Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.786065
Title: New techniques in wide-field interferometric imaging
Author: Mort, Benjamin James
Awarding Body: University of Oxford
Current Institution: University of Oxford
Date of Award: 2011
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Abstract:
With each new telescope that comes online, modern astronomy is ever moving towards a point where processing the resultant data deluge is arguably as much of a challenge as interpreting the science for which the instruments are designed. To this aim I have investigated the application of new processing techniques to two common problems in computationally intensive astronomical data reduction. I present a comparative study of the CLEAN deconvolution algorithm applied to the spectral analysis of jet speed data from the microquasar SS 433 recorded over the last 26 years to investigate a claim made in the recent literature that there was a phase shift in one of the key periodicities observed. This investigation demonstrates how the CLEAN algorithm can be used to great effect for the spectral analysis of unevenly sampled time-series data, revealing characteristics that would otherwise be unattainable. While using this technique it was possible to rule out any strong evidence for a steady systematic phase variation in the jet speed periodicity, due to the well known relationship between frequency instability and phase variation. It was not possible with the current data set to definitively exclude the possibility that a variable frequency dependence of the jet speed on orbital period is at play in this microquasar. Modern interferometric telescopes mandate innovative use of computers for data processing because (i) huge data rates and sensitive detectors require complex algorithms to optimally extract the valuable science results and (ii) the rising costs of electricity and the environmental impact of excessive power consumption means that simply allocating more and more computing resources to solve the problem is impractical. To this context I describe my implementation of the w-projection algorithm for imaging wide-field interferometric data on graphics processing units (GPU). In this study I identify the advantages and pitfalls of using GPUs for the processing of astronomical data and demonstrate that GPUs present a fast, cost effective and energy efficient solution to imaging for the next generation of radio interferometers where the value of FLOPS per watt is increasingly an important factor in developing new processing solutions. While investigating the application of imaging on the GPU, I have developed a modular, flexible testbed application for developing, studying and testing the deployment of GPU algorithms for interferometric imaging can I present this code.
Supervisor: Blundell, Katherine Sponsor: PPARC
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.786065  DOI: Not available
Share: