Multi-GPU maximum entropy image synthesis for radio astronomy

The maximum entropy method (MEM) is a well known deconvolution technique in radio-interferometry. This method solves a non-linear optimization problem with an entropy regularization term. Other heuristics such as CLEAN are faster but highly user dependent. Nevertheless, MEM has the following advantages: it is unsupervised, it has an statistical basis, it has a better resolution and better image quality under certain conditions. This work presents a high performance GPU version of non-gridded MEM, which is tested using interferometric and simulated data. We propose a single-GPU and a multi-GPU implementation for single and multi-spectral data, respectively. We also make use of the Peer-to-Peer and Unified Virtual Addressing features of newer GPUs which allows to exploit transparently and efficiently multiple GPUs. Several ALMA data sets are used to demonstrate the effectiveness in imaging and to evaluate GPU performance. The results show that a speedup from 1000 to 5000 times faster than a sequential version can be achieved, depending on data and image size. This has allowed us to reconstruct the HD142527 CO(6-5) short baseline data set in 2.1 minutes, instead of the 2.5 days that takes on CPU.

Comments: 11 pages, 13 figures

Similar Publications

Ring resonators provide a means of filtering specific wavelengths from a waveguide, and optionally dropping the filtered wavelengths into a second waveguide. Both of these features are potentially useful for astronomical instruments. In this paper we focus on their use as notch filters to remove the signal from atmospheric OH emission lines from astronomical spectra, however we also briefly discuss their use as frequency combs for wavelength calibration and as drop filters for Doppler planet searches. Read More


We present a method to classify initial conditions of a Long Gamma Ray Bursts model sourced by a single relativistic shock. It is based on the use of Artificial Neural Networks (ANNs) that are trained with Light Curves (LC) generated with Radiation Relativistic Hydrodynamics simulations. The model we use consists in a single shock with a highly relativistic injected beam into a stratified surrounding medium with profile 1/r2. Read More


This document describes a code to perform parameter estimation and model selection in targeted searches for continuous gravitational waves from known pulsars using data from ground-based gravitational wave detectors. We describe the general workings of the code and characterise it on simulated data containing both noise and simulated signals. We also show how it performs compared to a previous MCMC and grid-based approach to signal parameter estimation. Read More


The imminent advent of very large-scale optical sky surveys, such as Euclid and LSST, makes it important to find efficient ways of discovering rare objects such as strong gravitational lens systems, where a background object is multiply gravitationally imaged by a foreground mass. As well as finding the lens systems, it is important to reject false positives due to intrinsic structure in galaxies, and much work is in progress with machine learning algorithms such as neural networks in order to achieve both these aims. We present and discuss a Support Vector Machine (SVM) algorithm which makes use of a Gabor filterbank in order to provide learning criteria for separation of lenses and non-lenses, and demonstrate using blind challenges that under certain circumstances it is a particularly efficient algorithm for rejecting false positives. Read More


We examine in detail 15 Babylonian observations of lunar appulses and occultations made between 80 and 419 BC for the purpose of setting useful limits on Earth's clock error, as quantified by $\Delta$T, the difference between Terrestrial Time and Universal time. Our results are generally in agreement with reconstructions of $\Delta$T using untimed solar eclipse observations from the same period. We suggest a revised version of the simple quadratic fit to $\Delta$T in light of the new results. Read More


The Transiting Exoplanet Survey Satellite (TESS) will perform a two-year survey of nearly the entire sky, with the main goal of detecting exoplanets smaller than Neptune around bright and nearby stars. There do not appear to be any fundamental obstacles to continuing science operations for at least several years after the two-year Primary Mission. To provide a head start to those who will plan and propose for such a mission, we present simulations of exoplanet detections in a third year of TESS operations. Read More


We have conducted a large-scale survey of the northern plane using Kiso Wide Field Camera attached to Schmidt telescope at Kiso observatory. The KISOGP (KWFC Intensive Survey of the Galactic Plane) project have made 40-70 epoch observations in I band of about 320 sq. degrees for 5 years starting in 2012. Read More


During February 2016, CSIRO Astronomy and Space Science and the Max-Planck-Institute for Radio Astronomy installed, commissioned and carried out science observations with a phased array feed (PAF) receiver system on the 64m diameter Parkes radio telescope. Here we demonstrate that the PAF can be used for pulsar observations and we highlight some unique capabilities. We demonstrate that the pulse profiles obtained using the PAF can be calibrated and that multiple pulsars can be simultaneously observed. Read More


Radio interferometry most commonly involves antennas or antenna arrays of identical design. The identical antenna assumption leads to a convenient and useful mathematical simplification resulting in a scalar problem. An interesting variant to this is a "hybrid" interferometer involving two designs. Read More


We introduce ScannerBit, the statistics and sampling module of the public, open-source global fitting framework Gambit. ScannerBit provides a standardised interface to different sampling algorithms, enabling the use and comparison of multiple computational methods for inferring profile likelihoods, Bayesian posteriors, and other statistical quantities. The current version offers random, grid, raster, nested sampling, differential evolution, Markov Chain Monte Carlo (MCMC) and ensemble Monte Carlo samplers. Read More