Collection of projects for 2011

Gerneral trigger issues

  • Hlt1TrackMuon (Roel)
    • review velo track quality cuts
    • matching to L0 object
    • possibility to use NoIP line

  • review the 17 D2mm lines (Walter & Alksander)
    • use topo or TIS for control channels?
    • Note that the topo will explicitly veto prompt charm next year so probably not as useful here
    • agree on one trigger strategy (--> reduce the # of lines)
    • take only a reasonable bandwidth

  • L0/HLT1 Hadron strategy (Hamish (?) / Carla (?) / Sandra (?) Vava)
    • Review the possibility of having a hard-GEC/soft-ET L0 hadron trigger for charm
    • Review the possibility of a "soft" (di) hadron trigger which would subscribe to the above "hardGEC" L0 channel.
    • Review the optimum GEC level(s) for hadronic channels given the offline selection performance on the new (>23 Sept 2010) data.

  • Hlt2 Charm strategy (Carla/Sandra/Hamish)
    • Review the possibility of an inclusive Hlt2 charm trigger

  • Topological trigger (Mike/Chris Thomas)
    • Reoptimize the topological using the offline samples collected in 2010
    • Fisherize (and more?) the topological to improve efficiency to ~100%
    • Introduce confirmation with respect to Hlt1
    • Investigate adding Kshorts to the topological

  • Lifetime unbiased hadron triggers (Uli Kerzel/Vava)
    • Review lifetime unbiased hadron triggers for next year, compare efficiency to TIS efficiencies and conclude on an optimal strategy.

  • NN trigger for charm (Uli Kerzel/Harry Cliff)
    • Investigate a NeuroBayes based trigger for D->hh for next year

  • MC signal samples
    • force working groups to provide new MC2010 samples?

HLT tracking

  • set up line for two stage forward tracking w/o PT cut (Stephie/Johannes for the tracking part / ??? for the charm part)
  • prepare seeding for high PT tracks to recover efficiency for core channels (???)
    • a Pt cut of 0.5-1GeV in PatSeeding should allow us to run the seeding without major time penalty. This would make online and HLT2 tracking almost identical and recover the 1-3% tracking eff lost when only using PatForward.
  • review the efficiency/timing balance in the forward tracking (???)
    • if we use only inclusive triggers, it may be better to sacrifice a few percent per-track-efficiency for timing. Need to find the optimum balance for next year.
  • review the use of PatVeloGeneral

Operational issues

  • MonitorSvc:
    • Test bug fix for HLT trend plot values after run change (PM)
    • Test MonProfile::splitObject bug fix (PM)
    • Revisit Adder/Saver architecture and implementation (works, but we need to sleep ~1 minute at the start of the run) (PM) - Merge Adder & Saver? (PM)
    • Fix end of run bug for EFF: savesets at end of run are not marked with EOR -> this prevents Presenter from working in history mode for HLT shift pages (PM)
    • Modify top level Adder to use Olivier's Gaudi trend plot tool -> trend plots would be properly merged in history mode (PM)
  • Performance
    • Difference between "online" and "offline" understood? ("offline" factor 2 faster) (N)
    • tcmalloc (N)
    • configuration time (~7/8 mins): where can we gain (PM)
      • configuration time 1 job: 1 min 55 secs (PM)
    • monitoring tools to measure farm performance (PM)
  • Cond DB snapshots (N)
    • automate creation of snapshots after release
  • TCK management (N)
    • TCK creation. Procedure still takes 1 day. Can it be simplified/accelerated.
    • TCK management/ selection of TCKs - do we need to simplify
    • TCKsh simplification (scripts)
  • Documentation (N)
    • Documentation of routing bits (a TWIKI exists, do we need an EDMS article?)
  • Moore (perhaps this needs to go under a different heading)
    • Will the coming farm upgrade be big enough for 2011 running? Do we have the knobs to turn? (N)
    • GECs: relaxing Velo hits 3000->4000 (N)
    • Use new Velo tracking (N)
    • Christmas tree line (based on invers L0 cuts) - did Themis make a proposal? (N)
    • increase min PV multiplicity (N)
    • run off release area (N)
    • do we protect enough against bad online conditions (eg 0 B field) (N)
    • velolitecluster messages gone? (N)
  • Monitoring
    • "simple" C++ task for mass plots: hadronic mass plots. j/psi peak from DiMuonUnbiasedJPsiDecision. how to do this? (N)
      • the sel reports don't have enough information.
      • a flag to be used in the monitoring farm could signal to put this information back
      • run DaVinci on reconstruction farm Brunel output
    • make trend plot of mu (N)
      • in the pipeline; however, do we get enough L0 calo events in the MF?
    • monitor L0 event size. what else? (N)
    • rate to storage histogram ok? (PM)
    • review hlt shift pages (PM)
    • should we normalize against L0 physics instead of L0 calo? (PM)
    • make twiki pages for hlt shift pages (PM)

-- JohannesAlbrecht - 05-Oct-2010

Edit | Attach | Watch | Print version | History: r8 < r7 < r6 < r5 < r4 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r8 - 2010-11-18 - EricvanHerwijnen
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    LHCb All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2020 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback