Difference between revisions of "June 7, 2010 (11AM), Physics Working Group"

From GlueXWiki
Jump to: navigation, search
(Agenda)
(Minutes)
 
(2 intermediate revisions by 2 users not shown)
Line 22: Line 22:
 
#** [[Media:Jake_3pi.pdf|a few plots from Jake]]
 
#** [[Media:Jake_3pi.pdf|a few plots from Jake]]
 
#Updates on PAC report?
 
#Updates on PAC report?
 +
 +
== Minutes ==
 +
 +
 +
1.  Background MC
 +
 +
* We decided we will use the OSG for background MC generation and storage.
 +
 +
* The general idea will be to (i) generate "thrown" minimum-bias events with Pythia; (ii) run these events through HDGeant and mcsmear; (iii) do partial reconstruction (at least track reconstruction); and (iv) store some combination of output files on the grid.  Then, one can either (v.a) run analysis jobs on the grid to produce root files (which can then be read remotely by a locally running root executable using URL's); or (v.b) use "wget" (or some equivalent) to fetch files and run analysis jobs locally.
 +
 +
* Ryan will try to produce some specifications for this process, which can then be discussed.
 +
 +
* As a first step, Jake and Ryan will try to generate a small sample of Pythia events on the OSG and then run Jake's 3pi analysis code on the output (following [[HOWTO get your jobs to run on the Grid|Blake's howto on the wiki]]).
 +
 +
 +
2.  Update on IU's simulation of pi+pi-pi0
 +
 +
* Ryan showed a few of Jake's plots of signal MC.  The biggest difficulty right now seems to be with the pi0's, where there is a substantial combinatoric background.  Jake is working on ways to suppress this.
 +
 +
* The PID seems to be working well (but it's unclear how much the reconstruction code is "cheating").  Simon and David are still working on the PID code.  There was some discussion about which start time to use for TOF.  Richard has written up [[How HDGeant defines time-zero for physics events|a wiki page detailing how time-zero is currently defined for MonteCarlo-generated events simulated in HDGeant]] (true event start time vs machine time vs tagger time).

Latest revision as of 14:40, 7 June 2010

ESNET and EVO will be available:

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

2. An EVO room under GlueX will be reserved

3. Phone:

+1-800-377-8846 : US

+1-888-276-7715 : Canada

+1-302-709-8424 : International

then enter participant code: 77438230# (remember the "#")

Agenda

  1. Finding a common area for background MC data samples (Richard)
  2. Updates on simulations?
  3. Updates on PAC report?

Minutes

1. Background MC

  • We decided we will use the OSG for background MC generation and storage.
  • The general idea will be to (i) generate "thrown" minimum-bias events with Pythia; (ii) run these events through HDGeant and mcsmear; (iii) do partial reconstruction (at least track reconstruction); and (iv) store some combination of output files on the grid. Then, one can either (v.a) run analysis jobs on the grid to produce root files (which can then be read remotely by a locally running root executable using URL's); or (v.b) use "wget" (or some equivalent) to fetch files and run analysis jobs locally.
  • Ryan will try to produce some specifications for this process, which can then be discussed.
  • As a first step, Jake and Ryan will try to generate a small sample of Pythia events on the OSG and then run Jake's 3pi analysis code on the output (following Blake's howto on the wiki).


2. Update on IU's simulation of pi+pi-pi0

  • Ryan showed a few of Jake's plots of signal MC. The biggest difficulty right now seems to be with the pi0's, where there is a substantial combinatoric background. Jake is working on ways to suppress this.