Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “network performance”

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 271 records · Page 15
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Inference-Engine v0.1.0

Given a pre-trained neural network, Inference-Engine performs maps network inputs to outputs by executing the forward pass through the provided network. Although the predominant programming language for machine-learning is Python, most high-performance computing (HPC) applications are written in Fortran, C, or C++. Inference-Engine aims to support HPC programs and is written in Fortran, a language with a large feature set supporting interoperability with C. This software exposes concurrency in a portable way by using standard language features that some modern Fortran compilers can exploit with various optimizations, including offloading computation to a Graphics Processing Unit (GPU). In particular, this software makes extensive use of Fortran's "do concurrent" parallel loop construct, implicitly parallel array statements, and pure procedures that can be invoked inside "do concurrent" blocks. Inference-Engine also supports dynamic choice of inference methods at runtime. Two current options include one method that uses Fortran's "dot_product" intrinsic function inside "do concurrent" blocks and another method that instead uses Fortran' "matmul" array intrinsic function. We plan to investigate automatic compiler offloading of "do concurrent" calculations to GPUs and compile-time substitution of optimized libraries such as the Basic Linear Algebra Library (BLAS) for "matmul" invocations. We also envision the potential for the choice of which method to use could happen at program launch based on in situ performance measurements on any given platform.

Rouson, Damian↗
DOE OSTI2024

Ergodicity, lack thereof, and the performance of reservoir computing with memristive networks and nanowire

Networks composed of nanoscale memristive components, such as nanowire and nanoparticle networks, have recently received considerable attention because of their potential use as neuromorphic devices. In this study, we explore ergodicity in memristive networks, showing that the performance on machine leaning tasks improves when these networks are tuned to operate at the edge between two global stability points. We find this lack of ergodicity is associated with the emergence of memory in the system. We measure the level of ergodicity using the Thirumalai-Mountain metric, and we show that in the absence of ergodicity, two different memristive network systems show improved performance when utilized as reservoir computers (RC). We highlight that it is also important to let the system synchronize to the input signal in order for the performance of the RC to exhibit improvements over the baseline.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Ergodicity, lack thereof, and the performance of reservoir computing with memristive networks

Abstract Networks composed of nanoscale memristive components, such as nanowire and nanoparticle networks, have recently received considerable attention because of their potential use as neuromorphic devices. In this study, we explore ergodicity in memristive networks, showing that the performance on machine leaning tasks improves when these networks are tuned to operate at the edge between two global stability points. We find this lack of ergodicity is associated with the emergence of memory in the system. We measure the level of ergodicity using the Thirumalai-Mountain metric, and we show that in the absence of ergodicity, two different memristive network systems show improved performance when utilized as reservoir computers (RC). We highlight that it is also important to let the system synchronize to the input signal in order for the performance of the RC to exhibit improvements over the baseline.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

The optimal use of segmentation for sampling calorimeters

One of the key design choices of any sampling calorimeter is how fine to make the longitudinal and transverse segmentation. Here, to inform this choice, we study the impact of calorimeter segmentation on energy reconstruction. To ensure that the trends are due entirely to hardware and not to a sub-optimal use of segmentation, we deploy deep neural networks to perform the reconstruction. These networks make use of all available information by representing the calorimeter as a point cloud. To demonstrate our approach, we simulate a detector similar to the forward calorimeter system intended for use in the ePIC detector, which will operate at the upcoming Electron Ion Collider. We find that for the energy estimation of isolated charged pion showers, relatively fine longitudinal segmentation is key to achieving an energy resolution that is better than 10% across the full phase space. These results provide a valuable benchmark for ongoing EIC detector optimizations and may also inform future studies involving high-granularity calorimeters in other experiments at various facilities.

47 OTHER INSTRUMENTATION↗
DOE OSTI2024

Two-fermion negativity and confinement in the Schwinger model

We consider the fermionic (logarithmic) negativity between two fermionic modes in the Schwinger model. Recent results pointed out that fermionic systems can exhibit stronger entanglement than bosonic systems, exhibiting a negativity that decays only algebraically. The Schwinger model is described by fermionic excitations at short distances, while its asymptotic spectrum is the one of a bosonic theory. We show that the two-mode negativity detects this confining, fermion-to-boson transition, shifting from an algebraic decay to an exponential decay at distances of the order of the de Broglie wavelength of the first excited state. We derive analytical expressions in the massless Schwinger model and confront them with tensor network simulations. We also perform tensor network simulations in the massive model, which is not solvable analytically, and close to the Ising quantum critical point of the Schwinger model, where we show that the negativity behaves as its bosonic counterpart. Published by the American Physical Society 2024

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2025

Inference of three-dimensional hot-spot and shell morphology in inertial confinement fusion experiments using a convolutional neural network

