November 7, 2011 (11:30AM), Physics Working Group

From GlueXWiki
Jump to: navigation, search

ESNET and EVO will be available:

1. ESNET: 85-42553 (85-HALLD)

2. An EVO room under GlueX will be reserved

3. Phone:

  • +1-866-740-1260 : US+Canada
  • +1-303-248-0285 : International
  • then enter participant code: 3421244# (remember the "#"). Phone connection only upon request.
  • or www.readytalk.com (and code without the #)

Agenda

  1. Announcements
  2. Updates on reconstruction?
    1. Tracking
    2. Calorimetry
    3. PID
  3. Amplitude Analysis of b1pi (Igor)
    • Fits with real detector acceptance (no background yet).
    • Still low stats -> bad leaks.
  4. Cascade Simulations (Nathan)
  5. Mass production and storage of MC (UConn)

Minutes

I. Updates on Reconstruction Software:

A. Tracking: Simon fixed the recent crashes in the tracking code, which were apparently due to bad start counter timing information. Igor confirmed that this has fixed his problems.

B. Calorimetry: Will checked in tweaks to Matt's BCal algorithm. (Note that the KLOE algorithm is still the default algorithm.)

C. PID: No updates on PID, Paul is working on BCal segmentation studies.

II. Ongoing Analyses

A. b1pi (Igor): Igor showed the results of his first fits using full detector acceptance and no background. He generated two waves, and then used five in the fit. There was considerable leakage into non-signal waves, but he also used small samples of signal MC for the normalization integrals. This could, in principle, explain both the leakage and the apparent underestimation of the sizes of the error bars. The event generator (which uses accept-reject and runs around a few Hertz) is currently a bottle-neck for the signal MC production. Importance sampling on the omega could help a lot and Igor will investigate.

B. cascades (Nathan): Nathan is still looking at tracking issues. One test he performed was to seed tracks with truth information. This only resulted in minor improvements in efficiencies. Another test was to look at the difference between track efficiencies when the track is isolated (i.e. produced using a particle gun with the same kinematic distributions) and when the track is embedded in a full event. When isolated his pi- has an efficiency ~30%, but in a full event the same pi- has efficiency ~10%. So the background associated with additional curling tracks has a large effect on efficiency.

III. Mass MC Production and Storage

Richard may be able to give an update in two weeks. There is currently 60TB of pythia MC on the grid (in the older uncompressed format).