Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “fast algorithm”

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 1,117 records · Page 62
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

A Survey of Singular Value Decomposition Methods for Distributed Tall/Skinny Data

The Singular Value Decomposition (SVD) is one of the most important matrix factorizations, enjoying a wide variety of applications across numerous application domains. In statistics and data analysis, the common applications of SVD inclue Principal Components Analysis (PCA) and regression. Usually these applications arise on data that has far more rows than columns, so-called "tall/skinny" matrices. In the big data analytics context, this may take the form of hundreds of millions to billions of rows with only a few hundred columns. There is a need, therefore, for fast, accurate, and scalable tall/skinny SVD implementations which can fully utilize modern computing resources. To that end, we present a survey of three different algorithms for computing the SVD for these kinds of tall/skinny data layouts using MPI for communication. We contextualize these with common big data analytics techniques. Finally, we present both CPU and GPU timing results from the Summit supercomputer, and discuss possible alternative approaches.

Schmidt, Drew↗
DOE OSTI2022

Combining Spike Time Dependent Plasticity (STDP) and Backpropagation (BP) for Robust and Data Efficient Spiking Neural Networks (SNN)

National security applications require artificial neural networks (ANNs) that consume less power, are fast and dynamic online learners, are fault tolerant, and can learn from unlabeled and imbalanced data. We explore whether two fundamentally different, traditional learning algorithms from artificial intelligence and the biological brain can be merged. We tackle this problem from two directions. First, we start from a theoretical point of view and show that the spike time dependent plasticity (STDP) learning curve observed in biological networks can be derived using the mathematical framework of backpropagation through time. Second, we show that transmission delays, as observed in biological networks, improve the ability of spiking networks to perform classification when trained using a backpropagation of error (BP) method. These results provide evidence that STDP could be compatible with a BP learning rule. Combining these learning algorithms will likely lead to networks more capable of meeting our national security missions.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Fast Extraction and Characterization of Fundamental Frequency Events from a Large PMU Dataset using Big Data Analytics

A novel method for fast extraction of fundamental frequency events (FFE) based on measurements of frequency and rate of change of frequency by Phasor Measurement Units (PMU) is introduced. The method is designed to work with exceptionally large historical PMU datasets. Statistical analysis was used to extract the features and train Random Forest and Catboost classifiers. The method is capable of fast extraction of FFE from a historical dataset containing measurements from hundreds of PMUs captured over multiple years. The reported accuracy of the best algorithm for classification expressed as Area Under the receiver operating Characteristic curve reaches 0.98, which was obtained in out-of-sample evaluations on 109 system-wide events over 2 years observed at 43 PMUs. Then Minimum Volume Enclosing Ellipsoid Algorithm was used to further analyze the events. 93.72% events were correctly characterized, where average duration of the event as seen by the PMU was 9.93 sec.

Baembitov, Rashid↗
NASA NTRS1977

Variants and extensions of a fast direct numerical cauchy-riemann solver, with illustrative applications

Revised and extended versions of a fast, direct (noniterative) numerical Cauchy-Riemann solver are presented for solving finite difference approximations of first order systems of partial differential equations. Although the difference operators treated are linear and elliptic, one significant application of these extended direct Cauchy-Riemann solvers is in the fast, semidirect (iterative) solution of fluid dynamic problems governed by the nonlinear mixed elliptic-hyperbolic equations of transonic flow. Different versions of the algorithms are derived and the corresponding FORTRAN computer programs for a simple example problem are described and listed. The algorithms are demonstrated to be efficient and accurate.

