Run 2 Data Taking Run 2 Data Taking 50ns ramp (early measurement) - - PowerPoint PPT Presentation

SLIDE 1

Run 2 Data Taking

SLIDE 2

Run 2 Data Taking

50ns ramp (early measurement) 25ns data taking

…

SLIDE 3

Run 2 Data Taking

50ns ramp (early measurement) 25ns data taking

…

• –

wasn’t

• –
• –

SLIDE 4

Run 2 Data Taking

• –

wasn’t

• –
• –

SLIDE 5

Run 2 Data Taking

• 17
• –

wasn’t

• –
• –

SLIDE 6

Run 2 Data Taking

• b

– – – – – – – – – – ng

• –
• ur probes,

– … –

• “SMOG piquet“ every start and end of physics.
• b

– p (6.5 TeV) – Neon: 20h – p (6.5 TeV) – Helium: 20h – p (6.5 TeV) – Argon: 3 days – p (2.51 TeV) – Argon: 9 h – Pb (6.37Z TeV) – Argon: ongoing

• –

– … –

• “SMOG piquet“ every start and end of physics.

SLIDE 7

“Success is a journey, not a destination.”

Arthur Ashe

SLIDE 8

The evolution of LHCb in 2015

SLIDE 9

The evolution of the LHCb trigger in 2015

SLIDE 10

M1 M3 M2 M4 M5 RICH2 HCAL ECAL SPD/PS Magnet z 5m y 5m 10m 15m 20m TT T1 T2 T3 Vertex Locator

TRACKER

P of charged particles

VELO

Primary vertices Impact parameter

RICHES

K, pi particle ID

MUONS

Trigger and PID

E/HCAL

Trigger, p, e, gamma PID

Magnet

• At 13 TeV & L = 4 × 1032 cm-2s-1:

~45 kHz bb pairs produced ~ 1 MHz cc pairs produced Can only readout @ 1 MHz (must decide within 4 μs) Can only store O(10kHz) (decide using ~50K cores)

The Challenge

SLIDE 11

M1 M3 M2 M4 M5 RICH2 HCAL ECAL SPD/PS Magnet z 5m y 5m 10m 15m 20m TT T1 T2 T3 Vertex Locator

TRACKER

P of charged particles

VELO

Primary vertices Impact parameter

RICHES

K, pi particle ID

MUONS

Trigger and PID

E/HCAL

Trigger, p, e, gamma PID

Magnet

• At 13 TeV & L = 4 × 1032 cm-2s-1:

~45 kHz bb pairs produced ~ 1 MHz cc pairs produced Can only readout @ 1 MHz (must decide within 4 μs) Can only store O(10kHz) (decide using ~50K cores)

40 MHz bunch crossing rate

450 kHz h± 400 kHz µ/µµ 150 kHz e/γ

L0 Hardware Trigger : 1 MHz readout, high ET/PT signatures

Software High Level Trigger 29000 Logical CPU cores Offline reconstruction tuned to trigger time constraints Mixture of exclusive and inclusive selection algorithms 2 kHz Inclusive Topological

5 kHz Rate to storage

2 kHz Inclusive/ Exclusive Charm 1 kHz Muon and DiMuon

Run 1 Trigger

SLIDE 12

Hadronic Dimuon Mode D → hhh B → hh B+ → J/ K+ ✏(L0) [%] 27 62 93 ✏(HLT | L0) [%] 42 85 92 ✏(HLT × L0) [%] 11 52 84

D∗ → D0π [1211.1230]

]

2

c ) [GeV/

s +

π D ( M

2.005 2.01 2.015 2.02

)

2

c Candidates/(0.1 MeV/

0.2 0.4 0.6 0.8 1 1.2

6

10 × RS data Fit Background

LHCb

B0

s → Dsπ [1304.4741]

]

2

c ) invariant mass [MeV/

+

π

− s

(D

5350 5400 5450 5500 5550 )

