PYTHIA from 2017 to 2021: an overview

(1)

PYTHIA from 2017 to 2021: an overview

For the PYTHIA collaboration

Christian Bierlich, bierlich@thep.lu.se Lund University

Dec 7th 2021, 23rd MCnet Meeting, Manchester

1

(2)

PYTHIA: General purpose Monte Carlo

• General purpose MCEG for pp and much more.

• Versatility as a guiding principle.

The User

Input Main Program Output

Settings LHA...

LHEF

Event process Event event Info Pythia8Rivet HepMC Hist

ProcessLevel ProcessContainer PhaseSpace SLHAinterface ResonanceDecay

PartonLevel TimeShower SpaceShower Dire, Vincia MultipleInteractions BeamRemnants

HadronLevel StringFragmentation StringInteractions ParticleDecays BoseEinstein LowEnergySigma Merging

BeamParticle SigmaProcess SigmaTotal

Vec4, Rndm, ParticleData, PhysicsBase, UserHooks, HeavyIons, ...

Pythia 8.3 event generator

• Historically JETSET + LO ME and parton shower.

• Past 4 years: extensions to many collision systems, new

showers, shower uncertainties, many new soft physics models... ²

(3)

This talk

• Broad rather than deep overview.

• Not just developments on the physics side.

1. Physics developments. 2. Technical developments. 3. Organisatorial developments. 4. PYTHIA & the future.

3

(4)

This talk

PYTHIA 8.223, Jan 5 2017

• 10 authors (33% in Lund).

• Torbj¨orn Sj¨ostrand carrying most tasks and responsibilities.

• Recent physics focus: M&M.

• Mostly caught up with PYTHIA6, some new physics scope.

1. Physics developments.

2. Technical developments.

3. Organisatorial developments.

4. PYTHIA & the future.

3

(5)

This talk

PYTHIA 8.306, Jun 28, 2021

• 13 authors (38% in Lund).

• More distributed leadership structure.

• Recent physics focus: Soft QCD models & two new showers.

• Many benefits over PYTHIA6.

3

(6)

This talk

2015 2016 2017 2018 2019 2020 2021

140000 160000 180000 200000 220000 240000 260000 280000 300000

Lines of code

8.201* 8.205* 8.2098.2108.212 8.215 8.219 8.223 8.226

8.230 8.235 8.240 8.243* 8.244 8.245

8.301 8.302 8.303 8.305*8.306

* bugfix release PYTHIA 8.2 series PYTHIA 8.3 series

3

(7)

Main physics developments, 8.223 → 8.306

Not a complete list, but an overview of main physics extensions.

Most with published code, some only paper.

• Cross sections and diffraction:

1. Cross section calculations.

2. Diffraction with γ-beams

& UPCs.

• Parton showers: 1. VINCIA. 2. DIRE.

3. Automated shower variations.

• Heavy ion physics: 1. Angantyr.

• Hadronization:

1. String interactions. 2. Hadron vertices.

• After hadronization: 1. Hadronic rescattering. 2. Deuteron coalescence &

molecular states.

3. Extension to cosmic rays. And many, many more smaller

updates, fixes, convenience implementations etc. Apologies to those not

mentioned.

4

(8)

& UPCs.

• Parton showers:

1. VINCIA.

2. DIRE.

• Heavy ion physics: 1. Angantyr.

• Hadronization:

molecular states.

mentioned.

4

(9)

& UPCs.

• Parton showers:

1. VINCIA.

2. DIRE.

• Heavy ion physics:

1. Angantyr.

• Hadronization:

molecular states.

mentioned.

4

(10)

& UPCs.

• Parton showers:

1. VINCIA.

2. DIRE.

1. Angantyr.

• Hadronization:

1. String interactions.

2. Hadron vertices.

molecular states.

mentioned.

4

(11)

& UPCs.

• Parton showers:

1. VINCIA.

2. DIRE.

1. Angantyr.

• Hadronization:

2. Hadron vertices.

• After hadronization:

1. Hadronic rescattering.

2. Deuteron coalescence &

molecular states.

3. Extension to cosmic rays.

And many, many more smaller updates, fixes, convenience

implementations etc. Apologies to those not

mentioned.

4

(12)

& UPCs.

• Parton showers:

1. VINCIA.

2. DIRE.

1. Angantyr.

• Hadronization:

2. Hadron vertices.

• After hadronization:

1. Hadronic rescattering.

2. Deuteron coalescence &

molecular states.

3. Extension to cosmic rays.

And many, many more smaller updates, fixes, convenience

implementations etc.

