Search

This paper presents the GPU porting through OpenACC directives of the Dutch Atmospheric Large-Eddy Simulation (DALES) application, a high-resolution atmospheric model. The code is written in Fortran~90 and features parallel (distributed) execution through spatial domain decomposition. We assess the performance of the GPU offloading, comparing the time-to-solution on regular and accelerated HPC nodes. %comparing the computational time between distributed and accelerated nodes. A weak scaling analysis is conducted and portability across NVIDIA A100 and H100 hardware %and AMD hardware is discussed. Finally, we show how targeted kernels can benefit from further optimization with Kernel Tuner, a GPU kernels auto-tuning package.

Many studies have focused on developing and improving auto-tuning algorithms for Nvidia Graphics Processing Units (GPUs), but the effectiveness and efficiency of these approaches on AMD devices have hardly been studied. This paper aims to address this gap by introducing an auto-tuner for AMD's HIP. We do so by extending Kernel Tuner, an open-source Python library for auto-tuning GPU programs. We analyze the performance impact and tuning difficulty for four highly-tunable benchmark kernels on four different GPUs: two from Nvidia and two from AMD. Our results demonstrate that auto-tuning has a significantly higher impact on performance on AMD compared to Nvidia (10x vs 2x). Additionally, we show that applications tuned for Nvidia do not perform optimally on AMD, underscoring the importance of auto-tuning specifically for AMD to achieve high performance on these GPUs., Comment: 30th International European Conference on Parallel and Distributed Computing (Euro-PAR 2024) (Best Paper Candidates Session)

Understanding the origin of fast radio bursts (FRBs) has become the main science driver of recent dedicated FRB surveys. Between July 2019 and February 2022, we carried out ALERT, an FRB survey at 1370 MHz using the Apertif instrument installed at the Westerbork Synthesis Radio Telescope (WSRT). Here we report the detection of 18 new FRBs, and we study the properties of the entire 24 burst sample detected during the survey. For five bursts, we identify host galaxy candidates with >50% probability association. We observe an average linear polarisation fraction of $\sim$43% and an average circular polarisation fraction consistent with 0%. A third of the FRBs display multiple components. The sample next reveals a population of highly scattered bursts, which is most likely to have been produced in the immediate circumburst environment. Furthermore, two FRBs show evidence for high rotation measures, reaching |RM|>$10^3$ rad m$^{-2}$ in the source reference frames. Together, the scattering and rotation measures ALERT finds prove that a large fraction of FRBs are embedded in complex media such as star forming regions or supernova remnants. Through the discovery of the third most dispersed FRB so far, we show that one-off FRBs can emit at frequencies in excess of 6 GHz. Finally, we determine an FRB all-sky rate of $459^{+208}_{-155}$ sky$^{-1}$ day$^{-1}$ above a fluence limit of 4.1 Jy ms, and a fluence cumulative distribution with a power law index $\gamma=-1.23\pm0.06\pm0.2$, which is roughly consistent with the Euclidean Universe predictions. Through the high resolution in time, frequency, polarisation and localisation that ALERT featured, we were able to determine the morphological complexity, polarisation, local scattering and magnetic environment, and high-frequency luminosity of FRBs. We find all these strongly resemble those seen in young, energetic, highly magnetised neutron stars., Comment: 40 pages (including 11 of appendix), 37 figures, 5 tables. Accepted for publication in A&A

