sPHENIX tracking discussion

[Christof Roland <christof.roland AT cern.ch>]

Hmm, the resolution is probably still dominated by the scatter of the actual measurments, I don't think we can get any better than that by twiddling the errors.

Bad errors will wrongly weight the clusters, so this is bad, as the "fix' example shows. Correct errors should give you the best the measurments have to offer. 

There is still some fine tuning to do for n phi ==2 clusters and maybe a residual cluster charge/size dependence. 

Also I have not touched the z errors yet, but I assume this isn't that important. Will check though. 

cheers

   Christof

On 23. Mar 2022, at 14:37, Osborn, Joe <osbornjd AT ornl.gov> wrote:

Awesome, that is indeed great news! Is it possible to improve the resolution at all with the parameterized uncertainties? Maybe this is for a later discussion once we see how the parameterization actually works.

---------------------------

Joe Osborn, Ph.D.

Associate Research Scientist

Oak Ridge National Laboratory

osbornjd AT ornl.gov

(859)-433-8738

---

From: Christof Roland <christof.roland AT cern.ch>
Sent: Wednesday, March 23, 2022 9:32 AM
To: Osborn, Joe <osbornjd AT ornl.gov>
Cc: Hugo Pereira Da Costa via sPHENIX-tracking-l <sphenix-tracking-l AT lists.bnl.gov>
Subject: Re: [EXTERNAL] [Sphenix-tracking-l] Tracking efficiency with parameterized errors

 

Ha!

Cool. So the "Fix" plots shows that the errors actually do something. 

And the para ending up on top the default means that we can probably live

without storing the errors in the clusters. 

Thats pretty good news. 

Cheers

   Christof

On 23. Mar 2022, at 14:26, Osborn, Joe <osbornjd AT ornl.gov> wrote:

Hi Christof,

I chained the files together and plotted the pT resolution for quality<10 tracks, here's what I see.

Joe

---------------------------

Joe Osborn, Ph.D.

Associate Research Scientist

Oak Ridge National Laboratory

osbornjd AT ornl.gov

(859)-433-8738

---

From: Christof Roland <christof.roland AT cern.ch>
Sent: Wednesday, March 23, 2022 9:01 AM
To: Osborn, Joe <osbornjd AT ornl.gov>
Cc: Hugo Pereira Da Costa via sPHENIX-tracking-l <sphenix-tracking-l AT lists.bnl.gov>
Subject: Re: [EXTERNAL] [Sphenix-tracking-l] Tracking efficiency with parameterized errors

 

Hi Joe, 

currently I already have SvtxEvaluator output files. Check:

sphnx02->/phenix/u/bogui/data/MacrosFileSize/macros/detectors/sPHENIX/TracksPu50] 520) l *056*

-rw-r--r-- 1 bogui sphenix 2239689 Mar 18 12:51 TrackingDefault_1_1056.root

-rw-r--r-- 1 bogui sphenix 1178435 Mar 22 11:50 TrackingFix_1_1056.root

-rw-r--r-- 1 bogui sphenix 2352457 Mar 22 11:30 TrackingPara_1_1056.root

There are 1000 files, one event each.

Hugo's trackevaluation I can also run easily, if that is what you need. 

Cheers

On 23. Mar 2022, at 12:50, Osborn, Joe <osbornjd AT ornl.gov> wrote:

Hi Christof,

Cool! If you produce evaluation files, I can make the resolution plots pretty easily. Are you using the SvtxEvaluator or Hugo's evaluator?

Joe

---------------------------

Joe Osborn, Ph.D.

Associate Research Scientist

Oak Ridge National Laboratory

osbornjd AT ornl.gov

(859)-433-8738

---

From: sPHENIX-tracking-l <sphenix-tracking-l-bounces AT lists.bnl.gov> on behalf of Christof Roland via sPHENIX-tracking-l <sphenix-tracking-l AT lists.bnl.gov>
Sent: Wednesday, March 23, 2022 5:53 AM
To: Hugo Pereira Da Costa via sPHENIX-tracking-l <sphenix-tracking-l AT lists.bnl.gov>
Subject: [EXTERNAL] [Sphenix-tracking-l] Tracking efficiency with parameterized errors

 

Hi Everybody, 

I ran a few jobs resetting the errors in the TrkrClusters from the values assiged by the clusterizer to either fixed rphi errors or errors derived from a parametrization based on the track angles.

Below is a plot of the tracking efficiency using default errors, fixed errors and parmetrized errors. The efficiencies are identical.

Note of caution, I reset the errors in the PHTpcTrackSeedCircleFit module, so this happens after the cluster cleaner is run. 

I will rerun a version now setting the cluster error to a fixed value in the clusterizer already. This will then bypass the ClusterCleaner. 

@Joe I think it would be worthwhile to check the momentum resolution with the parameterized errors. 

What machinery do you use to produce your plots? It will take a while untill I can commit my stuff to github. Can I produce some 

output files you can use process to plot the resolution?

Cheers

   Christof 

<eff_error.jpg>

<ptres_paramUncert.pdf>