arXiv subjects

Alexander S. Szalay

Publications and source records attributed to Alexander S. Szalay.

At least 37 records · Page 2Linked to original sources

arXiv2016

Dispersal of tidal debris in a Milky-Way-sized dark matter halo

We simulate the tidal disruption of a collisionless N-body globular star cluster in a total of 300 different orbits selected to have galactocentric radii between 10 and 30 kpc in four dark matter halos: (a) a spherical halo with no subhalos, (b) a spherical halo with subhalos, (c) a realistic halo with no subhalos, and (d) a realistic halo with subhalos. This allows us to isolate and study how the halo's (lack of) dynamical symmetry and substructures affect the dispersal of tidal debris. The realistic halos are constructed from the snapshot of the Via Lactea II simulation at redshift zero. We find that the overall halo's lack of dynamical symmetry disperses tidal debris to make the streams fluffier, consistent with previous studies of tidal debris of dwarf galaxies in larger orbits than ours in this study. On the other hand, subhalos in realistic potentials can locally enhance the densities along streams, making streams denser than their counterparts in smooth potentials. We show that many long and thin streams can survive in a realistic and lumpy halo for a Hubble time. This suggests that upcoming stellar surveys will likely uncover more thin streams which may contain density gaps that have been shown to be promising probes for dark matter substructures.

astro-ph.GA↗

arXiv2016

An SSD-based eigensolver for spectral analysis on billion-node graphs

Many eigensolvers such as ARPACK and Anasazi have been developed to compute eigenvalues of a large sparse matrix. These eigensolvers are limited by the capacity of RAM. They run in memory of a single machine for smaller eigenvalue problems and require the distributed memory for larger problems. In contrast, we develop an SSD-based eigensolver framework called FlashEigen, which extends Anasazi eigensolvers to SSDs, to compute eigenvalues of a graph with hundreds of millions or even billions of vertices in a single machine. FlashEigen performs sparse matrix multiplication in a semi-external memory fashion, i.e., we keep the sparse matrix on SSDs and the dense matrix in memory. We store the entire vector subspace on SSDs and reduce I/O to improve performance through caching the most recent dense matrix. Our result shows that FlashEigen is able to achieve 40%-60% performance of its in-memory implementation and has performance comparable to the Anasazi eigensolvers on a machine with 48 CPU cores. Furthermore, it is capable of scaling to a graph with 3.4 billion vertices and 129 billion edges. It takes about four hours to compute eight eigenvalues of the billion-node graph using 120 GB memory.

cs.DC↗

arXiv2016

Doppler term in the galaxy two-point correlation function: wide-angle, velocity, Doppler lensing and cosmic acceleration effects

We study the parity-odd part (that we shall call Doppler term) of the linear galaxy two-point correlation function that arises from wide-angle, velocity, Doppler lensing and cosmic acceleration effects. As it is important at low redshift and at large angular separations, the Doppler term is usually neglected in the current generation of galaxy surveys. For future wide-angle galaxy surveys such as Euclid, SPHEREx and SKA, however, we show that the Doppler term must be included. The effect of these terms is dominated by the magnification due to relativistic aberration effects and the slope of the galaxy redshift distribution and it generally mimics the effect of the local type primordial non-Gaussianity with the effective nonlinearity parameter $f_{\rm NL}^{\rm eff}$ of a few, we show that this would affect forecasts on measurements of $f_{\rm NL}$ at low-redshift. Our results show that a survey at low redshift with large number density over a wide area of the sky could detect the Doppler term with a signal-to-noise ratio of $\sim 1-20$, depending on survey specifications.