2

c candidates / (15 MeV/ 2000 4000

data fit

+

π

− s

D →

s

B

+

K

− s

D →

s

B misid bkg. comb bkg.

a)

LHCb

−

π φ →

− s

D

]

2

c ) [GeV/

2.02

) c Candidates/(0.1 MeV/

6

10 RS data Fit Background

LHCb

B0

s → J/ψφ [1304.2600v3]

]

2

) [MeV/c

• K

+

K ψ m(J/

5320 5340 5360 5380 5400 5420

)

2

Candidates / (2.5 MeV/c

500 1000 1500 2000 2500 3000 3500 4000 4500

LHCb

] 2 ) [MeV/c π s m(D 5100 5200 5300 5400 5500 5600 5700 5800 ) 2 Candidates/(10 MeV/c 1000 2000 3000 4000

• 1

=1.0 fb int LHCb Preliminary L Data π s D → s Signal B ) ρ , π ( (*) s D → s B π D → d B π c Λ → b Λ ) ρ , π ( (*) (s) D → d B Combinatorial

Dsπ

]

2

c ) [GeV/

2.02

B0

s → µµ [1211.2674]

]

2

c [MeV/

− µ + µ

m 5000 5500 )

2

c Candidates / (44 MeV/ 2 4 6 8 10 12 14 16 LHCb BDT>0.7

• 1

3 fb

Very clean signals Large “dynamic range” Good trigger efficiencies …. except for charm …. but there is a lot of charm

Run 1 Performance

SLIDE 13

Run 2 Challenge

• Energy: 8 TeV → 13 TeV

+ σbb x 1.6

• σinelastic x 1.2
• multiplicity x 1.2
• Bunch spacing: 50 ns → 25 ns

+ constant lumi → pileup / 2

• 1 MHz L0/readout limit: 1/20 → 1/40
• spillover

SLIDE 14

Run 2 Challenge

• Energy: 8 TeV → 13 TeV

+ σbb x 1.6

• σinelastic x 1.2
• multiplicity x 1.2
• Bunch spacing: 50 ns → 25 ns

+ constant lumi → pileup / 2

• 1 MHz L0/readout limit: 1/20 → 1/40
• spillover

SLIDE 15

Run 2 Challenge

Can we maintain improve performance under more challenging conditions?

• Energy: 8 TeV → 13 TeV

+ σbb x 1.6

• σinelastic x 1.2
• multiplicity x 1.2
• Bunch spacing: 50 ns → 25 ns

+ constant lumi → pileup / 2

• 1 MHz L0/readout limit: 1/20 → 1/40
• spillover

SLIDE 16

“The formulation of the problem is

• ften more essential than its solution,

which may be merely a matter of mathematical or experimental skill.” “To raise new questions, new possibilities, to regard old questions from a new angle requires creative imagination and marks real advances…” — Albert Einstein

SLIDE 17

“The formulation of the problem is

• ften more essential than its solution,

which may be merely a matter of mathematical or experimental skill.” “To raise new questions, new possibilities, to regard old questions from a new angle requires creative imagination and marks real advances…” — Albert Einstein

What is the problem?

SLIDE 18

Some things are not rare…

0.1 1 10 10

• 7

10

• 6

10

• 5

10

• 4

10

• 3

10

• 2

10

• 1

10 10

1

10

2

10

3

10

4

10

5

10

6

10

7

10

8

10

9

10

• 7

10

• 6

10

• 5

10

• 4

10

• 3

10

• 2

10

• 1

10 10

1

10

2

10

3

10

4

10

5

10

6

10

7

10

8

10

9

σ σ σ σZZ σ σ σ σWW σ σ σ σWH σ σ σ σVBF MH=125 GeV

WJS2012

σ σ σ σjet(ET

jet > 100 GeV)

