Advancing radioastronomy into the era of High Performance Computing: The Case of Imaging
This is a hybrid event. It is organised at the Pawsey Supercomputing Research Centre but is also available online.
About the presenter
Claudio Gheller obtained his Ph.D. in astrophysics from the International School for Advanced Studies (SISSA/ISAS) in Trieste, Italy. His thesis focused on numerical cosmology. After that, he gained 12 years of experience as a computational scientist at CINECA, the Italian National Supercomputing Centre. In 2011, he relocated to ETH Zurich in Switzerland and assumed the role of leading the Scientific Community Engagement group at the CSCS Swiss National Supercomputing Centre. From 2018 to 2020 he worked as HPC specialist at the Swiss Plasma Center of EPFL (Switzerland). He joined the Italian National Institute for Astrophysics in 2020 and from July 2023 he acts as Project Manager for the National Institute for HPC, Big Data and Quantum Computing.
About the presentation
Since the last decade, radio astronomy has started a new era: the advent of the Square Kilometer Array (SKA), preceded by its precursors and pathfinders (like MWA, ASKAP or LOFAR), will produce a huge amount of data that will be hard to process with a traditional approach. To effectively handle the growing amount of data and computational demands, scientists must leverage modern high-performance computing architectures. In particular, heterogeneous systems, based on complex combinations of CPUs, accelerators, high-speed networks and composite storage devices need to be used in an efficient and effective way.
As part of the Italian National Centre for HPC and the SKA Regional Centre initiatives, we are working on the development of RICK (Radio Imaging Code Kernels) software. This software is designed to handle the massive amount of data that would be difficult to process using traditional methods by exploiting parallelism and accelerators. RICK represents the first example of radio imaging software enabled to GPUs and scalable to hundreds of tasks becoming a potential state-of-the-art work for the future SKA software suite. Portability (both code and performance) on multiple architectures is key for the RICK code.
I will provide an introduction to RICK, discussing its key architectural features and performance on different kind of processors and GPUs, highlighting the path toward its adoption to state-of-the-art radioastronomical data.