8
LHCb Lausanne Workshop, 21st March. 2006 1/8 Tracking Open Issues Tracking Open Issues E. Rodrigues, NIKHEF E. Rodrigues, NIKHEF LHCb Tracking and Alignment Workshop LHCb Tracking and Alignment Workshop Some topics to discuss … Some topics to discuss …

LHCb Lausanne Workshop, 21st March. 2006 1/8 Tracking Open Issues E. Rodrigues, NIKHEF LHCb Tracking and Alignment Workshop Some topics to discuss …

Embed Size (px)

Citation preview

Page 1: LHCb Lausanne Workshop, 21st March. 2006 1/8 Tracking Open Issues E. Rodrigues, NIKHEF LHCb Tracking and Alignment Workshop Some topics to discuss …

LHCb Lausanne Workshop, 21st March. 2006 1/8

Tracking Open IssuesTracking Open Issues

E. Rodrigues, NIKHEFE. Rodrigues, NIKHEF

LHCb Tracking and Alignment WorkshopLHCb Tracking and Alignment Workshop

Some topics to discuss …Some topics to discuss …

Page 2: LHCb Lausanne Workshop, 21st March. 2006 1/8 Tracking Open Issues E. Rodrigues, NIKHEF LHCb Tracking and Alignment Workshop Some topics to discuss …

LHCb Lausanne Workshop, 21st March. 2006 2/8

Tracks Locations in TESAt present:At present:

/Event/Rec//Event/Rec/Track/Track/IdealIdeal/Event/Rec/Track/Seed/Event/Rec/Track/Seed/Event/Rec/Track//Event/Rec/Track/VeloVelo/Event/Rec//Event/Rec/Track/Track/ForwardForward/Event/Rec/Track/VeloTT/Event/Rec/Track/VeloTT/Event/Rec/Track/Match/Event/Rec/Track/Match/Event/Rec/Track/KsTrack/Event/Rec/Track/KsTrack/Event/Rec/Track/Best/Event/Rec/Track/Best

But not enough … agree on following off-/on-line?But not enough … agree on following off-/on-line?

/Event/Rec/Track/Ideal/Event/Rec/Track/Ideal/Event/Rec/Track/Seed /Event/Hlt/Track/Seed /Event/Rec/Track/Seed /Event/Hlt/Track/Seed /Event/Hlt/Track/RZVelo /Event/Hlt/Track/RZVelo /Event/Rec/Track//Event/Rec/Track/Velo Velo /Event/Hlt/Track/Velo /Event/Hlt/Track/Velo /Event/Rec//Event/Rec/Track/Track/Forward Forward /Event/Hlt/Track/Forward /Event/Hlt/Track/Forward /Event/Rec/Track/VeloTT /Event/Hlt/Track/VeloTT/Event/Rec/Track/VeloTT /Event/Hlt/Track/VeloTT /Event/Rec/Track/Match/Event/Rec/Track/Match/Event/Rec/Track/KsTrack /Event/Hlt/Track/KsTrack/Event/Rec/Track/KsTrack /Event/Hlt/Track/KsTrack/Event/Rec/Track/Best /Event/Hlt/Track/Best/Event/Rec/Track/Best /Event/Hlt/Track/Best

Some locations not needed online/offline

Page 3: LHCb Lausanne Workshop, 21st March. 2006 1/8 Tracking Open Issues E. Rodrigues, NIKHEF LHCb Tracking and Alignment Workshop Some topics to discuss …

LHCb Lausanne Workshop, 21st March. 2006 3/8

2 Packages available:2 Packages available: PatChecker, TrackCheckersPatChecker, TrackCheckers

For sure a lot of double work:For sure a lot of double work: Could easily share tools and codeCould easily share tools and code Should be merged as much as possible, or even completely?Should be merged as much as possible, or even completely?

Do we have all we need for checking?Do we have all we need for checking? Tools neededTools needed Relations/linker tablesRelations/linker tables In .digi and .dst filesIn .digi and .dst files

Checking Packages

Page 4: LHCb Lausanne Workshop, 21st March. 2006 1/8 Tracking Open Issues E. Rodrigues, NIKHEF LHCb Tracking and Alignment Workshop Some topics to discuss …

LHCb Lausanne Workshop, 21st March. 2006 4/8

Tuning of Pattern Recognition

Versions for online:Versions for online: Default options for PatVelo, PatForward are for HLT, I believeDefault options for PatVelo, PatForward are for HLT, I believe Re-tuning for DC’06 is ongoingRe-tuning for DC’06 is ongoing

Versions to be used offline:Versions to be used offline: Use same VELO and forward tracking as online, but with different tuningUse same VELO and forward tracking as online, but with different tuning No such tuning yet availableNo such tuning yet available What is the timescale for having an offline tuning?What is the timescale for having an offline tuning?