Apologies to those not mentioned.

4

(13)

Cross section calculations (CB, Rasmussen & Sj¨ostrand: 1804.10373, 1907.12871)

• Old SaS default appended with several other models.

• Regge based parametrizations, includes LHC related updates.

• Alternative Mueller-dipole based → EIC & substructure.

p p

X

p p

gXp(0) g_Xp(0)

(total)

p p

X X

t t

p p

gXp(t) gXp(t)

g_Xp(t) g_Xp(t)

(elastic)

p p

X2 X2

X1

p p

t t

gX2p(t) gX2p(t) g^XX²1^X²(t) gX1p(0)

(single diff.)

p p

X1 X1

X2

p p

t t

gX1p(t) gX1p(t)

g^XX¹2^X¹(t)

gX2p(0)

(single diff.)

p X3 p

X2 X2

t X1

p p

gX3p(0) gX^X²3^X²(t) gX^X²1^X²(t) gX1p(0)

(double diff.)

p p

X2 X2

t2 t2

X3

X1 X1

p p

t1 t1

gX2p(t2) gX2p(t2) g^XX²3^X²(0) g^XX²3^X²(0) gX1p(t1) gX1p(t1)

(central diff.)

5

(14)

10²√ 10⁴

s (GeV) 0

50 100 150 200

σtot(mb)

COMPAS pp ABMST pp DL pp FMO pp PDG pp data

0 1 2 3 4 5

−t (GeV²) 10⁻⁶

10⁻⁴ 10⁻² 10⁰ 10²

dσel/dt(mb/GeV2) COMPAS pp

ABMST pp SaS pp pp data

√s = 7 TeV

10² 10⁴

√s (GeV) 0

10 20 30 40 50

σel(mb)

COMPAS pp ABMST pp SaS pp PDG pp data

10¹ 10² √10³ 10⁴ 10⁵

s (GeV) 0

5 10 15 20 25

σSD(mb)

SaS ABMST pp data

SD integrated cross section for ξ < 0.05

5

(15)

10² 10³ 10⁴ 10⁵

100 200 300 400

σtot[mb]

ABMST SaS+DL pp 7 TeV

PY8 dipoles unconfined PY8 dipoles confined Data

10² 10√³ 10⁴ 10⁵

s [GeV]

0.5 1.0 1.5

MC/Data

10² 10³

10 20 30 40 50

σel[mb]

ABMST SaS+DL pp 7 TeV

PY8 dipoles unconfined PY8 dipoles confined Data

10² √ 10³

s [GeV]

0.5 1.0 1.5

MC/Data

5

(16)

Hard and soft diffraction with γ-beams (Helenius & Rasmussen: 1901.05261)

• Important processes for DIS-type systems. Factorization breaking at HERA.

• Using MPIs to “fill the gap” of diffractive systems. Reject events where MPIs shroud the diffractive signature.

• Framework can also do UPCs! ⁶

(17)

VINCIA (Brooks, Preuss, Skands & Verheyen: 1907.08980, 2002.04393, 2003.00702, 2008.09468)

• Fully incorporated new shower, based on antenna formalism.

Interleaved evolution for ISR, FSR & coloured resonances.

Fully coherent soft interference for QED.

Includes module for electroweak shower(see also 2108.10786).

Technical: “sector” shower makes HO corrections easier.

Dedicated CKKW-L merging in VINCIA, exploiting power of sector showers.

NNLO matching in the pipeline (2108.07133).

7

(18)

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40

r 0.9

1.0 1.1 1.2

Ratioto ATLASdata 10⁰ 10¹

ρ(r)

pp→ t¯t → b¯bW⁺W⁻,√ s = 7 TeV

PS + MPI + had

pT bj∈ [70, 100] GeV

ATLAS data PYTHIA 8 VINCIA

• QCD: Vincias more narrow jet profile favoured by data

• b-jet profile in t¯t production. 7

(19)

r r r r rrrrrrrrrrrrrrrrrr rrrrrrr

|yZ| < 1 1 < |yZ| < 2

(×0.1)

|yZ| < 1

1 < |yZ| < 2

r D∅

Vincia (Sector) Vincia (Global) Pythia 8.3 10⁻⁵

10⁻⁴ 10⁻³ 10⁻² 10⁻¹ 1 10¹

φη^∗distribution, Z → e⁺e⁻, p ¯p,√s = 1.96 TeV

1/σdσ/dφ

∗ η

r r r r rrrrrrrrrrrrrrrrrr rrrrrrr

0.50.6 0.70.8 0.91 1.11.2 1.31.4

MC/Data r r r r rrrrrrrrrrrrrrrrrr rrrrrrr

10⁻² 10⁻¹ 1

0.50.6 0.70.8 0.91 1.11.2 1.31.4

φ^∗η

MC/Data

• ISR sector shower: Drell-Yan leptons opening angle.

