Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “data engineering”

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 775 records · Page 43
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

ExaFEL: extreme-scale real-time data processing for X-ray free electron laser science

ExaFEL is an HPC-capable X-ray Free Electron Laser (XFEL) data analysis software suite for both Serial Femtosecond Crystallography (SFX) and Single Particle Imaging (SPI) developed in collaboration with the Linac Coherent Lightsource (LCLS), Lawrence Berkeley National Laboratory (LBNL) and Los Alamos National Laboratory. ExaFEL supports real-time data analysis via a cross-facility workflow spanning LCLS and HPC centers such as NERSC and OLCF. Our work therefore constitutes initial path-finding for the US Department of Energy's (DOE) Integrated Research Infrastructure (IRI) program. We present the ExaFEL team's 7 years of experience in developing real-time XFEL data analysis software for the DOE's exascale supercomputers. We present our experiences and lessons learned with the Perlmutter and Frontier supercomputers. Furthermore we outline essential data center services (and the implications for institutional policy) required for real-time data analysis. Finally we summarize our software and performance engineering approaches and our experiences with NERSC's Perlmutter and OLCF's Frontier systems. This work is intended to be a practical blueprint for similar efforts in integrating exascale compute resources into other cross-facility workflows.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

High Explosive Testing Capabilities at Nevada National Security Site (NNSS) [Slides]

LANL is a non-profit Federally Funded Research & Development Center (FFRDC). LANL’s Integrated Weapons Experiments (J) Division addresses national security challenges by executing largescale, integrated, focused experiments, and tests of engineered devices and systems. J-NV plans, fields, and executes dynamic high explosive experiments at NNSS. This includes sub-critical experiments involving special nuclear material, as well as small- and largescale non-nuclear experiments. J-NV generates, analyzes, and disseminates unique, high-quality data.

42 ENGINEERING↗
DOE OSTI2020

Determination of Stress-Free Lattice Spacing (D₀) for Residual Stress Relaxation Measurement in Ni-based Superalloys by Neutron Diffraction

The residual stress and its relaxation in critical engineering components are key to structural materials reliability. The determination of residual stress in bulk engineering structures like Ni-based superalloys can be conducted by measuring the lattice strain with penetrating neutron diffraction. The state-of-the-art engineering diffraction beamline VULCAN at the Oak Ridge National Laboratory (ORNL) Spallation Neutron Source (SNS) provides the critical capability to evaluate the residual strain/stress relaxation in engineering component thanks to the high flux and event data features. By the means of residual stress/strain calculation, it is critical to accurately determine the stress-free lattice spacing d₀, which can be altered by the change of chemistry, unrelieved stress, and measurement scheme. Here we reported d₀ determination for determining residual stress relaxation and distribution in Alloy 718 superalloys after different quenching treatments. Selection of locations for d₀ measurement was discussed by considering the neutron path, attenuation and sample alignment. A dynamic d₀ that resulted from atom diffusion and chemistry change was estimated as function of time for the in-situ relaxation characterization. It demonstrated how the dynamic d₀ values may influence on the strain calculation in different thermally treated samples. The results highlight the importance of dynamic d₀ for in-situ relaxation involving chemical changes and provide guidance to the dynamic d₀ measurement and simulation.

36 MATERIALS SCIENCE↗
DOE OSTI2023

W-4 W88 Systems Engineering Summer 2023 Project [Slides]

The purpose is to create a simple, streamlined process to produce reports on out of tolerance conditions with CMM data. The methods are to develop the code or process within MATLAB to import the data (import tool can be used) and to develop the code to be able to plot the data of interest. The end states are to use MATLAB to analyze CMM data from the PA and document the process for a user with some knowledge of MATLAB and an engineering background.