Fast Radio Bursts must be powered by uniquely energetic emission mechanisms. This requirement has eliminated a number of possible source types, but several remain. Identifying the physical nature of Fast Radio Burst (FRB) emitters arguably requires good localisation of more detections, and broadband studies enabled by real-time alerting. We here present the Apertif Radio Transient System (ARTS), a supercomputing radio-telescope instrument that performs real-time FRB detection and localisation on the Westerbork Synthesis Radio Telescope (WSRT) interferometer. It reaches coherent-addition sensitivity over the entire field of the view of the primary dish beam. After commissioning results verified the system performed as planned, we initiated the Apertif FRB survey (ALERT). Over the first 5 weeks we observed at design sensitivity in 2019, we detected 5 new FRBs, and interferometrically localised each of these to 0.4--10 sq. arcmin. All detections are broad band and very narrow, of order 1 ms duration, and unscattered. Dispersion measures are generally high. Only through the very high time and frequency resolution of ARTS are these hard-to-find FRBs detected, producing an unbiased view of the intrinsic population properties. Most localisation regions are small enough to rule out the presence of associated persistent radio sources. Three FRBs cut through the halos of M31 and M33. We demonstrate that Apertif can localise one-off FRBs with an accuracy that maps magneto-ionic material along well-defined lines of sight. The rate of 1 every ~7 days next ensures a considerable number of new sources are detected for such study. The combination of detection rate and localisation accuracy exemplified by the 5 first ARTS FRBs thus marks a new phase in which a growing number of bursts can be used to probe our Universe., Comment: Accepted version

Fast radio bursts (FRBs) are extragalactic astrophysical transients whose brightness requires emitters that are highly energetic, yet compact enough to produce the short, millisecond-duration bursts. FRBs have thus far been detected between 300 MHz and 8 GHz, but lower-frequency emission has remained elusive. A subset of FRBs is known to repeat, and one of those sources, FRB 20180916B, does so with a 16.3 day activity period. Using simultaneous Apertif and LOFAR data, we show that FRB 20180916B emits down to 120 MHz, and that its activity window is both narrower and earlier at higher frequencies. Binary wind interaction models predict a narrower periodic activity window at lower frequencies, which is the opposite of our observations. Our detections establish that low-frequency FRB emission can escape the local medium. For bursts of the same fluence, FRB 20180916B is more active below 200 MHz than at 1.4 GHz. Combining our results with previous upper-limits on the all-sky FRB rate at 150 MHz, we find that there are 3-450 FRBs/sky/day above 50 Jy ms at 90% confidence. We are able to rule out the scenario in which companion winds cause FRB periodicity. We also demonstrate that some FRBs live in clean environments that do not absorb or scatter low-frequency radiation., Comment: 50 pages, 14 figures, 3 tables, submitted

We present the collaborative model of ESiWACE2 Services, where Research Software Engineers (RSEs) from the Netherlands eScience Center (NLeSC) and Atos offer their expertise to climate and earth system modeling groups across Europe. Within 6-month collaborative projects, the RSEs intend to provide guidance and advice regarding the performance, portability to new architectures, and scalability of selected applications. We present the four awarded projects as examples of this funding structure., Comment: RSE-HPC-2020 workshop

After years of using Graphics Processing Units (GPUs) to accelerate scientific applications in fields as varied as tomography, computer vision, climate modeling, digital forensics, geospatial databases, particle physics, radio astronomy, and localization microscopy, we noticed a number of technical, socio-technical, and non-technical challenges that Research Software Engineers (RSEs) may run into. While some of these challenges, such as managing different programming languages within a project, or having to deal with different memory spaces, are common to all software projects involving GPUs, others are more typical of scientific software projects. Among these challenges we include changing resolutions or scales, maintaining an application over time and making it sustainable, and evaluating both the obtained results and the achieved performance. %In this paper, we present the challenges and lessons learned from research software engineering GPU applications.

