Search NASA⌕ Search

SEARCH · Search NASA

Results for “Domain Distributed Framework”

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 289 records · Page 16

SMALE: Enhancing Scalability of Machine Learning Algorithms on Extreme-Scale Computing Platforms

Deployment and execution of machine learning tasks on extreme-scale computing platforms face several significant technical challenges: 1) High computing cost incurred by dense networks – The computing workload of deep networks with densely-connected topology increases rapidly with the network size, imposing a non-scalable computing model of extreme-scale computing platforms; 2) Non-optimized workload distribution – Many advanced deep learning algorithms, e.g., sparsification and irregular net-work topology, produce very unbalanced workload distribution on extreme-scale computing platforms. The computation efficiency is greatly hindered by the incurred data and computation redundancies as well as long tails of the node with extensive workload; 3) Constraints in data movement and I/O bottle-neck – Inter-node data movement in extreme-scale computing platforms are associated with high energy and latency costs, and subject to the constraints of I/O bandwidth; and 4) Generalization of algorithm realization and acceleration on computing platforms – The large varieties of machine learning algorithms and structures of extreme-scale computing platforms make the derivation of a generalized algorithm realization and acceleration method very challenging, which, however, is the requirement by domain scientists and interested users. We call the above challenges Smale’s Problems in Machine Learning and Understanding for High-Performance Computing Scientific Discovery. The objective of our three-year research project is to develop a holistic innovation set at structure, assembly, and acceleration layers of machine learning algorithms to address the above challenges in algorithm deployment and execution. Three tasks are particularly performed, including: At the algorithm structure level, we investigate the techniques that can structurally sparsify on the topology of deep networks for computing workload reduction. We also study clustering and pruning techniques that can optimize the workload distributions over the extreme-scale computing platforms; At the algorithm assembly level, we derive a unified learning framework for unsupervised transfer learning and dynamic growing capabilities. Novel training methods are also exploited to enhance the training efficiency of the proposed framework; At the algorithm acceleration level, we will develop a series of techniques that can accelerate the computation of sparse matrix operations, which are one of the core executions in deep learning and optimize memory access of the concerned platforms. Our proposed techniques attack the fundamental problems in machine learning algorithms running on extreme-scale computing platforms by vertically integrating the solutions at three closely entangled layers, paving the long-term scaling path of machine learning applications under DOE context. Three tasks corresponding to the above respective research orientations are performed during the three-year project period with our collaborators at ORNL. The outcome of the proposed project is anticipated to form a holistic solution set of novel algorithms and network topologies, efficient training techniques, and fast acceleration methods to promote the computing scalability of the machine learning applications of particular interest to DOE.

97 MATHEMATICS AND COMPUTING↗

Modeling aerosol growth by aqueous chemistry in nonprecipitating stratiform cloud

A new microphysics module based on a two-dimensional (2D) joint size distribution function representing both interstitial and cloud particles is developed and applied to studying aerosol processing in non-precipitating stratocumulus clouds. The module is implemented in a three-dimensional dynamical framework of a large-eddy simulation (LES) model and in a trajectory ensemble model (TEM). Both models are used to study the modification of sulfate aerosol by the activation - aqueous chemistry - resuspension cycle in shallow marine stratocumulus clouds. The effect of particle mixing and different size-distribution representations on modeled aerosol processing are studied in a comparison of the LES and TEM simulations with the identical microphysics treatment exposes and a comparison of TEM simulations with a 2D fixed and moving bin microphysics. Particle mixing which is represented in LES and neglected in the TEM leads to the mean relative per particle dry mass change in the TEM simulations being about 30% lower than in analogous subsample of LES domain. Particles in the final LES spectrum are mixed in from different “parcels”, some of which have experienced longer in-cloud residence times than the TEM parcels, all of which originated in the subcloud layer, have. The mean relative per particle dry mass change differs by 14% between TEM simulations with fixed and moving bin microphysics. Finally, the TEM model with the moving bin microphysics is used to evaluate assumptions about liquid water mass partitioning among activated cloud condensation nuclei (CCN) of different dry sizes. These assumptions are used in large-scale models to map the bulk aqueous chemistry sulfate production, which is largely proportional to the liquid water mass, to the changes in aerosol size distribution. It is shown that the commonly used assumptions that the droplet mass is independent of CCN size or that the droplet mass is proportional to the CCN size to the third power do not perform well in the considered case. The explicitly predicted water partitioning indicates that the mean mass of droplets participating in the models aqueous chemistry calculations is proportional to the dry CCN size.

