Photonics Status Report from the Photonics Working Group

1
Photonics Status Reportfrom the Photonics
Working Group

• The Photonics Team
• Uppsala Collaboration Meeting
• 12 October, 2004

2
The Work Plan

• Testing the tables
• Define the test to be performed on the tables
  (Stephan, Johan)
• Perform the test only with interface (no AMASIM)
• Compare PSI (Thomas B.) with scatserver (Stephan
  H.)
• Compare PTD-like photonics bulk tab with PTD tab
  (Stephan, Johan)
• Using the tables
• Preliminary tables binning optimization - size
  (Kyle, Paolo)
• NN-fit of tables (Daan)
• Zenith-wise simulation (Ignacio, Claudia)
• Producing the tables (UW, London, Wuppertal)
• Simulating events with AMASIM
• check with simple configurations (Rodin,
  Ignacio)
• Defining a list of benchmark plots

3
Photonics in Brief

• The tables size is affected by photon tracking
  binning
• The number of tables is affected by photon
  source granularity
• Source types
• ems light distribution
• muon pure Cherenkov
• Table Types
• Time-integrated Amplitude Tables (.abs)
• Integral Probability Tables for time delay
  (.prob)
• Differential Probability Tables (.diff)
• Tables for AMANDA contain
• Optical ice properties
• OM/PMT efficiencies
• Ice described by the Millenium Ice Model (Kurt)

4
The Interface
Reading the tables

• PSI allowed the development of an efficient
  testing tool
• compared with scatserver (AMASIM)
• PSI in AMASIM

5
Testing the Tables

• Efficient testing tool developed by work in
  Stockholm-Uppsala
• tables visualization tool PACMAN (Photon
  simulAtion CoMparison And aNalysis )
• ems amplitude tables tested

6
Testing the Tables

• Efficient testing tool developed by work in
  Stockholm-Uppsala
• tables visualization tool PACMAN (Photon
  simulAtion CoMparison And aNalysis )
• ems amplitude tables tested
• Debugging the Tables
• ems shower ? distribution Jacobian factor sin?
  was missing

7
Testing the Tables

• Efficient testing tool developed by work in
  Stockholm-Uppsala
• tables visualization tool PACMAN (Photon
  simulAtion CoMparison And aNalysis )
• ems amplitude tables tested
• Debugging the Tables
• ems shower ? distribution Jacobian factor sin?
  was missing
• area projection weighting (when area smaller
  than OM cross section)
• Similar feature observed in the PTD Tables

ems shower distribution reproduced by Kahae and
Suiuj (Christchurch) Check correction and
reproduce tables Understand how this
feature affects our present analyses
8
Testing the Tables

• Efficient testing tool developed by work in
  Stockholm-Uppsala
• tables visualization tool PACMAN (Photon
  simulAtion CoMparison And aNalysis )
• ems amplitude tables tested
• Debugging the Tables
• ems shower ? distribution Jacobian factor sin?
  was missing
• area projection weighting (when area smaller
  than OM cross section)
• PTD Tables affected by ems shower distribution
• Tracking binning effect on precision

no scattering included scattering smears
distributions How much importance for physics
parameters ?
9
Using the Tables

• binning affects the tables size (precision)
• binning compromise
• higher memory
• zenith-wise event simulation (1/9 of tables
  loaded)
• NN fit (15 KB vs xx MB) - only 5 slower than
  full tables

• fit muon diff tables
• compare with photonics
• improvement of precision
• properly weight bin content

Absolute error
Time delay
10
Simulating Events

• AMASIM simulation Photonics tables setup
• AMASIM Magnesium is working and we can use TWR
  information
• Use simulation to check a set of observables
• Check speed and usage efficiency
• Compare new simulation with what we are used to
  see
• List of benckmark plots

11
Benchmark Plots

• Categories
• Energy Regimes
• WIMP's
• Atmospheric Events (? E-3.7)
• muons and neutrinos (?µ and ?e)
• High Energy Events (e.g. ? E-2)
• UHE and EHE Events
• Event Types
• Track-like Events
• Cascade-like Events
• Parameters (reconstruction model-indipendent)
• Standard Parameters
• basic parameters TDC, ADC
• Volume parameters Nch, projected_smoothness(x,y,
  z), Nhits, COGz
• Parameters (reconstruction model-dependent)
• Standard Parameters
• Smoothness
• Time parameters Time Residual, Ndir, Nearly,
  Nlate
• High Level Parameters

12

• Evaluation of p.e. from
• Phit affected by ice properties
• Implementation
• comparison between PTD
• and photonics

