110 likes | 123 Vues
Photon(s)+X Group: HERWIG vs. MadGraph. Osamu Jinnouchi (Tokyo Tech), Fernando Monticelli (La Plate), Bruce Schumm (SCIPP), Harnan Wahlberg (La Plata) Incidental slides for the discussion of whether the Photon+X group should be asked to migrate to MadGraph for 2016 analyses.
E N D
Photon(s)+X Group: HERWIG vs. MadGraph Osamu Jinnouchi (Tokyo Tech), Fernando Monticelli (La Plate), Bruce Schumm (SCIPP), Harnan Wahlberg (La Plata) Incidental slides for the discussion of whether the Photon+X group should be asked to migrate to MadGraph for 2016 analyses
Unequivocal: the Photon+X group is committed to moving to MadGraphas long as it remains the recommendation of the SUSY group. • Osamu (supported by Fernando) has put significant time into the transition for diphotons; will continue to do so • La Plata has committed a student to the transition for models associated with photon+jets, photon+lepton However, the Photon+X group would prefer to use HERWIG for 2016 analyses (Diphoton strong and EW production; Photon+jets strong production). Why?
Photons+X: MadGraph vs. HERWIG • 13.3 fb-1 result (ICHEP) done w/ strong-production HERWIG signal model • Existing HERWIG grid only needs small extension for 40 fb-1 • Diphoton analysis has shown adequate understanding of HERWIGMadGraph for strong production gluino-bino model • But needs to be confirmed with generator-level study for gluino-bino/higgsino model Doing study will cause some delay; regeneration of MC “unnecessary”
Diphotons: MadGraph vs. HERWIG • 3.2 fb-1 result made use of HERWIG for gluino/bino model • When setting up analysis, we designed the HERWIG grid for ~100 fb-1; was a “feature” of our generation request • Subsequently we have achieved adequate understanding of gluino/bino model in MadGraph (Osamu) and submitted focus points • But no physics need to switch from HERWIG for strong prod. • EW model (Wino/bino) not understood; have worked on it for several months (see below); experts are not yet certain why MadGraph is not working (Emma Kuwertz)
Details on Diphoton wino/bino EW model problem (Slides from Osamu)
Conference Note public in early January 2014: Search for Supersymmetry in Diphoton Events with Large Missing Transverse Momentum in 8 TeV pp Collision Data with the ATLAS Detector ATLAS-CONF-2014-001 Final result in preparation
MadGraph Wino-bino sample (mw = 1000; mB = 100) Next slide: Makes no difference what wino state is being produced (chargino, neutralino):
A little more illumination: First three processes are 0-jet; second three are 1-jet; third three are 2-jet Shown to experts (Emma); not clear why this is happening. Doesn’t appear to be our implementation.
Summary of Options • Use HERWIG: gluino/bino grid done; straightforward to extend gluino/higgsino-bino grid and generate wino/bino grid (our request) • Use Madgraph: Regenerate “perfectly fine” gluino-bino grid and most of gluino/higgsino-bino grid; some delay for gluino/higgsino-bino grid, indefinite time delay for wino/bino grid • Use MadGraph, but go against recommendation and use only 1-jet for wino/bino grid MAY save time if 1-jet deemed to be adequate (what is standard? If HERWIG then odd to move away from it). Might that defeat any purpose of moving to MadGraph anyway?