The performance of inertial confinement fusion (ICF) implosions is sensitive to the three-dimensional (3D) morphology of the hot-spot and shell configurations. The ability to infer shell-mass uniformity and reconstruct 3D hot spots is crucial for quantifying the degradation of ignition criteria and improving symmetry in ICF implosion experiments. In this work, we present a deep-learning convolutional neural network (CNN) for reconstructing 3D hot-spot and shell structures for ICF capsules. The 3D geometry of the hot spot is reconstructed from x-ray images measured from multiple lines of sight on OMEGA. The shell configuration is inferred indirectly through machine learning using a convolutional neural network extensively trained on a dec3d simulation database. This simulation-dependent approach yields consistent agreement between reconstructed 3D shell densities and machine-learning optimized dec3d simulation results. This work demonstrates a CNN framework that successfully reconstructs 3D capsule structures from two-dimensional images in ICF implosions.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2023

Assessing parallel path cooling tower performance via artificial neural networks

Real-time monitoring of a research nuclear reactor, a system in which all generated power is dissipated to the environment, can be performed via analysis of the heat rejection from the cooling system. Given an inlet water temperature and flow rate, the reactor power can be well-approximated from the outlet water temperature; however, the instrumentation to measure outlet conditions may not be robust or accurate. If we know how a cooling tower performs from historical data, but cannot measure the outlet temperature, a mathematical representation of the system can be inverted to obtain the outlet water temperature that describes the cooling capacity. Unfortunately, model inversion processes are computationally expensive. To address this, an artificial neural network (ANN) is implemented to assess the performance of a multi-cell cooling tower for a nuclear reactor. This approach leverages the Merkel model to obtain an extensive data set describing performance of the cooling tower cells throughout a wide array of potential operating conditions. The Merkel model is expressed as a function of four parameters: the inlet and outlet water temperatures, inlet air wet bulb temperature, and ratio of liquid-to-gas mass flow rates (L/G), which together provide a non-dimensional number indicative of cooling tower performance, called the Merkel integral. Computing a 4-dimensional data structure that describes finite combinations of the Merkel integral, an inverse model is then generated using an ANN to determine the cell outlet water temperature from the other three model parameters along with the computed Merkel integral. Compared to traditional model inversion methods, the ANN reduces the computational time by approximately 4 orders of magnitude, with effectively no sacrifice to solution accuracy, and could be applied for different cooling towers in the event the performance curve is known. Finally, three use cases of the ANN are then reviewed: (1) determining the cell outlet water temperatures when gas flow at rated conditions (GFRC) is known, (2) performing the prior case without knowledge of the GRFC, and (3) assessing performance differences between the individual tower cells.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
NASA NTRS1988

Implementing direct, spatially isolated problems on transputer networks

Parametric studies were performed on transputer networks of up to 40 processors to determine how to implement and maximize the performance of the solution of problems where no processor-to-processor data transfer is required for the problem solution (spatially isolated). Two types of problems are investigated a computationally intensive problem where the solution required the transmission of 160 bytes of data through the parallel network, and a communication intensive example that required the transmission of 3 Mbytes of data through the network. This data consists of solutions being sent back to the host processor and not intermediate results for another processor to work on. Studies were performed on both integer and floating-point transputers. The latter features an on-chip floating-point math unit and offers approximately an order of magnitude performance increase over the integer transputer on real valued computations. The results indicate that a minimum amount of work is required on each node per communication to achieve high network speedups (efficiencies). The floating-point processor requires approximately an order of magnitude more work per communication than the integer processor because of the floating-point unit's increased computing capacity.

Ellis, Graham K.↗
DOE OSTI2021

Implementing a neural network interatomic model with performance portability for emerging exascale architectures

The two main thrusts of computational science are increasingly accurate predictions and faster calculations; to this end, the zeitgeist in molecular dynamics (MD) simulations is pursuing machine learned and data driven interatomic models, e.g. neural network potentials, and novel hardware architectures, e.g. GPUs. Current implementations of neural network potentials are orders of magnitude slower than traditional interatomic models and while looming exascale computing offers the ability to run large, accurate simulations with these models, achieving portable performance for MD with new and varied exascale hardware requires rethinking traditional algorithms, using novel data structures, and library solutions. We re-implement a neural network interatomic model in CabanaMD, an MD proxy application, built on libraries developed for performance portability. Our implementation shows significantly improved thread scaling in this complex kernel as compared to a current LAMMPS implementation, across both strong and weak scaling. Our single-source solution enables simulations up to 20 million atoms on a single CPU node and 4 million atoms with improved performance on a single GPU. Furthermore, we also explore parallelism and data layout choices (using flexible data structures called AoSoAs) and their effect on performance, seeing up to ~50% and ~5% improvements in performance on a GPU by choosing the right level of parallelism and data layout respectively.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1980

