-
Notifications
You must be signed in to change notification settings - Fork 4.6k
TICLv5a : Improve skeleton linking and parameters tuning #46539
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
TICLv5a : Improve skeleton linking and parameters tuning #46539
Conversation
|
cms-bot internal usage |
|
+code-checks Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-46539/42407 |
|
A new Pull Request was created by @waredjeb for master. It involves the following packages:
@Martin-Grunewald, @Moanwar, @cmsbuild, @jfernan2, @mandrenguyen, @mmusich, @srimanob, @subirsarkar can you please review it and eventually sign? Thanks. cms-bot commands are listed here |
|
test parameters:
|
|
@cmsbuild please test |
|
Hi, |
|
@smuzaffar do you have any idea why profiling tests are stuck? |
|
@felicepantaleo , the node which runs the profiling tests is busy running IB profiling jobs. Let me check if I can improve the situation |
|
please test I can not find any trace of previse profiling jobs , so lets restart the tests |
|
-1 Failed Tests: RelVals-INPUT
RelVals-INPUT
Expand to see more relval errors ...
Comparison SummarySummary:
|
|
ticlTrackstersMerge is doubling the time, do we understand why? Thanks |
Hi @jfernan2 . I'll be honest, I am quite confused, I am not touching the TracksterMergeProducer with this PR (as you can also see from the changes). ticlTrackstersMerge is part of TICLv4, in this PR I am only touching TICLv5 (workflows .203), so I am not expecting any changes on all the other workflows. |
|
I tried running wf 79481c79481,79482
< cylinder_radius_sqr = cms.vdouble(9, 9),
---
> cylinder_radius_sqr = cms.vdouble(9, 15),
> deltaRxy = cms.double(4.0),
79483,79484c79484,79486
< max_distance_projective_sqr = cms.vdouble(60, 60),
< max_distance_projective_sqr_closest_points = cms.vdouble(60, 60),
---
> lower_boundary = cms.vdouble(20, 10),
> lower_distance_projective_sqr = cms.vdouble(30, 60),
> lower_distance_projective_sqr_closest_points = cms.vdouble(10, 50),
79486,79488c79488,79490
< min_distance_z = cms.vdouble(30, 30),
< min_num_lcs = cms.uint32(7),
< min_trackster_energy = cms.double(10),
---
> min_distance_z = cms.vdouble(35, 35),
> min_num_lcs = cms.uint32(15),
> min_trackster_energy = cms.double(20),
79492c79494,79497
< wind = cms.double(0.036)
---
> upper_boundary = cms.vdouble(150, 100),
> upper_distance_projective_sqr = cms.vdouble(30, 60),
> upper_distance_projective_sqr_closest_points = cms.vdouble(5, 30),
> wind = cms.double(0.072)
79519,79520c79524,79525
< onnxEnergyModelPath = cms.FileInPath('RecoHGCal/TICL/data/ticlv5/onnx_models/linking/energy_v0.onnx'),
< onnxPIDModelPath = cms.FileInPath('RecoHGCal/TICL/data/ticlv5/onnx_models/linking/id_v0.onnx'),
---
> onnxEnergyModelPath = cms.FileInPath('RecoHGCal/TICL/data/ticlv5/onnx_models/patternrecognition/energy_v0.onnx'),
> onnxPIDModelPath = cms.FileInPath('RecoHGCal/TICL/data/ticlv5/onnx_models/patternrecognition/id_v0.onnx'), |
|
@cmsbuild, please test |
|
+1 Size: This PR adds an extra 12KB to repository
Comparison SummarySummary:
|
|
+hlt |
|
+1 |
|
Milestone for this pull request has been moved to CMSSW_15_0_X. Please open a backport if it should also go in to CMSSW_14_2_X. |
|
+Upgrade |
|
This pull request is fully signed and it will be integrated in one of the next master IBs (tests are also fine). This pull request will now be reviewed by the release team before it's merged. @sextonkennedy, @rappoccio, @antoniovilela, @mandrenguyen (and backports should be raised in the release meeting by the corresponding L2) |
|
+1 |
This PR improves the current Skeleton Linking
New Parameters: tuned parameters for hadronic reconstruction.
Code refactoring: simply code by refactoring pieces
Modified TICLGraph: modified logic for traversing the graph and build components
Small changes to TICLDumper: allow to run TICLDumper with TICLv4. It will consume SuperClusters (new with TICLv5) only in TICLv5
New parameters have been included also at HLT
The performance against TICLv4 can be found in the DPS approval talk https://indico.cern.ch/event/1462666/contributions/6158255/attachments/2940573/5170734/TICLv5_DPS_version3.pdf
Should be tested with a .203 workflow
E.g.
Edit 06/11/2024:
hltTiclTracksterLinks_cfiandhltTiclTracksterLinksUnseeded_cfi, Both of them were Unseeded with no differences between them, for consistency I removedhltTiclTracksterLinksUnseeded_cfifixing the Egamma sequence accordinglyhltTiclTracksterLinksL1Seeded_cfi