Challenges in exascale radio astronomy: Can the SKA ride the technology wave?

Abstract

The Square Kilometre Array (SKA) will be the most sensitive radio telescope in the world. This unprecedented sensitivity will be achieved by combining and analyzing signals from 262,144 antennas and 350 dishes at a raw datarate of petabits per second. The processing pipeline to create useful astronomical data will require hundreds of peta-operations per second, at a very limited power budget. We analyze the compute, memory and bandwidth requirements for the key algorithms used in the SKA. By studying their implementation on existing platforms, we show that most algorithms have properties that map inefficiently on current hardware, such as a low compute–bandwidth ratio and complex arithmetic. In addition, we estimate the power breakdown on CPUs and GPUs, analyze the cache behavior on CPUs, and discuss possible improvements. This work is complemented with an analysis of supercomputer trends, which demonstrates that current efforts to use commercial off-the-shelf accelerators results in a two to three times smaller improvement in compute capabilities and power efficiency than custom built machines. We conclude that waiting for new technology to arrive will not give us the instruments currently planned in 2018: one or two orders of magnitude better power efficiency and compute capabilities are required. Novel hardware and system architectures, to match the needs and features of this unique project, must be developed.

An energy-efficient custom architecture for the SKA1-low central signal processor

“…Accessing the main memory represents the biggest contribution to the power budget. In the case of the channelization, however, this contribution is higher in percentage, because of the relatively small exploitable input data locality of the filtering step, in particular when compared with the execution of the correlation, which presents a more compute-bound behavior [26]. As expected, at application level, the power consumption is dominated by the correlation.…”

Reduced-Precision Acceleration of Radio-Astronomical Imaging on Reconfigurable Hardware

“…Over the past couple of years, hardware technologies have significantly improved, and a fair amount of research has been done on optimizing radio-astronomy algorithms to satisfy the Square Kilometre Array requirements [104].…”

GPU-Based Data Processing for 2-D Microwave Imaging on MAST

“…Beam forming in radio astronomy has higher performance and is more energy efficient on a GPU system compared to a multi-core CPU system 12 and a recent analysis 13 concluded that for SKA, novel hardware and system architectures need to be developed to achieve the required power efficiency and compute capabilities. The European Organization for Nuclear Research (CERN) and LHC have investigated the potential of using GPUs in the high-level trigger algorithms used to select the interesting data to reduce the raw data obtained from this experiment at the nominal LHC collision rate of 40MHz or every 25 nanoseconds.…”