Search NASA⌕ Search

SEARCH · Search NASA

Results for “distributed parallelization”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,747 records · Page 97

Scalable line and plane relaxation in a parallel structured multigrid solver

The efficient solution of sparse, linear systems that arise through the discretization of partial differential equations remains a key challenge for a range of high performance scientific simulations. One approach for reducing data movement and improving performance is by exposing and exploiting structure in a problem through the use of robust structured multilevel solvers. By choosing coarsening that preserves the structure of the problem, these methods maintain efficient structured computation and communication throughout the multigrid hierarchy. However, when coarsening is not permitted to be dependent on the operator, anisotropy must be addressed by the smoother — producing error compatible for coarse-grid correction with structured coarsening. Here, the components required in a scalable parallel structured solver are described with a focus on memory and communication efficiency of robust smoothers. While the implementation of communication and memory reduction techniques in smoothers integrated in a complete 3D solver present a significant engineering challenge, a novel approach is proposed that addresses these challenges systematically through a change to the solver’s execution model. Enabled by user-level threading paired with a set of data and communication abstractions, this approach permits seamless aggregation of communication in plane smoothers — directly reusing code for a 2D distributed multilevel cycle. Results show an effective reduction in communication costs for coarse-grid problems, and result in a speedup of 8.7x in smoothing routines shown in Fig. 12 using this approach. This produces a significant improvement to strong scalability while maintaining favorable weak scaling behavior. Finally, a parallel scaling study using a series of refined meshes is included that demonstrates the effectiveness of this approach in an application of interest.

97 MATHEMATICS AND COMPUTING↗

Pore-network model of evaporation-induced salt precipitation in porous media: The effect of correlations and heterogeneity

Salt transport and precipitation in porous media constitute a set of complex and fascinating phenomena that are of considerable interest to several important problems, ranging from storage of CO 2 in geological formations, to soil fertility, and protection of pavements and roads, as well as historical monuments. The phenomena occur at the pore scale and are greatly influenced by the heterogeneity of the pore space morphology. In this paper, we present a pore-network (PN) model to study the phenomena. Vapor diffusion, capillary effect at the brine-vapor interface, flow of brine, and transport of salt and its precipitation in the pores that plug the pores partially or completely are all accounted for. The drying process is modeled by the invasion percolation, while transport of salt in brine is accounted for by the convective-diffusion equation. We demonstrate that the drying patterns, the clustering and connectivity of the pore throats in which salt precipitation occurs, the saturation distribution, and the drying rate are all strongly dependent upon the pore-size distribution, the correlations among the pore sizes, and the anisotropy of the pore space caused by stratification that most natural porous media contain. In particular, if the strata are more or less parallel to the direction of injection of the gas that dries out the pore space (air, for example) and/or causes salt precipitation (CO 2 , for example), the drying rate increases significantly. Moreover, salt tends to precipitate in clusters of neighboring pores that are parallel to the open surface of the porous medium.

42 ENGINEERING↗

Mechanisms of Furfural Reduction on Metal Electrodes: Distinguishing Pathways for Selective Hydrogenation of Bioderived Oxygenates

Electrochemical reduction of biomass-derived platform molecules is an emerging route for the sustainable production of fuels and chemicals. Understanding gaps between reaction conditions, underlying mechanisms, and product selectivity have limited the rational design of active, stable, and selective catalyst systems. Here, the mechanisms of electrochemical reduction of furfural, an important biobased platform molecule and model for aldehyde reduction, are explored through a combination of voltammetry, preparative electrolysis, thiol-electrode modifications, and kinetic isotope studies. It is demonstrated that two distinct mechanisms are operable on metallic Cu electrodes in acidic electrolytes: (i) electrocatalytic hydrogenation (ECH) and (ii) direct electroreduction. The contributions of each mechanism to the observed product distribution are clarified by evaluating the requirement for direct chemical interactions with the electrode surface and the role of adsorbed hydrogen. Further analysis reveals that hydrogenation and hydrogenolysis products are generated by parallel ECH pathways. By understanding the underlying mechanisms it enables the manipulation of furfural reduction by rationally tuning the electrode potential, electrolyte pH, and furfural concentration to promote selective formation of important biobased polymer precursors and fuels.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Three-dimensional cross-field flows at the plasma-material interface in an oblique magnetic field

