sPHENIX discussion of software

["Huang, Jin" <jhuang AT bnl.gov>]

Also available on agenda page: https://indico.bnl.gov/conferenceDisplay.py?confId=2688

 

 

Repository update

·       Carlos (TPC software team): TPC digitization, plan to merged (done)

o   Double checked the pull distribution of clusters

o   https://github.com/sPHENIX-Collaboration/coresoftware/pull/263

·       Chris: Stepping action updated to handle stopped particles.  

o   https://github.com/sPHENIX-Collaboration/coresoftware/pull/265 to 267

·       Tony: plan for new tracking Macro updates following Carlos’ pull request

 

 

New tracking software based on GenFit2 progressive tracking - Haiwang Yu

 

Summary: Prototype of new pattern recognition software based on GenFit2 progressive tracking is implemented and under test.

 

Slide 4: The new part is Trajectory Building. Not yet using Ambiguity cleaning. Rest part in the flow chart is using existing parts.

 

Slide 5:

Tony: what happen if multiple hit in a cluster bin

Haiwang: multi-map was used allows multiple hit in the same cluster window.

Carlos: shouldn’t have to use key if already binned for faster locating of smaller ranges of cluster. More details reported in cluster container design last week.

Tony: is clustering sorting and accessing time consuming?

Haiwang: Currently tracking code regroup clusters using key-map. This gives flexibility of pattern recognition software to choose binning. If container is binned, then we can directly use it. Will quantify the time consumption of key-window searching in future talks.

 

Slide 6:

Xin: why losing 10% clusters per layer

Haiwang: currently using 3-sigma searching window to balance time VS association efficiency. Need further tuning.

Xin: how materials are handled

Haiwang: full handling of material and geometry via ROOT TGeoManager utilities.

Xin: could use some approximation to help speed up

Haiwang: Yes. And the seeding is approximation with cylindrical hough transform.

Tony: real events will have very different pT spectrum.

TODO: change to use HIJING input, some are pre-generated: https://wiki.bnl.gov/sPHENIX/index.php/Event_generators#HIJING

 

Slide 7:

Haiwang: unknown how high pT track has less seeding efficiency. Could be from hough window optimization for intermediate to low pT tracks.

 

Slide 8:

Tony: much better resolution observed when compared with previous pattern recognition even in single track events with TPC.

Haiwang: part reason is to bring back efficiency of cluster association and recovers the full tracking level arm

 

Slide 9: profiling

Haiwang: using 10-track to perform the profile analysis, as the progressive pattern recognition was found to be quite linear to # of seeds.

Jin: many steps are not linear to number of tracks, therefore concern using 10 track profiling to indicate that of central HIJING. Suggest use central HIJING event to perform profiling.

 

Discussion:

 

Tony: goal is also to run this with ladders and recover the cluster association efficiency.

Haiwang: should be able to work with ladders. May need to add backward tracking to collect back double-hit-per-layer due to the overlapping ladders.  

 

TODOs

·       Test in Central HIJING events, instead of the current n-track events. If works, we could use this module for analysis and optimize for speed later

·       Work with Carlos when the new cluster container is available.

·       Test with ladder geometry, when new tracking macro is produced.

 

 

Discussion on sPHENIX coding convention- Chris Pinkenburg

 

Summary: discussing and composing the sPHENIX coding convention, which is aimed to improve readability and standardize allowed code handling.

 

Wiki document under construction by Chris: https://wiki.bnl.gov/sPHENIX/index.php/Codingconventions . Besides codling convention, it also addresses

·       include file ordering in the source code

·       Clang-format, a nice tool to format a source code according to predefined convention.

·       Examples of Not-to-do in MakeFile.am

 

Carlos: how do we handle data types cross platform, e.g. about float standard?

Chris: we only encounter problem variation of definition for long long. Not foresee problems in float storage changes. However, for IO objects that using the ROOT IO utilities, we should consider follow the ROOT format, e.g. in using Float_t instead of float.

 

Plan: make example code representing the coding style and an easy to follow standard for collaborator to copy-paste from.

 

 

 

______________________________

 

Jin HUANG

 

Associate Physicist

Brookhaven National Laboratory

Physics Department, Bldg 510 C

Upton, NY 11973-5000

 

Office: 631-344-5898

Cell:   757-604-9946

______________________________

 

From: Huang, Jin
Sent: Monday, April 10, 2017 11:52 AM
To: 'sphenix-software-l AT lists.bnl.gov' <sphenix-software-l AT lists.bnl.gov>
Cc: 'sphenix-l AT lists.bnl.gov' <sphenix-l AT lists.bnl.gov>
Subject: [reminder] sPHENIX simulation and software meeting, Tue Apr 11th, 1PM EST @ 2-219.

 

Hello everyone,

 

We will have a regular sPHENIX simulation and software meeting, Tue Apr 11th, 1PM EST @ 2-219.

 

Currently on agenda is:

•         New tracking software with GenFit-based progressive tracking - Haiwang

•         Follow up on software coding style convention - Chris

Further topics will be welcomed too  

 

The agenda page is at https://indico.bnl.gov/conferenceDisplay.py?confId=2688 , for which the common modification password is "1008"

•         For residents at BNL, we will meet at the 2-219 conference room if not closed for snow

•         To join the Meeting via phone bridge:  https://bluejeans.com/619960505 

•         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/numbers)

2) Enter Conference ID: 619960505

 

Cheers,

 

Jin

 

 

______________________________

 

Jin HUANG

 

Associate Physicist

Brookhaven National Laboratory

Physics Department, Bldg 510 C

Upton, NY 11973-5000

 

Office: 631-344-5898

Cell:   757-604-9946

______________________________