• Performance: VINCIA shower slow (oversampling) but sector merging faster (linear vs. factorial) due to limited histories.

7

(20)

DIRE (Gellersen, H¨oche & Prestel: 1506.05057, 2109.09706)

• Fully incorporated new shower, based on dipole formalism + collinear enhancements.

QCD and QED shower with automatic uncertainties.

Includes higher order corrections to kernels.

Focus on making merging easy, also for the user.

Option for Dark Matter emissions in shower.

8

(21)

DIRE (Gellersen, H¨oche & Prestel: 1506.05057, 2109.09706)

• Fully incorporated new shower, based on dipole formalism + collinear enhancements.

0 0.5 1

b

τ1

0 2 4 6 [nb]b 1τ/dσd

= 319 GeV s L = 351.6 pb-1

Preliminary H1

H1 data Djangoh Rapgap Pythia+Dire

0 0.5 1

b

τ1

0 2 4 6 [nb]b 1τ/dσd

= 319 GeV s L = 351.6 pb-1

Preliminary H1

H1 data

⊗ NP NNLO NNLO

⊗ NP NLO Pythia+Dire

(Figure credit: H1/Johannes Hessler)

• Well used (massive PR campaign) for ep → EIC.

• Here 1-jettiness event shape in new H1 analysis(2111.11364.)

8

(22)

Automated shower variations (Gellersen & Prestel: 2001.10476)

• Adding to previous PDF variation, one can now perform automatic renormalization scale variation in the CKKW-L, UMEPS, NL-3 and UNLOPS merging schemes.

• Completely unified weights scheme in progress, but difficult.

• Automating these tasks potentially improves users’ error

estimation significantly! Lots of potential and interest. ⁹

(23)

Angantyr (CB, Gustafson, L¨onnblad & Shah: 1806.10820)

• Framework for full heavy ion collisions.

Glauber calculation decides which nucleons hit each other.

PYTHIA pp, pn & nn events stacked on top of each other.

A clean slate for adding collective effects, no QGP.

Projectile Target η

dN dη

target wounded nucleon

projectile wounded nucleon pp collision

pA collision AAcollision

• Just specify your nuclear beams and run!

10

(24)

Angantyr (CB, Gustafson, L¨onnblad & Shah: 1806.10820)

• Framework for full heavy ion collisions.

Glauber calculation decides which nucleons hit each other.

PYTHIA pp, pn & nn events stacked on top of each other.

A clean slate for adding collective effects, no QGP.

bbb

bb bbb b

b bb b bb b b bb b b b b b bb b b b b b b b b b b

b b b b b b b b b b b b b b b b bb b b b b b b b b b b b b b b bb b b b b b b b bb bb

bbb bbbbbbbbbbbbbbbb

b Data

MC

10⁻⁷ 10⁻⁶ 10⁻⁵ 10⁻⁴ 10−3 10⁻²

Sum E^PbTdistribution, Pb–Pb √sNN=2.76 TeV

(1/Nevt)dN/d∑EPbT

bbb b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b bb b b bb b b b b b b b b b b b b b b b b b b b b b b

0 500 _{1.0 · 10}31.5 · 10³2.0 · 10³2.5 · 10³3.0 · 10³3.5 · 10³ 0.50.6

0.70.8 0.91 1.11.2 1.31.4

∑ E⊥

MC/Data b bbbb bbbb bbb bbbb bbbb bbb bbbb bbbb bbb

b bbbb bbbb bbb bbbb bbbb bbb bbbb bbbb bbb

Pythia8/Angantyr ALICE PbPbpS_{N N}= 5.02 TeV

b

-3 -2 -1 0 1 2 3 4 5

0 500 1000 1500 2000

2500(a) Centrality dependent η distribution PbPb,pSN N= 5.02 TeV

η (1/Nev)dNch/dη

• Just specify your nuclear beams and run!

10

(25)

String interactions (CB, Chakraborty, Gustafson & L¨onnblad: 1710.09725, 1807.05271, 1901.07447)

• Extending Lund strings’ abilities: interactions between strings.

String shoving generates flow.

Rope hadronization increases strangeness and baryons.

• Intended as an alternative to QGP models.

• Extensions to AA ongoing (2010.07595).

11

(26)

String interactions (CB, Chakraborty, Gustafson & L¨onnblad: 1710.09725, 1807.05271, 1901.07447)

• Extending Lund strings’ abilities: interactions between strings.

String shoving generates flow.

Rope hadronization increases strangeness and baryons.

1 0 1 2 3 4

1.000 1.005 1.010 1.015 1.020 1.025 1.030

S()/min(S) | | > 2.0| | < 2.50.5 GeV < p < 5.0 GeV

The ridge in Z-tagged events, Nch> 110 Pythia 8

Pythia 8 + Shoving (ATLAS pp high multiplicity)

10¹ 10² 10³

Nch, 2 <|η| <4, p⟂ > 0.15 GeV 10^-3

10^-2 10^-1

ratio to π+π−

2Ks⁰/π^± 6φ/π^± 2(Λ + ¯Λ)/π^±

6(Ξ + ¯Ξ)/π^± 16(Ω + ¯Ω)/π^±

e⁺e⁻ DIPSY pp Pythia8 + ropes DIPSY pPb DIPSY PbPb

• Intended as an alternative to QGP models.

• Extensions to AA ongoing (2010.07595).

11

(27)

Hadronic rescattering (CB, Ferreres-Sol´e, Sj¨ostrand & Utheim: 1808.04619, 2005.05658, 2103.09665)

• Hadrons may scatter again in the final state

• Some effects in pp, very important in ion collisions.

• Requires knowledge of hadron production vertices.

• ...a new framework for Low Energy QCD processes.

• ...with an extensive amount of cross sections!

(Rescatterings per 13 TeV ND pp event)

12

(28)

Hadronic rescattering (CB, Ferreres-Sol´e, Sj¨ostrand & Utheim: 1808.04619, 2005.05658, 2103.09665)

• Hadrons may scatter again in the final state

• Some effects in pp, very important in ion collisions.

b b bb b b b b b bb b bb bbbbbbbbbbbbbbbbbbbbb b b b b b b b b b b b

b Data

Rescattering off Rescattering on

10⁻⁴ 10⁻³ 10⁻² 10⁻¹

p + ¯p yield in INEL pp collisions at√s = 7 TeV in |y| < 0.5.

1Nineld2NdpTdy(c/GeV)

b b b b b b b b b b b b b b bbbbbbbbbbbbbbbbbbbbb b b b b b b b b b b b

1 2 3 4 5 6

0.50.6 0.70.8 0.91 1.11.2 1.31.4

pT(GeV/c)

MC/Data b b bb bbb bbbb b b b b b b b bbbbbb b

b bb

b bb bb

b bbbbbbbb

bbb b bb bb

b bb bb b b b b b b b ALICE Data

Pythia 8/Angantyr + rescattering

0 0.02 0.04 0.06 0.08 0.1 0.12 0.14

Pb-Pb √snn=5.02 TeV, v2{2, |∆η| > 1.4}

v2{2,|∆η|>1.4} b b bb bbb bbbb b b b b b b b bbbbbb b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b

10² 10³

0.50.6 0.70.8 0.91 1.11.2 1.31.4

Nch(|η| < 0.8)

MC/Data

• Inevitable for precision, even in min-bias.

• Low Energy framework very versatile, added bonus!

12

(29)

Deuteron coalescence & molecular states (Ilten & Utheim: 2108.03479)

• Existing model(s): Momentum space recombination of p + n⁰, p + p, n⁰+ n⁰→²H + X .

• Cross sections taken from experiments/shape only.

(Figure credit: ALICE/Alberto Caliva, Valentina Zaccolo)

• Extending to space-time in rescattering picture.

• Other molecular states; tetraquarks & pentaquarks. ¹³

(30)

Extension to cosmic rays ^(Sj¨ostrand & Utheim: 2108.03481)

• Building upon updated framework for low energy interactions.

• Proof-of-principle atmospheric cascade, a new playing field.

• Includes simplified model for pA interactions.

0 250 500 750 1000 1250 1500 1750

X (g/cm2) 10⁰

10¹ 10² 10³ 104

Probability

Atmospheric depth of interactions

Uniform p/n atmosphere Uniform nitrogen Exponential nitrogen Exponential nitrogen at 45

0 250 500 750 1000 1250 1500 1750

X (g cm²)

10² 10³ 10⁴ 10⁵

(1/nev)X 0dN

Number of hadrons at depth

Uniform p/n atmosphere Uniform nitrogen Exponential nitrogen Exponential nitrogen at 45

• 10⁸ GeV initiator proton through atmosphere. Left: number of interactions. Right: hadrons remaining above kinematic threashold.

