In the last few years, Input/Output (I/O) bandwidth limitation of legacy computer architectures forced us to reconsider where and how to store and compute data across a large range of applications. This shift has been made possible with the concurrent development of both smartNICs and programmable switches with a common programming language (P4), and the advent of attached High Bandwidth Memory within smartNICs/FPGAs. Recently, proposals to use this kind of technology have emerged to tackle computer science related issues such as fast consensus algorithm in the network, network accelerated key-value stores, machine learning, or data-center data aggregation. In this paper, we introduce a novel architecture that leverages these advancements to potentially accelerate and improve the processing of radio-astronomy Digital Signal Processing (DSP), such as correlators or beamformers, at unprecedented continuous rates in what we have called the "Atomic COTS" design. We give an overview of this new type of architecture to accelerate digital signal processing, leveraging programmable switches and HBM capable FPGAs. We also discuss how to handle radio astronomy data streams to pre-process this stream of data for astronomy science products such as pulsar timing and search. Finally, we illustrate, using a proof of concept, how we can process emulated data from the Square Kilometer Array (SKA) project to time pulsars.

Major digital signal processing tasks that take place at the site of the Square Kilometre Array (SKA) telescopes are beamforming and correlation. The additional task of searching for pulsars is not discussed in detail in this paper. The combined processing, for correlation, exceeds 10 Pops/s. Input data rates are up to 66 Tb/s (SURVEY array) and output data rates are up to 34 Tb/s (LOW array). The methods used to implement these correlators are discussed. For beamforming, LOW has the highest input data rate of 3 Pb/s coupled with a compute load of 5 Peta integer arithmetic operations per second (Pops/s). The purpose of the beamformer in each antenna array is different. In one case (LOW) the beamformer reduces the field of view, in another (SURVEY) it increases it and for the third array (MID) the beamforming is used to increase sensitivity to allow for the detection of pulsars.

The Boolardy Engineering Test Array (BETA) is the first prototype of the Australian Square Kilometer Array Pathfinder (ASKAP), a new radio telescope that is pioneering the use of Phased Array Feed (PAF) technology in radio astronomy to provide instantaneous wide field-of-view imaging. BETA consists of six 12 meter parabolic antennas, each equipped with a 188-element PAF at the focus and a signal processing system capable of imaging an instantaneous 7 square degree field-of-view with a bandwidth of 300MHz at a resolution of 18kHz. In this paper we describe the architecture of the digital signal processing chain and its implementation on reconfigurable hardware based on Field-Programmable Gate Array (FPGA) technology. We present early engineering results taken by the instrument during commissioning trials at the Murchison Radio Observatory (MRO) in remote Western Australia. In particular, we present the first multi-beam image of an astronomical source captured with a PAF-based interferometer.

Owing to the excessive demand on signal processing and space constraint, a full digital implementation of a large adaptive antenna array at millimeter wave frequencies is very challenging. Targeted at long range high data rate point-to-point link in the 70/80 GHz bands, a novel hybrid adaptive antenna array which consists of analogue subarrays followed by a digital beamformer is presented in this paper to overcome the digital implementation difficulty. Two subarray configurations, the interleaved subarray and the side-by-side subarray, are proposed, and two Doppler resilient adaptive angle-of-arrival estimation and beamforming algorithms, the differential beam tracking (DBT) and the differential beam search (DBS), are developed. Simulation results on the DBT and DBS performance are provided using a 64 element hybrid planar array of four 4 by 4 element subarrays with the two subarray configurations, respectively. Recursive mean square error (MSE) bounds of the developed algorithms are also analyzed and compared with simulated MSEs.