42 ENGINEERING↗
DOE OSTI2025

Progress Towards Developing Neutron Tolerant Magnetostrictive and Piezoelectric Transducers

Current generation light water reactors (LWRs), sodium cooled fast reactors (SFRs), small modular reactors (SMRs), and next generation nuclear plants (NGNPs) produce harsh environments in and near the reactor core that can severely tax material performance and limit component operational life. To address this issue, several Department of Energy Office of Nuclear Energy (DOE-NE) research programs are evaluating the long duration irradiation performance of fuel and structural materials used in existing and new reactors. In order to maximize the amount of information obtained from Material Testing Reactor (MTR) irradiations, DOE is also funding development of enhanced instrumentation that will be able to obtain in-situ, real-time data on key material characteristics and properties, with unprecedented accuracy and resolution. Such data are required to validate new multi-scale, multi-physics modeling tools under development as part of a science-based, engineering driven approach to reactor development. It is not feasible to obtain high resolution/microscale data with the current state of instrumentation technology. However, ultrasound-based sensors offer the ability to obtain such data if it is demonstrated that these sensors and their associated transducers are resistant to high neutron flux, high gamma radiation, and high temperature. To address this need, the Advanced Test Reactor National Scientific User Facility (ATR-NSUF) is funding an irradiation, led by PSU, at the Massachusetts Institute of Technology Research Reactor to test the survivability of ultrasound transducers. As part of this effort, PSU and collaborators have designed, fabricated, and provided piezoelectric and magnetostrictive transducers that are optimized to perform in harsh, high flux, environments. Four piezoelectric transducers were fabricated with either aluminum nitride, zinc oxide, or bismuth titanate as the active element that were coupled to either Kovar or aluminum waveguides and two magnetostrictive transducers were fabricated with Remendur or Galfenol as the active elements. Pulse-echo ultrasonic measurements of these transducers are made in-situ. This paper will present an overview of the test design including selection criteria for candidate materials and optimization of test assembly parameters, data obtained from both out-of-pile and in-pile testing at elevated temperatures, and an assessment based on initial data of the expected performance of ultrasonic devices in irradiation conditions

Reinhardt1, Brian↗
DOE OSTI2022

Heavy-Duty Vehicle Activity Updates for MOVES Using NREL Fleet DNA and CE-CERT Data

The U.S. Environmental Protection Agency's (EPA's) Motor Vehicle Emission Simulator (MOVES) is a publicly available tool used by researchers and policymakers to help understand motor vehicle emission sources at a national, county, and project level. Estimates of heavy-duty activity in the most recent version of the model at the time this work was conducted, MOVES2014, was identified as an area in need of improvement. The start activity in MOVES2014 is based on a limited and dated data set. In addition, MOVES2014 relies on drive cycles that represent on-network activity but do not account for idling activity that occurs on off-network roads, such as at a distribution center, while the truck is queuing or during loading and unloading. As a result, MOVES2014 may currently underestimate the number of starts and idle and soak time for heavy-duty trucks in real-world operation. The National Renewable Energy Laboratory (NREL) has previously leveraged its expansive Fleet DNA database of heavy-duty vehicles to idle and start activity for six of the nine heavy-duty vehicle source types of classes in the MOVES model. The data available in Fleet DNA from 416 conventional, diesel-powered vehicles provided activity estimates from more than 120,000 hours of operation throughout 14,682 vehicle days between October 2006 and January 2016. NREL calculated start fraction, starts per day, soak fraction, and idle fraction by hour of the day for each vehicle type, state, and vocation, and provided results in .CSV files that can be translated to MOVES table inputs. The idle and start activity from this initial analysis of Fleet DNA data was used to develop default idle and start data for heavy-duty vehicles in MOVES3. Satisfied with the results from the Fleet DNA data used for MOVES3, the EPA asked NREL to extend this start/soak/idle analysis using additional data from a larger number of vehicles for a potential future update to the MOVES model. Such a data set was achieved from a project led by the University of California at Riverside, College of Engineering, Center for Environmental Research & Technology (CE-CERT) and funded by California Air Resources Board. Specifically, this data set consists of 90 heavy-duty vehicles operated mainly in California, which can be separated into five of the nine heavy-duty vehicle classes in the MOVES model. In addition, the heavy-duty activity database collected by CE-CERT provided activity estimates from more than 44,000 hours of operation throughout 4,724 vehicle days between November 2014 and September 2016. This report details the analysis of the heavy-duty activity database collected from the University of California at Riverside by providing graphical analysis and context for the start, soak, and idle distributions. The comparison of the related results from both the Fleet DNA and CE-CERT data sets are documented as well.