This article describes experimental evidence that the magnetic presheath is a fully three-dimensional structure modified by ion–neutral collisions. Velocity distributions of both ions and neutrals, obtained via laser-induced fluorescence, show that cross field ion drifts do not result from entrainment of ions in a flowing neutral background. Ion flows parallel to E x B arise and accelerate to as much as 0.2c s within several ion gyroradii of the boundary surface, where c s is the sound speed. Within measurement resolution, the onset of the E x B aligned flow occurs at the same distance to the surface that ions begin to deflect from travel along magnetic field lines. Collisional fluid and particle-in-cell simulations of the boundary region are compared to the experimental measurements. We find that, in contrast to the classical collisionless Chodura model, collisional effects between the ions and the non-flowing neutral population are essential to quantitatively predict the observed ion drift velocities. No momentum coupling between ions and neutrals, separable from noise and other effects, is observed in either signal. Here, we discuss several explanations and implications of this observation.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Quantum two-photon emission in a photonic cavity

In this paper, we derive an expression for the two-photon spontaneous emission (TPSE) rate of an excited quantum emitter in the presence of arbitrary bodies in its vicinities. After investigating the influence of a perfectly conducting plate on the TPSE spectral distribution (Purcell effect), we demonstrate the equivalence of our expression with the more usual formula written in terms of the corresponding dyadic Green's function. We establish a convenient relation between the TPSE spectral distribution and the corresponding Purcell factors of the system. Next, we consider an emitter close to a dielectric medium and show that, in the near-field regime, the TPSE spectral distribution is substantially enhanced and changes abruptly at the resonance frequencies. Finally, motivated by the suppression that may occur in the one-photon spontaneous emission of an excited atom between two parallel conducting plates, we discuss the TPSE for this same situation and show that complete suppression can never occur for s → s transitions.

74 ATOMIC AND MOLECULAR PHYSICS↗

Scaling and Quality of Modularity Optimization Methods for Graph Clustering

Real-world graphs exhibit structures known as “communities” or “clusters” consisting of a group of vertices with relatively high connectivity between them, as compared to the rest of the vertices in the network. Graph clustering or community detection is a fundamental graph operation used to analyze real-world graphs occurring in the areas of computational biology, cybersecurity, electrical grids, etc. Similar to other graph algorithms, owing to irregular memory accesses and inherently sequential nature, current algorithms for community detection are challenging to parallelize. However, in order to analyze large networks, it is important to develop scalable parallel implementations of graph clustering that are capable of exploiting the architectural features of modern supercomputers. In response to the 2019 Streaming Graph Challenge, we present quality and performance analysis of our distributed- memory community detection using Vite, which is our distributed memory implementation of the popular Louvain method, on the ALCF Theta supercomputer. Clustering methods such as Louvain that rely on modularity maximization are known to suffer from the resolution limit problem, preventing identification of clusters of certain sizes. Hence, we also include quality analysis of our shared-memory implementation of the Fast-tracking Resistance method, in comparison with Louvain on the challenge datasets. Furthermore, we introduce an edge-balanced graph distribution for our distributed memory implementation, that significantly reduces communication, offering up to 80% improvement in the overall execution time. In addition to performance/quality analysis, we also include details on the power/energy consumption, and memory traffic of the distributed-memory clustering implementation using real-world graphs with over a billion edges.

