1 / 12

LHCb T-Rec Meeting

LHCb T-Rec Meeting. Clone Tracks Killing in DC’06 E. Rodrigues, NIKHEF. Clone Killing in DC06. Goal Find tracks that are clones of other tracks Definition of clones Two tracks are clones of each other if they share in the VELO and in the SEED stations at least 70% of hits

druce
Download Presentation

LHCb T-Rec Meeting

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. LHCb T-Rec Meeting Clone Tracks Killing in DC’06E. Rodrigues, NIKHEF LHCb T-Rec Meeting, 12th February. 2007

  2. Clone Killing in DC06 Goal • Find tracks that are clones of other tracks Definition of clones • Two tracks are clonesof each other if they share in the VELO and in the SEED stations at least 70% of hits Present usage in our tracking • Clone finder and killer run at the end of the tracking • Uses by default Long, Ttrack, Upstream and Downstream tracks • Outputs UNIQUE (= not clone) tracks to “best” container LHCb T-Rec Meeting, 12th February. 2007

  3. The bug (1/2) • Bug found in the TrackCloneFinder tool ! • It affects all DC06 DSTs produced so far with Brunel, • versions up to v31r0 Checking for clones on a buggy DST … with the fixed tool … Read back a buggy DST • check the “best” container • Check for clones with the fixed TrackCloneFinder tool • Plot the number of clones found per event … This is a lot ! The bug potentially affects the clone finding every time one of the tracks does not have OT hits -> IT and OT hits ignored in the comparisons … LHCb T-Rec Meeting, 12th February. 2007

  4. The bug (2/2) Looking at the type of these clone tracks … • Clones of all track types ! • NOTE: Velo tracks should not be included, as they are not considered by the clone killer in Brunel But can the TrackCloneFinder bug explain it all …? LHCb T-Rec Meeting, 12th February. 2007

  5. Clones found on fixed DST (1/3) All tracks considered Discarded: - Velo tracks - tracks from a same container Velo tracks discarded I.e. at level of P. R. tracks. This is important to keep in mind Read back a DST produced with the fixed tool • Check the “best” container • Check for clones with the fixed TrackCloneFinder tool • TrackCloneFinder.CompareAtLHCbIDsLevel = true LHCb T-Rec Meeting, 12th February. 2007

  6. Clones found on fixed DST (2/3) All tracks considered Velo tracks discarded Discarded: - Velo tracks - tracks from a same container LHCb T-Rec Meeting, 12th February. 2007

  7. Clones found on fixed DST (3/3) Observations & conclusions: • Almost all clones are from tracks of same type … • … and from the same container • The pattern recognition algorithms are providing sets of tracks with clones amongst them ! • “Reason”: by default the clone killer does not compare tracks from the same container (property SkipSameContainerTracks = true ) This is another « conceptual bug » ! LHCb T-Rec Meeting, 12th February. 2007

  8. Clones found on buggy DST (1/2) Read back a buggy DST with the fixed tool • Check the “best” container • TrackEventCloneKiller.SkipSameContainerTracks = false • Check for clones with the fixed TrackCloneFinder tool • TrackCloneFinder.CompareAtLHCbIDsLevel = true Discarded: - Velo tracks - tracks from a same container All tracks considered LHCb T-Rec Meeting, 12th February. 2007

  9. Clones found on buggy DST (2/2) All tracks considered Discarded: - Velo tracks - tracks from a same container LHCb T-Rec Meeting, 12th February. 2007

  10. A final check, after fixes Running Brunel: • Fixed TrackCloneFinder tool • Comparing fitted tracks at the level of the Measurements • Clone killer comparing tracks from the same container (property SkipSameContainerTracks = false , not the default now!) No clones found, obviously LHCb T-Rec Meeting, 12th February. 2007

  11. Proposed Solution (1/2) TrackTools • Code has been fixed • In CVS in the head version SIMPLE SOLUTION - in DaVinci • Add options at the very beginning of execution, to run a simple cleaner algorithm that removes clones from the best container • probably the easiest thing to do • Need to write the trivial algorithm Remarks • Fix requires a new DaVinci build with a new Rec release, in order to run on the DSTs produced so far • I would suggest to build a new Brunel as soon as possible to continue production with the bug fixed in the reconstruction LHCb T-Rec Meeting, 12th February. 2007

  12. Proposed Solution (2/2) Further remarks • Clone killer run in DaVinci will compare at the level of LHCbIDs, not the result of the fit (comparison of Measurements), as in Brunel • This is a small difference – but it is there FULL SOLUTION - in DaVinci • Add options at the very beginning of execution, to refit all the tracks in the “best” container • Run the clone killer afterwards, just as in Brunel • This does mean a significantly slower DaVinci! • Probably not what one really wants … LHCb T-Rec Meeting, 12th February. 2007

More Related