33 ADVANCED PROPULSION SYSTEMS↗
DOE OSTI2023

Genetic algorithm optimization of a chemical kinetic mechanism for propane at engine relevant conditions

Propane has demonstrated significant potential for reductions in greenhouse gas and pollutant emissions in medium- and heavy-duty engine applications, but further improvements require accurate, compact, and scalable chemical kinetic mechanisms to design the next generation of propane fueled engines, particularly at the boosted operating conditions necessary to meet the power density demand of medium- and heavy-duty applications. In this work, six key chemical reactions were identified in a reduced mechanism with 70 species and 352 reactions through a sensitivity analysis performed at conditions typical of thermodynamic trajectories observed in a high compression ratio, long stroke engine operated on propane from throttled to boosted operating conditions. While the original mechanism was validated against rapid compression machine (RCM) data, it was found to overpredict experimental autoignition tendencies in 2-zone, 0-D SI engine simulations performed in Chemkin Pro. Subsequently, a genetic algorithm approach was used to optimize the six reaction rate parameters within established uncertainty bounds by performing RCM simulations and comparing to two independent sets of literature ignition delay times for propane, thus generating two new kinetic mechanisms. The first optimization achieved a mean absolute percent error (MPE) reduction in 2nd stage ignition delay of 61.4% in seven generations, while the second optimization utilized a newer experimental RCM dataset, and achieved MPE reduction of 56.7% in seven generations, and further marginal improvement to 57.8% reduction in 34 generations. Finally, the two mechanisms were then evaluated again in the 2-zone 0-D SI engine model in Chemkin Pro comparing typical mean and knocking cycle trajectories, and it was found that the second optimized mechanism provided better prediction of knock onset at the representative conditions evaluated in this work, particularly for higher load operating conditions.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2020

DEEP CELLULAR RECURRENT NEURAL ARCHITECTURE FOR EFFICIENT MULTIDIMENSIONAL TIME-SERIES DATA PROCESSING

Efficient processing of time series data is a fundamental yet challenging problem in pattern recognition. Though recent developments in machine learning and deep learning have enabled remarkable improvements in processing large scale datasets in many application domains, most are designed and regulated to handle inputs that are static in time. Many real-world data, such as in biomedical, surveillance and security, financial, manufacturing and engineering applications, are rarely static in time, and demand models able to recognize patterns in both space and time. Current machine learning (ML) and deep learning (DL) models adapted for time series processing tend to grow in complexity and size to accommodate the additional dimensionality of time. Specifically, the biologically inspired learning based models known as artificial neural networks that have shown extraordinary success in pattern recognition, tend to grow prohibitively large and cumbersome in the presence of large scale multi-dimensional time series biomedical data such as EEG. Consequently, this work aims to develop representative ML and DL models for robust and efficient large scale time series processing. First, we design a novel ML pipeline with efficient feature engineering to process a large scale multi-channel scalp EEG dataset for automated detection of epileptic seizures. With the use of a sophisticated yet computationally efficient time-frequency analysis technique known as harmonic wavelet packet transform and an efficient self-similarity computation based on fractal dimension, we achieve state-of-the-art performance for automated seizure detection in EEG data. Subsequently, we investigate the development of a novel efficient deep recurrent learning model for large scale time series processing. For this, we first study the functionality and training of a biologically inspired neural network architecture known as cellular simultaneous recurrent neural network (CSRN). We obtain a generalization of this network for multiple topological image processing tasks and investigate the learning efficacy of the complex cellular architecture using several state-of-the?art training methods. Finally, we develop a novel deep cellular recurrent neural network (CDRNN) architecture based on the biologically inspired distributed processing used in CSRN for processing time series data. The proposed DCRNN leverages the cellular recurrent architecture to promote extensive weight sharing and efficient, individualized, synchronous processing of multi-source time series data. Experiments on a large scale multi-channel scalp EEG, and a machine fault detection dataset show that the proposed DCRNN offers state-of-the-art recognition performance while using substantially fewer trainable recurrent units.

Vidyaratne, Lasitha S.↗
DOE OSTI2020

Towards Trust-Augmented Visual Analytics for Data-Driven Energy Modeling

The promise of data-driven predictive modeling is being increasingly realized in various science and engineering disciplines, where experts are used to the more conventional, simulation-driven modeling practices. However, trust remains a bottleneck for greater adoption of machine learning-based models for domain experts, who might not be necessarily trained in data science. In this paper, we focus on the building energy domain, where physics-based simulations are being complemented or replaced by machine learning-based methods for forecasting energy supply and demand at various spatio-temporal scales. We study the trust problem in close collaboration with energy scientists and engineers and describe how visual analytics can be leveraged for alleviating this trust bottleneck for stakeholders with varying degrees of expertise and analytics goals in this domain.

Kandakatla, Akshith R.↗
DOE OSTI2022

Computer Model Calibration with Time Series Data Using Deep Learning and Quantile Regression

Computer models play a key role in many scientific and engineering problems. One major source of uncertainty in computer model experiments is input parameter uncertainty. Computer model calibration is a formal statistical procedure to infer input parameters by combining information from model runs and observational data. The existing standard calibration framework suffers from inferential issues when the model output and observational data are high-dimensional dependent data, such as large time series, due to the difficulty in building an emulator and the nonidentifiability between effects from input parameters and data-model discrepancy. To overcome these challenges, we, in this work, propose a new calibration framework based on a deep neural network (DNN) with long short-term memory layers that directly emulates the inverse relationship between the model output and input parameters. Adopting the “learning with noise” idea, we train our DNN model to filter out the effects from data-model discrepancy on input parameter inference. We also formulate a new way to construct interval predictions for DNN using quantile regression to quantify the uncertainty in input parameter estimates. Through a simulation study and real data application with the Weather Research and Forecasting Model Hydrological modeling system (WRF-Hydro), we show our approach can yield accurate point estimates and well-calibrated interval estimates for input parameters.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Sensitivity Analysis of Genome-Scale Metabolic Flux Prediction

TRIMER, Transcription Regulation Integrated with MEtabolic Regulation, is a genome-scale modeling pipeline targeting at metabolic engineering applications. Using TRIMER, regulated metabolic reactions can be effectively predicted by integrative modeling of metabolic reactions with a Transcription Factor (TF)-gene regulatory network (TRN), which is modeled via a Bayesian network (BN). In this paper, we focus on sensitivity analysis of metabolic flux prediction for uncertainty quantification of BN structures for TRN modeling in TRIMER. We propose a computational strategy to construct the uncertainty class of TRN models based on the inferred regulatory order uncertainty given transcriptomic expression data. With that, we analyze the prediction sensitivity of the TRIMER pipeline for the metabolite yields of interest. The obtained sensitivity analyses can guide Optimal Experimental Design (OED) to help acquire new data that can enhance TRN modeling and achieve specific metabolic engineering objectives, including metabolite yield alterations. Here we have performed small- and large-scale simulated experiments, demonstrating the effectiveness of our developed sensitivity analysis strategy for BN structure learning to quantify the edge importance in terms of metabolic flux prediction uncertainty reduction and its potential to effectively guide OED.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

Data Compression in the NEXT-100 Data Acquisition System

NEXT collaboration detectors are based on energy measured by an array of photomultipliers (PMT) and topological event filtering based on an array of silicon photomultipliers (SiPMs). The readout of the PMT sensors for low-frequency noise effects and detector safety issues requires a grounded cathode connection that makes the readout AC-couple with variations in the signal baseline. Strict detector requirements of energy resolution better than 1% FWHM require a precise baseline reconstruction that is performed offline for data analysis and detector performance characterization. Baseline variations make it inefficient to apply traditional lossy data compression techniques, such as zero-suppression, that help to minimize data throughput and, therefore, the dead time of the system. However, for the readout of the SiPM sensors with less demanding requirements in terms of accuracy, a traditional zero-suppression is currently applied with a configuration that allows for a compression ratio of around 71%. The third stage in the NEXT detectors program, the NEXT-100 detector, is a 100 kg detector that instruments approximately five times more PMT sensors and twice the number of SiPM sensors than its predecessor, the NEXT-White detector, putting more pressure in the DAQ throughput, expected to be over 900 MB/s with the current configuration, which will worsen the dead time of the acquisition data system. This paper describes the data compression techniques applied to the sensor data in the NEXT-100 detector, which reduces data throughput and minimizes dead time while maintaining the event rate to the level of its predecessor, around 50 Hz.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2020

Fatty Acid and Alcohol Metabolism in Pseudomonas putida: Functional Analysis Using Random Barcode Transposon Sequencing.

