Search NASA⌕ Search

SEARCH · Search NASA

Results for “Parallelization efficiency”

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 1,567 records · Page 87

Plasma Dynamics and Nonthermal Particle Acceleration in 3D Nonrelativistic Magnetic Reconnection

Understanding plasma dynamics and nonthermal particle acceleration in 3D magnetic reconnection has been a long-standing challenge. In this paper, we explore these problems by performing large-scale fully kinetic simulations of multi-X-line plasmoid reconnection with various parameters in both the weak- and strong-guide-field regimes. In each regime, we have identified its unique 3D dynamics that lead to field-line chaos and efficient acceleration, and we have achieved nonthermal acceleration of both electrons and protons into power-law spectra. The spectral indices agree well with a simple Fermi acceleration theory that includes guide-field dependence. In the low-guide-field regime, the flux rope kink instability governs the 3D dynamics for efficient acceleration. The weak dependence of the spectra on the ion-to-electron mass ratio and β ($\ll$1) implies that the particles are sufficiently magnetized for Fermi acceleration in our simulations. While both electrons and protons are injected at reconnection exhausts, protons are primarily injected by perpendicular electric fields through Fermi reflections and electrons are injected by a combination of perpendicular and parallel electric fields. The magnetic power spectra agree with in situ magnetotail observations, and the spectral index may reflect a reconnection-driven size distribution of plasmoids instead of the Goldreich–Sridhar vortex cascade. As the guide field becomes stronger, the oblique flux ropes of large sizes capture the main 3D dynamics for efficient acceleration. Intriguingly, the oblique flux ropes can also experience flux rope kink instability, to drive extra 3D dynamics. This work has broad implications for 3D reconnection dynamics and particle acceleration in heliophysics and astrophysics.

79 ASTRONOMY AND ASTROPHYSICS↗

Optimal interconnection of three-terminal tandem solar cells

Three-terminal (3T) tandem solar cells require an adapted module integration scheme in order to explore their full efficiency potential. The three terminals allow to extract the power of the top and bottom cell separately. In a cell string, the wide bandgap top cells are interconnected in parallel to multiple bottom cells resulting in a parallel/series interconnection. This interconnection scheme affects the operation of the subcells, the resulting current path between the subcells, the layout of the cell interconnects, and the system level. Here, we analyze by simulations and experiments the aspects of the module integration of series- and reverse-connected 3T cells with their practical impact on module processes and performance as well as the effect of varying voltage ratios on the string-end losses. If the subcells are connected in series, the module integration requires insulation layers and significantly longer interconnects compared to devices with reverse-connected subcells. Tandem devices with a reverse connection and a voltage ratio between top and bottom cell of 2:1 allow a lean interconnection design and low integration losses. Here, we present an approach for the integration of bypass diodes for the protection against shading effects that allow to minimize string-end and shading losses for a system of modules featuring 3T cells.

14 SOLAR ENERGY↗

Robust foil regenerator flow loss and heat transfer tests under oscillating flow condition

A regenerator works as an energy storage component in a Stirling engine. Thus, it is one of the most critical components that makes the Stirling engine highly efficient. Traditionally, the regenerators are made of wire media such as woven screens and random fibers. These wire regenerators experience considerably high flow losses through the regenerator due to the flow separation and the resulting vortices. The high flow losses associated with the flow separation and vortices result in a lower engine efficiency. This research is to test a new robust foil regenerator to investigate its feasibility for free-piston Stirling engine applications. The robust foil regenerator is configured with parallel flow channels and stiffening ribs. Comparing to complex flows patterns in woven screen and random fiber regenerators, the flows in the parallel plates are simple and the flow losses are considerably small, and the heat transfer coefficient is also relatively small. To test the regenerator experimentally, a regenerator test rig was manufactured. To verify the accuracy of the test rig, a woven screen regenerator was made and tested prior to the robust foil regenerator test. The friction factor correlation of the woven screen regenerator shows a good agreement with the published regenerator correlations. The robust foil regenerator tests were conducted under the oscillating flow condition with different Reynolds numbers. The flow losses were measured with dynamic pressure transducers and the friction coefficient was derived from measured pressure drop. The measured temperature difference over the regenerator was used to calculate the heat transfer coefficient and Nusselt number. The figure of merit with two methods of NPH/NTU ratio and j/f ratio was calculated to evaluate the suitability of the robust foil regenerator. The results show that the robust foil regenerator has considerably higher figure of merit and therefore, better performance than other types of regenerator.

