17 8 7 One e & one µ, pT > 8, 25 GeV 20 (µ) 7, 24 13 2 One loose e & one µ, pT > 18, 15 GeV 15, 10 17, 14 1.5 2.6 Two τ’s, pT > 40, 30 GeV 20 (i), 12 (i) (+jets) 35, 25 6 35 One τ & one isolated µ, pT > 30, 15 GeV 12 (i), 10 (+jets) 25, 14 (i) 1.5 7 One τ & one isolated e, pT > 30, 18 GeV 12 (i), 15 (i) (+jets) 25, 17 (i) 3 9 Three leptons Three loose e’s, pT > 18, 11, 11 GeV 15, 2 × 8 17, 2 × 10 15 < 0.1 Three µ’s, each pT > 7 GeV 3 × 6 3 × 6 0.1 3 Three µ’s, pT > 21, 2 × 5 GeV 20 20, 2 × 4 13 4 Two µ’s & one loose e, pT > 2 × 11, 13 GeV 2 × 10 (µ’s) 2 × 10, 12 1.5 0.2 Two loose e’s & one µ, pT > 2 × 13, 11 GeV 2 × 8, 10 2 × 12, 10 1.1 0.1 One photon One loose γ, pT > 145 GeV 22 (i) 140 20 30 Two photons Two loose γ’s, pT > 40, 30 GeV 2 × 15 35, 25 8 40 Two tight γ’s, pT > 27, 27 GeV 2 × 15 2 × 22 8 16 Single jet Jet (R = 0.4), pT > 420 GeV 100 380 3 38 Jet (R = 1.0), pT > 460 GeV 100 420 3 35 Emiss

T

Emiss

T

> 200 GeV 50 110 6 230 Multi-jets Four jets, each pT > 110 GeV 3 × 50 4 × 100 0.4 18 Five jets, each pT > 80 GeV 4 × 15 5 × 70 3.5 14 Six jets, each pT > 70 GeV 4 × 15 6 × 60 3.5 5 Six jets, each pT > 55 GeV, |η| < 2.4 4 × 15 6 × 45 3.5 18 b−jets One b (ǫ = 60%), pT > 235 GeV 100 225 3 24 Two b’s (ǫ = 60%), pT > 160, 60 GeV 100 150, 50 3 20 One b (ǫ = 70%) & three jets, each pT > 85 GeV 4 × 15 4 × 75 3.5 19 Two b (ǫ = 60%) & one jet, pT > 65, 65, 110 GeV 2 × 20, 75 2 × 55, 100 2.7 25 Two b (ǫ = 60%) & two jets, each pT > 45 GeV 4 × 15 4 × 35 3.5 56 b−physics Two µ’s, pT > 6, 6 GeV 6, 6 6, 6 4.7 20 plus dedicated b-physics selections Total 85 1500

ATL-DAQ-PUB-2017-001 Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 32 / 29

slide-46
SLIDE 46

Trigger Rates from 2017 @ 1.7 × 1034cm−2s−1

Trigger Typical of ine selection Trigger Selection Level-1 Peak HLT Peak Level-1 (GeV) HLT (GeV) Rate (kHz) Rate (Hz) L =1.7 ×1034 cm− 2s− 1 Single leptons Single isolated µ, pT >27 GeV 20 26 (i) 16 187 Single isolated tight e, pT >27 GeV 22 (i) 26 (i) 26 178 Single µ, pT >52 GeV 20 50 16 65 Single e, pT >61 GeV 22 (i) 60 26 17 Single τ, pT >170 GeV 100 160 1.2 49 Two leptons Two µ’s, each pT >15 GeV 2 ×10 2 ×14 2.0 30 Two µ’s, pT >23, 9 GeV 20 22, 8 16 42 Two very loose e’s, each pT >18 GeV 2 ×15 (i) 2 ×17 1.6 11 One e & one µ, pT >8, 25 GeV 20 (µ) 7, 24 16 5 One e & one µ, pT >18, 15 GeV 15, 10 17, 14 2.0 4 One e & one µ, pT >27, 9 GeV 22 (e, i) 26, 8 26 2 Two τ’s, pT >40, 30 GeV 20 (i), 12 (i) (+jets, topo) 35, 25 5.1 59 One τ & one isolated µ, pT >30, 15 GeV 12 (i), 10 (+jets) 25, 14 (i) 2.1 9 One τ & one isolated e, pT >30, 18 GeV 12 (i), 15 (i) (+jets) 25, 17 (i) 3.9 16 Three leptons Three loose e’s, pT >25, 13, 13 GeV 20, 2 ×10 24, 2 ×12 1.2 <0.1 Three µ’s, each pT >7 GeV 3 ×6 3 ×6 0.2 8 Three µ’s, pT >21, 2 ×5 GeV 20 20, 2 ×4 16 8 Two µ’s & one loose e, pT >2 ×11, 13 GeV 2 ×10 (µ’s) 2 ×10, 12 2.0 0.3 Two loose e’s & one µ, pT >2 ×13, 11 GeV 2 ×8, 10 2 ×12, 10 1.6 0.2 One photon One loose γ, pT >145 GeV 22 (i) 140 26 46 Two photons Two loose γ’s, pT >55, 55 GeV 2 ×20 50, 50 2.4 6 Two medium γ’s, pT >40, 30 GeV 2 ×20 35, 25 2.4 18 Two tight γ’s, pT >25, 25 GeV 2 ×15 (i) 2 ×20 (i) 2.4 15 Single jet Jet (R =0.4), pT >435 GeV 100 420 3.4 33 Jet (R =1.0), pT >480 GeV 100 460 3.4 24 E miss

T

E miss

T

>200 GeV 50 110 4.4 100 Multi-jets Four jets, each pT >125 GeV 3 ×50 4 ×115 0.5 16 Five jets, each pT >95 GeV 4 ×15 5 ×85 4.9 10 Six jets, each pT >80 GeV 4 ×15 6 ×70 4.9 4 Six jets, each pT >60 GeV, |η| <2.0 4 ×15 6 ×55, |η| <2.4 4.9 15 b− jets One b ( =40%), pT >235 GeV 100 225 3.4 15 Two b’s ( =60%), pT >185, 70 GeV 100 175, 60 3.4 12 One b ( =40%) & three jets, each pT >85 GeV 4 ×15 4 ×75 4.9 15 Two b’s ( =70%) & one jet, pT >65, 65, 160 GeV 2 ×30, 85 2 ×55, 150 2.7 15 Two b’s ( =60%) & two jets, each pT >45 GeV 4 ×15 4 ×35 4.9 13 B-Physics Two µ’s, pT >11, 6 GeV 11, 6 11, 6 (di-µ) 3.1 50 Two µ’s, pT >6, 6 GeV, 2.5 <m(µ, µ) <4.0 GeV 2 ×6 (J /ψ, topo) 2 ×6 (J /ψ) 1.8 59 Two µ’s, pT >6, 6 GeV, 4.7 <m(µ, µ) <5.9 GeV 2 ×6 (B, topo) 2 ×6 (B) 1.8 7 Two µ’s, pT >6, 6 GeV, 7 <m(µ, µ) <12 GeV 2 ×6 (Υ, topo) 2 ×6 (Υ) 1.5 10 Total Rate 85 1550

TriggerPublicResults Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 33 / 29

slide-47
SLIDE 47

Full Size Event Display

Dijet event collected in 2017, with mjj = 9.3TeV .

EventDisplayRun2Physics Rhys Owen (University of Birmingham) The ATLAS Trigger System in Run-2 14 Feb 2018 34 / 29