DataIntensive Processing for a Pulsar Virtual Observatory - PowerPoint PPT Presentation

1 / 11
About This Presentation
Title:

DataIntensive Processing for a Pulsar Virtual Observatory

Description:

Advanced Network technology. Mid range: NW-GRID and local clusters ... Jodrell Bank Radio Observatory searches for fast pulsars in binary systems. ... – PowerPoint PPT presentation

Number of Views:19
Avg rating:3.0/5.0
Slides: 12
Provided by: Caro110
Category:

less

Transcript and Presenter's Notes

Title: DataIntensive Processing for a Pulsar Virtual Observatory


1
  • Data-Intensive Processing for a Pulsar Virtual
    Observatory
  • June 17 2008 Liverpool
  • Astronet Infrastructure Roadmap Symposium
  • John Brooke - University of Manchester

http//www.esnw.ac.uk
2
E-Science North-West (ESNW) research
Bio-Medical
Physics and Astronomy
Social science
Chemistry
Semantic and Knowledge Technologies
Database Technologies
Generic Grid and e-Science technologies
3
North-West Gridadvancing Grid Technologies and
Applications
http//www.nw-grid.ac.uk
Hooks to other Grid consortia NGS, WRG
Acknowledgement Rob Allan STFC Daresbury and
NW-GRID consortium.
Top end HPCx
Applications and industry
Pharma, meds, bio, social, env, astro, CCPs
Portals, client toolkits, active overlays
Mid range NW-GRID and local clusters
User interfaces
Sensor networks and experimental facilities
Advanced Network technology
Desktop pools Condor etc.
4
The Jodrell Bank Pulsar Search Code
  • Jodrell Bank Radio Observatory searches for fast
    pulsars in binary systems.
  • In order to obtain a high signal-to-noise ratio
    they observe across a wide band of radio
    frequencies and integrate.
  • However the interstellar medium is dispersive
    (the speed of the radio waves depends on
    frequency)
  • The parameter space is huge, interstellar
    dispersion, orbital parameters, acceleration are
    unknown.
  • In searching for candidates we have a very
    data-intensive processing task and then the use
    of intuition in selecting candidates for further
    observation.

5
Dispersion across RF band
Diagram supplied by Michael Kramer Jodrell Bank
Observatory
6
Pulsar Virtual Observatory process
7
Scheduling pipeline
Data
Master
Slaves
8
Architecture for distributed data processing
9
Current status
  • Rewrite of pulsar detection tools in Michael
    Keiths PhD thesis. Re-analysis of the Parkes
    Multi-beam Survey led to the discovery of 30 new
    pulsars and re-analysis of the Perseus arm survey
    produced 10 more.
  • An architecture for distributing the searching
    algorithms over clusters and computational grids
    has been developed in the Grid1D project. Details
    at
  • http//www.rcs.manchester.ac.uk/research/projects
    /pulsargrid/
  • An extensive search for binary pulsars has been
    carried out on NorthGrid. The results are
    currently being analyzed.

10
SKA pulsar survey
expect 20000 psrs incl. 1000 MSPs and
BH-pulsar binaries
Thanks to Michael Kramer JBO
11
Pulsar Virtual Observatory
  • At present it still requires expertise in pulsar
    astronomy to use the searching methods.
  • The ultimate aim is to automate this process as
    much as possible since pulsar searching will
    become harder to find fainter pulsars in regions
    of high dispersion measure.
  • Pattern recognition techniques may be able to be
    utilised in the candidate identification stage.
  • Pulsar searching can be linked with observations
    at other wavelengths, e.g. X-ray.
Write a Comment
User Comments (0)
About PowerShow.com