An analysis of network deployment for a dedicated navigation tracking system

Stringent navigation accuracy requirements for future deep space missions are expected to be met by conventional two way radio metric data supplemented with Very Long Baseline Interferometric (VLBI) data provided by a dedicated tracking network. The performance of several proposed VLBI network configurations have been evaluated by comparing navigation capability for the probe delivery and post flight probe trajectory phases of the Galileo mission. Candidate configurations include stations located at the current intercontinental DSN sites, a network with sites in Hawaii, Florida and Washington, Regional (300 km) and a Micro (30 km) baseline systems. Sensitivity to systematic error sources, network deployment, viewing period and sampling strategy are investigated. The dominant factor in performance is shown to be the baseline length.

Ellis, J.↗
NASA NTRS1991

Performance analysis of LAN bridges and routers

Bridges and routers are used to interconnect Local Area Networks (LANs). The performance of these devices is important since they can become bottlenecks in large multi-segment networks. Performance metrics and test methodology for bridges and routers were not standardized. Performance data reported by vendors is not applicable to the actual scenarios encountered in an operational network. However, vendor-provided data can be used to calibrate models of bridges and routers that, along with other models, yield performance data for a network. Several tools are available for modeling bridges and routers - Network II.5 was used. The results of the analysis of some bridges and routers are presented.

Hajare, Ankur R.↗
NASA NTRS1996

Performance Evaluation in Network-Based Parallel Computing

Network-based parallel computing is emerging as a cost-effective alternative for solving many problems which require use of supercomputers or massively parallel computers. The primary objective of this project has been to conduct experimental research on performance evaluation for clustered parallel computing. First, a testbed was established by augmenting our existing SUNSPARCs' network with PVM (Parallel Virtual Machine) which is a software system for linking clusters of machines. Second, a set of three basic applications were selected. The applications consist of a parallel search, a parallel sort, a parallel matrix multiplication. These application programs were implemented in C programming language under PVM. Third, we conducted performance evaluation under various configurations and problem sizes. Alternative parallel computing models and workload allocations for application programs were explored. The performance metric was limited to elapsed time or response time which in the context of parallel computing can be expressed in terms of speedup. The results reveal that the overhead of communication latency between processes in many cases is the restricting factor to performance. That is, coarse-grain parallelism which requires less frequent communication between processes will result in higher performance in network-based computing. Finally, we are in the final stages of installing an Asynchronous Transfer Mode (ATM) switch and four ATM interfaces (each 155 Mbps) which will allow us to extend our study to newer applications, performance metrics, and configurations.

Dezhgosha, Kamyar↗
NASA NTRS2006

Simulation Modeling and Performance Evaluation of Space Networks

In space exploration missions, the coordinated use of spacecraft as communication relays increases the efficiency of the endeavors. To conduct trade-off studies of the performance and resource usage of different communication protocols and network designs, JPL designed a comprehensive extendable tool, the Multi-mission Advanced Communications Hybrid Environment for Test and Evaluation (MACHETE). The design and development of MACHETE began in 2000 and is constantly evolving. Currently, MACHETE contains Consultative Committee for Space Data Systems (CCSDS) protocol standards such as Proximity-1, Advanced Orbiting Systems (AOS), Packet Telemetry/Telecommand, Space Communications Protocol Specification (SCPS), and the CCSDS File Delivery Protocol (CFDP). MACHETE uses the Aerospace Corporation s Satellite Orbital Analysis Program (SOAP) to generate the orbital geometry information and contact opportunities. Matlab scripts provide the link characteristics. At the core of MACHETE is a discrete event simulator, QualNet. Delay Tolerant Networking (DTN) is an end-to-end architecture providing communication in and/or through highly stressed networking environments. Stressed networking environments include those with intermittent connectivity, large and/or variable delays, and high bit error rates. To provide its services, the DTN protocols reside at the application layer of the constituent internets, forming a store-and-forward overlay network. The key capabilities of the bundling protocols include custody-based reliability, ability to cope with intermittent connectivity, ability to take advantage of scheduled and opportunistic connectivity, and late binding of names to addresses. In this presentation, we report on the addition of MACHETE models needed to support DTN, namely: the Bundle Protocol (BP) model. To illustrate the use of MACHETE with the additional DTN model, we provide an example simulation to benchmark its performance. We demonstrate the use of the DTN protocol and discuss statistics gathered concerning the total time needed to simulate numerous bundle transmissions

network protocols↗
DOE OSTI2024

How to Design and Implement an Equitable Building Performance Standard: Lessons from the Building Performance Standards Technical Assistance Network