σ σ σ σjet(ET

jet > √

√ √ √s/20) σ σ σ σggH

LHC Tevatron

events / sec for L = 10

33 cm

• 2s
• 1

σ σ σ σb σ σ σ σtot

proton - (anti)proton cross sections

σ σ σ σW σ σ σ σZ σ σ σ σt

σ σ σ σ ( ( ( (nb) ) ) ) √ √ √ √s (TeV)

{

8

SLIDE 19

Some things are not rare…

0.1 1 10 10

• 7

10

• 6

10

• 5

10

• 4

10

• 3

10

• 2

10

• 1

10 10

1

10

2

10

3

10

4

10

5

10

6

10

7

10

8

10

9

10

• 7

10

• 6

10

• 5

10

• 4

10

• 3

10

• 2

10

• 1

10 10

1

10

2

10

3

10

4

10

5

10

6

10

7

10

8

10

9

σ σ σ σZZ σ σ σ σWW σ σ σ σWH σ σ σ σVBF MH=125 GeV

WJS2012

σ σ σ σjet(ET

jet > 100 GeV)

σ σ σ σjet(ET

jet > √

√ √ √s/20) σ σ σ σggH

LHC Tevatron

events / sec for L = 10

33 cm

• 2s
• 1

σ σ σ σb σ σ σ σtot

proton - (anti)proton cross sections

σ σ σ σW σ σ σ σZ σ σ σ σt

σ σ σ σ ( ( ( (nb) ) ) ) √ √ √ √s (TeV)

{

8

]

2

c ) [GeV/

s +

π D ( M

2.005 2.01 2.015 2.02

)

2

c Candidates/(0.1 MeV/

0.2 0.4 0.6 0.8 1 1.2

6

10 × RS data Fit Background

LHCb

]

2

c ) [GeV/

s +

π D ( M

2.005 2.01 2.015 2.02

)

2

c Candidates/(0.1 MeV/

2 4 6 8 10

3

10 × WS data Fit Background

LHCb

8.4M events

3.6K events

D0→K-π+ D0→π-K+

Observation of D0 D0 Oscillations

• R. Aaij et al.*

(LHCb Collaboration)

(Received 6 November 2012; published 5 March 2013) We report a measurement of the time-dependent ratio of D0 ! Kþ to D0 ! Kþ decay rates in Dþ-tagged events using 1:0 fb1 of integrated luminosity recorded by the LHCb experiment. We measure the mixing parameters x02 ¼ ð0:9 1:3Þ 104, y0 ¼ ð7:2 2:4Þ 103, and the ratio of doubly-Cabibbo-suppressed to Cabibbo-favored decay rates RD ¼ ð3:52 0:15Þ 103, where the uncertainties include statistical and systematic sources. The result excludes the no-mixing hypothesis with a probability corresponding to 9.1 standard deviations and represents the first observation of D0 D0

• scillations from a single measurement.

PRL 110, 101802 (2013) Selected for a Viewpoint in Physics P H Y S I C A L R E V I E W L E T T E R S

week ending 8 MARCH 2013

SLIDE 20

“The problem is not the problem. The problem is your attitude about the problem”

SLIDE 21

Offline → Online!

• Do “Online” what used to be

done “Offline”

• Calibrate in “Real Time”
• Run offline reconstruction
• nline
• Skip offline reconstruction /

skimming

• Don’t store events / information

that you won’t really use…

SLIDE 22

40 MHz bunch crossing rate

450 kHz h± 400 kHz µ/µµ 150 kHz e/γ

L0 Hardware Trigger : 1 MHz readout, high ET/PT signatures

Software High Level Trigger 29000 Logical CPU cores Offline reconstruction tuned to trigger time constraints Mixture of exclusive and inclusive selection algorithms 2 kHz Inclusive Topological

