sPHENIX tracking discussion

[Anthony Frawley <afrawley AT fsu.edu>]

Hi All,

Sorry for the delay getting these out. Here are a few notes from the tracking meeting we had last Tuesday, February 13.

Cheers

Tony

---------

Darren presented some slides updating his proposal for new storage objects following the discussion at the February 6th meeting. See his posted slides for details.

https://indico.bnl.gov/conferenceDisplay.py?confId=4268

For context, we had agreed on February 6 that the hit container (TrkrHitSet) would contain hit information from a detector element which would contain a cluster. 

There was discussion about slide 3, where there were open questions about what was the appropriate detector element for the INTT and TPC. The conclusion was:

INTT:  A sensor (not a chip).

TPC:  An arc of the readout comprising 1/12 of a ring.

Also on slide 3, there was discussion about moving the layer to the top of the key holding the hitset identifying information, since we sort hits by layer. In the end we decided to leave the subsystem ID at the top of the key, and have the subsystem local layer next.

Slide 5:

Keep the summed ADC value in the cluster? Yes, needed for sanity checks, cuts, dE/dx.

Mapping from clusters to hits will be needed for finding dead pixels, for example. The cluster should have an additional pointer to a map of (cluster -> readout hits (i.e. cells)) OR the cluster key allows one to go to TrkrHitSet to get a pointer to the hit/cluster map
 - We decided to go with the latter, put it in TrkrHitSet.
We decided we should have 2 separate objects for tracker details and truth information, to facilitate the switch to real data.

Should we implement the new storage objects before/during/after the code reorganization? We agreed they should be set up in parallel, and we should switch to them when ready. Best not to wait too long, or code gets out of sync.

Discussion about function naming. Jin sent a link on Bluejeans for Darren to look at.

We think we now have the storage objects at a point where they are ready to implement.

Who will do what?
Darren has the MVTX part almost ready.
Darren will make a skeleton for the INTT and TPC and ask experts to fill in functions.
Tony will fill it out for the TPC.

We briefly discussed code reorganization. We will make that the main focus of the tracking meeting on February 20.  We agreed that Haiwang should start on the separation of pattern recognition and tracking into separate modules now. Haiwang will make some slides for the meeting on the 20th.