sPHENIX tracking discussion

["Frawley, Anthony" <afrawley AT fsu.edu>]

Hi All,

Here are the minutes of today's (March 31) tracking meeting:

Cheers
Tony


Redesign of hit/digit/cluster structure and containers - Carlos Perez
-----------------------------------------------------------------------------------------
Carlos described his new scheme for data structures for the TPC. This includes a new object, vHit, a light version of the hit class, that would be used for all tracking detectors if it is suitable.

Slide 3:
The proposed vHit object, which records the information for G4Step, keeps only the average position and length of track in the TPC layer.
There was discussion of whether for other detectors we need to keep exit and exit points. For the maps we do, yes, because of the way the charge sharing needs to be done. Not clear for the INTT.
Jin argued also that large angle tracks may cause problems due to multiple steps in a layer if we store only the average point.

slide 4:
Can't assume cylindrical symmetry - the maps and intt ladders are tilted!

Stores clusters in local coordinates with identifying detector information. Maybe not necessary, but there is a need to make cluster finding and pattern recognition fast by organization.
Haiwang commented that this would not match very well with what Genfit needs.
Christof: there is a short term need to stay in global coords for Genfit, but later, for data handling, we may want to use local coords.

It was noted that a generalized hit container should be accompanied by generalized cluster container, for easy cross reference. Carlos is still thinking about the cluster container.

Jin: Recommend against inventing a new hit container. The existing class PHG4Hit is written to allow use of any new parameters you need, and you can use it without any parameters you don't want. Suggest using PHG4Hit for what you are planning to do with vHit. Then you only have to change 2 files, and it does not affect anyone else.  This sounded reasonable to all. Jin and Carlos will discuss offline. We also decided that we needed to keep the entrance and exit points for many reasons.

Haiwang: Comment on clusterizer speed - the Hough transform in the existing tracking code (which includes all tracking layers!) costs 97% of the total time in a Hijing event, so the clusterizer is not a big effect (1.6% at present). Christof: the Hough speed will be reduced a lot by limiting the number of layers used to find the track seeds, but we do not have results yet to quantify this.

Christof: We will need hits sorted by layer for pattern recognition, and by eta, phi. That is how we would handle real data.

Other Reports:
--------------------
Tony: MAPS charge sharing and INTT cluster fix are both in place, and we can check them off.

Haiwang: made a pull request for the truth pattern reconstruction and Genfit refitting to make it easily usable by others, and added macros that include the hooks needed to run them.

Tony: I looked at Haiwang's macros, and they have up-to-date ladder and cylinder cell models of the silicon - just a couple of tweaks needed. So I will use Haiwang's macros as the basis for the new ladder silicon and cylinder silicon macros. I have a pointer to an up to date TPC macro from Veronica, and will move the TPC part into the new macros. Then we can retire all of the obsolete macros, so they don't have to be maintained.

Christof said that he got Haiwang's code, and matched it with the toy model.

Sourav: is working on using a reduced number of layers in the Hough transform, nothing to report yet.

---

From: Frawley, Anthony
Sent: Tuesday, March 28, 2017 3:38 PM
To: sphenix-tracking-l AT lists.bnl.gov
Cc: Frawley, Anthony
Subject: sPHENIX tracking meeting, Friday March 31 at 9:00 am EDT

 

Hello All,

We will have an sPHENIX tracking meeting on Friday, March 31, at 9:00 am EDT.

The focus will be on the tracking software upgrade. The agenda page is:

https://indico.bnl.gov/conferenceDisplay.py?confId=2985

sPHENIX tracking meeting (31 March 2017) indico.bnl.gov To join the Meeting: https://bluejeans.com/822435626 To join via Browser: https://bluejeans.com/822435626/browser To join with Lync: https://bluejeans.com/822435626/lync To join via Cisco Jabber Video: https://bluejeans.com/822435626/jabber To join via Room System: Video Conferencing System: bjn.vc -or- 199.48.152.152 Meeting ID: 822435626 To join via Phone: 1) Dial: +1 408 740 7256 +1 888 240 2560(US Toll Free) +1 408 317 9253(Alternate Number) (see all numbers - http://bluejeans.com/

Best regards

Tony