13
(No Transcript)
14
Action Items
yes ! This is how Kurt likes to call it

• testing the tables
• extend test to time tables (Johan new student
  in UW ?)
• extend test to infinite muon table (Johan new
  student in UW ?)
• different binning for muon and ems (fine mu
  tables but changing level2 integration binning)
  (UW)
• isotropic sources to derive back the scattering
  and abs lengths test algorithms (asked by Kurt)
  bins (UW)
• PTD tables (feature-corrected)
• needs check (Jennis student Kahae Han, Suiuj)
• correct ems bug on them (ask P. Doksus ?)
• regenerate new PTD tables (UW)
• do the tests with PTD tables (PSI, UW)
• NN tables
• do the timing tests infinite muon (PSI, Daan)
• check if NN fit needs finer or coarser binning
  differential tables (Daan)
• generation of AMASIM events (Rodin, Ignacio, )
• simulation tree update after new releases should
  be faster (simuperl) (Paolos substitute ?)
• PSI in AMASIM (fortran interface and use it in
  AMASIM, StephanThomas)
• one time-binning (integrate photoint)

15
Action Items
yes ! This is how Kurt likes to call it

• source code intervention
• statistical info in photo2amasim (Jenni)
• decouple amplitude and time tables in loading
• unload function needed to dynamically free the
  tables from memory
• check the effect of 10 change in ice properties
  correlated with NN fits (requested by Kurt)
• reading differential tables get only photons
  within a short time delay range (req by Rodin)
• more people (AMANDA/IceCube) will be interested
  in photonics tables soon or later
• soon is better !
• we really encourage people (PIs) to invite
  people to help

16
Conclusions

• the tables test started and is very successful
• Stockholm/Uppsala collaboration extremely
  productive
• powerful tables visualization tool (PSI Root
  Macros)
• useful debugging power (tables understanding)
• need to extend test to muon tables and time
  tables
• what is the optimal size ? Compromise
• speed issues
• NN fit improved Now simulation needed to check
  !
• PSI interface in AMASIM will allow use of
  photonics/PTD/NN
• AMASIM simulated events on the way (also with
  TWR info)
• need of Collaboration-wide effort to check
  simulation data

17
Web Pages

• Photonics portal
• http//amanda.berkeley.edu/kurt/photonics/
• Photonics production
• http//amanda.wisc.edu/simulation/photoproduction
  /tables/
• PACMAN page
• http//www.physto.se/johan/PACMAN/index.html
• PSI page
• http//www.physto.se/burgess/icecube/PSInterface
  /
• AMASIM Magnesium Page
• http//www.physto.se/amanda/amasim/magnesium/mag
  nesium.html

18

• testing tools
• ems amplitude tests from Johan PACMAN Photon
  simulAtion Comparison And aNalysis (web page)
• extend test to time tables (Johan new student
  in UW ?)
• extend test to infinite muon table (Johan new
  student in UW ?)
• different binning for muon and ems (fine mu
  tables but changing level2 integration binning)
  (UW)
• isotropic sources to derive back the scattering
  and abs lengths test algorithms (asked by Kurt)
  bins ? (UW)
• PTD tables (bug-corrected) needs check (Jenni
  student Kahae Han, Suiuj)
• correct ems bug on them (ask P. Doksus ?)
• regenerate new PTD tables
• do the tests with PTD tables (PSI, UW)
• NN tables
• do the timing tests infinite muon (PSI, Daan)
• generation of AMASIM events (Rodin, Ignacio ?)
• simulation tree update after new releases should
  be faster (simuperl) (Paolos substitute ?)
• PSI in AMASIM (fortran interface and use it in
  AMASIM, StephanThomas)
• zenith-wise simulation (unload function needed)
  (photonics source code, PTD P. Doksus)
• one time-binning (integrate photoint)
• decouple amplitudes and time tables (change
  photonics loading code)
• statistical error info in photo2amasim (Jenni)

19
List of Benchmark Plots

• The list of benchmark plot is intended to provide
  a guide for the process of acquire that
  confidence on the new simulation tools. The
  IceCube experiment will benefit of all this
  effort.
• Categories
• Energy Regimes
• WIMP's
• Atmospheric Events (? E-3.7)
• muons and neutrinos (?µ and ?e)
• High Energy Events (e.g. ? E-2)
• UHE and EHE Events
• Event Types
• Track-like Events
• Cascade-like Events
• Parameters (model-indipendent)
• Standard Parameters
• basic parameters TDC, ADC
• Volume parameters Nch, Nch(depth,x,y,z),
  projected_smoothness(x,y,z), Nhits, COGz,
  Smoothness (likelihood)
• Time parameters Time Residual, Ndir, Nearly,
  Nlate (likelihood-dependent)
• High Level Parameters

20

• PSI Thomas summary page. Because of PSI the test
  tools could be quickly developed
• Debugging Johan figure for shower and
  weighting-area bugs (before and after)
• Isotropic distribution (as another effect of
  bugs)
• The bug affect PTD (plot) analyses affected
• Stephans 2 plots (no scattering) binning
• Table size (bin compromize)
• NN-fit (13KB vs MB tavles) speed issue NN only
  5 slower ! plot
• Zenith-wise simulation
• Peter Steffen talk
• Tests collaboration success
• Visualization found bugs (shower) and other bugs.
  Ped was informed. Johan made
• Future
• see previous slide