58 GEOSCIENCES↗

Electromagnetic resonant modes of dielectric sphere bilayers

Sphere bilayers have been proposed as promising structures for electromagnetic management in photonic crystal devices. These arrangements are made of two intertwined subsets of spheres of different size and refractive index, one subset filling the interstitial sites of the other. We present a systematic study of the electromagnetic resonant modes of the bilayers, in comparison with those of the constituent subsets of spheres. Three samples were built with glass and Teflon spheres and their transmission spectra measured in the microwave range (10–25 GHz). Simulations with finite integration time-domain method are in good agreement with experiments. Results show that the bilayer presents the same resonances as one of the subsets but modified by the presence of the other in its resonant frequencies and in the electric field distributions. As this distortion is not very large, the number of resonances in a selected spectral region is determined by the dominant subset. The degree of freedom that offers the bilayer could be useful to fine tune the resonances of the structure for different applications. A map of modes useful to guide this design is also presented. Scale invariance of Maxwell equations allows the translation of these results in the microwave range to the visible region; hence, some possible applications are discussed in this framework.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Preparing For Advanced Air Mobility: A System-Level Framework For Infrastructure, Investment, and Deployment

Advanced air mobility (AAM) is moving from demonstration toward early deployment, supported by a growing national strategy that outlines how these systems may evolve across airspace, infrastructure, and operations. As this transition takes shape, a more practical question comes into focus: what does it mean to be ready? This paper introduces a Capability Maturity Model (CMM) as a structured way to think about that challenge. Rather than treating readiness as a fixed condition, it frames it as a progression - one that develops across multiple, interdependent domains over time.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Fast Learning for Immersive Engagement in Energy Simulations

The fast computation which is critical for immersive engagement with and learning from energy simulations would be furthered by developing a general method for creating rapidly computed simplified versions of NREL's computation-intensive energy simulations. Created using machine learning techniques, these 'reduced form' simulations can provide statistically sound estimates of the results of the full simulations at a fraction of the computational cost with response times - typically less than one minute of wall-clock time - suitable for real-time human-in-the-loop design and analysis. Additionally, uncertainty quantification techniques can document the accuracy of the approximate models and their domain of validity. Approximation methods are applicable to a wide range of computational models, including supply-chain models, electric power grid simulations, and building models. These reduced-form representations cannot replace or re-implement existing simulations, but instead supplement them by enabling rapid scenario design and quality assurance for large sets of simulations. We present an overview of the framework and methods we have implemented for developing these reduced-form representations.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Parallel Implicit Algorithms for CFD

The main goal of this project was efficient distributed parallel and workstation cluster implementations of Newton-Krylov-Schwarz (NKS) solvers for implicit Computational Fluid Dynamics (CFD.) "Newton" refers to a quadratically convergent nonlinear iteration using gradient information based on the true residual, "Krylov" to an inner linear iteration that accesses the Jacobian matrix only through highly parallelizable sparse matrix-vector products, and "Schwarz" to a domain decomposition form of preconditioning the inner Krylov iterations with primarily neighbor-only exchange of data between the processors. Prior experience has established that Newton-Krylov methods are competitive solvers in the CFD context and that Krylov-Schwarz methods port well to distributed memory computers. The combination of the techniques into Newton-Krylov-Schwarz was implemented on 2D and 3D unstructured Euler codes on the parallel testbeds that used to be at LaRC and on several other parallel computers operated by other agencies or made available by the vendors. Early implementations were made directly in Massively Parallel Integration (MPI) with parallel solvers we adapted from legacy NASA codes and enhanced for full NKS functionality. Later implementations were made in the framework of the PETSC library from Argonne National Laboratory, which now includes pseudo-transient continuation Newton-Krylov-Schwarz solver capability (as a result of demands we made upon PETSC during our early porting experiences). A secondary project pursued with funding from this contract was parallel implicit solvers in acoustics, specifically in the Helmholtz formulation. A 2D acoustic inverse problem has been solved in parallel within the PETSC framework.