astro-ph.CO↗

arXiv2016

Quantifying correlations between galaxy emission lines and stellar continua

We analyse the correlations between continuum properties and emission line equivalent widths of star-forming and active galaxies from the Sloan Digital Sky Survey. Since upcoming large sky surveys will make broad-band observations only, including strong emission lines into theoretical modelling of spectra will be essential to estimate physical properties of photometric galaxies. We show that emission line equivalent widths can be fairly well reconstructed from the stellar continuum using local multiple linear regression in the continuum principal component analysis (PCA) space. Line reconstruction is good for star-forming galaxies and reasonable for galaxies with active nuclei. We propose a practical method to combine stellar population synthesis models with empirical modelling of emission lines. The technique will help generate more accurate model spectra and mock catalogues of galaxies to fit observations of the new surveys. More accurate modelling of emission lines is also expected to improve template-based photometric redshift estimation methods. We also show that, by combining PCA coefficients from the pure continuum and the emission lines, automatic distinction between hosts of weak active galactic nuclei (AGNs) and quiescent star-forming galaxies can be made. The classification method is based on a training set consisting of high-confidence starburst galaxies and AGNs, and allows for the similar separation of active and star-forming galaxies as the empirical curve found by Kauffmann et al. We demonstrate the use of three important machine learning algorithms in the paper: k-nearest neighbour finding, k-means clustering and support vector machines.

astro-ph.GA↗

arXiv2015

Toward Accurate Modeling of the Nonlinear Matter Bispectrum: Standard Perturbation Theory and Transients from Initial Conditions

Accurate modeling of nonlinearities in the galaxy bispectrum, the Fourier transform of the galaxy three-point correlation function, is essential to fully exploit it as a cosmological probe. In this paper, we present numerical and theoretical challenges in modeling the nonlinear bispectrum. First, we test the robustness of the matter bispectrum measured from N-body simulations using different initial conditions generators. We run a suite of N-body simulations using the Zel'dovich approximation and second-order Lagrangian perturbation theory (2LPT) at different starting redshifts, and find that transients from initial decaying modes systematically reduce the nonlinearities in the matter bispectrum. To achieve 1% accuracy in the matter bispectrum for $z\le3$ on scales $k<1$ $h$/Mpc, 2LPT initial conditions generator with initial redshift of $z\gtrsim100$ is required. We then compare various analytical formulas and empirical fitting functions for modeling the nonlinear matter bispectrum, and discuss the regimes for which each is valid. We find that the next-to-leading order (one-loop) correction from standard perturbation theory matches with N-body results on quasi-linear scales for $z\ge1$. We find that the fitting formula in Gil-Mar\'ın et al. (2012) accurately predicts the matter bispectrum for $z\le1$ on a wide range of scales, but at higher redshifts, the fitting formula given in Scoccimarro & Couchman (2001) gives the best agreement with measurements from N-body simulations.

astro-ph.CO↗

arXiv2015

Optimize Unsynchronized Garbage Collection in an SSD Array

Solid state disks (SSDs) have advanced to outperform traditional hard drives significantly in both random reads and writes. However, heavy random writes trigger fre- quent garbage collection and decrease the performance of SSDs. In an SSD array, garbage collection of individ- ual SSDs is not synchronized, leading to underutilization of some of the SSDs. We propose a software solution to tackle the unsyn- chronized garbage collection in an SSD array installed in a host bus adaptor (HBA), where individual SSDs are exposed to an operating system. We maintain a long I/O queue for each SSD and flush dirty pages intelligently to fill the long I/O queues so that we hide the performance imbalance among SSDs even when there are few parallel application writes. We further define a policy of select- ing dirty pages to flush and a policy of taking out stale flush requests to reduce the amount of data written to SSDs. We evaluate our solution in a real system. Experi- ments show that our solution fully utilizes all SSDs in an array under random write-heavy workloads. It improves I/O throughput by up to 62% under random workloads of mixed reads and writes when SSDs are under active garbage collection. It causes little extra data writeback and increases the cache hit rate.

cs.OS↗

arXiv2015

Simulating tidal streams in a high resolution dark matter halo

We simulate tidal streams in the presence and absence of substructures inside the zero redshift snapshot of the Via Lactea II (VL-2) simulation. A halo finder is used to remove and isolate the subhalos found inside the high resolution dark matter halo of VL-2, and the potentials for both the main halo and all the subhalos are constructed individually using the self-consistent field (SCF) method. This allows us to make direct comparison of tidal streams between a smooth halo and a lumpy halo without assuming idealized profiles or triaxial fits. We simulate the kinematics of a star cluster starting with the same orbital position but two different velocities. Although these two orbits are only moderately eccentric and have similar apo- and pericentric distances, we find that the two streams have very different morphologies. We conclude that our model of the potential of VL-2 can provide insights about tidal streams that have not been explored by previous studies using idealized or axisymmetric models.

