Physics in CMS

1
Physics in CMS
Jim Branson DOE/NSF Review November 27, 2001

• Physics Reconstruction and Simulation
  Organization
• Current PRS Activities
• HLT results
• Interaction with Core Software and Computing

2
A Great Challenge

• Millions of channels of fast, rad-hard detectors.
• Electronics to cope with 40 MHz rate.
• Software to process 100 kHz of highly complex
  events for HLT.
• Physics selection of 100 best events per second.
• Computing Facilities and Analysis Tools to use
  multi-Petabyte data base.
• Instant Results in first month of running.
• I f HLT software (same as offline) is not ready
  we will write much more to mass store.

100 kHz
3
Luminosity for a 5s Higgs Signal

• 5 sigma signal possible in one year of running
  from 100 GeV to 600 GeV.
• From 150 to 450, 5 sigma effect could come in a
  few months.
• WW and ZZ() using leptons
• Hardware trigger and software must be ready at
  the beginning of running.
• CMS has (resolution) advantage for most difficult
  region.

4
SUSY in One Week???

• High energy can pay off immediately.
• Squark gluino search starts the first day of
  running.
• Missing energy plus jets.
• Not the CMS strong suit.
• Need tracker to measure jets and missing energy
  well.

5
PRS Organization

• Physics Reconstruction Selection (PRS) groups
• Groups were put in place in April 99, formalized
  within each detector group with the start of CPT,
  end of 2000.
• Today, since Jan 2001, four groups
• ECAL-Electron/Photon C. Seez
• Muons U. Gasparini/D. Acosta
• HCAL-Jet/Missing ET S. Eno/S. Kunori
• TRACKER-b/t (vertex) M. Mannelli/L. Silvestris
• Charge create CMS detector, reconstruction and
  simulation software evaluate full chain (from
  Lvl-1 to offline) of physics selection for CMS.
  First priority the HLT
• study, design implement algorithms and code to
  provide the rejection/efficiency needed for CMS
  physics plan
• Primary Physics activities OO software to study
  HLT.
• Complementary to ATLAS
• Physics TDR later with final software.

6
US in CMS
USA
7
US CMS Physics Coordination

• Start effort in US.
• Time zone difference.
• Develop our facilities.
• More available manpower.
• All physics groups have made their meetings
  available to us.
• US participation is low.
• We are well integrated into CMS-wide PRS
  activities.
• Monthly(?) US meetings to coordinate our
  activities.
• Input to Software and Computing project.
• US is active but has not reached its potential in
  PRS.
• Strong in muon group.
• HCAL needs US based effort increase.
• e/g and tracker interest and smaller effort.
• US PRS users have been very satisfied with
  support from Software and Computing project.

8
PRS Work

• Have been working with OO software (C
  reconstruction code ORCA)
• Stick with OO code even if this would mean delays
• CMS decision was to postpone Physics TDR until
  software ready
• Original plan start with proof-of-principle for
  high-L physics
• Highest priority is the discovery physics
  channels
• Need to study staging scenarios forces study of 2
  X 1033 first
• Major dates/milestones
• Level-2 work complete last year
• This year first Level-3 (Level-2tracking)
  results
• End of year first breakdown of rate to storage
  (the O(102)Hz)
• DAQ TDR Dec 2002 (HLT)
• Physics TDR Dec 2004

9
ORCA/HLT Production

• HLT studies use full simulation and
  reconstruction of high luminosity LHC events.
• 17.3 minimum bias events per crossing (in time
  pileup).
• 9 crossings contribute to digitization (out of
  time pileup).
• Event sample sizes 1 million background
  signal events.
• Massive data moving exercise 150 minimum bias
  events randomly chosen from a sample of 100k and
  with a size of 35MB for every signal event (1MB).
• One CMS event requires of order 100-400 times
  more computing to digitize than one RUN II
  event.
• Produced about 1 million for JetMET and for Muon.
• We bias background generation toward high Pt,
  improving effectiveness.
• Production has been difficult this year.
• Hardware and software problems at CERN and
  elsewhere.
• Worldwide organized production now working well.

25 nsec
25 nsec
25 nsec
25 nsec
25 nsec
25 nsec
10
The Twofold Way
11
OSCAR CMSIM Comparison
Very preliminary!!
DDD Project ready in spring 02.
12
e/g Level 2.5
13
e/g Triggers
14
Electron Tracking Efficiency with E/p cut
15
Electron Position Matching in h
16
Code Timing on GHz PIII
17
Pixel Isolation (not yet included)
18
Electron Rates and Efficiency
19
Tuning Global Muon Trigger
20
Muon plus Calorimeter Rates
21
(No Transcript)
22
(No Transcript)
23
(No Transcript)
24
(No Transcript)
25
Physics Content of Muon Triggers
PTgt20
26
Track Reconstruction
27
Track Reconstruction Global Efficiency
28
Track Parameter Resolution
29
Track Reconstruction Efficiency in b Jets
30
Jet Correction Calorimeter Only
31
Example of a 45 GeV ET Jet

• Strong CMS B-field bends charged particles out of
  jet.
• Separate clusters are visible in calorimeter.
• Can use tracks to improve jet measurement.

32
Jet Improvements Using Tracks
Improved Resolution
Improved raw Energy Scale
33
Improvements for 100 GeV Jets
34
(No Transcript)
35
(No Transcript)
36
High Level Trigger summary

• This year
• Consolidation and improvements in Lvl-2
• Pixels are now officialy Lvl-2.5 looking at
  possibility of applying pixel requirements before
  anything else
• First results from full track matching for
  electrons, muons and b/t
• Muons very good, now a question of b vs W/Z
• Electrons tracker material next limitation need
  to tune track finding for electrons no results
  on photons yet
• b/t tagging first results ts in excellent
  shape
• Ongoing work understand/improve on jet and
  missing ET
• Starting to look at energy flow and possible
  tracker help
• Goal a first breakdown of the output to storage
  at 2x1033 by Dec 01
• Next big milestone DAQ TDR for end 2002
• After that Physics TDR at end 2004

37
Grand Summary

• CMS Physics activities now focus on HLT studies
  with prototype OO software.
• US is well integrated into PRS work but the level
  of effort must grow.
• Big contrast to Europe where there are few major
  experiments running.
• Simulation and Reconstruction software is in good
  shape.
• Best Ive ever seen for 5 (or even 2) years in
  advance of an experiment.
• Analysis tools also available.
• HLT results validate concept of triggering in
  large farm of computers.
• Core Software and Computing has served us well.
• Good start, but a lot of changes coming now.