2019/12/16 by Cees Carels, Karel Adámek, Carels, Cees +5
Engineering · Environmental Science · Physics and Astronomy · #Antenna Design and Optimization #Astrophysics and Cosmic Phenomena #FOS: Physical sciences #Instrumentation and Methods for Astrophysics (astro-ph.IM) #Radio Astronomy Observations and Technology #Soil Moisture and Remote Sensing
paper · pdf · doi:10.48550/arxiv.1912.07704
openalex publication_date 2019/12/16 · openalex created_date 2021/07/05 · openalex updated_date 2026/07/28
Upcoming large scale telescope projects such as the Square Kilometre Array\n(SKA) will see high data rates and large data volumes; requiring tools that can\nanalyse telescope event data quickly and accurately. In modern radio\ntelescopes, analysis software forms a core part of the data read out, and\nlong-term software stability and maintainability are essential. AstroAccelerate\nis a many core accelerated software package that uses NVIDIA(R) GPUs to perform\nrealtime analysis of radio telescope data, and it has been shown to be\nsubstantially faster than realtime at processing simulated SKA-like data.\nAstroAccelerate contains optimised GPU implementations of signal processing\ntools used in radio astronomy including dedispersion, Fourier domain\nacceleration search, single pulse detection, and others. This article describes\nthe transformation of AstroAccelerate from a C-like prototype code to a\nproduction-ready software library with a C++ API and a Python interface; while\npreserving compatibility with legacy software that is implemented in C. The\ndesign of the software library interfaces, refactoring aspects, and coding\ntechniques are discussed.\n