astro-ph.GA↗

arXiv2015

Simulating Deep Hubble Images With Semi-empirical Models of Galaxy Formation

We simulate deep images from the Hubble Space Telescope (HST) using semi-empirical models of galaxy formation with only a few basic assumptions and parameters. We project our simulations all the way to the observational domain, adding cosmological and instrumental effects to the images, and analyze them in the same way as real HST images ("forward modeling"). This is a powerful tool for testing and comparing galaxy evolution models, since it allows us to make unbiased comparisons between the predicted and observed distributions of galaxy properties, while automatically taking into account all relevant selection effects. Our semi-empirical models populate each dark matter halo with a galaxy of determined stellar mass and scale radius. We compute the luminosity and spectrum of each simulated galaxy from its evolving stellar mass using stellar population synthesis models. We calculate the intrinsic scatter in the stellar mass-halo mass relation that naturally results from enforcing a monotonically increasing stellar mass along the merger history of each halo. The simulated galaxy images are drawn from cutouts of real galaxies from the Sloan Digital Sky Survey, with sizes and fluxes rescaled to match those of the model galaxies. The distributions of galaxy luminosities, sizes, and surface brightnesses depend on the adjustable parameters in the models, and they agree well with observations for reasonable values of those parameters. Measured galaxy magnitudes and sizes have significant magnitude-dependent biases, with both being underestimated near the magnitude detection limit. The fraction of galaxies detected and fraction of light detected also depend sensitively on the details of the model.

astro-ph.GA↗

arXiv2015

FlashGraph: Processing Billion-Node Graphs on an Array of Commodity SSDs

Graph analysis performs many random reads and writes, thus, these workloads are typically performed in memory. Traditionally, analyzing large graphs requires a cluster of machines so the aggregate memory exceeds the graph size. We demonstrate that a multicore server can process graphs with billions of vertices and hundreds of billions of edges, utilizing commodity SSDs with minimal performance loss. We do so by implementing a graph-processing engine on top of a user-space SSD file system designed for high IOPS and extreme parallelism. Our semi-external memory graph engine called FlashGraph stores vertex state in memory and edge lists on SSDs. It hides latency by overlapping computation with I/O. To save I/O bandwidth, FlashGraph only accesses edge lists requested by applications from SSDs; to increase I/O throughput and reduce CPU overhead for I/O, it conservatively merges I/O requests. These designs maximize performance for applications with different I/O characteristics. FlashGraph exposes a general and flexible vertex-centric programming interface that can express a wide variety of graph algorithms and their optimizations. We demonstrate that FlashGraph in semi-external memory performs many algorithms with performance up to 80% of its in-memory implementation and significantly outperforms PowerGraph, a popular distributed in-memory graph engine.

cs.DC↗

arXiv2014

Nonlinear Behavior of Baryon Acoustic Oscillations in Redshift Space from the Zel'dovich Approximation

Baryon acoustic oscillations (BAO) are a powerful probe of the expansion history of the universe, which can tell us about the nature of dark energy. In order to accurately characterize the dark energy equation of state using BAO, we must understand the effects of both nonlinearities and redshift space distortions on the location and shape of the acoustic peak. In a previous paper we introduced a novel approach to 2nd order perturbation theory in configuration space using the Zel'dovich approximation, and presented a simple result for the first nonlinear term of the correlation function. In this paper, we extend this approach to redshift space. We show how perform the computation, and present the analytic result for the first nonlinear term in the correlation function. Finally, we validate our result through comparison to numerical simulations.

astro-ph.CO↗

arXiv2014

Efficient classification of billions of points into complex geographic regions using hierarchical triangular mesh

We present a case study about the spatial indexing and regional classification of billions of geographic coordinates from geo-tagged social network data using Hierarchical Triangular Mesh (HTM) implemented for Microsoft SQL Server. Due to the lack of certain features of the HTM library, we use it in conjunction with the GIS functions of SQL Server to significantly increase the efficiency of pre-filtering of spatial filter and join queries. For example, we implemented a new algorithm to compute the HTM tessellation of complex geographic regions and precomputed the intersections of HTM triangles and geographic regions for faster false-positive filtering. With full control over the index structure, HTM-based pre-filtering of simple containment searches outperforms SQL Server spatial indices by a factor of ten and HTM-based spatial joins run about a hundred times faster.