Keyes, David E.↗

Graphical Optimization of Spectral Shift Reconstructions for Optical Backscatter Reflectometry

Optical backscatter reflectometry (OBR) is an interferometric technique that can be used to measure local changes in temperature and mechanical strain based on spectral analyses of backscattered light from a singlemode optical fiber. The technique uses Fourier analyses to resolve spectra resulting from reflections occurring over a discrete region along the fiber. These spectra are cross-correlated with reference spectra to calculate the relative spectral shifts between measurements. The maximum of the cross-correlated spectra—termed quality—is a metric that quantifies the degree of correlation between the two measurements. Recently, this quality metric was incorporated into an adaptive algorithm to (1) selectively vary the reference measurement until the quality exceeds a predefined threshold and (2) calculate incremental spectral shifts that can be summed to determine the spectral shift relative to the initial reference. Using a graphical (network) framework, this effort demonstrated the optimal reconstruction of distributed OBR measurements for all sensing locations using a maximum spanning tree (MST). By allowing the reference to vary as a function of both time and sensing location, the MST and other adaptive algorithms could resolve spectral shifts at some locations, even if others can no longer be resolved.

47 OTHER INSTRUMENTATION↗

Survey of Cybersecurity Governance, Threats, and Countermeasures for the Power Grid

The convergence of Information Technologies and Operational Technology systems in industrial networks presents many challenges related to availability, integrity, and confidentiality. In this paper, we evaluate the various cybersecurity risks in industrial control systems and how they may affect these areas of concern, with a particular focus on energy-sector Operational Technology systems. There are multiple threats and countermeasures that Operational Technology and Information Technology systems share. Since Information Technology cybersecurity is a relatively mature field, this paper emphasizes on threats with particular applicability to Operational Technology and their respective countermeasures. We identify regulations, standards, frameworks and typical system architectures associated with this domain. We review relevant challenges, threats, and countermeasures, as well as critical differences in priorities between Information and Operational Technology cybersecurity efforts and implications. These results are then examined against the recommended National Institute of Standards and Technology framework for gap analysis to provide a complete approach to energy sector cybersecurity. We provide analysis of countermeasure implementation to align with the continuous functions recommended for a sound cybersecurity framework.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

Transitioning from File-Based HPC Workflows to Streaming Data Pipelines with openPMD and ADIOS2

This paper aims to create a transition path from file-based IO to streaming-based workflows for scientific applications in an HPC environment. By using the openPMP-api, traditional workflows limited by filesystem bottlenecks can be overcome and flexibly extended for in situ analysis. The openPMD-api is a library for the description of scientific data according to the Open Standard for Particle-Mesh Data (openPMD). Its approach towards recent challenges posed by hardware heterogeneity lies in the decoupling of data description in domain sciences, such as plasma physics simulations, from concrete implementations in hardware and IO. The streaming backend is provided by the ADIOS2 framework, developed at Oak Ridge National Laboratory. This paper surveys two openPMD-based loosely-coupled setups to demonstrate flexible applicability and to evaluate performance. In loose coupling, as opposed to tight coupling, two (or more) applications are executed separately, e.g. in individual MPI contexts, yet cooperate by exchanging data. This way, a streaming-based workflow allows for standalone codes instead of tightly-coupled plugins, using a unified streaming-aware API and leveraging high-speed communication infrastructure available in modern compute clusters for massive data exchange. We determine new challenges in resource allocation and in the need of strategies for a flexible data distribution, demonstrating their influence on efficiency and scaling on the Summit compute system. The presented setups show the potential for a more flexible use of compute resources brought by streaming IO as well as the ability to increase throughput by avoiding filesystem bottlenecks.

Poeschel, Franz↗

Uncertainty Analysis via Failure Domain Characterization: Polynomial Requirement Functions

This paper proposes an uncertainty analysis framework based on the characterization of the uncertain parameter space. This characterization enables the identification of worst-case uncertainty combinations and the approximation of the failure and safe domains with a high level of accuracy. Because these approximations are comprised of subsets of readily computable probability, they enable the calculation of arbitrarily tight upper and lower bounds to the failure probability. A Bernstein expansion approach is used to size hyper-rectangular subsets while a sum of squares programming approach is used to size quasi-ellipsoidal subsets. These methods are applicable to requirement functions whose functional dependency on the uncertainty is a known polynomial. Some of the most prominent features of the methodology are the substantial desensitization of the calculations from the uncertainty model assumed (i.e., the probability distribution describing the uncertainty) as well as the accommodation for changes in such a model with a practically insignificant amount of computational effort.