With its ability to catabolize a wide variety of carbon sources and a growing engineering toolkit, Pseudomonas putida KT2440 is emerging as an important chassis organism for metabolic engineering. Despite advances in our understanding of the organism, many gaps remain in our knowledge of the genetic basis of its metabolic capabilities. The gaps are particularly noticeable in our understanding of both fatty acid and alcohol catabolism, where many paralogs putatively coding for similar enzymes coexist, making biochemical assignment via sequence homology difficult. To rapidly assign function to the enzymes responsible for these metabolisms, we leveraged random barcode transposon sequencing (RB-Tn-Seq). Global fitness analyses of transposon libraries grown on 13 fatty acids and 10 alcohols produced strong phenotypes for hundreds of genes. Fitness data from mutant pools grown on fatty acids of varying chain lengths indicated specific enzyme substrate preferences and enabled us to hypothesize that DUF1302/DUF1329 family proteins potentially function as esterases. From the data, we also postulate catabolic routes for the two biogasoline molecules isoprenol and isopentanol, which are catabolized via leucine metabolism after initial oxidation and activation with coenzyme A (CoA). Because fatty acids and alcohols may serve as both feedstocks and final products of metabolic-engineering efforts, the fitness data presented here will help guide future genomic modifications toward higher titers, rates, and yields. IMPORTANCE To engineer novel metabolic pathways into P. putida , a comprehensive understanding of the genetic basis of its versatile metabolism is essential. Here, we provide functional evidence for the putative roles of hundreds of genes involved in the fatty acid and alcohol metabolism of the bacterium. These data provide a framework facilitating precise genetic changes to prevent product degradation and to channel the flux of specific pathway intermediates as desired.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

Quantitative texture analysis at the WAND 2 and HIDRA diffractometers

Data collection and analysis strategies have been developed for efficient and reliable crystallographic texture measurements at two recently upgraded neutron diffractometers: the Wide Angle Neutron Diffractometer Squared (WAND 2 ) and the High Intensity Diffractometer for Residual Stress Analysis (HIDRA) at the High Flux Isotope Reactor located at Oak Ridge National Laboratory. These methods are demonstrated using measurements on a variety of textured samples, including multi-phase steel composites and polycrystalline calcite (CaCO 3 ). Reference measurements were also made at VULCAN, the engineering diffractometer located at the Spallation Neutron Source. The texture data obtained on the different instruments are in agreement, and WAND 2 is more time efficient than HIDRA. Two analysis methods were investigated, single-peak fitting to obtain individual pole figures for inversion and Rietveld texture analysis using MAUD. Here, the impact of the differences between the various textures obtained was evaluated through the calculation of diffraction elastic constants, which is one application of the texture data collected. Both instruments were found to provide texture data that are suitable for complementing other analyses, such as residual stress mapping.

47 OTHER INSTRUMENTATION↗
DOE OSTI2023

Status of HEU-Pb in the International Criticality Safety Benchmark Evaluation Project (ICSBEP) Handbook

The Department of Energy invests tens of millions of dollars each year to develop the next generation of nuclear engineering modeling & simulation (M&S) tools. These tools are used to analyze advanced reactor designs and the safety of current nuclear operations. As computers become more powerful, we are able to enhance resolution in our calculations. This improved resolution is taking us to a point where the limitations of simulation capability are in the quality of data, including our ability to quantify the uncertainty and sensitivity of the data. In order to model systems of interest with increasing accuracy, the industry must improve key nuclear data measurements. Thus, M&S tools need evaluated and quality-assured experimental data for validation purposes. The International Criticality Safety Benchmark Evaluation Project (ICSBEP) compiles and evaluates experiment data in a handbook that can be used by criticality safety engineers and others to validate computer codes and cross-section libraries at nuclear facilities. Both critical and subcritical experiments are included in the handbook. Figure 1 organizes all the benchmark evaluations that have been performed by the isotope of interest, in this case Pb, and the average neutron energy the system. Compared to other isotopes of interest for nuclear applications, there are few benchmark evaluations for Pb systems. The lack of integral measurements to determine errors in Pb cross-section data has caused the latest nuclear cross-section libraries to over/underestimate k eff compared to experimental results. Therefore, this evaluation fills an important knowledge gap in benchmark evaluations.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2021