Martin, E. D.↗
DOE OSTI2023

Fast model-based scenario optimization in NSTX-U enabled by analytic gradient computation

Model-based optimization offers a systematic approach to advanced scenario planning. In this case, the feedforward-control inputs (actuator trajectories) that are needed to attain and sustain a desired scenario are obtained by solving a nonlinear constrained optimization problem. This class of problems generally minimize a cost function that measures the difference between desired and actual plasma states. Several numerical optimization algorithms, such as sequential quadratic programming, require repeated calculation of the cost function gradients with respect to the input trajectories. Calculating these gradients numerically can be computationally intensive, increasing the time needed to solve the feedforward-control optimization problem. Here, this work introduces a method to analytically calculate these cost function gradients from the current profile evolution model. This can significantly reduce the computational time and allow for fast feedforward-control optimization, which would eventually enable optimal scenario planning between discharges. The performance of the feedforward optimizer with analytical gradients is compared to a traditional optimization algorithm based on numerical gradients for different NSTX-U scenarios. The plasma dynamics in the optimization algorithm are simulated using the Control Oriented Transport SIMulator (COTSIM). Results of the work show that analytical gradients consistently reduce the computation time while achieving trajectories that are comparable to those obtained by traditional optimization algorithms based on numerical gradients.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2024

Modifying the Asynchronous Jacobi Method for Data Corruption Resilience

Moving scientific computation from high-performance computing (HPC) and cloud computing (CC) environments to devices on the edge, i.e., physically near instruments of interest, has received tremendous interest in recent years. Such edge computing environments can operate on data in situ, offering enticing benefits over data aggregation to HPC and CC facilities that include avoiding costs of transmission, increased data privacy, and real-time data analysis. Because of the inherent unreliability of edge computing environments, new fault-tolerant approaches must be developed before the benefits of edge computing can be realized. Motivated by algorithm-based fault tolerance, a variant of the asynchronous Jacobi (ASJ) method is developed that achieves resilience to data corruption by rejecting solution approximations from neighbor devices according to a bound derived from convergence theory. Numerical results on a two-dimensional Poisson problem show that the new rejection criterion, along with a novel approximation to the shortest path length on which the criterion depends, restores convergence for the ASJ variant in the presence of certain types data corruption. Numerical results are obtained for when the singular values in the analytic bound are approximated. Additional linear systems are also explored, one with a more dense sparsity pattern and one that includes advection. All results indicate that successful resilience to data corruption depends on whether the bound tightens fast enough to reject corrupted data before the iteration evolution deviates significantly from that predicted by the convergence theory defining the bound. This observation generalizes to future work on algorithm-based fault tolerance for other asynchronous algorithms, including upcoming approaches that leverage Krylov subspaces.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

A scalable multidimensional fully implicit solver for Hall magnetohydrodynamics

We propose an optimally performant fully implicit algorithm for the Hall magnetohydrodynamics (HMHD) equations based on multigrid-preconditioned Jacobian-free Newton-Krylov methods. HMHD is a challenging system to solve numerically because it supports stiff fast dispersive waves. The preconditioner is formulated using an operator-split approximate block factorization (Schur complement), informed by physics insight. We use a vector-potential formulation (instead of a magnetic field one) to allow a clean segregation of the problematic $\nabla$ x $\nabla$ x operator in the electron Ohm's law subsystem. This segregation allows the formulation of an effective damped block-Jacobi smoother for multigrid. We demonstrate by analysis that our proposed block-Jacobi iteration is convergent and has the smoothing property. The resulting HMHD solver is verified linearly with wave propagation examples, and nonlinearly with the GEM challenge reconnection problem by comparison against another HMHD code. We demonstrate the excellent algorithmic and parallel performance of the algorithm up to 16384 MPI tasks in two dimensions.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
NASA NTRS1995

