DCR

1
DCR

• Plan of presentation
• Soumya Mohanty Overview, aims work done
• R. Balasubramanian Details of Hardware, Database
• Using the database
• R. Balasubramanian
• Soma Mukherjee
• S. Babak

2
What is DCR?
3
Using a database Data Mining Data Exploration

• Different but complementary approaches.
• Data exploration
• I want to see the time series corresponding to a
  bunch of triggers that I selected from a
  database. (Then do more analysis on this selected
  data.)
• Typically, Follow up data is short, Quick look
  environment needed, no specific queries
• Data Mining
• Can the transients seen over a month be
  classified into groups? What was the rate of
  transients in each group as a function of time
  (Maybe some types occur in the day, some occur in
  the night). (Then use this information to
  quantify the quality of long data stretches).
• Purely database based Re-analysis of raw data
  may be impractical

4
What is special about DCR?

• Any such transformation will introduce errors
• Spurious information
• Misses genuine stuff

Information Transformer
Raw noisy data
5
Control on False Alarm Rate

• Important for Data mining
• Statistical analysis done on database itself
  since reanalysis of long stretch of data
  expensive
• Need to put error bars
• Not so important for Data exploration
• Looking for information about specific events
• Each explorer will work with his/her own short
  data stretch

6
Initial Design of DCR
7
rDCR project Beginnings

• Group formed Sep01
• AEI Soumya Mohanty (lead), Soma Mukherjee
• Cardiff R.Balasubramanian, S.Babak, D.Churches,
  B.S.Sathyaprakash
• Birmingham A.Vecchio, R.Ingley (new)
• Decided to include Kalman Filter also
• Made list of software components GEO and
  support
• Much thinking on the Design of GEO
• Coding started Oct01

8
Aspects of rDCR project, Status

• Building GEO
• Integration of GEO, Frames, MPI, MySQL
• Integration of entire chain on Hannover Beowulf
• Implementing DCR algorithms
• MBLT existed as a mixture of C and matlab codes
• PSDCD exists as a C (DMT ready) code
• Simple Non-stationarity detector
• Database Analysis
• Administrative CVS, Web Site, Coding conventions
  etc.

9
GEO

• Major component of all the work done so far
• All GEO codes planned have been completed and
  tested

10
Frame, MPI integration

• Reading Frames adopted a simple model for data
  requests
• Status Done
• Purpose of MPI code to encapsulate a GEO based
  code and run it on a Beowulf cluster
• One slave one channel dataflow model
  implemented
• Status Done

11
Integration with MySQL

• Creating an elegant and easy to use OO design
  took most of the time
• Writing to database done

12
Run entire chain on the Hannover Beowulf

• Making Beowulf ready, installing software took
  about a month
• Done

Run Simple monitor on all channels
Database
Get Data
Read Frames
13
rDCR Algorithms Implementation

• MBLT code in C written
• Completion of MBLT was dependent on a number of
  lower level classes
• Filter, Filter Design, Demod, Modulate, Running
  Median.
• Code now complete and under testing
• Did not get a significant increase in performance
  over Matlab code
• Mainly because the most time consuming part is
  coded in C and the same is used here.
• This is quite preliminary

14
Database Analysis

• A simple monitor was built to test IO and as many
  classes as possible while the DCR algorithms were
  being coded
• Main aim start exercising the post-database
  analysis chain (software, logistics)
• Software Matlab, Triana
• Logistics Need to transfer tables to local
  machines

15
DCR on the Web

• http//www.aei.mpg.de/mohanty/DCR/DCRindex.html

16
Current Status Plans

• All software originally planned in place
• MBLT code ready testing
• Database analysis exercises done
• Plans
• Construct a simple non-stationarity test
• Complete rDCR
• Complete Kalman Filter
• Run on all of coincidence run data
• Extend GEO
• Analyse rDCR database