About NCSA    Projects    User Info    News     

University of Illinois
at Urbana-Champaign


HPC in radio astronomy


The Radio Astronomy Imaging group are involved in several projects involving the use of high-performance computing (HPC) in radio interferometry. This builds on the specific synergies possible at NCSA and UIUC in this area, as well as the consortium collaborations that we are engaged in.

Petascale imaging and calibration in radio interferometry: This project concerns the domain algorithmic challenges and computational science and engineering advances needed to scale radio-interferometric image formation and calibration imaging algorithms to the petascale regime. This work is closely related to our role in the TDP Calibration and Processing Group.

HPC frameworks for radio interferometry reduction codes: This project concerns the re-factoring of large domain community data analysis codes for scalable deployment on HPC architectures. This approach is described in a recent paper by Kemball, Crutcher, and Hasan, 2007, Sofware Engineering & Practice, in press. This effort uses the tools provided by the Common Component Architecture HPC project, including the language interoperability tool, Babel.

HPC processing workflows for radio interferometry data: As described above, community access to contemporary radio interferometers require aggressive development of automated workflows (often described as pipelines within radio astronomy) to provide reference reduction of radio astronomical data. Current efforts are focused on developing pipeline data reduction capabilities for the CARMA telescope.

HPC computational statistics in radio interferometry: Modern developments in computational statistics, together with advances in available HPC computing resources, allow new approaches to important unsolved problems in this radio interferometry, including imaging fidelity assessment. This project is a collaboration with faculty in the UIUC Department of Statistics. Recent results are published by Kemball & Martinsek (2005).

Innovative hardware architectures for HPC computing: This work involves research into hardware acceleration for HPC computing in radio astronomy, and includes collaboration with faculty at UIUC in Engineering and Astronomy, as well as internal NCSA collaborations.