The NAS Computational Aerosciences Archive

In order to further the state-of-the-art in computational aerosciences (CAS) technology, researchers must be able to gather and understand existing work in the field. One aspect of this information gathering is studying published work available in scientific journals and conference proceedings. However, current scientific publications are very limited in the type and amount of information that they can disseminate. Information is typically restricted to text, a few images, and a bibliography list. Additional information that might be useful to the researcher, such as additional visual results, referenced papers, and datasets, are not available. New forms of electronic publication, such as the World Wide Web (WWW), limit publication size only by available disk space and data transmission bandwidth, both of which are improving rapidly. The Numerical Aerodynamic Simulation (NAS) Systems Division at NASA Ames Research Center is in the process of creating an archive of CAS information on the WWW. This archive will be based on the large amount of information produced by researchers associated with the NAS facility. The archive will contain technical summaries and reports of research performed on NAS supercomputers, visual results (images, animations, visualization system scripts), datasets, and any other supporting meta-information. This information will be available via the WWW through the NAS homepage, located at http://www.nas.nasa.gov/, fully indexed for searching. The main components of the archive are technical summaries and reports, visual results, and datasets. Technical summaries are gathered every year by researchers who have been allotted resources on NAS supercomputers. These summaries, together with supporting visual results and references, are browsable by interested researchers. Referenced papers made available by researchers can be accessed through hypertext links. Technical reports are in-depth accounts of tools and applications research projects performed by NAS staff members and collaborators. Visual results, which may be available in the form of images, animations, and/or visualization scripts, are generated by researchers with respect to a certain research project, depicting dataset features that were determined important by the investigating researcher. For example, script files for visualization systems (e.g. FAST, PLOT3D, AVS) are provided to create visualizations on the user's local workstation to elucidate the key points of the numerical study. Users can then interact with the data starting where the investigator left off. Datasets are intended to give researchers an opportunity to understand previous work, 'mine' solutions for new information (for example, have you ever read a paper thinking "I wonder what the helicity density looks like?"), compare new techniques with older results, collaborate with remote colleagues, and perform validation. Supporting meta-information associated with the research projects is also important to provide additional context for research projects. This may include information such as the software used in the simulation (e.g. grid generators, flow solvers, visualization). In addition to serving the CAS research community, the information archive will also be helpful to students, visualization system developers and researchers, and management. Students (of any age) can use the data to study fluid dynamics, compare results from different flow solvers, learn about meshing techniques, etc., leading to better informed individuals. For these users it is particularly important that visualization be integrated into dataset archives. Visualization researchers can use dataset archives to test algorithms and techniques, leading to better visualization systems, Management can use the data to figure what is really going on behind the viewgraphs. All users will benefit from fast, easy, and convenient access to CFD datasets. The CAS information archive hopes to serve as a useful resource to those interested in computational sciences. At present, only information that may be distributed internationally is made available via the archive. Studies are underway to determine security requirements and solutions to make additional information available. By providing access to the archive via the WWW, the process of information gathering can be more productive and fruitful due to ease of access and ability to manage many different types of information. As the archive grows, additional resources from outside NAS will be added, providing a dynamic source of research results.

Miceli, Kristina D.↗
DOE OSTI2020

Fast GPU 3D diffeomorphic image registration

3D image registration is one of the most fundamental and computationally expensive operations in medical image analysis. Here, we present a mixed-precision, Gauss–Newton–Krylov solver for diffeomorphic registration of two images. Our work extends the publicly available CLAIRE library to GPU architectures. Despite the importance of image registration, only a few implementations of large deformation diffeomorphic registration packages support GPUs. Our contributions are new algorithms to significantly reduce the run time of the two main computational kernels in CLAIRE: calculation of derivatives and scattered-data interpolation. Additionally, we deploy (i) highly-optimized, mixed-precision GPU-kernels for the evaluation of scattered-data interpolation, (ii) replace Fast-Fourier-Transform (FFT)-based first-order derivatives with optimized 8th-order finite differences, and (iii) compare with state-of-the-art CPU and GPU implementations. As a highlight, we demonstrate that we can register clinical images in less than 6 s on a single NVIDIA Tesla V100. This amounts to over 20 speed-up over the current version of CLAIRE and over 30 speed-up over existing GPU implementations.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Rapid CFD Using Machine Learning Algorithms (CRADA Final Report)

This is a collaborative effort between Lawrence Livermore National Security, LLC as manager and operator of Lawrence Livermore National Laboratory (“LLNL”) and Guardian Glass, LLC ("Guardian Glass") to develop a fast-running emulator of the reactive Computational Fluid Dynamics (“CFD”) simulations needed to understand the complex reactions and flows in the glass melting, fining, and forming subprocesses. This CRADA project is sponsored under the High-Performance Computing for Manufacturing (“HPC4Mfg”) Program of the Department of Energy’s Advanced Manufacturing Office (“AMO”) within the Energy Efficiency and Renewable Energy (“EERE”) Office.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Data reduction through optimized scalar quantization for more compact neural networks

