Introduce a single LST-seeded, CKF-building iteration based on Patatrack pixel tracks for Phase 2 HLT #47891
Conversation
|
cms-bot internal usage |
|
A new Pull Request was created by @VourMa for master. It involves the following packages:
@AdrianoDee, @Martin-Grunewald, @Moanwar, @cmsbuild, @DickyChant, @miquork, @mmusich, @srimanob, @subirsarkar can you please review it and eventually sign? Thanks. cms-bot commands are listed here |
|
test parameters:
|
|
please test |
|
+1 Size: This PR adds an extra 72KB to repository
Comparison SummarySummary:
|
OK, let's wait for the presentation before the review. |
| OriginTransverseErrorMultiplier = cms.double(1), | ||
| MinOneOverPtError = cms.double(1), | ||
| magneticField = cms.string(''), | ||
| TTRHBuilder = cms.string('WithTrackAngle'), |
There was a problem hiding this comment.
just a curiosity.
Can one use here hltESPTTRHBuilderWithTrackAngle instead?
I think we have a proliferation of ESProducers doing the same thing in the menu.
There was a problem hiding this comment.
Yes, should be possible (my oversight not using the proper HLT one). Let me do it in one go after the presentation, together with any potential feedback I get there.
…n, LST seeding + CKF building variant
|
test parameters:
|
|
@cmsbuild, please test |
|
+1 Size: This PR adds an extra 68KB to repository
Comparison SummarySummary:
|
|
There are differences in: In general tracking efficiency w.r.t. TPs goes down. Is this understood? |
This should be a purely technical PR for those configurations, so I can check again whether something is changed by mistake for them. |
Nevermind, a physics change is expected for those two workflows specifically from the fact that the single iteration configuration is now switched to use only 4-hit or longer tracks only (i.e. removing the triplets from the Patatrack seeds), as was advertized in the presented slides. This can be seen from the number of layers with measurement for pixel tracks: I think that we want to switch to configurations without Patatracks triplets, and the benefits/drawbacks of this change can be seen, for the configuration similar to .753, when comparing red vs. black in the presented slides. Personally, I think that this is the direction we want to move towards. That said, I will also let @rovere comment whether we want this change now. |
| _hltPhase2PixelTracksSoASingleIterPatatrack = hltPhase2PixelTracksSoA.clone( minHitsPerNtuplet = 3 ) | ||
|
|
||
| from Configuration.ProcessModifiers.singleIterPatatrack_cff import singleIterPatatrack | ||
| singleIterPatatrack.toReplaceWith(hltPhase2PixelTracksSoA, _hltPhase2PixelTracksSoASingleIterPatatrack) |
There was a problem hiding this comment.
for future record, this is responsible of the changes discussed at #47891 (comment)
this is indeed the direction that I think we agreed we should take (together with the OT extension of the CA algorithm for pixel tracks). Thanks for clarifying this is proposed already in this PR. |
|
+hlt
|
|
+Upgrade |
|
@cms-sw/pdmv-l2 Is there anything from your side that you would like me to follow up on this PR? Thank you in advance! |
|
+pdmv |
cmsbuild
added
fully-signed
pdmv-approved
and removed
pending-signatures
pdmv-pending
labels
|
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. @mandrenguyen, @rappoccio, @sextonkennedy, @antoniovilela (and backports should be raised in the release meeting by the corresponding L2) |
|
+1 |
6 participants
This PR introduces the new configuration described in the title. The configuration looks like this:
The implementation utilizes already available
procModifiers, and can be run with theprocModifiersequencealpaka,singleIterPatatrack,trackingLST,seedingLST. Workflow 0.757 is being introduced to monitor this configuration.It should be noted that this PR also updates the single iteration configuration to use Patatrack pixel tracks without triplets.
As part of the validation, the
procModifiers mentioned above have been tested separately and in their "allowed" combinations, and seem to be giving expected results. In any case, it would be good to test the following, previously introduced workflows:Slides on this were presented at the HLT Upgrade meeting of April 22nd.
FYI @slava77