Page 5: LHCb Lausanne Workshop, 21st March. 2006 1/8 Tracking Open Issues E. Rodrigues, NIKHEF LHCb Tracking and Alignment Workshop Some topics to discuss …

LHCb Lausanne Workshop, 21st March. 2006 5/8

Track filters:Track filters: Alignment developpers have already raised the question of how to easily get e.g. a Alignment developpers have already raised the question of how to easily get e.g. a

long track with only VELO / seed hits (LHCbIDs)long track with only VELO / seed hits (LHCbIDs) Seems like a “stripping” tool could be handy …Seems like a “stripping” tool could be handy … We have already some very simple track selector tool, butWe have already some very simple track selector tool, but Actually one could make a set of filter tools similar to the FilterCriterion family of Actually one could make a set of filter tools similar to the FilterCriterion family of

tools of DaVinci!tools of DaVinci!

Clone killer and “tracks cleaner” before DST writer:Clone killer and “tracks cleaner” before DST writer: Clone killer algorithm provides “best” set of tracks for physicsClone killer algorithm provides “best” set of tracks for physics These tracks still have too much info that is not to be stored on DSTThese tracks still have too much info that is not to be stored on DST Need a procedure to clean these tracks: all but first state, all measurements, etc.Need a procedure to clean these tracks: all but first state, all measurements, etc. Have to see relation to old “PackXxx” algorithms …Have to see relation to old “PackXxx” algorithms …

Non-const accessors :Non-const accessors : Can we please make sure we have a clear idea of the problem, the possible solutions Can we please make sure we have a clear idea of the problem, the possible solutions

and the pros and cons? With some quantitative timing justifications …and the pros and cons? With some quantitative timing justifications …

Miscellaneous

Page 6: LHCb Lausanne Workshop, 21st March. 2006 1/8 Tracking Open Issues E. Rodrigues, NIKHEF LHCb Tracking and Alignment Workshop Some topics to discuss …

LHCb Lausanne Workshop, 21st March. 2006 6/8

Fitting and OT Ambiguities

Outcome of the event model reviewOutcome of the event model review

Most people wanted a very simple LHCbIDMost people wanted a very simple LHCbID Means no more “spare bits”, where OT ambiguities could be storedMeans no more “spare bits”, where OT ambiguities could be stored

LHCbID simplifiedLHCbID simplified Contains simply some bits for the detector type and the bits for the XxxChannelIDsContains simply some bits for the detector type and the bits for the XxxChannelIDs Bye-bye OT ambiguity …Bye-bye OT ambiguity …

IssuesIssues

OT ambiguities only available in the pattern recognition algorithmsOT ambiguities only available in the pattern recognition algorithms How to pass that info on to the fitting? Seems impossible now …How to pass that info on to the fitting? Seems impossible now …

Can we do (fit) without the OT ambiguities?Can we do (fit) without the OT ambiguities? Not clear. Nobody really knows … will be investigated with DC’06 …Not clear. Nobody really knows … will be investigated with DC’06 …

The problem in 1 sentence:The problem in 1 sentence:How to fit without OT ambiguities?How to fit without OT ambiguities?

Page 7: LHCb Lausanne Workshop, 21st March. 2006 1/8 Tracking Open Issues E. Rodrigues, NIKHEF LHCb Tracking and Alignment Workshop Some topics to discuss …

LHCb Lausanne Workshop, 21st March. 2006 7/8

One possible solutionOne possible solution

Build the Measurements inside the PR algorithmsBuild the Measurements inside the PR algorithms Probably OK for offline trackingProbably OK for offline tracking Easy recipe “on demand”. E.g.: Henk Jan has in hands a few lines to do this …Easy recipe “on demand”. E.g.: Henk Jan has in hands a few lines to do this …

could even introduce a new algorithm property “MakeMeasurements” …could even introduce a new algorithm property “MakeMeasurements” … I would propose this solution offlineI would propose this solution offline

What about online?What about online? Information lost if the Measurements are not built in the “Pat” packages …Information lost if the Measurements are not built in the “Pat” packages … … … unless we make e.g. some linker-like table à la MCTrackInfo/whatever … or ...unless we make e.g. some linker-like table à la MCTrackInfo/whatever … or ...

Otherwise …Otherwise …

Need to know the impact of not having the ambiguities at the start of the fittingNeed to know the impact of not having the ambiguities at the start of the fitting Are we stuck?Are we stuck? Any brilliant ideas / alternatives? The contest is open …Any brilliant ideas / alternatives? The contest is open …

Fitting and OT Ambiguities

Page 8: LHCb Lausanne Workshop, 21st March. 2006 1/8 Tracking Open Issues E. Rodrigues, NIKHEF LHCb Tracking and Alignment Workshop Some topics to discuss …

LHCb Lausanne Workshop, 21st March. 2006 8/8

Bottom line …

• Having a robust / flexible / beautiful tracking is not yet a realityHaving a robust / flexible / beautiful tracking is not yet a reality

• Many issues have to and will be addressed during our DC’06Many issues have to and will be addressed during our DC’06

• Looking forward to the real challenge – DATA!Looking forward to the real challenge – DATA!

• Disclaimer: apologies for the issues I’ve probably forgotten to state …Disclaimer: apologies for the issues I’ve probably forgotten to state …