42 ENGINEERING↗

Record acceleration of the two-dimensional Ising model using a high-performance wafer-scale engine

The versatility and wide-ranging applicability of the Ising model, originally introduced to study phase transitions in magnetic materials, have made it a cornerstone in statistical physics and a valuable tool for evaluating the performance of emerging computer hardware. Here, we present a novel implementation of the two-dimensional Ising model on Cerebras Wafer-Scale Engine (WSE) – a revolutionary processor that is opening new frontiers in computing. In our deployment of the checkerboard algorithm, we optimized the Ising model to take advantage of the unique WSE architecture. Specifically, we employed a compressed bit representation storing 16 spins on each int16 word, and efficiently distributed the spins over the processing units enabling seamless weak scaling and limiting communications to only immediate neighboring units. Our implementation can handle up to 754 simulations in parallel, achieving an aggregate of over 61.8 trillion flip attempts per second for Ising models with up to 200 million spins. This represents a gain of up to 148 times over previously reported single-devices with a highly optimized implementation on NVIDIA V100 and up to 88 times in productivity compared to NVIDIA H100. Our findings highlight the significant potential of the WSE in scientific computing, particularly in the field of materials modeling.

Ising model↗

OpenEdge: A collaborative, open-source, multi-purpose direct simulation Monte Carlo for plasma simulation in magnetic fusion environments

OpenEdge is a collaborative, open-source, object-oriented Direct Simulation Monte Carlo (DSMC) code, designed specifically for plasma simulations in magnetic fusion environments. Here, the code features include advanced structures, robust capabilities, and an effective parallelization strategy, all of which significantly enhance performance. It includes specialized modules for managing complex particle interactions, including collisions, ionization/recombination, and reflection/sputtering. Benchmarks and performance analyses have confirmed its efficiency and scalability. Versatile and adaptable, OpenEdge is applied across a broad spectrum of plasma-material interaction studies and charged particle transport in various fusion research settings.

Boundary plasma↗

Generation of lower hybrid and whistler waves by an ion velocity ring distribution

Using fully kinetic simulations in two and three spatial dimensions, we consider the generation and nonlinear evolution of lower hybrid waves produced by a cold ion ring velocity distribution in a low beta plasma. We show that the initial development of the instability is very similar in two and three dimensions and not significantly modified by electromagnetic effects, consistent with linear theory. At saturation, the level of electric field fluctuations is a small fraction of the background thermal energy; the electric field and corresponding density fluctuations consist of long, field-aligned striations. Energy extracted from the ring goes primarily into heating the background ions and the electrons at comparable rates. The initial growth and saturation of the magnetic components of the lower hybrid waves are related to the electric field components, consistent with linear theory. As the growing electric field fluctuations saturate, parallel propagating whistler waves develop by the interaction of two lower hybrid waves. At later times, these whistlers are replaced by longer wavelength, parallel propagating whistlers that grow through the decay of the lower hybrid fluctuations. Wave matching conditions demonstrate these conversion processes of lower hybrid waves to whistler waves. The conversion efficiency (=ratio of the whistler wave energy to the energy in the saturated lower hybrid waves) is computed and found to be significant ({approx}15%) for the parameters of the three-dimensional simulation (and even larger in the two-dimensional simulation), although when normalized in terms of the initial kinetic energy in the ring ions the overall efficiency is very small (<10{sup -4}). The results are compared with relevant linear and nonlinear theory.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Parallel parametric amplification of coherently excited propagating spin waves in a microscopic Ni{sub 81}Fe{sub 19} waveguide

We present parallel parametric amplification of coherently excited, propagating spin waves in a microstructured magnonic Ni{sub 81}Fe{sub 19} waveguide. Amplification is achieved by the pumping field generated by a microwave current flowing through a Cu micro-stripline underneath the waveguide. By employing microfocussed Brillouin light scattering spectroscopy, we investigate the spatial decay of the propagating spin waves and their amplification by means of parallel pumping. We analyze the dependence of the intensity of the amplified spin waves on the spin-wave excitation power, pumping power, and pumping duration, revealing the most efficient working point for a noise-free amplification. This paves the way for a frequency selective amplification of spin waves in microstructured magnonic circuits.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Scalable chemical oxygen - iodine laser

The problem of scaling chemical oxygen - iodine lasers (COILs) is discussed. The results of experimental study of a twisted-aerosol singlet oxygen generator meeting the COIL scalability requirements are presented. The energy characteristics of a supersonic COIL with singlet oxygen and iodine mixing in parallel flows are also experimentally studied. The output power of {approx}7.5 kW, corresponding to a specific power of 230 W cm{sup -2}, is achieved. The maximum chemical efficiency of the COIL is {approx}30%.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Mode Multiplexing for Scalable Cavity-Enhanced Operations in Neutral-Atom Arrays

Neutral-atom arrays provide a versatile platform for quantum information processing. However, in large-scale arrays, efficient photon collection remains a bottleneck for key tasks such as fast, nondestructive qubit readout and remote entanglement distribution. We propose a cavity-based approach that enables fast, parallel operations over many atoms using multiple modes of a single optical cavity. By selectively shifting the relevant atomic transitions, each atom can be coupled to a distinct cavity mode, allowing independent simultaneous processing. We present practical system designs that support cavity-mode multiplexing with up to 50 modes, enabling rapid mid-circuit syndrome extraction and significantly enhancing entanglement distribution rates between remote atom arrays. This approach offers a scalable solution to core challenges in neutral-atom arrays, advancing the development of practical quantum technologies.

