fix: ParticleID collision in Geant4 #2039

benjaminhuth · 2023-04-18T16:40:17Z

This attempts to fix particle collisions in Geant4. These occur I think when the bits in the ActsFatras::Barcode are not enough to capture all results. This way it could happen that multiple particles where associated with a single ID:

I think the correct fix for this to check for collision and ignoring particles that collide. Furthermore, the barcode generation is improved, so that collisions should happen less frequently.

Since this happens quite frequently with default settings for Geant4+Pythia8(ttbar), I think all so far generated events using this configuration are suffering from this problem

andiwand · 2023-04-18T17:15:18Z

tagging @Corentin-Allaire as this might impact the ambi performance eval?

codecov · 2023-04-18T17:15:45Z

Codecov Report

Merging #2039 (822abb8) into main (74df04d) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##             main    #2039   +/-   ##
=======================================
  Coverage   49.83%   49.83%           
=======================================
  Files         420      420           
  Lines       23962    23962           
  Branches    10859    10859           
=======================================
  Hits        11942    11942           
  Misses       4390     4390           
  Partials     7630     7630

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

andiwand

looks good from my side

I wonder if we can optimize the barcode or the utilization of the barcode in the G4 translation so we avoid overflows for ttbar

Corentin-Allaire · 2023-04-18T17:21:33Z

tagging @Corentin-Allaire as this might impact the ambi performance eval?

It definitely could affect it, if we have multiple track with the same id then one of them will be treated as a duplicate

github-actions · 2023-04-18T17:42:05Z

📊 Physics performance monitoring for `411a5e6`

Full report
Seeding: seeded, truth estimated, orthogonal
CKF: seeded, truth smeared, truth estimated, orthogonal
IVF: seeded, truth smeared, truth estimated, orthogonal
Ambiguity resolution: seeded, orthogonal
Truth tracking
Truth tracking (GSF)

Vertexing

Vertexing vs. mu

IVF seeded

IVF truth_smeared

IVF truth_estimated

IVF orthogonal

Seeding

Seeding seeded

Seeding truth_estimated

Seeding orthogonal

CKF

CKF seeded

CKF truth_smeared

CKF truth_estimated

CKF orthogonal

Ambiguity resolution

seeded

Truth tracking (Kalman Filter)

Truth tracking

Truth tracking (GSF)

Truth tracking

benjaminhuth · 2023-04-18T18:49:03Z

Actually there are quite a lot of collisions in one event:

20:47:59    Geant4Simula   WARNING   2718782 particles missing in the output data because of particle ID collissions

Corentin-Allaire · 2023-04-18T19:04:48Z

Actually there are quite a lot of collisions in one event:
20:47:59    Geant4Simula   WARNING   2718782 particles missing in the output data because of particle ID collissions

That number seems a bit too high, for how many event is that ?

andiwand · 2023-04-18T19:22:47Z

That number seems a bit too high, for how many event is that ?

I dont actually know but it might be plausible considering all the secondary particles? I once use G4 with 100 muon event and got like 10k secondaries

benjaminhuth · 2023-04-18T19:30:26Z

That number seems a bit too high, for how many event is that ?

Its for one event, but without post selection. And probably you need to divide this number by two, because right now it counts initial and final particles.

Corentin-Allaire · 2023-04-18T19:35:09Z

That number seems a bit too high, for how many event is that ?

Its for one event, but without post selection. And probably you need to divide this number by two, because right now it counts initial and final particles.

I just run the job with the default selector and I see 0 overlap so this probably doesn't change much in the end ?

asalzburger · 2023-04-19T05:17:36Z

looks good from my side

I wonder if we can optimize the barcode or the utilization of the barcode in the G4 translation so we avoid overflows for ttbar

Sure, if we can change the barcode behaviour to be safe against this, let's do it!

asalzburger · 2023-04-19T07:11:23Z

Since this happens quite frequently with default settings for Geant4+Pythia8(ttbar), I think all so far generated events using this configuration are suffering from this problem

Can you specify what this means? Which particles are disregarded?
And how does this free enough space in the barcode bits?

andiwand

thanks for turning this around so quickly 👍

looks good - just two questions

Examples/Algorithms/Geant4/include/ActsExamples/Geant4/EventStoreRegistry.hpp

Examples/Algorithms/Geant4/src/ParticleTrackingAction.cpp

benjaminhuth · 2023-04-20T18:28:12Z

Can you specify what this means? Which particles are disregarded? And how does this free enough space in the barcode bits?

Finally, I have some plots @asalzburger
That is 1 event ttbar+Geant4. But I still don't understand every detail of the results.

asalzburger · 2023-04-21T09:26:28Z

@benjaminhuth - can you fix the format, please ?

benjaminhuth · 2023-04-21T09:27:34Z

Yeah and I need to do some clean-up still

This reverts commit ad6edd7.

This reverts commit dbad4c2.

benjaminhuth · 2023-04-21T13:36:01Z

@asalzburger The CI should run through now, could you reapprove?