cs.DB↗

arXiv2014

CANDELS/GOODS-S, CDFS, ECDFS: Photometric Redshifts For Normal and for X-Ray-Detected Galaxies

We present photometric redshifts and associated probability distributions for all detected sources in the Extended Chandra Deep Field South (ECDFS). The work makes use of the most up-to-date data from the Cosmic Assembly Near-IR Deep Legacy Survey (CANDELS) and the Taiwan ECDFS Near-Infrared Survey (TENIS) in addition to other data. We also revisit multi-wavelength counterparts for published X-ray sources from the 4Ms-CDFS and 250ks-ECDFS surveys, finding reliable counterparts for 1207 out of 1259 sources ($\sim 96\%$). Data used for photometric redshifts include intermediate-band photometry deblended using the TFIT method, which is used for the first time in this work. Photometric redshifts for X-ray source counterparts are based on a new library of AGN/galaxy hybrid templates appropriate for the faint X-ray population in the CDFS. Photometric redshift accuracy for normal galaxies is 0.010 and for X-ray sources is 0.014, and outlier fractions are $4\%$ and $5.4\%$ respectively. The results within the CANDELS coverage area are even better as demonstrated both by spectroscopic comparison and by galaxy-pair statistics. Intermediate-band photometry, even if shallow, is valuable when combined with deep broad-band photometry. For best accuracy, templates must include emission lines.

astro-ph.GA↗

arXiv2014

The redshift-space galaxy two-point correlation function and baryon acoustic oscillations

Future galaxy surveys will measure baryon acoustic oscillations (BAOs) with high significance, and a complete understanding of the anisotropies of BAOs in redshift space will be important to exploit the cosmological information in BAOs. Here we describe the anisotropies that arise in the redshift-space galaxy two-point correlation function (2PCF) and elucidate the origin of features that arise in the dependence of the BAOs on the angle between the orientation of the galaxy pair and the line of sight. We do so with a derivation of the configuration-space 2PCF using streaming model. We find that, contrary to common belief, the locations of BAO peaks in the redshift-space 2PCF are anisotropic even in the linear theory. Anisotropies in BAO depend strongly on the method of extracting the peak, showing maximum 3 % angular variation. We also find that extracting the BAO peak of $r^2ξ(r,μ)$ significantly reduces the anisotropy to sub-percent level angular variation. When subtracting the tilt due to the broadband behavior of the 2PCF, the BAO bump is enhanced along the line of sight because of local infall velocities toward the BAO bump. Precise measurement of the angular dependence of the redshift-space 2PCF will allow new geometrical tests of dark energy beyond the BAO.

astro-ph.CO↗

arXiv2014

The Dark Matter Contribution to Galactic Diffuse Gamma Ray Emission

Observations of diffuse Galactic gamma ray emission (DGE) by the Fermi Large Area Telescope (LAT) allow a detailed study of cosmic rays and the interstellar medium. However, diffuse emission models of the inner Galaxy underpredict the Fermi-LAT data at energies above a few GeV and hint at possible non-astrophysical sources including dark matter (DM) annihilations or decays. We present a study of the possible emission components from DM using the high-resolution Via Lactea II N-body simulation of a Milky Way-sized DM halo. We generate full-sky maps of DM annihilation and decay signals that include modeling of the adiabatic contraction of the host density profile, Sommerfeld enhanced DM annihilations, $p$-wave annihilations, and decaying DM. We compare our results with the DGE models produced by the Fermi-LAT team over different sky regions, including the Galactic center, high Galactic latitudes, and the Galactic anti-center. This work provides possible templates to fit the observational data that includes the contribution of the subhalo population to DM gamma-ray emission, with the significance depending on the annihilation/decay channels and the Galactic regions being considered.

astro-ph.CO↗

arXiv2014

Efficient Catalog Matching with Dropout Detection

Not only source catalogs are extracted from astronomy observations. Their sky coverage is always carefully recorded and used in statistical analyses, such as correlation and luminosity function studies. Here we present a novel method for catalog matching, which inherently builds on the coverage information for better performance and completeness. A modified version of the Zones Algorithm is introduced for matching partially overlapping observations, where irrelevant parts of the data are excluded up front for efficiency. Our design enables searches to focus on specific areas on the sky to further speed up the process. Another important advantage of the new method over traditional techniques is its ability to quickly detect dropouts, i.e., the missing components that are in the observed regions of the celestial sphere but did not reach the detection limit in some observations. These often provide invaluable insight into the spectral energy distribution of the matched sources but rarely available in traditional associations.