5 kHz Rate to storage

2 kHz Inclusive/ Exclusive Charm 1 kHz Muon and DiMuon

Trigger Evolution

• 2011: increased bandwidth
• 2 kHz → 5 kHz to accommodate charm
• 29K CPU cores
• 2012: add deferred triggering to utilize farm

between fills

• 20% deferral → 25% extra capacity
• 2015: split HLT
• 50K CPU cores
• buffer full HLT1 output (150 kHz) to 5PB of disk
• HLT2 uses “offline quality” calibrations

LHCb 2011

SLIDE 23

40 MHz bunch crossing rate

450 kHz h± 400 kHz µ/µµ 150 kHz e/γ

L0 Hardware Trigger : 1 MHz readout, high ET/PT signatures

Software High Level Trigger 29000 Logical CPU cores Offline reconstruction tuned to trigger time constraints Mixture of exclusive and inclusive selection algorithms

5 kHz Rate to storage Defer 20% to disk

• Trigger Evolution
• 2011: increased bandwidth
• 2 kHz → 5 kHz to accommodate charm
• 29K CPU cores
• 2012: add deferred triggering to utilize farm

between fills

• 20% deferral → 25% extra capacity
• 2015: split HLT
• 50K CPU cores
• buffer full HLT1 output (150 kHz) to 5PB of disk
• HLT2 uses “offline quality” calibrations

LHCb 2012

SLIDE 24

Trigger Evolution

• 2011: increased bandwidth
• 2 kHz → 5 kHz to accommodate charm
• 29K CPU cores
• 2012: add deferred triggering to utilize farm

between fills

• 20% deferral → 25% extra capacity
• 2015: split HLT
• 50K CPU cores
• buffer full HLT1 output (150 kHz) to 5PB of disk
• HLT2 uses “offline quality” calibrations

40 MHz bunch crossing rate

450 kHz h± 400 kHz µ/µµ 150 kHz e/γ

L0 Hardware Trigger : 1 MHz readout, high ET/PT signatures

Software High Level Trigger

12.5 kHz Rate to storage

Partial event reconstruction, select displaced tracks/vertices and dimuons Buffer events to disk, perform online detector calibration and alignment Full offline-like event selection, mixture

• f inclusive and exclusive triggers
• LHCb 2015

SLIDE 25

Software Improvements

• Equivalent to ‘a few MCHF’ of hardware
• Unified online and offline reconstruction!
• PT threshold: 1.3 GeV/c → 0.5 GeV/c
• Drop (IP | muon match) requirement in

HLT1

• εHLT(charm): +50%
• εHLT(B+ → D0π+): +20% (75% → 90%)

Run 2 software Run 2 configuration

v48r1 (2015 reco)

Run 2 software Run 1 configuration

v48r1

⇓ ⇓

v45r1

Run 1 software Run 1 configuration

Area ∝ cycle count

“Start where you are. Use what you have. Do what you can.” — Arthur Ashe

SLIDE 26

Software Improvements

• Equivalent to ‘a few MCHF’ of hardware
• Unified online and offline reconstruction!
• PT threshold: 1.3 GeV/c → 0.5 GeV/c
• Drop (IP | muon match) requirement in

HLT1

• εHLT(charm): +50%
• εHLT(B+ → D0π+): +20% (75% → 90%)

Run 2 software Run 2 configuration

v48r1 (2015 reco)

Run 2 software Run 1 configuration

v48r1

⇓ ⇓

v45r1

Run 1 software Run 1 configuration

Area ∝ cycle count

“Start where you are. Use what you have. Do what you can.” — Arthur Ashe

SLIDE 27

RICH PID

Track efficiency Trigger effic

IP resolution rigger efficiency

εHLT(B+->D0π+) ~ 90% εHLT(B+->D0π+) ~ 75%

Performance: Run 1 vs. Run 2

SLIDE 28

“Turbo” Output

• Online reconstruction == Offline

reconstruction

• Online calibration == Offline calibration
• Turbo: store Trigger Data only
• For a given bandwidth, increases the

event rate by an order of magnitude

• Ideal for high-yield analysis
• 185 out of 374 HLT2 selections go to “Turbo”

⇓

SLIDE 29

“Turbo” Charm

Candidates / (0.992366 MeV/c2)

0.02 0.04 0.06 0.08 0.1 0.12 ×106

Data Fit Signal Background

NSig 2171646.45 ± 2207.31 NBkg 736641.05 ± 1853.92 χ2/DoF 3.39

LHCb preliminary √s = 13 TeV

m(K −π+) [MeV/c2]

1.8 1.82 1.84 1.86 1.88 1.9 1.92 ×103

∆/σ

• 4
• 2

2 4

D0 → K−π+

Candidates / (0.992366 MeV/c2)

20 40 60 80 100 ×103

Data Fit Signal Background

NSig 2014696.44 ± 2854.52 NBkg 787659.46 ± 2630.81 χ2/DoF 3.83

LHCb preliminary √s = 13 TeV

m(K −π+π+) [MeV/c2]

1.82 1.84 1.86 1.88 1.9 1.92 1.94 ×103

∆/σ

• 4
• 2

2 4

D+ → K−π+π+

Candidates / (0.992366 MeV/c2)

1 2 3 4 5 6 ×103

Data Fit Signal Background

NSig 91956.53 ± 345.58 NBkg 25375.76 ± 229.89 χ2/DoF 1.28

LHCb preliminary √s = 13 TeV

m(K −K +π+) [MeV/c2]

1.92 1.94 1.96 1.98 2 2.02 2.04 ×103

∆/σ

• 4
• 2

2 4

D+

s → K−K+π+

SLIDE 30

Measurements of prompt charm production cross-sections in pp collisions at √s = 13 TeV

The LHCb collaboration†

Abstract Production cross-sections of prompt charm mesons are measured with the first data from pp collisions at the LHC at a centre-of-mass energy of 13 TeV. The data sample corresponds to an integrated luminosity of 4.98 ± 0.19 pb−1 collected by the LHCb

• experiment. The production cross-sections of D0, D+, D+

s , and D∗+ mesons are

measured in bins of charm meson transverse momentum, pT, and rapidity, y, and cover the range 0 < pT < 15 GeV /c and 2.0 < y < 4.5. The ratios of the integrated

cross-sections between charm mesons agree with previously measured fragmentation

• fractions. The inclusive cc cross-section within the range of 0 < pT < 8 GeV

/c is found to be σ(pp → ccX) = 2940 ± 3 ± 180 ± 160 µb, where the uncertainties are due to statistical, systematic and fragmentation fraction uncertainties, respectively.

arXiv:1510.01707v1 [hep-ex] 6 Oct 2015

SLIDE 31

The prompt atmospheric neutrino flux in the light of LHCb

Rhorry Gauld,a Juan Rojo,b Luca Rottoli,b Subir Sarkarb,c and Jim Talbertb

aInstitute for Particle Physics Phenomenology, Durham University, Durham DH1 3LE, UK bRudolf Peierls Centre for Theoretical Physics, 1 Keble Road, University of Oxford, OX1 3NP

Oxford, UK

cNiels Bohr International Academy, Copenhagen University, Blegdamsvej 17, 2100 Copenhagen,

Denmark

E-mail: rhorry.gauld@durham.ac.uk, juan.rojo@physics.ox.ac.uk, luca.rottoli@physics.ox.ac.uk, subir.sarkar@physics.ox.ac.uk, jim.talbert@physics.ox.ac.uk Abstract: The recent observation of very high energy cosmic neutrinos by IceCube heralds the beginning of neutrino astronomy. At these energies, the dominant background to the astrophysical signal is the flux of ‘prompt’ neutrinos, arising from the decay of charmed mesons produced by cosmic ray collisions in the atmosphere. In this work we provide predictions for the prompt atmospheric neutrino flux in the framework of perturbative QCD, using state-of-the-art Monte Carlo event generators. Our calculation includes the constraints set by charm production measurements from the LHCb experiment at 7 TeV, recently validated with the corresponding 13 TeV data. Our result for the prompt flux is a factor of about 2 below the previous benchmark calculation, in general agreement with