Summary form only given. The international radio astronomy community are working toward the next generation telescope the square kilometre array (SKA). Australia's contribution is the Australian SKA Pathfinder which is being built by CSIRO in Western Australia. ASKAP is exploring the concept of placing a phased array feed at the focus of a 12 m parabolic reflector. This can be thought of a radio camera with the feeds of the phased array being the receptors and the dish the optics. However to optimize performance the receptor data must be calibrated and summed in a beamformer. The array generates beams that cover a large area of the sky. For ASKAP there are 192 inputs to the beamformer each sampled at 768 MS/s. For each of 36 antennas this is processed to generate up to 30 dual polarisation antenna beams which are decimated into 16384 frequency channels. This processing is implemented in an FPGA system that computes at greater than 1000 Tera arithmetic operations per second. Internal data flow in the system is 70 Tb/s. For ASKAP, FPGAs are the optimum choice in terms of total system cost which includes hardware and firmware development and total power consumption over the life of the instrument. Design of this system present a number of challenges including how to partition the system, efficient data flow and command and control.

A new type of hybrid antenna array consisting of analogue subarrays followed by a digital beamformer is proposed for practical implementation of long range high data rate millimetre wave communications systems. An adaptive algorithm, referred to as the differential beam search (DBS), is proposed for the angle of arrival (AoA) estimation to control the phase shifters in the analogue subarrays and to perform digital beamforming. This algorithm does not need the knowledge of a reference signal and effectively solves the phase ambiguity problem in AoA estimation inherent to the practical subarray configuration. The performance of the proposed DBS algorithms is demonstrated by simulations.

The Australia SKA Pathfinder (ASKAP) is a new telescope under development as a world-class high-dynamic-range wide-field-of-view survey instrument. It will utilize focal plane phased array feeds on the 36 12-m antennas that will compose the array. The large amounts of data present a huge computing challenge, and ASKAP will store data products in an archive after near real-time pipeline processing. This powerful instrument will be deployed at a new radio-quiet observatory, the Murchison Radio-astronomy Observatory in the midwest region of Western Australia, to enable sensitive surveys of the entire sky to address some of the big questions in contemporary physics. As a pathfinder for the SKA, ASKAP will demonstrate field of view enhancement and computing/processing technology as well as the operation of a large-scale radio array in a remote and radio-quiet region of Australia.

The Murchison Widefield Array is a dipole-based aperture array synthesis telescope designed to operate in the 80-300 MHz frequency range. It is capable of a wide range of science investigations but is initially focused on three key science projects: detection and characterization of three-dimensional brightness temperature fluctuations in the 21 cm line of neutral hydrogen during the epoch of reionization (EoR) at redshifts from six to ten; solar imaging and remote sensing of the inner heliosphere via propagation effects on signals from distant background sources; and high-sensitivity exploration of the variable radio sky. The array design features 8192 dual-polarization broadband active dipoles, arranged into 512 ldquotilesrdquo comprising 16 dipoles each. The tiles are quasi-randomly distributed over an aperture 1.5 km in diameter, with a small number of outliers extending to 3 km. All tile-tile baselines are correlated in custom field-programmable gate array based hardware, yielding a Nyquist-sampled instantaneous monochromaticuvcoverage and unprecedented point spread function quality. The correlated data are calibrated in real time using novel position-dependent self-calibration algorithms. The array is located in the Murchison region of outback Western Australia. This region is characterized by extremely low population density and a superbly radio-quiet environment, allowing full exploitation of the instrumental capabilities.

This paper describes a correlator that is optimized for the Xilinx Virtex-4 SX FPGA, and its application in the SKAMP radio telescope at the Molonglo Radio Observatory. The digital backend of the SKAMP telescope consists of more than 800 Virtex-4 FPGAs. Correlation is performed between each and every pairing of antenna inputs, so the SKAMP telescope, with its 384 inputs, has approximately 74,000 antenna correlations; with 100 MHz of input bandwidth from each antenna this requires real-time processing of more than 7 tera complex multiply-accumulates per second. The correlation cell described takes advantage of the hard IP blocks found within the Virtex-4 FPGA to perform one 4+4-bit complex correlation per cycle at a clock rate exceeding 256 MHz. At the core of each cell is an efficient 4-bit signed complex multiplier, implemented using the 18-bit signed multiplier of the Virtex-4 DSP slice, and a short term accumulator, implemented using the adjacent Block RAM. Nearly 30,000 correlation cells are instantiated across 192 Virtex-4 SX35 devices in order to process all the data from the SKAMP telescope.