Talks, verbal reports

• Update on p11.02 certification for muons: Martin Wegner.
  Efficiencies on single muon MC events are 50% for tight and 77% for loose averaged over |eta|<2. Rejection factor of 2 can be obtained by requiring a loose muon for single muon triggers. Higher rejection factor possible if pt cut made. We need to know how good the pt resolution for loose muons is, in order to know whether or not it makes sense to apply a pt cut on loose muons. Main worry is the number of loose muons that fail the chisquared cut in the track fit and thus fail to become tight. Is this also a problem with the offline muons? (It was when p11.02 was first being tested). In gold plated J/psi data events efficiencies look reasonable for forward but rather low for central muons. We need rejection rates separately for central and forward regions.
• Update on dynamic unpacking for muons: Christophe Clement.
  Dynamic unpacker + a completely nominal cfg.dat file give performance identical to that obtained with the old unpacker + correct cfg.dat. Timing OK. Memory consumption and running with full trigger list are main things that still need to be tested. Code needs to be put into CVS. Fantastic progress!
• Update on Global tracking: Daniel Whiteson
  With offline geometry the number of SMT hits picked up by 8-layer CFT tracks improves dramatically. One possibility to make use of L3 global tracking would be to run a single track filter on single muon triggers. For example, a rejection factor of about 20 could be obtained for an efficiency of 50%. (The efficiency is measured by seeing how many events with a tight offline muon have a L3 global track with pt above a given cut.) Daniel will investigate by how much the efficiency/rejection can be improved if 7-hit CFT tracks are included if they have matched SMT hits. Again, we need rejection rates separately for central and forward regions. Issue: how do we get "offline" quality geometry into L3 online?
• Timing for the tau tool: Yann Coudau
  Making the calorimeter clusters passed by pointer rather than by value decreased the average time per event taken by the tau tool from 180 ms to 20 ms!

Plan

• p11.03 release to L3 farm (or p11.02 if p11.03 has problems)
• single electron filters as proposed two weeks ago (!)
• muon tool/filter either in a test run or in standard production (depending on progress in the next few days)
• (global tracking may possibly be ready too, or may need another week?)

Scribe: Terry Wyatt.