The STACK-SLIDE Search Initial Report: PULG F2F UWM June 2003 - PowerPoint PPT Presentation

1 / 14
About This Presentation
Title:

The STACK-SLIDE Search Initial Report: PULG F2F UWM June 2003

Description:

Title: The R-modes: New Theoretical Results Author: Gregory Mendell Last modified by: Gregory Mendell Created Date: 3/9/2001 12:39:06 AM Document presentation format – PowerPoint PPT presentation

Number of Views:57
Avg rating:3.0/5.0
Slides: 15
Provided by: Gregory245
Category:

less

Transcript and Presenter's Notes

Title: The STACK-SLIDE Search Initial Report: PULG F2F UWM June 2003


1
The STACK-SLIDE SearchInitial Report PULG F2F
UWM June 2003
  • Gregory Mendell, Mike Landry
  • LIGO Hanford Observatory

2
Initial STACK-SLIDE Proposal
  • Mendell/Landry at LHO to code stack slide
    algorithm based on Brady/Creighton gr-qc/9812014
  • We are in the very preliminary stages of this
    project, trying to answer questions like
  • Is anyone else planning on working on this
    (emailed Creighton/Brady/Riles/Chin)?
  • Is there any known problem with stack slide? Is
    Hough all we need?
  • Do we need to modify B/C algorithm? Stack slide
    SFTs, DeFTs, or F-stat?
  • Algorithm will be coded under LAL, and search run
    under LDAS
  • Driver code will be available to all (detailed
    understanding of LDAS not required)
  • Parallel search exploit THz LDAS computing
    power
  • Efficient I/O of SFTs, results database (or frame
    output) inherent in LDAS
  • We will have a more complete proposal/progress
    report at the June F2F
  • Code written and working by the August LSC meeting

3
Definitions
  • T0 observation time.
  • M number of data segments (number of stacks).
  • N number of data points in one segment.
  • NT total number of data points in T0.
  • A signal amplitude.
  • ? square root variance of the noise.

4
Coherent Search
For A2NM gt 4?2
Detection if
False Alarm Rate
5
Coherent Power vs. Freq
6
STACK-SLIDE Search
One stack
For A2N lt ?2
(Ave M stacks)
Detection if
False Alarm Rate
(Brady/Creighton gr-qc/9812014)
? Incomplete Gamma Function
7
STACK-SLIDE Power vs. Freq
8
Stacking and Sliding
9
Complexity?
Complexity STACK-SLIDE O(MNlog2N) O( NM2)?
Complexity SFT Coherent Per Sky Position O(
MNlog2N) O(?kNM2)?
Complexity Coherent FFT Resampled Time Series Per
Sky Position O( MNlog2MN)?
10
We are still working to
  • Understand the Algorithm.
  • Understand the Computational Complexity.
  • Understand the Statistics.

11
Brady/Creighton algorithm and potential
modifications
Resample? Or use SFTs, DeFTs, F-stat From LALDemod
  • The algorithm shown in the Flowchart
  • is iterated in a hierarchical approach

12
Will Provide LDAS, LAL, Driver Code
  • LDAS code will be based on knownpulsardemod DSO.
  • LAL parameter space metric code exists.
  • Need LAL STACK-SLIDE function.
  • User-friendly driver code that anyone can run.

13
LDAS Code To Write
14
Example Driver getFstat
Usage ./getFstat_v1.tclsh ltsitegt ltifogt
ltstart_timegt ltdurationgt lttemplateFreq0gt
ltband_widthgt ltsft_timegt ltframecachefilegt
lttemplatefilegt ltoutputdatabasegt ltoutfilegt
ltrungt Example ./getFstat_v1.tclsh lho H1
729976096 8192 1283.5 2 2048 SFTFrameCache_H1_0402
03.txt testJ1939.params ldas_tst testFstat.xml
run Starting Thu Jun 12 125154 PDT 2003 LDAS
job with LJrun at lho User gmendell Job id
LDAS-WA890236. LDAS job succeeded. Total time to
run job 76 seconds. gmendell_at_vulcan scripts
less testFstat.xml ... ltLIGO_LW
Name"dataContainer(Vect)Frame"gt ltDim
Scale"1.2207031250000000e-04" Start"1.2825000000
000000e03" Unit"hz"gt16388lt/Dimgt Name"data"
Type"Local"gt 2.6003204903121175e00,2.38348086002
41506e00,2.7261332694330522e00,...
Write a Comment
User Comments (0)
About PowerShow.com