L2Cal Upgrade

1
L2Cal Upgrade

• L2cal team
• 041807

2
Refresher on existing L2 clustering

• Hardware based
• Pacman looks for seeds then clusters all
  contiguous towers that satisfy a shoulder
  threshold
• Problems-Jets high rate due to jets being
  artificially energetic at high lumi, position of
  seed is used as position of jet
• -Photons/electrons- 2 high Et passes

3
Current clustering algorithms

• Pass0 -Look for seed tower Em gt 2 GeV and
  Had/Em lt 0.125-One tower per cluster
• Pass1-Seed has Em gt 8GeV-Shoulder Em gt7.5
  GeV ( no limit on number of towers , marked
  towers as used and dont use again )
• Pass2 jets
• Pass4 (alpha clustering )-Use Pass1 seed
  position then cluster using pass0 tower info
  (towers have Had/Em cut )-In central ( clusters
  do not open in phi eta-1 wrt to seed)-Plug (
  3x3 box )-No crossing eta zero boundary etc

4
Current isolation

• Calculated for Pass1, Pass4
• Start with seed position-find iso for 3X3 and
  (4) 3x4 tower configurations-Use smallest of
  the 5 sums

5
Status of algorithms ( upgrade )

• Pass0 is unchanged
• Pass 1 is no more
• Pass4 ( alpha ) is unchanged
• Isolation is unchanged (Pass 4)
• Jets will be new (pacman?cones)
• Missing Et will be new
• If there are other improvements now is a good
  time to let us know

6
Offline tools

• Have L2Cal simulation ( TrigSim ) which includes
  Pass0,Jets,Pass4, Iso, MET
• Also have Stntuple version of code which will be
  available within HTTF tools-can also make
  available independent of Higgs tools-If your
  ntuples have DCAS info you can run new
  clustering
• Tools already available for testing/optimizing
  new triggers

7
What is the status of the upgrade

• Have all hardware in hand
• About half of the hardware is already installed
• Software (as described) is written- beginning to
  run it parasitically
• So things are starting to come together

8
What happens when pass 1 is used in L2 code

• Rewritten level-2 code to use alpha clustering if
  passed 1 or 4, 4 is used in the code.