Stefan Westerlund

Thesis

A Parallel Source Finder For Searching Radio Astronomy Images using High Performance Computing

Summary

Source Finding is the practice of searching astronomy data for the sources of signal in that data. This project will create a source finder capable of searching a full resolution Australia SKA Pathfinder (ASKAP) data cube, approximately 4 TB in size, in significantly less time than the observation and processing time required to make that data cube, using the computing resources available. Existing source finding efforts for surveys involve manually confirming the output produced by source finder programs. The Wallaby survey, using ASKAP, is expected to find around half a million galaxies, up from the 4,315 found in the HIPASS survey, which used the Parkes telescope. It is extremely undesirable to spend the time and money required to manually confirm each galaxy that is output by a source finder on this scale of data. Therefore the source finder will need to be able to produce an output catalogue of sufficiently good quality that manual confirmation of this output is unnecessary. As such, this source finder will require improved accuracy, compared to previous efforts, as well as a high performance implementation suitable for processing the large quantities of data produced by ASKAP.

Why my research is important

This research is important because new telescopes, such as ASKAP, will produce much more data than previous installations. Current source finding techniques do not scale well to this amount of data. Therefore, new techniques are needed to efficiently and accurately search the data produced by these new instruments.

Funding

• Australia Postgraduate Award