Ghosh, Sayan↗

SU-E-J-164: An Investigation of a Low-Cost ‘dry’ Optical-CT Scanning System for 3D Dosimetry

Purpose: To characterize and explore the efficacy of a novel low-cost, lowfluid, broad-beam optical-CT system for 3D-dosimetry in radiochromic Presage dosimeters. Leading current optical-CT systems incorporate expensive glass-based telecentric lens technology, and a fluid bath with substantial amounts of fluid (which introduces an inconvenience factor) to minimize refraction artifacts. Here we introduce a novel system which addresses both these limitations by: (1) the use of Fresnel lenses in a telecentric arrangement, and (2) a ‘solid’ fluid bath which dramatically reduces the amount of fluid required for refractive-index (RI) matching. Materials Methods: A fresnel based telecentric optical-CT system was constructed which expands light from a single red LED source into a nominally parallel beam into which a cubic ‘dry-tank’ is placed. The drytank consists of a solid polyurethane cube (with the same RI as Presage) but containing a cylindrical cavity (11.5cm diameter × 11cm ) into which the dosimeter is placed for imaging. A narrow (1-3mm) gap between the walls of the dosimeter and dry-tank is filled with a fluid of similar RI. This arrangement reduces the amount of RI fluid from about 1000cc to 75cc, yielding substantial practical benefit in convenience and cost. The new system was evaluated in direct comparison against Eclipse planning system from a 4-field parallel-opposed treatmen Results: Gamma calculations of dose from DFOS-dry system versus Eclipse showed 92% and 97% agreement with 4mm/4% and 5mm/5% criteria, respectively, in the central 80% of dose distribution. Reconstructions showed some edge artifacts, as well as some dose underestimation towards the dosimeter edge. Conclusion: The implementation of Fresnel based ‘dry’ optical-CT for 3Ddosimetry would represent an important advance enhancing costeffectiveness and practical viability. The performance of the prototype presented here is not yet comparable to the state-of-the-art, but shows sufficient promise for further investigation to elevate image quality to match gold-standard optical-CT systems. This work was supported by NIH R01CA100835.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Percolator: Scalable Pattern Discovery in Dynamic Graphs

We demonstrate Percolator, a distributed system for graph pattern discovery in dynamic graphs. In contrast to conventional mining systems, Percolator advocates efficient pattern mining schemes that (1) support pattern detection with keywords; (2) integrate incremental and parallel pattern mining; and (3) support analytical queries such as trend analysis. The core idea of Percolator is to dynamically decide and verify a small fraction of patterns and their in- stances that must be inspected in response to buffered updates in dynamic graphs, with a total mining cost independent of graph size. We demonstrate a) the feasibility of incremental pattern mining by walking through each component of Percolator, b) the efficiency and scalability of Percolator over the sheer size of real-world dynamic graphs, and c) how the user-friendly GUI of Percolator inter- acts with users to support keyword-based queries that detect, browse and inspect trending patterns. We also demonstrate two user cases of Percolator, in social media trend analysis and academic collaboration analysis, respectively.

Choudhury, Sutanay↗

Interplay of Electrode Heterogeneity and Lithium Plating

Lateral microstructure heterogeneity in anodes is known to induce nonuniform current density, state of charge, and lithium plating. This means that such electrode heterogeneity can limit the fast charging of lithium-ion batteries. In this work, a combination of experiments and simulation is employed to understand the effect of mm scale lateral heterogeneity on cell aging. A previously developed model was extended to efficiently simulate SEI formation and Li plating for independent regions of an electrode. The model consists of three parallel regions each described under a P2D framework and with a distinct ionic resistance and possibly active material loading. The results suggest that during fast charge when the active material is uniformly distributed across the three regions, the region with the highest resistance reaches the end of life sooner than the other regions. There is also positive feedback from Li metal filling the pores near the separator interface that further accelerates lithium plating. Lastly, when there is a non-uniform active material distribution associated with the ionic resistance heterogeneity, tight competition between regions can occur, leading to less overall lithium plating and plating that is more uniform between regions.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

CORRLA-RS

The CORRLA-RS package provides a suite of statistical methods for sampling multidimensional distributions and to conduct sensitivity and correlation analysis of large scale data in the Rust programming language. The software provides a unique solution to multidimensional constrained sampling problems utilizing a combination of parallelized Markov Chain Monte Carlo methods and traditional rejection sampling. The sensitivity and correlation analysis methods are backed by a high performance randomized singular value decomposition implementation which enables datasets larger than the random access memory (RAM) size to be analyzed. Additionally, CORRLA-RS implements the active subspace identification method using a KD-Tree and the randomized singular value decomposition acting in concert.

