[DO NOT MERGE] Single Iteration Patatrack and Patatrack-only-seeded LST in HLT#105
Conversation
| trackingLST.toReplaceWith(hltGeneralTracks, _hltGeneralTracksLST) | ||
|
|
||
| _hltGeneralTracksLSTSingleIterPatatrack = hltGeneralTracks.clone( | ||
| TrackProducers = cms.VInputTag("hltInitialStepTrackSelectionHighPuritypTTCLST", "hltInitialStepTracksT5TCLST", "hltHighPtTripletStepTrackSelectionHighPurity"), |
There was a problem hiding this comment.
looking closer : what is in hltHighPtTripletStepTrackSelectionHighPurity ? this looks like a legacy recovery unmodified.
I interpreted the plan for the "singleIter" to be literally a single (general) tracking iteration.
There was a problem hiding this comment.
The LST single iteration is actually "a single iteration of seeds going into LST as input", hence the "Patatrack-only-seeded LST" configuration. An actual single iteration with LST means no building in the endcaps, and I don't think this makes sense at this point (unless Patatrack pixel tracks magically become amazingly better).
There was a problem hiding this comment.
An actual single iteration with LST means no building in the endcaps, and I don't think this makes sense at this point
OK, maybe an actual single iteration needs to be added.
What we talked about yesterday or in the past as an ultimate goal is to have a single iteration pata+LST+finalfit and eventually get this to work. Even if it works poorly now, the goal is to make it work and start with having this configuration available.
There was a problem hiding this comment.
Such a configuration has never been tested but I can test it and show it on Friday. After we see what we get, I can close this PR and open another one with the commit about the single iteration setup with CKF and with an actual single iteration LST. Then we merge that on top of the other PR and we are done (for now). Does this sound like a plan?
|
should this be closed? |
Yes |
* Use phase2 strip cluster size when creating mkfit::Hit. * Add checks for existence of the ph2_clustSize branch.
2 participants
As per title, with the single iteration seeded by >3 hit Patatrack pixel tracks being controlled by the newly introduced
singleIterPatatrackprocModifier.This is a follow up on PR #94, the same description on the PR also holds for this one.
In terms of performance, this is shown in the "Tests on LST Configurations - Baseline vs. Patatrack-only-seeded LST in InitialStep" section of this presentation.