Search NASA⌕ Search

Engineering topics

Chang, Philip

Publications and source records attributed to Chang, Philip.

At least 19 records

Deployment of inference as a service at the US CMS Tier-2 data centers

Coprocessors, especially GPUs, will be a vital ingredient of data production workflows at the HL-LHC. At CMS, the GPU-as-a-service approach for production workflows is implemented by the SONIC project (Services for Optimized Network Inference on Coprocessors). SONIC provides a mechanism for outsourcing computationally demanding algorithms, such as neural network inference, to remote servers, where requests from multiple clients are intelligently distributed across multiple GPUs by a load-balancing service. This talk highlights the recent progress in deploying SONIC at selected U.S. CMS Tier-2 data centers. Using realistic CMS Run3 data processing workflows, such as those containing transformer-based algorithms, we demonstrate how SONIC is integrated into the production-like environment to enable accelerated inference offloading. We will present developments from both the client and server sides, including production job and data center configurations for NVIDIA and AMD GPUs. We will also present performance scaling benchmarks and discuss the challenges of operating SONIC in CMS production, such as server discovery, GPU saturation, fallback server logic, etc.

Holzman, Burt↗

Snowmass2021 Cosmic Frontier White Paper: Observational Facilities to Study Dark Matter

We present an overview of future observational facilities that will significantly enhance our understanding of the fundamental nature of dark matter. These facilities span a range of observational techniques including optical/near-infrared imaging and spectroscopy, measurements of the cosmic microwave background, pulsar timing, 21-cm observations of neutral hydrogen at high redshift, and the measurement of gravitational waves. Such facilities are a critical component of a multi-pronged experimental program to uncover the nature of dark matter, while often providing complementary measurements of dark energy, neutrino physics, and inflation.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

A Roadmap for HEP Software and Computing R&D for the 2020s

Particle physics has an ambitious and broad experimental programme for the coming decades. This programme requires large investments in detector hardware, either to build new facilities and experiments, or to upgrade existing ones. Similarly, it requires commensurate investment in the R&D of software to acquire, manage, process, and analyse the shear amounts of data to be recorded. In planning for the HL-LHC in particular, it is critical that all of the collaborating stakeholders agree on the software goals and priorities, and that the efforts complement each other. In this spirit, this white paper describes the R&D activities required to prepare for this software upgrade.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Importance of Resolving the Spectral Support of Beam-plasma Instabilities in Simulations

Many astrophysical plasmas are prone to beam-plasma instabilities. For relativistic and dilute beams, the spectral support of the beam-plasma instabilities is narrow, i.e., the linearly unstable modes that grow with rates comparable to the maximum growth rate occupy a narrow range of wavenumbers. This places stringent requirements on the box-sizes when simulating the evolution of the instabilities. We identify the implied lower limits on the box size imposed by the longitudinal beam plasma instability, i.e., typically the most stringent condition required to correctly capture the linear evolution of the instabilities in multidimensional simulations. We find that sizes many orders of magnitude larger than the resonant wavelength are typically required. Using one-dimensional particle-in-cell simulations, we show that the failure to sufficiently resolve the spectral support of the longitudinal instability yields slower growth and lower levels of saturation, potentially leading to erroneous physical conclusion.

79 ASTRONOMY AND ASTROPHYSICS↗

Discovery of a Group of Receding, Variable Halo Stars toward Norma

We present results from spectroscopic observations of a trio of Cepheid candidates identified from K {sub s} -band light curves toward Norma. The spectra show that these stars are moving with a large and similar radial velocity—the heliocentric velocities are 171 ± 32 km s{sup −1}, 164 ± 37 km s{sup −1}, and 173 ± 20 km s{sup −1}. The average radial velocity is ∼169 km s{sup −1}, which is large and distinct from typical stars in the Galaxy’s stellar disk. Given the radial velocities and associated 1 σ error, we find that the combined probability that these three stars are foreground Milky Way disk stars is ∼7 × 10{sup −4}%, and the probability that these are large-amplitude spotted stars in a binary is ∼10{sup −5}%. These objects at l ∼ 333° and b ∼ −1° are therefore associated with the stellar halo. The identification of these sources as Type I Cepheids is not certain, and thus the distances of these sources are not yet well established. Assuming the 3.6 μ m period–luminosity relation of Type I Cepheids gives a distance of ∼78 kpc for these sources.

79 ASTRONOMY AND ASTROPHYSICS↗

SHARP: A Spatially Higher-order, Relativistic Particle-in-cell Code

Numerical heating in particle-in-cell (PIC) codes currently precludes the accurate simulation of cold, relativistic plasma over long periods, severely limiting their applications in astrophysical environments. We present a spatially higher-order accurate relativistic PIC algorithm in one spatial dimension, which conserves charge and momentum exactly. We utilize the smoothness implied by the usage of higher-order interpolation functions to achieve a spatially higher-order accurate algorithm (up to the fifth order). We validate our algorithm against several test problems—thermal stability of stationary plasma, stability of linear plasma waves, and two-stream instability in the relativistic and non-relativistic regimes. Comparing our simulations to exact solutions of the dispersion relations, we demonstrate that SHARP can quantitatively reproduce important kinetic features of the linear regime. Our simulations have a superior ability to control energy non-conservation and avoid numerical heating in comparison to common second-order schemes. We provide a natural definition for convergence of a general PIC algorithm: the complement of physical modes captured by the simulation, i.e., those that lie above the Poisson noise, must grow commensurately with the resolution. This implies that it is necessary to simultaneously increase the number of particles per cell and decrease the cell size. We demonstrate that traditional ways for testing for convergence fail, leading to plateauing of the energy error. This new PIC code enables us to faithfully study the long-term evolution of plasma problems that require absolute control of the energy and momentum conservation.

79 ASTRONOMY AND ASTROPHYSICS↗

A Semi-analytic Criterion for the Spontaneous Initiation of Carbon Detonations in White Dwarfs

Despite over 40 years of active research, the nature of the white dwarf progenitors of SNe Ia remains unclear. However, in the last decade, various progenitor scenarios have highlighted the need for detonations to be the primary mechanism by which these white dwarfs are consumed, but it is unclear how these detonations are triggered. In this paper we study how detonations are spontaneously initiated due to temperature inhomogeneities, e.g., hotspots, in burning nuclear fuel in a simplified physical scenario. Following the earlier work by Zel’Dovich, we describe the physics of detonation initiation in terms of the comparison between the spontaneous wave speed and the Chapman–Jouguet speed. We develop an analytic expression for the spontaneous wave speed and utilize it to determine a semi-analytic criterion for the minimum size of a hotspot with a linear temperature gradient between a peak and base temperature for which detonations in burning carbon–oxygen material can occur. Our results suggest that spontaneous detonations may easily form under a diverse range of conditions, likely allowing a number of progenitor scenarios to initiate detonations that burn up the star.

79 ASTRONOMY AND ASTROPHYSICS↗

The linear instability of dilute ultrarelativistic e{sup ±} pair beams

The annihilation of TeV photons from extragalactic TeV sources and the extragalactic background light produces ultrarelativistic e{sup ±} beams, which are subject to powerful plasma instabilities that sap their kinetic energy. Here we study the linear phase of the plasma instabilities that these pair beams drive. To this end, we calculate the linear growth rate of the beam-plasma and oblique instability in the electrostatic approximation in both the reactive and kinetic regimes, assuming a Maxwell–Jüttner distribution for the pair beam. We reproduce the well-known reactive and kinetic growth rates for both the beam-plasma and oblique mode. We demonstrate for the oblique instability that there is a broad spectrum of unstable modes that grow at the maximum rate for a wide range of beam temperatures and wave-vector orientations relative to the beam. We also delineate the conditions for applicability for the reactive and kinetic regimes and find that the beam-plasma mode transitions to the reactive regime at a lower Lorentz factor than the oblique mode due to a combination of their different scalings and the anisotropy of the velocity dispersions. Applying these results to the ultrarelativistic e{sup ±} beams from TeV blazars, we confirm that these beams are unstable to both the kinetic oblique mode and the reactive beam-plasma mode. These results are important in understanding how powerful plasma instabilities may sap the energy of the ultrarelativistic e{sup ±} beams as they propagate through intergalactic space.

79 ASTRONOMY AND ASTROPHYSICS↗

BOW TIES IN THE SKY. I. THE ANGULAR STRUCTURE OF INVERSE COMPTON GAMMA-RAY HALOS IN THE FERMI SKY

Extended inverse Compton halos are generally anticipated around extragalactic sources of gamma rays with energies above 100 GeV. These result from inverse Compton scattered cosmic microwave background photons by a population of high-energy electron/positron pairs produced by the annihilation of the high-energy gamma rays on the infrared background. Despite the observed attenuation of the high-energy gamma rays, the halo emission has yet to be directly detected. Here, we demonstrate that in most cases these halos are expected to be highly anisotropic, distributing the upscattered gamma rays along axes defined either by the radio jets of the sources or oriented perpendicular to a global magnetic field. We present a pedagogical derivation of the angular structure in the inverse Compton halo and provide an analytic formalism that facilitates the generation of mock images. We discuss exploiting this fact for the purpose of detecting gamma-ray halos in a set of companion papers.