14

(31)

Technical developments

• Still a “library style” generator written in C++ with no intent to change. But:

Gradual transition to C++11 since 8.3X.

Mature double–sided Python interface (PyBind11).

More external interfaces, notably Rivet and HepMC3.

• Transition from svn → gitlab.com

More possibilities for collaboration on issues.

Automatic checks (both technical and physics) at commit-level, merge level and release.

Still some manual checks (PVS).

Strong gatekeeper → distributed code checks (with a codemaster to oversee).

Main repo private. Have https://pythia.org for code tarballs, historic code (dating back to 1986!) and online manual.

• Technical changes supporting organisatorial changes.

15

(32)

15

(33)

15

(34)

15

(35)

Organisatorial developments

• TS withdrawing from managerial roles → rotating triumvirate consisting of spokesperson, codemaster and webmaster taking over.

Aim to spread responsibilities, increase bus-factor.

Could a future network help with clerical tasks?

• Yearly in-person meetings (or week-long COVID distance meetings).

Time to discuss physics and know your collaborator!

Time for organisatorial update, eg. working on more transparent author requirements.

Major sprints for code and manuscript writing.

Rotating organizers at PYTHIA nodes.

• One such outcome: Git repo with 1h topical tutorials.

Need for consistent and maintained tutorial material for special topics!

Also a chance to try each others’ work.

Already used for summer schools, might be extended with video in the future.

16

(36)

16

(37)

16

(38)

16

(39)

Organisatorial cont’d

• Mailing list retired, and replaced with issue desk.

Historic questions → less time spent?

In reality it still takes a lot of time!

Importance of low-bar MCnet summer schools can not be overstated!

• Volounteer bug-finding increased.

Large efforts by a few volounteers dramatically increased code quality.

Discussion about formalized acknowledgements in progress.

Maybe an untapped resource for the wider community?

• Communication with experiments.

Contact persons for all LHC experiments (and others) are helpful. Pragmatically different in approach.

PHENOmenal discussion meetings started (mainly PYTHIA/ALICE, but others also present).

17

(40)

17

(41)

17

(42)

PYTHIA and the future

• Many developments in the pipeline, here just a selection!

• A dedicated PYTHIA 8.3 paper with physics and guides for new users.

Complete replacement of the PYTHIA6 physics manual.

Aiming for SciPost Physics Codebases, together with code releases.

• New physics/technical ventures on several fronts.

• Driven by individuals’ interests, little common strategy.

Cosmic ray physics, coherent framework for HI physics, eA support, NNLO matching, more electroweak shower options, ...

PYTHIA contrib, better ME interfacing, HPC compatibility, ...

• Active involvement in EIC community.

18

(43)

PYTHIA and the future

• Many developments in the pipeline, here just a selection!

• A dedicated PYTHIA 8.3 paper with physics and guides for new users.

Complete replacement of the PYTHIA6 physics manual.

Aiming for SciPost Physics Codebases, together with code releases.

• New physics/technical ventures on several fronts.

• Driven by individuals’ interests, little common strategy.

Cosmic ray physics, coherent framework for HI physics, eA support, NNLO matching, more electroweak shower options, ...

PYTHIA contrib, better ME interfacing, HPC compatibility, ...

• Active involvement in EIC community.

18

(44)

The PYTHIA collaboration

• CB, Lund, hadronization, HI, ALICE. webmaster.

• Nishita Desai, Tata Inst, SUSY, SLHA, BSM.

• Leif Gellersen, Lund, scale uncertainties, matching/merging.

• Ilkka Helenius, Jyv¨askyla, photoproduction, γ − γ, diffraction.

deputy spokesperson.

• Philip Ilten, Cincinnati, τ’s, onia, LHCb. codemaster.

• Leif L¨onnblad, Lund, HI, hadronization.

• Stephen Mrenna, Fermilab, SUSY, matching/merging, CMS.

• Stefan Prestel, Lund, matching/merging, DIRE, ATLAS.

• Christian Preuss, Z¨urich, VINCIA, ext ME, matching/merging.

• Torbj¨orn Sj¨ostrand, Lund, SM, parton showers, MPIs, CR, hadronization, core structure.

• Peter Skands, Monash, VINCIA, MPIs, CR, tuning, hadronization. spokesperson.

• Marius Utheim, Jyv¨askyla, hadronic rescattering.

• Rob Verheyen, UCL, weak showers, VINCIA.

19