GALARIO: a GPU accelerated library for analysing radio interferometer observations

Abstract

We present GALARIO, a computational library that exploits the power of modern graphical processing units (GPUs) to accelerate the analysis of observations from radio interferometers like ALMA or the VLA. GALARIO speeds up the computation of synthetic visibilities from a generic 2D model image or a radial brightness profile (for axisymmetric sources). On a GPU, GALARIO is 150 faster than standard Python and 10 times faster than serial C++ code on a CPU. Highly modular, easy to use and to adopt in existing code, GALARIO comes as two compiled libraries, one for Nvidia GPUs and one for multicore CPUs, where both have the same functions with identical interfaces. GALARIO comes with Python bindings but can also be directly used in C or C++. The versatility and the speed of GALARIO open new analysis pathways that otherwise would be prohibitively time consuming, e.g. fitting high resolution observations of large number of objects, or entire spectral cubes of molecular gas emission. It is a general tool that can be applied to any field that uses radio interferometer observations. The source code is available online at http://github.com/mtazzari/galario under the open source GNU Lesser General Public License v3.Comparing a model image computed on a regular grid to observed visibilities that are scattered across the Fourier plane involves a series of 1D and 2D array operations such as Fourier transforms, transpositions, and interpolations . The size of the arrays used to properly model the visibilities are set by the prop-

Homogeneous Analysis of the Dust Morphology of Transition Disks Observed with ALMA: Investigating Dust Trapping and the Origin of the Cavities

“…We explored four free parameters (r peak , σ int , σ ext , and F total ) with 200 walkers and 1000 steps in each case, while the center, PA, and inclination are fixed. We adopted a set of uniform prior probability distributions for the free parameters explored by the Markov chain, such that As tests, we randomly chose three targets of our sample (J15583692, DoAr 44, and SR 21) and perform the fit of the visibilities leaving the inclination, PA, and center as free parameters, and using the publicly available code GALARIO (Tazzari et al 2018). These tests gave similar results in all three cases, providing confidence about the accuracy of our procedure.…”

“…Software: CASA (McMullin et al 2007), GALARIO (Tazzari et al 2018), emcee (Foreman-Mackey et al 2013). …”

Are inner disc misalignments common? ALMA reveals an isotropic outer disc inclination distribution for young dipper stars

“…We are most fortunate to have the opportunity to conduct observations from this mountain. We used the following software: lightkurve (Lightkurve Collaboration et al 2018), Astropy (Astropy Collaboration et al 2013Collaboration et al , 2018, Matplotlib (Hunter 2007), uvplot (Tazzari 2017), GALARIO (Tazzari et al 2018). Table C1 presents the detection limits derived from highcontrast imaging of the dippers in our sample, when available.…”

“…However, the former method is not appropriate for discs with large inner cavities (there are a few of these "transition discs" in our sample; see Figure 1) and also does not provide posterior distributions that are useful for our analysis of the disc inclination distribution (see Section 4.1). Thus we (re)derive the disc inclinations for our entire sample using the GALARIO computational library (Tazzari et al 2018) combined with the emcee package (Foreman-Mackey et al. 2013), which allows us to quickly fit 2D disc models to the ALMA visibilities by enabling the rapid exploration of parameter space.…”

“…We use the GPU version of GALARIO, which is ∼150× faster than standard Python implementations that rely on scipy and numpy packages. Tazzari et al (2018) provides a detailed explanation of employing GALARIO to fit interferometric visibilities like those from ALMA and we follow their general implementation procedure. The fits typically use 120 walkers and 5000 steps for the MCMC; the mean accepted fraction of steps are all between 0.2 and 0.5, implying that the chains are converging.…”