2017/11/29 by Karel Adámek, Sofia Dimoudi, Adámek, Karel +4
Environmental Science · Physics and Astronomy · #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.1711.10855
openalex publication_date 2017/11/29 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
We present an improvement of our implementation of the Correlation Technique\nfor the Fourier Domain Acceleration Search (FDAS) algorithm on Graphics\nProcessor Units (GPUs) (Dimoudi & Armour 2015; Dimoudi et al. 2017). Our new\nimproved convolution code which uses our custom GPU FFT code is between 2.5 and\n3.9 times faster the than our cuFFT-based implementation (on an NVIDIA P100)\nand allows for a wider range of filter sizes then our previous version. By\nusing this new version of our convolution code in FDAS we have achieved 44%\nperformance increase over our previous best implementation. It is also\napproximately 8 times faster than the existing PRESTO GPU implementation of\nFDAS (Luo 2013). This work is part of the AstroAccelerate project (Armour et\nal. 2002), a many-core accelerated time-domain signal processing library for\nradio astronomy.\n