Current and upcoming radio telescopes are being designed with increasing sensitivity to detect new and mysterious radio sources of astrophysical origin. While this increased sensitivity improves the likelihood of discoveries, it also makes these instruments more susceptible to the deleterious effects of Radio Frequency Interference (RFI). The challenge posed by RFI is exacerbated by the high data-rates achieved by modern radio telescopes, which require real-time processing to keep up with the data. Furthermore, the high data-rates do not allow for permanent storage of observations at high resolution. Offline RFI mitigation is therefore not possible anymore. The real-time requirement makes RFI mitigation even more challenging because, on one side, the techniques used for mitigation need to be fast and simple, and on the other side they also need to be robust enough to cope with just a partial view of the data. The Apertif Radio Transient System (ARTS) is the real-time, time-domain, transient detection instrument of the Westerbork Synthesis Radio Telescope (WSRT), processing 73 Gb of data per second. Even with a deep learning classifier, the ARTS pipeline requires state-of-the-art real-time RFI mitigation to reduce the number of false-positive detections. Our solution to this challenge is RFIm, a high-performance, open-source, tuned, and extensible RFI mitigation library. The goal of this library is to provide users with RFI mitigation routines that are designed to run in real-time on many-core accelerators, such as Graphics Processing Units, and that can be highly-tuned to achieve code and performance portability to different hardware platforms and scientific use-cases. Results on the ARTS show that we can achieve real-time RFI mitigation, with a minimal impact on the total execution time of the search pipeline, and considerably reduce the number of false-positives., Comment: 6 pages, 10 figures. To appear in Proceedings from the 2019 Radio Frequency Interference workshop (RFI 2019), Toulouse, France (23-26 September)

Real-time searches for faint radio pulses from unknown radio transients are computationally challenging. Detections become further complicated due to continuously increasing technical capabilities of transient surveys: telescope sensitivity, searched area of the sky, number of antennas or dishes, temporal and frequency resolution. The new Apertif transient survey on the Westerbork telescope happens in real-time on GPUs by means of the single-pulse search pipeline AMBER (Sclocco, 2017). AMBER initially carries out auto tuning: it finds the most optimal configuration of user-controlled parameters per each of four pipeline kernels so that each kernel performs its task as fast as possible. The pipeline uses a brute-force (BF) exhaustive search which in total takes 5 - 24 hours to run depending on the processing cluster architecture. We apply more heuristic, biologically driven genetic algorithms (GAs) to limit the exploration of the total parameter space, tune all four kernels together and reduce the tuning time to few hours. Our results show that after only few hours of tuning, GAs always find similar or even better configurations for all kernels together than the combination of single kernel configurations tuned by the BF approach. At the same time, by means of their genetic operators, GAs converge into better solutions than those obtained by pure random searches. The explored multi-dimensional parameter space is very complex and has multiple local optima as the evolution of randomly generated configurations does not always guarantee global solution., Comment: 13 pages, 8 figures, 3 tables

In this paper, we study the parallelization of the dedispersion algorithm on many-core accelerators, including GPUs from AMD and NVIDIA, and the Intel Xeon Phi. An important contribution is the computational analysis of the algorithm, from which we conclude that dedispersion is inherently memory-bound in any realistic scenario, in contrast to earlier reports. We also provide empirical proof that, even in unrealistic scenarios, hardware limitations keep the arithmetic intensity low, thus limiting performance. We exploit auto-tuning to adapt the algorithm, not only to different accelerators, but also to different observations, and even telescopes. Our experiments show how the algorithm is tuned automatically for different scenarios and how it exploits and highlights the underlying specificities of the hardware: in some observations, the tuner automatically optimizes device occupancy, while in others it optimizes memory bandwidth. We quantitatively analyze the problem space, and by comparing the results of optimal auto-tuned versions against the best performing fixed codes, we show the impact that auto-tuning has on performance, and conclude that it is statistically relevant., Comment: 10 pages, published in the proceedings of IPDPS 2014

Dedispersion, the removal of deleterious smearing of impulsive signals by the interstellar matter, is one of the most intensive processing steps in any radio survey for pulsars and fast transients. We here present a study of the parallelization of this algorithm on many-core accelerators, including GPUs from AMD and NVIDIA, and the Intel Xeon Phi. We find that dedispersion is inherently memory-bound. Even in a perfect scenario, hardware limitations keep the arithmetic intensity low, thus limiting performance. We next exploit auto-tuning to adapt dedispersion to different accelerators, observations, and even telescopes. We demonstrate that the optimal settings differ between observational setups, and that auto-tuning significantly improves performance. This impacts time-domain surveys from Apertif to SKA., Comment: 8 pages, accepted for publication in Astronomy and Computing