Aqua, Ziv [Massachusetts Institute of Technology (↗

FatMan vs. LittleBoy: Scaling Up Linear Algebraic Operations in Scale-Out Data Platforms

Linear algebraic operations such as matrix manipulations form the kernel of many machine learning and other crucial algorithms. Scaling up as well as scaling out such algorithms are highly desirable to enable efficient processing over millions of data points. To this end, we present a matrix manipulation approach to effectively scale-up each node in a scale-out data parallel platform such as Apache Spark. Specifically, we enable hardware acceleration for matrix multiplications in a distributed Spark setup without user intervention. Our approach supports both dense and sparse distributed matrices, and provides flexible control of acceleration by matrix density. We demonstrate the benefit of our approach for generalized matrix multiplication operations over large matrices with up to four billion elements. To connect the effectiveness of our approach with machine learning applications, we performed Gramian matrix computation via generalized matrix multiplications. Our experiments show that our approach achieves more than 2× performance speed-up, and up to 96.1% computation improvement, compared to a state of the art Spark MLlib for dense matrices.

97 MATHEMATICS AND COMPUTING↗

UMAMI: A Recipe for Generating Meaningful Metrics through Holistic I/O Performance Analysis

I/O efficiency is essential to productivity in scientific computing, especially as many scientific domains become more data-intensive. Many characterization tools have been used to elucidate specific aspects of parallel I/O performance, but analyzing components of complex I/O subsystems in isolation fails to provide insight into critical questions: how do the I/O components interact, what are reasonable expectations for application performance, and what are the underlying causes of I/O performance problems? To address these questions while capitalizing on existing component-level characterization tools, we propose an approach that combines on-demand, modular synthesis of I/O characterization data into a unified monitoring and metrics interface (UMAMI) to provide a normalized, holistic view of I/O behavior. We evaluate the feasibility of this approach by applying it to a month-long benchmarking study on two distinct largescale computing platforms. We present three case studies that highlight the importance of analyzing application I/O performance in context with both contemporaneous and historical component metrics, and we provide new insights into the factors affecting I/O performance. By demonstrating the generality of our approach, we lay the groundwork for a production-grade framework for holistic I/O analysis.

97 MATHEMATICS AND COMPUTING↗

DIMPLES: Distributed Influence Maximization for Pandemic pLanning on Exascale Systems

We study exascale parallel algorithms for the selection of intervention or monitoring strategies in massive realistic socio-technical networks through scalable Influence Maximization (InfMax) algorithms. We employ novel techniques to enable efficient scaling on up to 8k nodes of OLCF Frontier, with 65k AMD GPUs and 458k AMD CPU cores. Current state-of-the-art InfMax tools are limited to networks with only a few million actors (vertices) and a few hundred million interactions (edges). By overcoming these limitations, we show that our approach is capable of processing a realistic social contact network of the United States with 285 million nodes and about 8 billion edges. This two orders-of-magnitude improvement over the previous state-of-the-art is obtained by leveraging algorithmic advancements for the InfMax problem and designing several problem-specific approaches to overlap communication with computation, improve GPU efficiency, and lower the application’s memory requirements. We evaluate strong scaling for computing 10k most influential seeds using up to 8k nodes of an exascale system, and weak scaling from 128 to 8k system nodes for seed sets ranging from 625 to 40k seeds. We achieve the fastest-known runtime of 25 minutes while performing 48 million diffusion simulations totaling 2.31 petabytes to identify 40k influential seeds using 8k nodes, and take 5.75 minutes to identify 10k seeds while using 4k nodes.

Minutoli, Marco [Pacific Northwest National Labora↗

FOX: A Fault-Oblivious Extreme-Scale Execution Environment Boston University Final Report Project Number: DE-SC0005365

Exascale computing systems will provide a thousand-fold increase in parallelism and a proportional increase in failure rate relative to today's machines. Systems software for exascale machines must provide the infrastructure to support existing applications while simultaneously enabling efficient execution of new programming models that naturally express dynamic, adaptive, irregular computation; coupled simulations; and massive data analysis in a highly unreliable hardware environment with billions of threads of execution. The FOX project explored systems software and runtime support for a new approach to the data and work distribution for fault oblivious application execution. Our major OS work at Boston University focused on developing a new light-weight operating systems model that provides an appropriate context for both multi-core and multi-node application development. This work is discussed in section 1. Early on in the FOX project BU developed infrastructure for prototyping dynamic HPC environments in which the sets of nodes that an application is run on can be dynamically grown or shrunk. This work was an extension of the Kittyhawk project and is discussed in section 2. Section 3 documents the publications and software repositories that we have produced. To put our work in context of the complete FOX project contribution we include in section 4 an extended version of a paper that documents the complete work of the FOX team.