• ther recent estimates, but with an improved estimate of the uncertainty. This alleviates

the existing tension between the theoretical prediction and IceCube limits, and suggests that a direct direction of the prompt flux is imminent.

arXiv:1511.06346v2 [hep-ph] 23 Nov 2015

SLIDE 32

]

2

c [MeV/

• µ

+

µ

m

2950 3000 3050 3100 3150 3200

2

c Candidates per 5 MeV/

2 4 6 8 10 12

3

10 ×

• 1

=3.05 pb

int

L = 13 TeV, s LHCb c < 3 GeV/

T

p 2 < < 3.5 y 3 <

[ps]

z

t

• 10
• 8
• 6
• 4
• 2

2 4 6 8 10

Candidates per 0.2 ps

1 10

2

10

3

10

4

10

5

10

• 1

=3.05 pb

int

L = 13 TeV, s LHCb c < 3 GeV/

T

p 2 < < 3.5 y 3 <

Data Total fit b

• from-

ψ J/ ψ Prompt J/ Wrong PV Background

Abstract: The production of J/ψ mesons in proton-proton collisions at a centre-of-mass energy of √s = 13 TeV is studied with the LHCb detector. Cross-section measurements are performed as a function of the transverse momentum pT and the rapidity y of the J/ψ meson in the region pT < 14 GeV/c and 2.0 < y < 4.5, for both prompt J/ψ mesons and J/ψ mesons from b-hadron decays. The production cross-sections integrated over the kinematic coverage are 15.30 ± 0.03 ± 0.86 µb for prompt J/ψ and 2.34 ± 0.01 ± 0.13 µb for J/ψ from b-hadron decays, assuming zero polarization of the J/ψ meson. The first uncertainties are statistical and the second systematic. The cross-section reported for J/ψ mesons from b-hadron decays is used to extrapolate to a total b¯ b cross-section. The ratios

• f the cross-sections with respect to √s = 8 TeV are also determined.

Published for SISSA by Springer

Received: September 3, 2015 Accepted: October 5, 2015 Published: October 26, 2015

Measurement of forward J/ψ production cross-sections in pp collisions at √s = 13 TeV

The LHCb collaboration

SLIDE 33

Abstract: The production of J/ψ mesons in proton-proton collisions at a centre-of-mass energy of √s = 13 TeV is studied with the LHCb detector. Cross-section measurements are performed as a function of the transverse momentum pT and the rapidity y of the J/ψ meson in the region pT < 14 GeV/c and 2.0 < y < 4.5, for both prompt J/ψ mesons and J/ψ mesons from b-hadron decays. The production cross-sections integrated over the kinematic coverage are 15.30 ± 0.03 ± 0.86 µb for prompt J/ψ and 2.34 ± 0.01 ± 0.13 µb for J/ψ from b-hadron decays, assuming zero polarization of the J/ψ meson. The first uncertainties are statistical and the second systematic. The cross-section reported for J/ψ mesons from b-hadron decays is used to extrapolate to a total b¯ b cross-section. The ratios

• f the cross-sections with respect to √s = 8 TeV are also determined.