Gurecky, William [Oak Ridge National Laboratory (O↗

Efficient First-Order Algorithms for Large-Scale, Non-Smooth Maximum Entropy Models with Application to Wildfire Science

Maximum entropy (MaxEnt) models are a class of statistical models that use the maximum entropy principle to estimate probability distributions from data. Due to the size of modern data sets, MaxEnt models need efficient optimization algorithms to scale well for big data applications. State-of-the-art algorithms for MaxEnt models, however, were not originally designed to handle big data sets; these algorithms either rely on technical devices that may yield unreliable numerical results, scale poorly, or require smoothness assumptions that many practical MaxEnt models lack. In this paper, we present novel optimization algorithms that overcome the shortcomings of state-of-the-art algorithms for training large-scale, non-smooth MaxEnt models. Our proposed first-order algorithms leverage the Kullback–Leibler divergence to train large-scale and non-smooth MaxEnt models efficiently. For MaxEnt models with discrete probability distribution of n elements built from samples, each containing m features, the stepsize parameter estimation and iterations in our algorithms scale on the order of O(mn) operations and can be trivially parallelized. Moreover, the strong ℓ1 convexity of the Kullback–Leibler divergence allows for larger stepsize parameters, thereby speeding up the convergence rate of our algorithms. To illustrate the efficiency of our novel algorithms, we consider the problem of estimating probabilities of fire occurrences as a function of ecological features in the Western US MTBS-Interagency wildfire data set. Our numerical results show that our algorithms outperform the state of the art by one order of magnitude and yield results that agree with physical models of wildfire occurrence and previous statistical analyses of wildfire drivers.

Physics↗

The Peculiar Filamentary H i Structure of NGC 6145

In this paper, we report the peculiar H i morphology of the cluster spiral galaxy NGC 6145, which has a 150 kpc H i filament on one side that is nearly parallel to its major axis. This filament is made up of several H i clouds and the diffuse H i gas between them, with no optical counterparts. We compare its H i distribution with other one-sided H i distributions in the literature and find that the overall H i distribution is very different from the typical tidal and ram-pressure stripped H i shape, and that its morphology is inconsistent with that of a pure accretion event. Only ∼30% of the total H i gas is anchored on the stellar disk, while most of the H i gas forms the filament in the west. At a projected distance of 122 kpc, we find a massive elliptical companion (NGC 6146) with extended radio emission whose axis points to an H i gap in NGC 6145. The velocity of the H i filament shows an overall line-of-sight motion of 80–180 km s{sup −1} with respect to NGC 6145. Using the long-slit spectra of NGC 6145 along its major stellar axis, we find that some outer regions show enhanced star formation, while in contrast, almost no star formation activities are found in its center (<2 kpc). Pure accretion, tidal, or ram-pressure stripping are not likely to produce the observed H i filament. An alternative explanation is the jet stripping from NGC 6146, although direct evidence for a jet-cold gas interaction has not been found.

