Kemp et al. describe a trial the use of commercial high performance computing (HPC) for a large scale processing task in astronomy — processing data from the GASKAP-HI pilot surveys. This was done in part to provide reference images for an upgrade to ASKAPSoft (the ASKAP data processing software), and to provide science images for the GASKAP collaboration, using the joint deconvolution capability of WSClean.

The standard approach to imaging a single ASKAP PAF footprint (of 36 beams) is a linear mosaic, wherein distinct deconvolved images, (i.e., with the inherent point-spread-function response from the sky-brightness distribution removed), produced from each individual beam, are combined in a pointed mosaic. However, for GASKAP-HI observations of fields with extended diffuse emission, a joint deconvolution approach is necessary – in which the images from each beam are stitched together in a single image before moving on to deconvolution. However, this imaging approach is computationally intensive, and Pawsey’s first supercomputer did not possess sufficient node memory, hampering efforts to develop this imaging mode in ASKAPSoft. Fortunately, the increased capabilities of the new Setonix supercomputer have enabled the development of joint deconvolution in ASKAPSoft.

The team found the key advantage of using commercial HPC was the immediate access and high availability, with the main disadvantage being the need for improved HPC knowledge to take best advantage of the facility. The image above shows the the fields in the GASKAP-HI Pilot Surveys. Cyan boxes denote the approximate ASKAP PAF footprint for the pilot phase I fields (20 hr integrations); yellow boxes are those for the pilot phase II fields (10 hr integrations).