79 ASTRONOMY AND ASTROPHYSICS↗

PATCHY BLAZAR HEATING: DIVERSIFYING THE THERMAL HISTORY OF THE INTERGALACTIC MEDIUM

TeV-blazars potentially heat the intergalactic medium (IGM) as their gamma rays interact with photons of the extragalactic background light to produce electron–positron pairs, which lose their kinetic energy to the surrounding medium through plasma instabilities. This results in a heating mechanism that is only weakly sensitive to the local density, and therefore approximately spatially uniform, naturally producing an inverted temperature–density relation in underdense regions. In this paper we go beyond the approximation of uniform heating and quantify the heating rate fluctuations due to the clustered distribution of blazars and how this impacts the thermal history of the IGM. We analytically compute a filtering function that relates the heating rate fluctuations to the underlying dark matter density field. We implement it in the cosmological code GADGET-3 and perform large-scale simulations to determine the impact of inhomogeneous heating. We show that because of blazar clustering, blazar heating is inhomogeneous for z ≳ 2. At high redshift, the temperature–density relation shows an important scatter and presents a low temperature envelope of unheated regions, in particular at low densities and within voids. However, the median temperature of the IGM is close to that in the uniform case, albeit slightly lower at low redshift. We find that blazar heating is more complex than initially assumed and that the temperature–density relation is not unique. Our analytic model for the heating rate fluctuations couples well with large-scale simulations and provides a cost-effective alternative to subgrid models.

79 ASTRONOMY AND ASTROPHYSICS↗

Star formation in self-gravitating turbulent fluids

We present a model of star formation in self-gravitating turbulent gas in which the turbulent velocity, v{sub T}, is a dynamical variable and is adiabatically heated by the collapse. The theory predicts the run of density, infall, and turbulent velocity and the rate of star formation in compact massive clouds. The adiabatic heating ensures that the turbulent pressure is dynamically important at all radii. The system evolves toward a coherent spatial structure with a fixed run of density, ρ(r,t)→ρ(r); mass flows through this structure onto the central star or star cluster. We define the sphere of influence of the accreted matter by m{sub ∗}=M{sub g}(r{sub ∗}), where m{sub *} is the stellar plus disk mass in the nascent star cluster and M{sub g}(r) is the gas mass inside radius r. Both v{sub T} and the infall velocity, |u{sub r}|, decrease with decreasing r for r>r{sub ∗}; v{sub T}(r)∼r{sup p}, the size–line-width relation, with p≈0.2−0.3, explaining the observation that Larson’s Law is altered in massive star-forming regions. The infall velocity is generally smaller than the turbulent velocity at r>r{sub ∗}. For r<r{sub ∗}, the infall and turbulent velocities are again similar, and both increase with decreasing r as r{sup −1/2}, with a magnitude about half of the free-fall velocity. The accreted (stellar) mass grows superlinearly with time, M-dot {sub ∗}=ϕM{sub cl}(t/τ{sub ff}){sup 2}, with ϕ a dimensionless number somewhat less than unity, M{sub cl} the clump mass, and τ{sub ff} the free-fall time of the clump. We suggest that small values of p can be used as a tracer of convergent collapsing flows.

79 ASTRONOMY AND ASTROPHYSICS↗

TIME-VARYING DYNAMICAL STAR FORMATION RATE

We present numerical evidence of dynamic star formation in which the accreted stellar mass grows superlinearly with time, roughly as t {sup 2}. We perform simulations of star formation in self-gravitating hydrodynamic and magnetohydrodynamic turbulence that is continuously driven. By turning the self-gravity of the gas in the simulations on or off, we demonstrate that self-gravity is the dominant physical effect setting the mass accretion rate at early times before feedback effects take over, contrary to theories of turbulence-regulated star formation. We find that gravitational collapse steepens the density profile around stars, generating the power-law tail on what is otherwise a lognormal density probability distribution function. Furthermore, we find turbulent velocity profiles to flatten inside collapsing regions, altering the size-line width relation. This local flattening reflects enhancements of turbulent velocity on small scales, as verified by changes to the velocity power spectra. Our results indicate that gas self-gravity dynamically alters both density and velocity structures in clouds, giving rise to a time-varying star formation rate. We find that a substantial fraction of the gas that forms stars arrives via low-density flows, as opposed to accreting through high-density filaments.