79 ASTRONOMY AND ASTROPHYSICS↗

ADAPT

The purpose of the ADAPT code is to generate Dynamic Event Trees (DET) using a user specified set of simulators. ADAPT can utilize any simulation tool which meets a minimal set of requirements. ADAPT is based on the concept of DET which uses explicit modeling of the deterministic dynamic processes that take place during a nuclear reactor plant system (or other complex system) evolution along with stochastic modeling. When DET are used to model various aspects of Probabilistic Risk Assessment (PRA), all accident progression scenarios starting from an initiating event are considered simultaneously. The DET branching occurs at user specified times and/or when an action is required by the system and/or the operator. These outcomes then decide how the dynamic system variables will evolve in time for each DET branch. Since two different outcomes at a DET branching may lead to completely different paths for system evolution, the next branching for these paths may occur not only at separate times, but can be based on different branching criteria. The computational infrastructure allows for flexibility in ADAPT to link with different system simulation codes, parallel processing of the scenarios under consideration, on-line scenario management (initiation as well as termination), analysis of results, and user friendly graphical capabilities. The ADAPT system is designed for a distributed computing environment; the scheduler can track multiple concurrent branches simultaneously. The scheduler is modularized so that the DET branching strategy can be modified (e.g. biasing towards the worst-case scenario/event). Independent database systems store data from the simulation tasks and the DET structure so that the event tree can be constructed and analyzed later. ADAPT is provided with a user-friendly client which can easily sort through and display the results of an experiment, precluding the need for the user to manually inspect individual simulator runs.

Reynolds, John↗

Moving multiple sinks through wireless sensor networks for lifetime maximization.

Unattended sensor networks typically watch for some phenomena such as volcanic events, forest fires, pollution, or movements in animal populations. Sensors report to a collection point periodically or when they observe reportable events. When sensors are too far from the collection point to communicate directly, other sensors relay messages for them. If the collection point location is static, sensor nodes that are closer to the collection point relay far more messages than those on the periphery. Assuming all sensor nodes have roughly the same capabilities, those with high relay burden experience battery failure much faster than the rest of the network. However, since their death disconnects the live nodes from the collection point, the whole network is then dead. We consider the problem of moving a set of collectors (sinks) through a wireless sensor network to balance the energy used for relaying messages, maximizing the lifetime of the network. We show how to compute an upper bound on the lifetime for any instance using linear and integer programming. We present a centralized heuristic that produces sink movement schedules that produce network lifetimes within 1.4% of the upper bound for realistic settings. We also present a distributed heuristic that produces lifetimes at most 25:3% below the upper bound. More specifically, we formulate a linear program (LP) that is a relaxation of the scheduling problem. The variables are naturally continuous, but the LP relaxes some constraints. The LP has an exponential number of constraints, but we can satisfy them all by enforcing only a polynomial number using a separation algorithm. This separation algorithm is a p-median facility location problem, which we can solve efficiently in practice for huge instances using integer programming technology. This LP selects a set of good sensor configurations. Given the solution to the LP, we can find a feasible schedule by selecting a subset of these configurations, ordering them via a traveling salesman heuristic, and computing feasible transitions using matching algorithms. This algorithm assumes sinks can get a schedule from a central server or a leader sink. If the network owner prefers the sinks make independent decisions, they can use our distributed heuristic. In this heuristic, sinks maintain estimates of the energy distribution in the network and move greedily (with some coordination) based on local search. This application uses the new SUCASA (Solver Utility for Customization with Automatic Symbol Access) facility within the PICO (Parallel Integer and Combinatorial Optimizer) integer programming solver system. SUCASA allows rapid development of customized math programming (search-based) solvers using a problem's natural multidimensional representation. In this case, SUCASA also significantly improves runtime compared to implementations in the ampl math programming language or in perl.