astro-ph.IM↗

arXiv2013

The AGORA High-Resolution Galaxy Simulations Comparison Project

We introduce the AGORA project, a comprehensive numerical study of well-resolved galaxies within the LCDM cosmology. Cosmological hydrodynamic simulations with force resolutions of ~100 proper pc or better will be run with a variety of code platforms to follow the hierarchical growth, star formation history, morphological transformation, and the cycle of baryons in and out of 8 galaxies with halo masses M_vir ~= 1e10, 1e11, 1e12, and 1e13 Msun at z=0 and two different ("violent" and "quiescent") assembly histories. The numerical techniques and implementations used in this project include the smoothed particle hydrodynamics codes GADGET and GASOLINE, and the adaptive mesh refinement codes ART, ENZO, and RAMSES. The codes will share common initial conditions and common astrophysics packages including UV background, metal-dependent radiative cooling, metal and energy yields of supernovae, and stellar initial mass function. These are described in detail in the present paper. Subgrid star formation and feedback prescriptions will be tuned to provide a realistic interstellar and circumgalactic medium using a non-cosmological disk galaxy simulation. Cosmological runs will be systematically compared with each other using a common analysis toolkit, and validated against observations to verify that the solutions are robust - i.e., that the astrophysical assumptions are responsible for any success, rather than artifacts of particular implementations. The goals of the AGORA project are, broadly speaking, to raise the realism and predictive power of galaxy simulations and the understanding of the feedback processes that regulate galaxy "metabolism." The proof-of-concept dark matter-only test of the formation of a galactic halo with a z=0 mass of M_vir ~= 1.7e11 Msun by 9 different versions of the participating codes is also presented to validate the infrastructure of the project.

astro-ph.GA↗

arXiv2013

Graywulf: A platform for federated scientific databases and services

Many fields of science rely on relational database management systems to analyze, publish and share data. Since RDBMS are originally designed for, and their development directions are primarily driven by, business use cases they often lack features very important for scientific applications. Horizontal scalability is probably the most important missing feature which makes it challenging to adapt traditional relational database systems to the ever growing data sizes. Due to the limited support of array data types and metadata management, successful application of RDBMS in science usually requires the development of custom extensions. While some of these extensions are specific to the field of science, the majority of them could easily be generalized and reused in other disciplines. With the Graywulf project we intend to target several goals. We are building a generic platform that offers reusable components for efficient storage, transformation, statistical analysis and presentation of scientific data stored in Microsoft SQL Server. Graywulf also addresses the distributed computational issues arising from current RDBMS technologies. The current version supports load balancing of simple queries and parallel execution of partitioned queries over a set of mirrored databases. Uniform user access to the data is provided through a web based query interface and a data surface for software clients. Queries are formulated in a slightly modified syntax of SQL that offers a transparent view of the distributed data. The software library consists of several components that can be reused to develop complex scientific data warehouses: a system registry, administration tools to manage entire database server clusters, a sophisticated workflow execution framework, and a SQL parser library.

cs.DB↗

arXiv2013

The Open Connectome Project Data Cluster: Scalable Analysis and Vision for High-Throughput Neuroscience

We describe a scalable database cluster for the spatial analysis and annotation of high-throughput brain imaging data, initially for 3-d electron microscopy image stacks, but for time-series and multi-channel data as well. The system was designed primarily for workloads that build connectomes---neural connectivity maps of the brain---using the parallel execution of computer vision algorithms on high-performance compute clusters. These services and open-science data sets are publicly available at http://openconnecto.me. The system design inherits much from NoSQL scale-out and data-intensive computing architectures. We distribute data to cluster nodes by partitioning a spatial index. We direct I/O to different systems---reads to parallel disk arrays and writes to solid-state storage---to avoid I/O interference and maximize throughput. All programming interfaces are RESTful Web services, which are simple and stateless, improving scalability and usability. We include a performance evaluation of the production system, highlighting the effectiveness of spatial data organization.

cs.DC↗