Published for SISSA by Springer

Received: September 3, 2015 Accepted: October 5, 2015 Published: October 26, 2015

Measurement of forward J/ψ production cross-sections in pp collisions at √s = 13 TeV

The LHCb collaboration

] c ) [GeV/ ψ J/ (

T

p

5 10

)

T

p /d σ (d

13/8

R

1 2 3

= 8 TeV cross-section ratio s = 13 TeV/ s LHCb LHCb FONLL

SLIDE 34

The Future…

L0/Readout limit @ 1 MHz

SLIDE 35

The Future…

L0/Readout limit @ 1 MHz

40 MHz bunch crossing rate

450 kHz h± 400 kHz µ/µµ 150 kHz e/γ

L0 Hardware Trigger : 1 MHz readout, high ET/PT signatures

Software High Level Trigger

12.5 kHz (0.6 GB/s) to storage

Partial event reconstruction, select displaced tracks/vertices and dimuons Buffer events to disk, perform online detector calibration and alignment Full offline-like event selection, mixture

• f inclusive and exclusive triggers

LHCb 2015 Trigger Diagram

SLIDE 36

The Future…

L0/Readout limit @ 1 MHz

30 MHz inelastic event rate (full rate event building) Software High Level Trigger 2-5 GB/s to storage

Full event reconstruction, inclusive and exclusive kinematic/geometric selections Add offline precision particle identification and track quality information to selections Output full event information for inclusive triggers, trigger candidates and related primary vertices for exclusive triggers

LHCb Upgrade Trigger Diagram

Buffer events to disk, perform online detector calibration and alignment

40 MHz bunch crossing rate

450 kHz h± 400 kHz µ/µµ 150 kHz e/γ

L0 Hardware Trigger : 1 MHz readout, high ET/PT signatures

Software High Level Trigger

12.5 kHz (0.6 GB/s) to storage

Partial event reconstruction, select displaced tracks/vertices and dimuons Buffer events to disk, perform online detector calibration and alignment Full offline-like event selection, mixture

• f inclusive and exclusive triggers

LHCb 2015 Trigger Diagram

Take what ye can! — Jack Sparrow

SLIDE 37

“The Journey of a thousand miles begins with a single step” — Lao Tzu

SLIDE 38

SLIDE 39

“If you cannot explain it simply, you do not understand it well enough”

— Albert Einstein

SLIDE 40

D0 mixing

d,s,b ‘in the loop’ instead of u,c,t ⇒ GIM (almost) kills this amplitude...

D0 π− K+ W − u c u d u s Vcd V ∗

us

D0 K+ π− W + u c u s u d V ∗

cs

Vud

D0 D0 K+π−

D0 q = d, s, b W q = d, s, b W c u u c Vcq Vuq V ∗

uq

V ∗

cq

|V ∗

csVud| = O(1)

|VcdV ∗

us| = O(λ2) ≈ 0.04

Look for ‘wrong sign’ D0 decays

SLIDE 41

What needs to be improved?

• Tracking:
• faster/better algorithms
• More CPU time
• Real-time calibration
• Particle ID:
• Faster algorithms
• More CPU time
• Real-time calibration
• [mm]

ρ

2 4 6

VELO

ε

0.7 0.75 0.8 0.85 0.9 0.95 1

LHCb

• ffline
• nline

Candidates / ( 0.2 ps )

1 10

2

10

3

10

4

10

LHCb

t [ps]

5 10

pull

• 4
• 2

2 4

• 0.4

0.2 0.0 0.2 0.4 0.06 0.08 0.10 0.12 0.14 LHCb ATLAS 19.2 fb 1 CMS 20 fb 1 CDF 9.6 fb 1 DØ 8 fb 1 SM

68% CL contours ( )

Combined 3 fb 1

SLIDE 42

Farm Node

100%

For DQ only

Local disk
 Buffer

1 MHz

HLT1 EvtBuilder

150 KHz

Online “Real-Time” Calibration

SLIDE 43

Farm Node

100%

For DQ only

Conditions
 DB Calib&Align Selected
 Events Local disk
 Buffer

1 MHz

HLT1 EvtBuilder

150 KHz

Online “Real-Time” Calibration

VELO & Tracker Alignment OT Timing RICH refractive index RICH image

SLIDE 44

Farm Node

100%

For DQ only

Conditions
 DB Calib&Align Selected
 Events Local disk
 Buffer

1 MHz

HLT1 EvtBuilder

150 KHz

Online “Real-Time” Calibration

SLIDE 45

Farm Node

100%

For DQ only

Conditions
 DB Local disk
 Buffer

1 MHz 12.5 KHz

HLT1 HLT2 EvtBuilder

150 KHz

Online “Real-Time” Calibration

SLIDE 46

SV

IP p p

• Topological N-body Triggers
• Utilizes excellent vertex and

momentum resolution to compute:

• Uses a dedicated “Bonzai”

Boosted Decision Tree [JINST 8 (2013) P02013 ] with

• PT, IP𝝍2, FD𝝍2, minv, mcorr
• Capable of filling its allotted

bandwidth with ~100% pure generic bb events

mcorr ≡ q m2

inv + |PT miss|2 + |PT miss|

• mass (GeV)

5 10 50 100

HLT2 2-Body Topo measured corrected

mass (GeV)

5 10 200 400 600

HLT2 3-Body Topo measured corrected

mass (GeV)

5 10 500 1000

HLT2 4-Body Topo measured corrected

Example: 4-body B decay, minv and mcorr for 2, 3 and 4 body selections

minv

minv

minv

mcorr mcorr mcorr

HLT2 4-body HLT2 3-body HLT2 2-body

SLIDE 47

SV

IP p p

• Topological N-body Triggers
• Utilizes excellent vertex and

momentum resolution to compute:

• Uses a dedicated “Bonzai”

Boosted Decision Tree [JINST 8 (2013) P02013 ] with

• PT, IP𝝍2, FD𝝍2, minv, mcorr
• Capable of filling its allotted

bandwidth with ~100% pure generic bb events

mcorr ≡ q m2

inv + |PT miss|2 + |PT miss|

SLIDE 48

• Same principle as Run 1 :

preselect displaced tracks with ∑ PT, followed by BBDT

• Timing: <0.1 ms (*)
• At 25-50 kHz output rate, large

efficiency gains over Run 1

• red: run 1 efficiency
• green: 2x run 1 efficiency
• LHCb-PUB-2014-031

TOPO Rate [kHz] 20 40 60 80 100 Efficiency 0.2 0.4 0.6 0.8 1

• µ

+

µ ]

• π

+

[K

*

K → B LHCb Simulation

TOPO Rate [kHz] 20 40 60 80 100 Efficiency 0.2 0.4 0.6 0.8 1

]

• K

+

[K φ ]

• K

+

[K φ →

s

B LHCb Simulation

TOPO Rate [kHz] 20 40 60 80 100 Efficiency 0.2 0.4 0.6 0.8 1

]

• K

+

[K φ ]

• µ

+

µ (1S)[ ψ →

s

B LHCb Simulation

TOPO Rate [kHz] 20 40 60 80 100 Efficiency 0.2 0.4 0.6 0.8 1

π

+

K

• K

→

s

B LHCb Simulation

TOPO Rate [kHz] 20 40 60 80 100 Efficiency 0.2 0.4 0.6 0.8 1

• π

]

• π

+

K

+

[p

+ c

Λ →

b

Λ LHCb Simulation

TOPO Rate [kHz] 20 40 60 80 100 Efficiency 0.2 0.4 0.6 0.8 1

+

π

• π

+

]K

• π

+

[K D →

+

B LHCb Simulation

(*) on our 2011 reference machine: Intel X5650 (Westmere) @ 2.67 GHz

The Upgrade Trigger

SLIDE 49

• Algorithm Optimizations
• HLT1 adds tracking in VErtex

LOcator (VELO) and primary vertex reconstruction

track µ µ µ

• ther
• min. pT [ GeV]

1.0 0.5 1.6

• VELO tracks, either matched to

muon hits, or with large IP are extended through the magnet

• PT dependent search windows:

Really bad for charm physics