97 MATHEMATICS AND COMPUTING↗

Load Balancing Scientific Applications

The largest supercomputers have millions of independent processors, and concurrency levels are rapidly increasing. For ideal efficiency, developers of the simulations that run on these machines must ensure that computational work is evenly balanced among processors. Assigning work evenly is challenging because many large modern parallel codes simulate behavior of physical systems that evolve over time, and their workloads change over time. Furthermore, the cost of imbalanced load increases with scale because most large-scale scientific simulations today use a Single Program Multiple Data (SPMD) parallel programming model, and an increasing number of processors will wait for the slowest one at the synchronization points. To address load imbalance, many large-scale parallel applications use dynamic load balance algorithms to redistribute work evenly. The research objective of this dissertation is to develop methods to decide when and how to load balance the application, and to balance it effectively and affordably. We measure and evaluate the computational load of the application, and develop strategies to decide when and how to correct the imbalance. Depending on the simulation, a fast, local load balance algorithm may be suitable, or a more sophisticated and expensive algorithm may be required. We developed a model for comparison of load balance algorithms for a specific state of the simulation that enables the selection of a balancing algorithm that will minimize overall runtime.

97 MATHEMATICS AND COMPUTING↗

A Fault-oblivious Extreme-scale Execution Environment

Exascale computing systems will provide a thousand-fold increase in parallelism and a proportional increase in failure rate relative to today's machines. Systems software for exascale machines must provide the infrastructure to support existing applications while simultaneously enabling efficient execution of new programming models that naturally express dynamic, adaptive, irregular computation; coupled simulations; and massive data analysis in a highly unreliable hardware environment with billions of threads of execution. We propose a new approach to the data and work distribution model provided by system software based on the unifying formalism of an abstract file system. The proposed hierarchical data model provides simple, familiar visibility and access to data structures through the file system hierarchy, while providing fault tolerance through selective redundancy. The hierarchical task model features work queues whose form and organization are represented as file system objects. Data and work are both first class entities. By exposing the relationships between data and work to the runtime system, information is available to optimize execution time and provide fault tolerance. The data distribution scheme provides replication (where desirable and possible) for fault tolerance and efficiency, and it is hierarchical to make it possible to take advantage of locality. The user, tools, and applications, including legacy applications, can interface with the data, work queues, and one another through the abstract file model. This runtime environment will provide multiple interfaces to support traditional Message Passing Interface applications, languages developed under DARPA's High Productivity Computing Systems program, as well as other, experimental programming models. We will validate our runtime system with pilot codes on existing platforms and will use simulation to validate for exascale-class platforms. In this final report, we summarize research results from the work done at the Ohio State University towards the larger goals of the project listed above.

97 MATHEMATICS AND COMPUTING↗

Polarization-dependent conversion efficiency of carbon nanotube-Si heterojunction solar cells based on aligned carbon nanotube films

We have fabricated solar cells that implement heterojunctions of Si and aligned carbon nanotube (CNT) films. Polarization-dependent optical absorption of highly aligned CNTs led to polarization-dependent conversion efficiencies, which in turn provided insight into the role of the CNT layer in the power generation mechanism in these heterojunction solar cells. When the incident light polarization was parallel to the CNT alignment direction so that the light absorption in the CNT layer was maximized, the short circuit photocurrent decreased by ~25%. This indicates that electron–hole pair generation in the CNT layer does not directly lead to power generation and that the CNT-Si heterojunctions in our devices are Schottky junctions rather than p–n junctions.

Physics↗

FatMan vs. LittleBoy: Scaling up Linear Algebraic Operations in Scale-out Data Platforms

Linear algebraic operations such as matrix manipulations form the kernel of many machine learning and other crucial algorithms. Scaling up as well as scaling out such algorithms are highly desirable to enable efficient processing over millions of data points. To this end, we present a matrix manipulation approach to effectively scale-up each node in a scale-out data parallel platform such as Apache Spark. Specifically, we enable hardware acceleration for matrix multiplications in a distributed Spark setup without user intervention. Our approach supports both dense and sparse distributed matrices, and provides flexible control of acceleration by matrix density. We demonstrate the benefit of our approach for generalized matrix multiplication operations over large matrices with up to four billion elements. To connect the effectiveness of our approach with machine learning applications, we performed Gramian matrix computation via generalized matrix multiplications. Our experiments show that our approach achieves more than 2x performance speed-up, and up to 96.1% computation improvement, compared to a state of the art Spark MLlib for dense matrices.

Lim, Seung-Hwan↗