Raw data generation for several existing and planned large physics experiments now exceeds TB/s rates, generating untenable data sets in very little time. Those data often demonstrate high dimensionality while containing limited information. Meanwhile, Machine Learning algorithms are now becoming an essential part of data processing and data analysis. Those algorithms can be used offline for post processing and post data analysis, or they can be used online for real time processing providing ultra low latency experiment monitoring. Both use cases would benefit from data throughput reduction while preserving relevant information: one by reducing the offline storage requirements by several orders of magnitude and the other by allowing ultra fast online inferencing with low complexity Machine Learning models. Moreover, reducing the data source throughput also reduces material cost, power and data management requirements. In this work we demonstrate optimized nonuniform scalar quantization for data source reduction. This data reduction allows lower dimensional representations while preserving the relevant information of the data, thus enabling high accuracy Tiny Machine Learning classifier models for online fast inferences. We demonstrate this approach with an initial proof of concept targeting the CookieBox, an array of electron spectrometers used for angular streaking, that was developed for LCLS-II as an online beam diagnostic tool. We used the Lloyd-Max algorithm with the CookieBox dataset to design an optimized nonuniform scalar quantizer. Optimized quantization lets us reduce input data volume by 69% with no significant impact on inference accuracy. When we tolerate a 2% loss on inference accuracy, we achieved 81% of input data reduction. Finally, the change from a 7-bit to a 3-bit input data quantization reduces our neural network size by 38%.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Scaling pair count to next galaxy surveys

ABSTRACT Counting pairs of galaxies or stars according to their distance is at the core of real-space correlation analyses performed in astrophysics and cosmology. Upcoming galaxy surveys (LSST, Euclid) will measure properties of billions of galaxies challenging our ability to perform such counting in a minute-scale time relevant for the usage of simulations. The problem is only limited by efficient access to the data, hence belongs to the big data category. We use the popular Apache Spark framework to address it and design an efficient high-throughput algorithm to deal with hundreds of millions to billions of input data. To optimize it, we revisit the question of non-hierarchical sphere pixelization based on cube symmetries and develop a new one dubbed the ‘Similar Radius Sphere Pixelization’ (SARSPix) with very close to square pixels. It provides the most adapted indexing over the sphere for all distance-related computations. Using LSST-like fast simulations, we compute autocorrelation functions on tomographic bins containing between a hundred million to one billion data points. In each case, we achieve the construction of a standard pair-distance histogram in about 2 min, using a simple algorithm that is shown to scale, over a moderate number of nodes (16–64). This illustrates the potential of this new techniques in the field of astronomy where data access is becoming the main bottleneck. They can be easily adapted to other use-cases as nearest-neighbours search, catalogue cross-match or cluster finding. The software is publicly available from https://github.com/astrolabsoftware/SparkCorr.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2020

Evaluation of smart charging for electric vehicle-to-building integration: A case study

Higher electric vehicle (EV) adoption will stress the importance of demand flexibility to achieve more economic, efficient, and reliable grid operation. Charging technologies will be paramount in shifting temporally to better fit the variable generation of wind and solar. As such, analysis is warranted on the benefits of EV charge scheduling with respect to installation cost, operation cost, difficulty of implementation, and grid flexibility. We tackle this by analyzing the cost savings of implementing an EV charge scheduling infrastructure to reduce demand charges and installation costs. In this paper, we analyze a case study for operation of 16 level 2 chargers and 1 fast charger for two different building types. We then evaluate various test phases for controlling building and charging loads using an adaptive charging network (ACN) algorithm to characterize the ACN’s potential to reduce overall project cost.

