No abstract
We study the benefits and limits of parallelised Markov chain Monte Carlo (MCMC) sampling in cosmology. MCMC methods are widely used for the estimation of cosmological parameters from a given set of observations and are typically based on the Metropolis-Hastings algorithm. Some of the required calculations can however be computationally intensive, meaning that a single long chain can take several hours or days to calculate. In practice, this can be limiting, since the MCMC process needs to be performed many times to test the impact of possible systematics and to understand the robustness of the measurements being made. To achieve greater speed through parallelisation, MCMC algorithms need to have short auto-correlation times and minimal overheads caused by tuning and burn-in. The resulting scalability is hence influenced by two factors, the MCMC overheads and the parallelisation costs. In order to efficiently distribute the MCMC sampling over thousands of cores on modern cloud computing infrastructure, we developed a Python framework called CosmoHammer which embeds emcee, an implementation by Foreman-Mackey et al. (2012) of the affine invariant ensemble sampler by Goodman and Weare (2010). We test the performance of CosmoHammer for cosmological parameter estimation from cosmic microwave background data. While Metropolis-Hastings is dominated by overheads, CosmoHammer is able to accelerate the sampling process from a wall time of 30 hours on a dual core notebook to 16 minutes by scaling out to 2048 cores. Such short wall times for complex data sets opens possibilities for extensive model testing and control of systematics.Comment: Published version. 17 pages, 6 figures. The code is available at http://www.astro.ethz.ch/refregier/research/Software/cosmohamme
Context. Solar observatories are providing the world-wide community with a wealth of data, covering wide time ranges (e.g. Solar and Heliospheric Observatory, SOHO), multiple viewpoints (Solar TErrestrial RElations Observatory, STEREO), and returning large amounts of data (Solar Dynamics Observatory, SDO). In particular, the large volume of SDO data presents challenges; the data are available only from a few repositories, and full-disk, full-cadence data for reasonable durations of scientific interest are difficult to download, due to their size and the download rates available to most users. From a scientist's perspective this poses three problems: accessing, browsing, and finding interesting data as efficiently as possible. Aims. To address these challenges, we have developed JHelioviewer, a visualisation tool for solar data based on the JPEG 2000 compression standard and part of the open source ESA/NASA Helioviewer Project. Since the first release of JHelioviewer in 2009, the scientific functionality of the software has been extended significantly, and the objective of this paper is to highlight these improvements. Methods. The JPEG 2000 standard offers useful new features that facilitate the dissemination and analysis of high-resolution image data and offers a solution to the challenge of efficiently browsing petabyte-scale image archives. The JHelioviewer software is open source, platform independent, and extendable via a plug-in architecture. Results. With JHelioviewer, users can visualise the Sun for any time period between September 1991 and today; they can perform basic image processing in real time, track features on the Sun, and interactively overlay magnetic field extrapolations. The software integrates solar event data and a timeline display. Once an interesting event has been identified, science quality data can be accessed for in-depth analysis. As a first step towards supporting science planning of the upcoming Solar Orbiter mission, JHelioviewer offers a virtual camera model that enables users to set the vantage point to the location of a spacecraft or celestial body at any given time.
Aims. The Spectrometer Telescope for Imaging X-rays (STIX) on Solar Orbiter is a hard X-ray imaging spectrometer, which covers the energy range from 4 to 150 keV. STIX observes hard X-ray bremsstrahlung emissions from solar flares and therefore provides diagnostics of the hottest (⪆10 MK) flare plasma while quantifying the location, spectrum, and energy content of flare-accelerated nonthermal electrons. Methods. To accomplish this, STIX applies an indirect bigrid Fourier imaging technique using a set of tungsten grids (at pitches from 0.038 to 1 mm) in front of 32 coarsely pixelated CdTe detectors to provide information on angular scales from 7 to 180 arcsec with 1 keV energy resolution (at 6 keV). The imaging concept of STIX has intrinsically low telemetry and it is therefore well-suited to the limited resources available to the Solar Orbiter payload. To further reduce the downlinked data volume, STIX data are binned on board into 32 selectable energy bins and dynamically-adjusted time bins with a typical duration of 1 s during flares. Results. Through hard X-ray diagnostics, STIX provides critical information for understanding the acceleration of electrons at the Sun and their transport into interplanetary space and for determining the magnetic connection of Solar Orbiter back to the Sun. In this way, STIX serves to link Solar Orbiter’s remote and in-situ measurements.
The radio emission during 201 selected X-ray solar flares was surveyed from 100 MHz to 4 GHz with the Phoenix-2 spectrometer of ETH Zürich. The selection includes all RHESSI flares larger than C5.0 jointly observed from launch until June 30, 2003. Detailed association rates of radio emission during X-ray flares are reported. In the decimeter wavelength range, type III bursts and the genuinely decimetric emissions (pulsations, continua, and narrowband spikes) were found equally frequently. Both occur predominantly in the peak phase of hard X-ray (HXR) emission, but are less in tune with HXRs than the high-frequency continuum exceeding 4 GHz, attributed to gyrosynchrotron radiation. In 10% of the HXR flares, an intense radiation of the above genuine decimetric types followed in the decay phase or later. Classic meter-wave type III bursts are associated in 33% of all HXR flares, but only in 4% are they the exclusive radio emission. Noise storms were the only radio emission in 5% of the HXR flares, some of them with extended duration. Despite the spatial association (same active region), the noise storm variations are found to be only loosely correlated in time with the X-ray flux. In a surprising 17% of the HXR flares, no coherent radio emission was found in the extremely broad band surveyed. The association but loose correlation between HXR and coherent radio emission is interpreted by multiple reconnection sites connected by common field lines.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.