79 ASTRONOMY AND ASTROPHYSICS↗

SUPPRESSION OF STAR FORMATION IN NGC 1266

NGC 1266 is a nearby lenticular galaxy that harbors a massive outflow of molecular gas powered by the mechanical energy of an active galactic nucleus (AGN). It has been speculated that such outflows hinder star formation (SF) in their host galaxies, providing a form of feedback to the process of galaxy formation. Previous studies, however, indicated that only jets from extremely rare, high-power quasars or radio galaxies could impart significant feedback on their hosts. Here we present detailed observations of the gas and dust continuum of NGC 1266 at millimeter wavelengths. Our observations show that molecular gas is being driven out of the nuclear region at M-dot {sub out}≈110 M{sub ⊙} yr{sup –1}, of which the vast majority cannot escape the nucleus. Only 2 M {sub ☉} yr{sup –1} is actually capable of escaping the galaxy. Most of the molecular gas that remains is very inefficient at forming stars. The far-infrared emission is dominated by an ultra-compact (≲ 50 pc) source that could either be powered by an AGN or by an ultra-compact starburst. The ratio of the SF surface density (Σ{sub SFR}) to the gas surface density (Σ{sub H{sub 2}}) indicates that SF is suppressed by a factor of ≈50 compared to normal star-forming galaxies if all gas is forming stars, and ≈150 for the outskirt (98%) dense molecular gas if the central region is powered by an ultra-compact starburst. The AGN-driven bulk outflow could account for this extreme suppression by hindering the fragmentation and gravitational collapse necessary to form stars through a process of turbulent injection. This result suggests that even relatively common, low-power AGNs are able to alter the evolution of their host galaxies as their black holes grow onto the M-σ relation.

79 ASTRONOMY AND ASTROPHYSICS↗

THE EFFECT OF NONLINEAR LANDAU DAMPING ON ULTRARELATIVISTIC BEAM PLASMA INSTABILITIES

Very high energy gamma-rays from extragalactic sources produce pairs from the extragalactic background light, yielding an electron-positron pair beam. This pair beam is unstable to various plasma instabilities, especially the ''oblique'' instability, which can be the dominant cooling mechanism for the beam. However, recently, it has been claimed that nonlinear Landau damping renders it physically irrelevant by reducing the effective damping rate to a low level. Here we show with numerical calculations that the effective damping rate is 8 × 10{sup –4} the growth rate of the linear instability, which is sufficient for the ''oblique'' instability to be the dominant cooling mechanism of these pair beams. In particular, we show that previous estimates of this rate ignored the exponential cutoff in the scattering amplitude at large wave numbers and assumed that the damping of scattered waves entirely depends on collisions, ignoring collisionless processes. We find that the total wave energy eventually grows to approximate equipartition with the beam by increasingly depositing energy into long-wavelength modes. As we have not included the effect of nonlinear wave-wave interactions on these long-wavelength modes, this scenario represents the ''worst case'' scenario for the oblique instability. As it continues to drain energy from the beam at a faster rate than other processes, we conclude that the ''oblique'' instability is sufficiently strong to make it the physically dominant cooling mechanism for high-energy pair beams in the intergalactic medium.

79 ASTRONOMY AND ASTROPHYSICS↗

Lower limits on the anisotropy of the extragalactic gamma-ray background implied by the 2FGL and 1FHL catalogs

In principle, the angular anisotropy in the extragalactic gamma-ray background (EGRB) places severe constraints upon putative populations of unresolved gamma-ray point sources. Existing estimates of the EGRB anisotropy have been constructed by excising known point sources, e.g., taken from the First or Two Year Fermi-LAT Source Catalog (1FGL or 2FGL, respectively) and statistically analyzing the residual gamma-ray sky maps. We perform an independent check of the EGRB anisotropy limits by comparing the values obtained from the 1FGL-masked sky maps to the signal implied by sources that lie below the 1FGL detection threshold in the more sensitive 2FGL and 1FHL (First Fermi-LAT catalog of >10 GeV sources). As such, our analysis provides an internal consistency check of implications for source counts and spectral index distributions of gamma-ray bright active galactic nuclei obtained from Fermi-LAT data. Based on this, we find evidence for substantially larger anisotropies than those previously reported at energies above 5 GeV, where BL Lac objects are likely to provide the bulk of their contribution to the EGRB. This uncertainty in the EGRB anisotropy cautions against using it as an independent constraint for the high-redshift gamma-ray universe. Moreover, this would suggest that contrary to previous claims, smooth extensions of the resolved point-source population may be able to simultaneously explain both the isotropic and anisotropic components of the EGRB.

