Home » PANDA » PandaRoot » Fast Simulations » First test with fast sim: Ds(2536)+
| First test with fast sim: Ds(2536)+ [message #16216] |
Thu, 03 April 2014 16:40  |
Elisabetta Prencipe (2)
Messages: 214
Registered: February 2013
|
first-grade participant |
From: *ikp.kfa-juelich.de
|
|
Dear all,
I am trying to do again my simulations with the fast sim (trunk-rev-24275), and compare the results obtained with the full simulation.
I am not sure if I am doing the proper comparison, this is why I post plots and numbers which I obtain and ask to the physics convenors.
So, the channel under exam is:
pbar p ----> Ds- Ds1'(2536)+
Ds- --->K+K-pi-
Ds1'--->D*0 K+
D*0 --->D0 gamma
D0 ---->K-pi+
plab = 9.83 GeV/c
N_generated events with EvtGen = 10 000
I show you a comparison between the number of entries obtained in case I plot the mass (full simulation) and "m" (fast simulation). Here attached you can look at some distributions.
---------------------------------------------------
-------- D0 ---- | Ds ----| Ds1'
---------------------------------------------------
Fast | 9962 | 15253 | 2960
Full | 5127 | 2450 | 349 (ftm)
Full | 8842 | 5307 | 2331 (best pid)
Full | 10506 | 12204 | 1140 (true pid)
Looking at the distributions obtained from the fast simulation, their shape look similar to the ones obtained from the "tr_pid" in the full simulation. So, I guess this is what we should compare. Correct?
Here attached are some distributions of interest. Please do not be worry because of the fact that the efficiency of Ds1' looks so low: in the analysis I will use the missing mass of the Ds- to reconstruct it., and it improves a lot actually. Here I am just checking consistency with the full reconstruction of all decay products in this chain.
Best regards, Elisabetta
|
|
|
|
| Re: First test with fast sim: Ds(2536)+ [message #16217 is a reply to message #16216] |
Thu, 03 April 2014 16:53  |
Elisabetta Prencipe (2)
Messages: 214
Registered: February 2013
|
first-grade participant |
From: *ikp.kfa-juelich.de
|
|
Some more addings to my previous post: for people who like to reproduce these results, this is how I filled up the lists:
// *** Select with no PID info ('All'); type and mass are set
theAnalysis->FillList(kplus, "KaonBestPlus", pidalg);
theAnalysis->FillList(kminus, "KaonBestMinus", pidalg);
theAnalysis->FillList(piplus, "PionBestPlus", pidalg);
theAnalysis->FillList(piminus,"PionBestMinus", pidalg);
theAnalysis->FillList(gammas, "Neutral",pidalg0);
where:
TString pidalg = "PidChargedProbability";
TString pidalg0 ="PidNeutralProbability";
Best regards, Elisabetta
|
|
|
|
| Re: First test with fast sim: Ds(2536)+ [message #16224 is a reply to message #16216] |
Fri, 04 April 2014 07:51 |
Klaus Götzen
Messages: 293
Registered: June 2006
Location: GSI
|
first-grade participant |
From: *gsi.de
|
|
Hi Elisabetta,
that are fast first results! 
I think in general, that the consistency between fast and full sim in terms of absolute numbers is not so important. In your case of selection I'm quite sure that the PID works better in fast sim, so that the 'BestPid' leads to something comparable to 'True PID' in full sim.
However, the question which now comes is: What detectors can be switched off or how worse resolutions and efficiencies for individual detector components are allowed to be without a too strong degradation of your figure of merits. That could be e.g. signal efficiency, signal to noise ratio, but maybe also something completely different.
You could e.g. turn off the barrel EMC, the Forward Spectrometer, the DIRC etc, and look, what your results are compared to the full setup.
Best,
Klaus
|
|
|
|
| Re: First test with fast sim: Ds(2536)+ [message #16225 is a reply to message #16224] |
Fri, 04 April 2014 08:43 |
Elisabetta Prencipe (2)
Messages: 214
Registered: February 2013
|
first-grade participant |
From: *ikp.kfa-juelich.de
|
|
Hi Klaus,
OK I'll try and I will let you know what changes if I switch off this or that detector.
Let's say, the steps now are:
a) update the release mar14
b) run full- and fast- simulation for the same channel, in the same release
c) check consistency of evt shape/distributions
this with all detectors included, in the fast simulation.
d) try to switch off this/that detector and see what changes
e) give feedback to this forum and/or to the regular meetings
Reading the email in the analysis forum, I would ask (as secondary order question, for the time being) what shall we do with the background: I mean, the main way to reject it is:
1) cut on the photon momentum
2) use the kinematic fit
in the charm analysis, at least.
Point 1) in fast simulations is OK. What about point 2) ?
Thank you for your reply,
Elisabetta
|
|
|
|
|
|
|
|
|
|
Goto Forum:
Current Time: Wed Dec 04 20:54:29 CET 2024
Total time taken to generate the page: 0.00653 seconds
|
GSI Forum
Search
Help
Members
Login
Home
