CMS / ILC PFA contact meeting

10377183 (Vidyo)



Vincent Boudry (LLR – École polytechnique, CNRS/IN2P3)
Gettogether meeting: Quick presentation of actors & work.
Notes taken By Roberto Salerno
Edited by Vincent Boudry

Benevoly organised this "first"-contact meeting to give an overview for everyone.
To be organised and managed (volonteers?).
Some part of AIDA"-1" (≠ AIDA2) works involves running developped SW on CMS upgrade.

Pedro + Valeri: 
They are providing the inputs for PandoraPFA algos (e.g. the full chain Sim > Digi > RecHit)
They are facing some technical problems mainly related to the adaptation of Run1 geometry in CMSSW (Lindsey can help on this part)
*) During the week they could provide the full chain up to the RecHit to be run in Pandora 
*) Once we have the RecHits, estimation to have another week to implement all the thing in Pandora.

slide 1 names of people that made most the work up to now
Q: Do you plan to have full PU event simulation or do you plan to have different paths?
No idea, but we should be able to run it following the official paths 
Feeding Pandora with the new detector should run out of the box (the code run on CLIC events) confident that they will be no problems 
Q: Did you already start to modify the model and play with the parameters (cell size/wafers size/etc.)?
The optimisation is done with the standalone simulation, now we have somehow a concept that is more or less working 

Integration of Arbor inside Pandora
1. Work between Manqi + Rémi Été in Lyon started but maybe 1 or 2 weeks to have the work finished
2. Work to implement Arbor in the current CMSSW framework on going. 
Remark of slide 14: Comparison of ARBOR and Pandora shown to be similar but (remark from Hieu) this is due to diffrent digitisation. 
Q: Performance of the Arbor algo in presence of PU?
We should need an interaction with Remi that is making the work

General discussion:
Q: Pandora is making matching tracks and Clusters? Is there a re-cluster in Pandora?
Pandora re-runs various algo and re-does the clustering taking as input the tracker and first cluster
Finally we want to the take the Pandora stuff and implement in CMSSW.

Q: What should be done next? Technical discussion with the people involved here?
Help to put people together… things that can be shared in the both directions
AIDA2… proposal is going to be written (first draft end of June) and submitted early September

First milestone performance potential for the TP to be submitted in the second week of september 
Beginning of august CWR 
Draft to start the review process mid-june
For the TP most of the material from the standalone but we need the full simulation for the end of the year. Collaboration with ILC is critical 
End of this year or beginning of next year more answers about the physics performance 

Lindsey: Tuning the PF takes 1-2 years.  First we need something fully integrate in CMSSW 

CMS should have discussion on how we can continue to benefit form the support from Mark and the ILC folks Arbor in CMSSW and Pandora.  Possible return might be the integration of CMS code on tracking into PFA SW. 

AIDA to be discuss in the next meeting.

There are minutes attached to this event. Show them.
    • 2:00 PM 2:05 PM
      CMS PFlow Present Design 5m 10377183



      Speaker: Lindsey Gray (Fermi National Accelerator Lab. (US))
    • 2:05 PM 2:15 PM
      PandoraPFA in HGCAL 10m 10377183



      Speaker: Niki Saoulidou (University of Athens (GR))
    • 2:15 PM 2:25 PM
      ARBOR 10m
      Speaker: Mr Manqi RUAN (IHEP/LLR)