SubTER ZTEM (Z-Axis Tipper Electromagnetic) Survey Results, Mineral Mountains Area, Utah

This data was acquired as part of the Subsurface Science, Technology and Engineering Research, and Development (SubTER) Crosscut which is a collaboration across the Department of Energy offices involved in research activities in energy production/extraction, subsurface storage, and environmental remediation. The contents of this zip contains the ZTEM survey measured and processed results for the Mineral Mountains area, Utah. The data were acquired through subcontract to Geotech Inc of Aurora, ON, Canada and their narrative report is included. There are also two readme.txt file included that further describe the data.

15 GEOTHERMAL ENERGY↗
DOE OSTI2026

Project Phase 1 Report: Reducing Data Center Peak Cooling Demand and Energy Costs With Cold Underground Thermal Energy Storage (Cold UTES)

Cold Underground Thermal Energy Storage (Cold UTES) is an ultra-long duration grid energy storage technology. With Cold-UTES, low-cost grid power is converted to cold thermal energy and stored in the native subsurface rock at the point of use. Cold UTES is one approach within the general category of engineered geothermal systems. Cold UTES for peak-hour cooling of data centers (DCs) was studied for deployment in Maricopa County, Arizona and Loudoun County, Virgina using thermal storage capacities from 4 GWh-th to over 1,000 GWh-th (>1 Terawatt-hour). The two sites have different power and transmission systems, available grid energy resources, daily and seasonal load profiles, weather conditions, and grid regulatory requirements. The study results indicate high value for both locations and because of this, likely indicates value across most of the US and the world. The basis of the study was a 1,000 MW-e hourly electric use of DC computing and auxiliary loads, which was modeled as 1 GW-th of thermal load to a dry cooled heat rejection system - i.e. a cooling system that does not consume water. The electric power required for cooling the DC varies as the air temperature changes. In cold weather only the dry-coolers are used, with an electrical load for cooling load as low as 10 MW-e. In hot summer hours, chillers and dry- coolers are required, which raises the electrical load for cooling load to as much as 300 MW-e. The continuous and peak cooling electrical loads result in a grid interconnection requirement of no less than 1,300 MW-e. From both a grid and thermal design modeling perspective the 1.3 GW-e could either be a single facility or result from the total load at multiple sites.

15 GEOTHERMAL ENERGY↗
DOE OSTI2020

Electronic Bottleneck Suppression in Next‐Generation Networks with Integrated Photonic Digital‐to‐Analog Converters

Digital‐to‐analog converters (DAC) are indispensable functional units in signal processing instrumentation and wide‐band telecommunication links for both civil and military applications. As photonic systems are capable of high data throughput and low latency, an increasingly found system limitation stems from the required domain crossing such as digital to analog and electronic to optical. A photonic DAC implementation, in contrast, enables a seamless signal conversion with respect to both energy efficiency and short signal delay, often requiring bulky discrete optical components and electric–optic transformation, hence introducing inefficiencies. Herein, a novel coherent parallel photonic DAC concept along with a 4‐bit experimental prototype capable of performing this DAC without optic–electric–optic domain crossing is introduced. This new paradigm guarantees a linear intensity weighting among bits when operating at high sampling rates (50 GHz), featuring an exceptional sampling efficiency (> 100 GS ) and small footprint (≈1 mm 2 ) in an 8‐bit implementation. Importantly, this photonic DAC enables seamless interfaces of next‐generation data processing hardware with high relevance in data centers, task‐specific compute accelerators such as neuromorphic engines, and network edge processing applications.

Meng, Jiawei↗