This attempts to fix particle collisions in Geant4. These occur I think when the bits in the `ActsFatras::Barcode` are not enough to capture all results. This way it could happen that multiple particles where associated with a single ID: ![image](https://user-images.githubusercontent.com/37871400/232845631-9b36bcb3-47a1-49bf-91a2-313305d7314b.png) I think the correct fix for this to check for collision and ignoring particles that collide. Furthermore, the barcode generation is improved, so that collisions should happen less frequently. Since this happens quite frequently with default settings for Geant4+Pythia8(ttbar), I think all so far generated events using this configuration are suffering from this problem

benjaminhuth and others added 2 commits April 18, 2023 17:56

first version

3589fb4

update

95cad8c

benjaminhuth added the 🚧 WIP Work-in-progress label Apr 18, 2023

benjaminhuth added this to the next milestone Apr 18, 2023

andiwand approved these changes Apr 18, 2023

View reviewed changes

benjaminhuth marked this pull request as draft April 18, 2023 18:23

update

b4f1e12

Merge branch 'main' into fix/particle-id-collision

487e1d4

Corentin-Allaire mentioned this pull request Apr 19, 2023

ParticleID collision in Geant4 #2041

Closed

benjaminhuth added 3 commits April 19, 2023 15:02

update

2120156

update

411525a

update

6695d73

benjaminhuth marked this pull request as ready for review April 20, 2023 08:16

benjaminhuth requested a review from andiwand April 20, 2023 08:17

andiwand approved these changes Apr 20, 2023

View reviewed changes

Examples/Algorithms/Geant4/include/ActsExamples/Geant4/EventStoreRegistry.hpp Show resolved Hide resolved

Examples/Algorithms/Geant4/src/ParticleTrackingAction.cpp Show resolved Hide resolved

Merge branch 'main' into fix/particle-id-collision

9d4fbd8

asalzburger added automerge and removed 🚧 WIP Work-in-progress labels Apr 21, 2023

benjaminhuth and others added 3 commits April 21, 2023 10:45

update

dbad4c2

skip neutral

ad6edd7

Merge branch 'main' into fix/particle-id-collision

35b5bcc

benjaminhuth added 🚧 WIP Work-in-progress and removed automerge labels Apr 21, 2023

benjaminhuth added 3 commits April 21, 2023 11:39

Revert "skip neutral"

942c807

This reverts commit ad6edd7.

Revert "update"

9a2f632

This reverts commit dbad4c2.

update

5a59427

Corentin-Allaire mentioned this pull request Apr 21, 2023

fix: enable G4 secondary hits #2049

Merged

Update root_file_hashes.txt

411a5e6

benjaminhuth added Component - Examples Affects the Examples module Impact - Major Significant bug and/or affects a lot of modules and removed 🚧 WIP Work-in-progress Impact - Major Significant bug and/or affects a lot of modules labels Apr 21, 2023

Merge branch 'main' into fix/particle-id-collision

822abb8

asalzburger added the automerge label Apr 21, 2023

Corentin-Allaire approved these changes Apr 21, 2023

View reviewed changes

kodiakhq bot merged commit f2d8c25 into acts-project:main Apr 21, 2023

github-actions bot removed the automerge label Apr 21, 2023

benjaminhuth deleted the fix/particle-id-collision branch May 10, 2023 07:15

paulgessinger modified the milestones: next, v26.0.0 May 22, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

fix: ParticleID collision in Geant4 #2039

fix: ParticleID collision in Geant4 #2039

benjaminhuth commented Apr 18, 2023 •

edited

Loading

andiwand commented Apr 18, 2023

codecov bot commented Apr 18, 2023 •

edited

Loading

andiwand left a comment

Corentin-Allaire commented Apr 18, 2023

github-actions bot commented Apr 18, 2023 •

edited

Loading

benjaminhuth commented Apr 18, 2023 •

edited

Loading

Corentin-Allaire commented Apr 18, 2023

andiwand commented Apr 18, 2023 •

edited

Loading

benjaminhuth commented Apr 18, 2023

Corentin-Allaire commented Apr 18, 2023

asalzburger commented Apr 19, 2023

asalzburger commented Apr 19, 2023

andiwand left a comment

benjaminhuth commented Apr 20, 2023 •

edited

Loading

asalzburger commented Apr 21, 2023

benjaminhuth commented Apr 21, 2023

benjaminhuth commented Apr 21, 2023

fix: ParticleID collision in Geant4 #2039

fix: ParticleID collision in Geant4 #2039

Conversation

benjaminhuth commented Apr 18, 2023 • edited Loading

andiwand commented Apr 18, 2023

codecov bot commented Apr 18, 2023 • edited Loading

Codecov Report

andiwand left a comment

Choose a reason for hiding this comment

Corentin-Allaire commented Apr 18, 2023

github-actions bot commented Apr 18, 2023 • edited Loading

📊 Physics performance monitoring for 411a5e6

Vertexing

Seeding

CKF

Ambiguity resolution

Truth tracking (Kalman Filter)

Truth tracking (GSF)

benjaminhuth commented Apr 18, 2023 • edited Loading

Corentin-Allaire commented Apr 18, 2023

andiwand commented Apr 18, 2023 • edited Loading

benjaminhuth commented Apr 18, 2023

Corentin-Allaire commented Apr 18, 2023

asalzburger commented Apr 19, 2023

asalzburger commented Apr 19, 2023

andiwand left a comment

Choose a reason for hiding this comment

benjaminhuth commented Apr 20, 2023 • edited Loading

asalzburger commented Apr 21, 2023

benjaminhuth commented Apr 21, 2023

benjaminhuth commented Apr 21, 2023

benjaminhuth commented Apr 18, 2023 •

edited

Loading

codecov bot commented Apr 18, 2023 •

edited

Loading

github-actions bot commented Apr 18, 2023 •

edited

Loading

📊 Physics performance monitoring for `411a5e6`

benjaminhuth commented Apr 18, 2023 •

edited

Loading

andiwand commented Apr 18, 2023 •

edited

Loading

benjaminhuth commented Apr 20, 2023 •

edited

Loading