sPHENIX discussion of software

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

Also available on agenda https://indico.bnl.gov/conferenceDisplay.py?confId=1343

 

HCal simulation - John Lajoie

 

Page 4:

Jin: Eta ~ 0.9 has smaller high side tail?

John: general observation to say so. Unclear reason yet.

 

Page 6:

Mike: what is the eta for the leakage study?

John: usually eta~0.3

Mike: so, we should not expect side leakage.

 

Page 11-12:

John L.: energy resolution and leakage did NOT change much after sampling fraction tuning on the G4hit level.

Mike: ALSO found very little difference in previous study when implementing light collection model in the HCal code.

John L./John H.: Suggest resolution is dominated by fluctuation of shower sampling, rather than variation in sampling fraction. Suggest not deserve correcting SF.

John H., John L., Chris: Could be verified with a high SF calorimeter with standalone simulation or cylinder calorimeter.

 

Abhisek: sigma[E]/E would be different after correcting for the missing energy?

John L.: leading order correction would be a scale up, not change sigma[E]/E

Mike: beyond linear scale up, a complicated correction to treat missing energy would be difficult in HI occupancies. Building tools (jet and calorimeter evaluators) to allow such study in the next stage of full detector simulations.

 

Page 16:

Chris: in the production, number of crossing for inner and outer Hcal is locked to be equal up to 10 crossing. At 15 crossing, inner Hcal use 10 crossing and outer Hcal use 15 crossing due to geometrical constraint.

 

Page 26-33:

Chris: low side tail at higher crossing could from that Hcal is thinner in interaction length at high crossing.

Mike: 0% high side tail suggest fitting issue.

 

Abhisek: use tower to do such summary plots.

John: not for this particular study, which use G4hit level information to illustrate performance VS intrinsic structure for the HCal.

 

Jin: high side tail seems high for 1/2 crossings?

John: need to follow up with higher stat. Could be fit issue at low stat.

 

Chris:

1)      Eta = 0 points are possible too.

2)      Newer high stat prod. Do not have absorber hits and we can use SF from first production.

3)      High stat production will be done in about 24 hours

 

 

Preliminary conclusions:

John L. / John H.: purpose of tilt would be for non-channeling for zero-field runs. Crossing choice has a wide range of acceptable choices. Final decision dominated by ease of engineering.

 

Suggestions on pre-CDR digested plots:

Mike: good to plot 24GeV pion for various crossing as summary plots.

Megan: resolution in stat. and const. term VS crossing are good for summary too.

Action Item [John L.]: Agree. Good suggestions. Try to get by Friday for pre-CDR.

 

 

 

Production status: Chris Pinkenburg

 

Current production outputs:

/gpfs02/phenix/prod/sPHENIX/preCDR/pro.1-beta.5/

 

Three set of data being generated/recently finished:

·       High stat. Hcal study should be done in

·       Finished embedding for gamma

·       Embedding for electron-ID is on-going

 

No more request for pre-CDR related productions.

 

 

Analysis software:

 

Pull request 69 from Jin: discussion on a few issue for track projection. Mike plan to put in a permanent fix. This pull request will NOT be merged.

Pull request 70 from Mike: evaluator accepts strict/non-strict mode. Will also add features to report # of errors.

Pull request 73 from Nils: Keep the CaloTower development branch up-to-date. Currently used in forward calorimeter work. Plan to have another focused discussion after pre-CDR before merging to main repository.

Pull request 74 from Alan: TPC updates. Verified by Mike.

Pull request 75 from Mike: handle primary and full particle list without assumption they share ids. Decide to merge. Jin will add documentation to PHG4Particle.  

 

Mike: is magnetic field info saved on DST?

Jin/Achim: no.

Mike/Achim: it would be very large to save the whole field map to DST T1 tree.

Mike/Jin: good to add the central field strength / sign to a simple compact node.

 

 

TODO after pre-CDR:

Jin: Tower -> truth association?

Chris: don't think PHENIX PISA save raw Geant3 hits in EMCal. Suggest tower -> primary particle contributor tracing.

Conclusion: Mike plan to save a list of primary particle -> energy contributor to the RawTower. Then evaluator judging on this list to decide whether go through PHG4Hit tracing.  

 

 

New Detector Display via ROOT/TGeo: Abhisek Sen

 

Page 3:

Need XML lib for the GDML exportation. Currently under private builds.

 

Jin: file size for xml?

Abhisek: about ~10MB

 

Jin: Does the geometry exportation also save tracks?

Abhisek: tracks not yet exported.

 

Mike: Does the display know the magnetic field and track bending.

Abhisek: not yet added magnetic field. But possible.  

Jin: One can add magnetic field via TEveMagField:

 

Jin: it will be nice to import events from evaluator to draw the reconstructed events. One easy, ROOT-only interface to fetch reconstructed object would be via evaluator NTuples.  

Mike: NTuples based code require heavy interface maintenance in the future. Suggest to modularize the interface function to localize future changes to the input file format.

 

Jin: Suggest borrow event display experience from FVTX (Geant/PISA -> TGeo -> Event display) experience. The latest version is authored by Aaron Key: 

 http://www.phenix.bnl.gov/viewvc/viewvc.cgi/phenix/offline/display/fvtx/fvtx_event_display.C

 

Jin: what is the numbers after the Geometry node name?

Abhisek: part of exportation. Need to find out.  

 

Abhisek: hint, EMCal daughter volume has too many fibers (20M), which may exhaust resource of host computer if enable to display fibers.

 

 

Plan:

Evolve this nice new development into event display for sPHENIX.

 

 

 

______________________________

 

Jin HUANG

 

Brookhaven National Laboratory

Physics Department, Bldg 510 C

Upton, NY 11973-5000

 

Office: 631-344-5898

Cell:   757-604-9946

______________________________

 

> -----Original Message-----

> From: Huang, Jin

> Sent: Monday, October 12, 2015 3:36 PM

> To: 'sphenix-l AT lists.bnl.gov' <sphenix-l AT lists.bnl.gov>; 'sphenix-software-

> l AT lists.bnl.gov' <sphenix-software-l AT lists.bnl.gov>

> Subject: [reminder] sPHENIX simulation meetings, Tue Oct 13th, 1-3PM @ 3-

> 192

>

> Dear All,

>

> This is a reminder for our weekly simulation meeting Tue Oct 13th, 1-3PM @

> 3-192

>

> Let's start the meeting with production and software updates, which will be

> followed by simulation studies by subsystems developers. More

> contributions welcomed.

>

> The agenda page is at

> https://indico.bnl.gov/conferenceDisplay.py?confId=1343   , for which the

> modification password is "1008"

> •         For residents at BNL, we will meet in the 3-192 conference room.

> •         To join the Meeting via bluejeans: https://bluejeans.com/468481110

> •         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: 468481110

>

> Cheers,

>

> Jin

>

>

> ______________________________

>

> Jin HUANG

>

> Brookhaven National Laboratory

> Physics Department, Bldg 510 C

> Upton, NY 11973-5000

>

> Office: 631-344-5898

> Cell:   757-604-9946

> ______________________________