States and local governments seek to accomplish the intersecting goals of reducing greenhouse gas emissions, improving building operations, and bettering the daily lives of their communities. Building Performance Standards (BPS) have emerged as a critical policy lever to reach these intertwined climate and societal goals. These policies, if shaped and implemented well, have the chance to not only help reach our nation's climate, energy, and livability goals, but to do so with the active participation of those traditionally excluded from policy processes. This nascent policy movement provides jurisdictions across the country the opportunity to shape these policies from the outset to deliver comfort, health, and safety in our built environment for all. The US Department of Energy in partnership with our National Laboratories have been providing technical assistance for jurisdictions interested in, adopting, and implementing Building Performance Standards. Through this work, the BPS Technical Assistance Network (TA Network) has tracked and documented the innovative and equitable approaches to BPS across the country. The TA Network has crafted foundational technical analysis, such as building stock and emissions impacts, equity prioritization and peak load impacts, and aggregated cost benefit analysis. And by combining powerful technical analysis with dissemination of best practices resources to support equitable implementation, the TA Network provides jurisdictions with the tools and support necessary to embark on their ambitious policy goals.

building performance standards↗
DOE OSTI2021

Accelerating Binarized Neural Networks via Bit-Tensor-Cores in Turing GPUs

Despite foreseeing tremendous speedups over conventional deep neural networks, the performance advantage of binarized neural networks (BNNs) has merely been showcased on general-purpose processors such as CPUs and GPUs. In fact, due to being unable to leverage bit-level-parallelism with a word-based architecture, GPUs have been criticized for extremely low utilization (1%) when executing BNNs. Consequently, the latest tensorcores in NVIDIA Turing GPUs start to experimentally support bit computation. In this work, we look into this brand new bit computation capability and characterize its unique features. We show that the stride of memory access can significantly affect performance delivery and a data-format co-design is highly desired to support the tensorcores for achieving superior performance than existing software solutions without tensorcores. We realize the tensorcore-accelerated BNN design, particularly the major functions for fully-connect and convolution layers — bit matrix multiplication and bit convolution. Evaluations on two NVIDIA Turing GPUs show that, with ResNet-18, our BTC-BNN design can process ImageNet at a rate of 5.6K images per second, 77% faster than state-of-the-art. Our BNN approach is released on https://github.com/pnnl/TCBNN.

Li, Ang↗
DOE OSTI2022

Performance trade-offs in reconfigurable networks for HPC

Designing efficient interconnects to support high-bandwidth and low-latency communication is critical toward realizing high performance computing (HPC) and data center (DC) systems in the exascale era. At extreme computing scales, providing the requisite bandwidth through overprovisioning becomes impractical. These challenges have motivated studies exploring reconfigurable network architectures that can adapt to traffic patterns at runtime using optical circuit switching. Despite the plethora of proposed architectures, surprisingly little is known about the relative performances and trade-offs among different reconfigurable network designs. We aim to bridge this gap by tackling two key issues in reconfigurable network design. First, we study how cost, power consumption, network performance, and scalability vary based on optical circuit switch (OCS) placement in the physical topology. Specifically, we consider two classes of reconfigurable architectures: one that places OCSs between top-of-rack (ToR) switches—ToR-reconfigurable networks (TRNs)—and one that places OCSs between pods of racks—pod-reconfigurable networks (PRNs). Second, we tackle the effects of reconfiguration frequency on network performance. Our results, based on network simulations driven by real HPC and DC workloads, show that while TRNs are optimized for low fan-out communication patterns, they are less suited for carrying high fan-out workloads. PRNs exhibit better overall trade-off, capable of performing comparably to a fully non-blocking fat tree for low fan-out workloads, and significantly outperform TRNs for high fan-out communication patterns.

Teh, Min Yee↗
DOE OSTI2020

Low Size, Weight, and Power Neuromorphic Computing to Improve Combustion Engine Efficiency

Neuromorphic computing offers one path forward for AI at the edge. However, accessing and effectively utilizing a neuromorphic hardware platform is non-trivial. In this work, we present a complete pipeline for neuromorphic computing at the edge, including a small, inexpensive, low-power, FPGA-based neuromorphic hardware platform, a training algorithm for designing spiking neural networks for neuromorphic hardware, and a software framework for connecting those components. We demonstrate this pipeline on a real-world application, engine control for a spark-ignition internal combustion engine. We illustrate how we connect engine simulations with neuromorphic hardware simulations and training software to produce hardware-compatible spiking neural networks that perform engine control to improve fuel efficiency. We present initial results on the performance of these spiking neural networks and illustrate that they outperform open-loop engine control. We also give size, weight, and power estimates for a deployed solution of this type.

Schuman, Catherine↗