54 ENVIRONMENTAL SCIENCES↗

Photoelectron holographic interferences from multiple returning in strong-field tunneling ionization

By numerically solving the time-dependent Schrödinger equation, we theoretically investigate the holographic structure of the multiple-returning rescattering recorded in the photoelectron momentum distribution. In the orthogonally polarized two-color fields, we show that the interference patterns periodically oscillate with the relative phase. Tracing the relative phase where the oscillation minimizes, we find that the dependences of the phase on the parallel momentum for the first-returning and the second-returning rescattering holographic interferences are different. With the quantum-orbit model, we demonstrated that the difference is related to the ionization and rescattering times of the electron. Moreover, we show that the influence of the Coulomb interaction on the oscillation of both the first-returning rescattering and the second-returning rescattering holographic interferences can be ignored. This suggests us that the temporal characteristics of the electron can be resolved by analyzing the relative phase’s dependence of the multiple-returning holographic interference.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Internal structure of the nanogratings generated inside bulk fused silica by ultrafast laser direct writing

The aim of the present work was to characterize the internal structure of nanogratings generated inside bulk fused silica by ultrafast laser processing and to study the influence of diluted hydrofluoric acid etching on their structure. The nanogratings were inscribed at a depth of 100 μm within fused silica wafers by a direct writing method, using 1030 nm radiation wavelength and the following processing parameters: E = 5 μJ, τ = 560 fs, f = 10 kHz, and v = 100 μm/s. The results achieved show that the laser-affected regions are elongated ellipsoids with a typical major diameter of about 30 μm and a minor diameter of about 6 μm. The nanogratings within these regions are composed of alternating nanoplanes of damaged and undamaged material, with an average periodicity of 351 ± 21 nm. The damaged nanoplanes contain nanopores randomly dispersed in a material containing a large density of defects. These nanopores present a roughly bimodal size distribution with average dimensions for each class of pores 65 ± 20 × 16 ± 8 × 69 ± 16 nm{sup 3} and 367 ± 239 × 16 ± 8 × 360 ± 194 nm{sup 3}, respectively. The number and size of the nanopores increases drastically when an hydrofluoric acid treatment is performed, leading to the coalescence of these voids into large planar discontinuities parallel to the nanoplanes. The preferential etching of the damaged material by the hydrofluoric acid solution, which is responsible for the pores growth and coalescence, confirms its high defect density.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

A Novel Application of Parallel Betweenness Centrality to Power Grid Contingency Analysis

In Energy Management Systems, contingency analysis is commonly performed for identifying and mitigating potentially harmful power grid component failures. The exponentially increasing combinatorial number of failure modes imposes a significant computational burden for massive contingency analysis. It is critical to select a limited set of high-impact contingency cases within the constraint of computing power and time requirements to make it possible for real-time power system vulnerability assessment. In this paper, we present a novel application of parallel betweenness centrality to power grid contingency selection. We cross-validate the proposed method using the model and data of the western US power grid, and implement it on a Cray XMT system - a massively multithreaded architecture - leveraging its advantages for parallel execution of irregular algorithms, such as graph analysis. We achieve a speedup of 55 times (on 64 processors) compared against the single-processor version of the same code running on the Cray XMT. We also compare an OpenMP-based version of the same code running on an HP Superdome shared-memory machine. The performance of the Cray XMT code shows better scalability and resource utilization, and shorter execution time for large-scale power grids. This proposed approach has been evaluated in PNNL’s Electricity Infrastructure Operations Center (EIOC). It is expected to provide a quick and efficient solution to massive contingency selection problems to help power grid operators to identify and mitigate potential widespread cascading power grid failures in real time.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Grain in weakly ionized plasma in the presence of an external magnetic field: Charging by plasma currents and effective potential

The problem of grain screening is solved numerically for the case of weakly ionized plasma in the presence of an external magnetic field. The plasma dynamics is described within the drift-diffusion approximation under the assumption that the grain absorbs all encountered electrons and ions. We also assume that the plasma current through the grain surface is equal to zero in the stationary state. This condition is used to perform self-consistent calculations of the grain charge. The spatial distribution of the screened grain potential is studied and compared with the analytical estimates. It is shown that at the distances larger than the Debye length such potential has the Coulomb-like asymptotics with the effective charge dependent on the angle between the radius vector and the external magnetic field direction. The numerical solutions show that in the direction parallel to the external magnetic field the effective potential can have nonmonotonic behavior.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