79 ASTRONOMY AND ASTROPHYSICS↗

Implications of plasma beam instabilities for the statistics of the Fermi hard gamma-ray blazars and the origin of the extragalactic gamma-ray background

Fermi has been instrumental in constraining the luminosity function and redshift evolution of gamma-ray bright BL Lac objects, a subpopulation of blazars with almost featureless optical spectra. This includes limits on the spectrum and anisotropy of the extragalactic gamma-ray background (EGRB), redshift distribution of nearby Fermi active galactic nuclei (AGNs), and the construction of a logN-log S relation. Based on these, it has been argued that the evolution of the gamma-ray bright BL Lac population must be much less dramatic than that of other AGNs. However, critical to such claims is the assumption that inverse Compton cascades reprocess emission above a TeV into the Fermi energy range, substantially enhancing the strength of the observed limits. Here we demonstrate that in the absence of such a process, due, e.g., to the presence of virulent plasma beam instabilities that preempt the cascade, a population of TeV-bright BL Lac objects that evolve similarly to quasars is consistent with the population of hard gamma-ray BL Lac objects observed by Fermi. Specifically, we show that a simple model for the properties and luminosity function is simultaneously able to reproduce their logN-log S relation, local redshift distribution, and contribution to the EGRB and its anisotropy without any free parameters. Insofar as the naturalness of a picture in which the hard gamma-ray BL Lac population exhibits the strong redshift evolution observed in other tracers of the cosmological history of accretion onto halos is desirable, this lends support for the absence of the inverse Compton cascades and the existence of the beam plasma instabilities.

79 ASTRONOMY AND ASTROPHYSICS↗

NGC 1266 As a local candidate for rapid cessation of star formation

We present new Spectrographic Areal Unit for Research on Optical Nebulae (SAURON) integral-field spectroscopy and Swift Ultraviolet Optical Telescope (UVOT) observations of molecular outflow host galaxy NGC 1266 that indicate NGC 1266 has experienced a rapid cessation of star formation. Both the SAURON maps of stellar population age and the Swift UVOT observations demonstrate the presence of young (<1 Gyr) stellar populations within the central 1 kpc, while existing Combined Array for Research in Millimeter-Wave Astronomy CO(1-0) maps indicate that the sites of current star formation are constrained to only the inner few hundred parsecs of the galaxy. The optical spectrum of NGC 1266 from Moustakas and Kennicutt reveal a characteristic poststarburst (K+A) stellar population, and Davis et al. confirm that ionized gas emission in the system originate from a shock. Galaxies with K+A spectra and shock-like ionized gas line ratios may comprise an important, overlooked segment of the poststarburst population, containing exactly those objects in which the active galactic nucleus (AGN) is actively expelling the star-forming material. While AGN activity is not the likely driver of the poststarburst event that occurred 500 Myr ago, the faint spiral structure seen in the Hubble Space Telescope Wide-field Camera 3 Y-, J- and H-band imaging seems to point to the possibility of gravitational torques being the culprit. If the molecular gas were driven into the center at the same time as the larger scale galaxy disk underwent quenching, the AGN might be able to sustain the presence of molecular gas for ≳ 1 Gyr by cyclically injecting turbulent energy into the dense molecular gas via a radio jet, inhibiting star formation.

79 ASTRONOMY AND ASTROPHYSICS↗

A PARAMETER-SPACE STUDY OF CARBON-OXYGEN WHITE DWARF MERGERS

The merger of two carbon-oxygen white dwarfs can lead to either a spectacular transient, stable nuclear burning or a massive, rapidly rotating white dwarf. Simulations of mergers have shown that the outcome strongly depends on whether the white dwarfs are similar or dissimilar in mass. In the similar-mass case, both white dwarfs merge fully and the remnant is hot throughout, while in the dissimilar case, the more massive, denser white dwarf remains cold and essentially intact, with the disrupted lower mass one wrapped around it in a hot envelope and disk. In order to determine what constitutes ''similar in mass'' and more generally how the properties of the merger remnant depend on the input masses, we simulated unsynchronized carbon-oxygen white dwarf mergers for a large range of masses using smoothed-particle hydrodynamics. We find that the structure of the merger remnant varies smoothly as a function of the ratio of the central densities of the two white dwarfs. A density ratio of 0.6 approximately separates similar and dissimilar-mass mergers. Confirming previous work, we find that the temperatures of most merger remnants are not high enough to immediately ignite carbon fusion. During subsequent viscous evolution, however, the interior will likely be compressed and heated as the disk accretes and the remnant spins down. We find from simple estimates that this evolution can lead to ignition for many remnants. For similar-mass mergers, this would likely occur under sufficiently degenerate conditions that a thermonuclear runaway would ensue.

79 ASTRONOMY AND ASTROPHYSICS↗