High-Time Resolution GPU Imager for FRB searches at low radio frequencies

M. Sokolowski,G. Aniruddha,C. Di Pietrantonio,C. Harris,D. C. Price,S. McSweeney,R. B. Wayth,N. D. R. Bhat

2024-05-22

Abstract:Fast Radio Bursts (FRBs) are millisecond dispersed radio pulses of predominately extra-galactic origin. Although originally discovered at GHz frequencies, most FRBs have been detected between 400 to 800 MHz. Nevertheless, only a handful of FRBs were detected below 400 MHz. Searching for FRBs at low frequencies is computationally challenging due to increased dispersive delay that must be accounted for. However, the wide field of view (FoV) of low-frequency telescopes - such as the the Murchison Widefield Array (MWA), and prototype stations of the low-frequency Square Kilometre Array (SKA-Low) - makes them promising instruments to open a low-frequency window on FRB event rates, and constrain FRB emission models. The standard approach, inherited from high-frequencies, is to form multiple tied-array beams to tessellate the entire FoV and perform the search on the resulting time series. This approach, however, may not be optimal for low-frequency interferometers due to their large FoVs and high spatial resolutions leading to a large number of beams. Consequently, there are regions of parameter space in terms of number of antennas and resolution elements (pixels) where interferometric imaging is computationally more efficient. Here we present a new high-time resolution imager BLINK implemented on modern Graphical Processing Units (GPUs) and intended for radio astronomy. The main goal for this imager is to become part of a fully GPU-accelerated FRB search pipeline. We describe the imager and present its verification on real and simulated data processed to form all-sky and widefield images from the MWA and prototype SKA-Low stations. We also present and compare benchmarks of the GPU and CPU code executed on laptops, desktop computers, and Australian supercomputers. The code is publicly available at

Instrumentation and Methods for Astrophysics

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the computational challenges faced in searching for Fast Radio Bursts (FRBs) at low radio frequencies. Specifically:

1. **Computational Complexity**: At low radio frequencies, due to the increase in dispersion delay, the computational complexity of processing data is significantly increased. Traditional high - frequency methods (such as forming multiple array beams to cover the entire field of view) may not be applicable in low - frequency interferometers because these interferometers have a large field of view and high spatial resolution, resulting in the need to form a large number of beams, thereby increasing the computational burden.

2. **Data Processing Efficiency**: Existing software packages are less efficient in processing low - frequency data, limiting the amount of data that can be processed. Therefore, developing an efficient real - time data processing and search pipeline is crucial for increasing the detection rate of low - frequency FRBs.

3. **Energy Efficiency Problem**: As high - performance computing centers increasingly adopt GPUs, developing more energy - efficient GPU software has become an important goal. This not only helps to improve computational efficiency but also reduces the carbon footprint, meeting the requirements of the astronomical community for environmental sustainability.

To address these problems, the paper introduces a new high - time - resolution GPU imager (BLINK), which aims to be part of a fully GPU - based FRB search pipeline. This imager can run efficiently on modern GPUs and is suitable for data from multiple radio telescopes. By optimizing computational efficiency and energy efficiency, this imager is expected to make image - domain FRB searches in low - frequency interferometers computationally feasible, economical, and successful.