30 DIRECT ENERGY CONVERSION↗
DOE OSTI2023

Writing the Programs of Programmable Catalysis

It has long been known that non-steady state and periodic catalytic reactor operation in terms of temperature, pressure, and composition can lead to higher overall productivity and/or product selectivity than the best steady operation. Recently, the emergence of catalysts whose intrinsic properties can be made to oscillate with time, introduces advanced forcing capabilities that can be “programmed” into the catalysts to broaden the scope and applicability of periodic operation to surface chemistry. In this work, an algorithmic approach is implemented to significantly accelerate the discovery and optimization of periodic steady states of catalytic reactors. Decomposition of complex dynamics into fundamental mechanistic fast–slow steps is seen to improve conceptual understanding of the relationship between binding energy oscillation protocols and overall catalytic rates. Finding structured forcing protocols, optimally tailored to the multiple time scales of a given individual mechanism, requires an efficient search of high-dimensional parameter spaces. Here, this is enabled here through active learning (Bayesian optimization, enhanced by our proposed Bayesian continuation). Implementation of these methods is shown to accelerate the evaluation of catalyst programs by up to several orders of magnitude. Faster screening of programmable catalysts to discover periodic steady states enables the optimization of catalytic operating protocols and thus opens the possibility for catalyst engineering based on optimal forcing programs to control rate and product selectivity, even for complex multistep catalytic mechanisms.

catalysis↗
DOE OSTI2022

A Novel Approach to Quantum Circuit Partitioning

Quantum synthesis presents an effective method of circuit optimization, but scales exponentially with the number of qubits in the circuit. This problem can be addressed by partitioning the circuit into blocks with a limited number of qubits. Existing partitioning algorithms make large trade-offs to achieve either high speed or quality. We propose a method of circuit partitioning which is competitive with existing algorithms for both metrics. The proposed method is compared with two existing methods across common circuit architectures, matching an exhaustive solution in performance and a fast solution on time.

Clark, Joseph↗
DOE OSTI2025

NuFast-LBL

NuFast is designed to calculate all nine neutrino oscillation probabilities in matter for long-baseline accelerator (e.g. NOvA, T2K, DUNE, HK) and reactor experiments (e.g. JUNO) very quickly, using the algorithms optimized for realistic oscillation scenarios. NuFast is provided in Fortran, C++, and Python, although no particular guarantees are made that the Python code is "fast".

Denton, PeterB. [Brookhaven National Laboratory (B↗
DOE OSTI2023

Series FACTS Devices for Increasing Resiliency in Severe Weather Conditions

Severe weather conditions are low-probability, high-impact events that affect grid operations. The majority of power outages are caused by severe weather conditions. Grid resiliency to weather events can be enhanced by decreasing the reliance on its affected sections. One way to do this is to reduce the power flow through lines vulnerable to severe weather. If a line is disconnected, its initial power flow is distributed through the neighbor lines, which may cause congestion in the grid. FACTS devices can be used to control the power flow of lines that have a higher chance of power outages. Most previous works do not consider weather events in power flow control. In this work, a linearized optimal power flow (OPF)–based algorithm is developed to minimize the real power flow of vulnerable lines considering the thermal limits of lines to prevent infeasible solutions; the simulation is fast, making it suitable for large-scale systems. The proposed optimization problem is presented as a mixed-integer linear program (MILP), making it capable of using short-term load forecasting due to its high solution speed. The proposed optimization problem considers multiple lines with different outage probabilities and the uncertainties of the weather forecast. Moreover, it estimates the power reduction in vulnerable lines due to changes in the series FACTS devices. The performance of the proposed optimization problem is tested on IEEE 14-, 30-, and 118-bus systems for several scenarios. The results are validated with the AC power flow results from MATPOWER.

24 POWER TRANSMISSION AND DISTRIBUTION↗
NASA NTRS1969

Fast Fourier Transform Spectral Analysis Program

Fast Fourier Transform Spectral Analysis Program is used in frequency spectrum analysis of postflight, space vehicle telemetered trajectory data. This computer program with a digital algorithm can calculate power spectrum rms amplitudes and cross spectrum of sampled parameters at even time increments.

Daniel, J. A., Jr.↗