Crespo, Luis G.↗

Framework for Processing Citizens Science Data for Applications to NASA Earth Science Missions

Citizen science (or crowdsourcing) has drawn much high-level recent and ongoing interest and support. It is poised to be applied, beyond the by-now fairly familiar use of, e.g., Twitter for natural hazards monitoring, to science research, such as augmenting the validation of NASA earth science mission data. This interest and support is seen in the 2014 National Plan for Civil Earth Observations, the 2015 White House forum on citizen science and crowdsourcing, the ongoing Senate Bill 2013 (Crowdsourcing and Citizen Science Act of 2015), the recent (August 2016) Open Geospatial Consortium (OGC) call for public participation in its newly-established Citizen Science Domain Working Group, and NASA's initiation of a new Citizen Science for Earth Systems Program (along with its first citizen science-focused solicitation for proposals). Over the past several years, we have been exploring the feasibility of extracting from the Twitter data stream useful information for application to NASA precipitation research, with both "passive" and "active" participation by the twitterers. The Twitter database, which recently passed its tenth anniversary, is potentially a rich source of real-time and historical global information for science applications. The time-varying set of "precipitation" tweets can be thought of as an organic network of rain gauges, potentially providing a widespread view of precipitation occurrence. The validation of satellite precipitation estimates is challenging, because many regions lack data or access to data, especially outside of the U.S. and in remote and developing areas. Mining the Twitter stream could augment these validation programs and, potentially, help tune existing algorithms. Our ongoing work, though exploratory, has resulted in key components for processing and managing tweets, including the capabilities to filter the Twitter stream in real time, to extract location information, to filter for exact phrases, and to plot tweet distributions. The key step is to process the "precipitation" tweets to be compatible with satellite-retrieved precipitation data. These key components for processing and managing "precipitation" tweets (and additional ones to be developed) are not limited to precipitation, nor are they limited to the Twitter social medium. Indeed, to maximize the value of our work for NASA earth science programs, these components should be generalized and be part of an overall framework for processing citizen science data for science research. In this paper, we outline such a framework.

earth science satellite data↗

Automatic Halo Management for the Uintah GPU-Heterogeneous Asynchronous Many-Task Runtime

The Uintah computational framework is used for the parallel solution of partial differential equations on adaptive mesh refinement grids using modern supercomputers. Uintah is structured with an application layer and a separate runtime system. Uintah is based on a distributed directed acyclic graph of computational tasks, with a task scheduler that efficiently schedules and executes these tasks on both CPU cores and on-node accelerators. The runtime system identifies task dependencies, creates a task graph prior to the execution of these tasks, automatically generates MPI message tags, and automatically performs halo transfers for simulation variables. Automating halo transfers in a heterogeneous environment poses significant challenges when tasks compute within a few milliseconds, as runtime overhead affects wall time execution, or when simulation variables require large halos spanning most or all of the computational domain, as task dependencies become expensive to process. These challenges are magnified at production scale when application developers require each compute node perform thousands of different halo transfers among thousands simulation variables. The principal contribution of this work is to (1) identify and address inefficiencies that arise when mapping tasks onto the GPU in the presence of automated halo transfers, (2) implement new schemes to reduce runtime system overhead, (3) minimize application developer involvement with the runtime, and (4) show overhead reduction results from these improvements.

Concurrency↗

Large-Scale Multi-Dimensional Document Clustering on GPU Clusters

Document clustering plays an important role in data mining systems. Recently, a flocking-based document clustering algorithm has been proposed to solve the problem through simulation resembling the flocking behavior of birds in nature. This method is superior to other clustering algorithms, including k-means, in the sense that the outcome is not sensitive to the initial state. One limitation of this approach is that the algorithmic complexity is inherently quadratic in the number of documents. As a result, execution time becomes a bottleneck with large number of documents. In this paper, we assess the benefits of exploiting the computational power of Beowulf-like clusters equipped with contemporary Graphics Processing Units (GPUs) as a means to significantly reduce the runtime of flocking-based document clustering. Our framework scales up to over one million documents processed simultaneously in a sixteennode GPU cluster. Results are also compared to a four-node cluster with higher-end GPUs. On these clusters, we observe 30X-50X speedups, which demonstrates the potential of GPU clusters to efficiently solve massive data mining problems. Such speedups combined with the scalability potential and accelerator-based parallelization are unique in the domain of document-based data mining, to the best of our knowledge.

99 GENERAL AND MISCELLANEOUS↗

Representativeness-based Sampling Network Design for the State of Alaska

This data set collection consists of data products described in Hoffman et. al., 2013. Resource and logistical constraints limit the frequency and extent of environmental observations, particularly in the Arctic, necessitating the development of a systematic sampling strategy to maximize coverage and objectively represent environmental variability at desired scales. A quantitative methodology for stratifying sampling domains, informing site selection, and determining the representativeness of measurement sites and networks is described here. Multivariate spatiotemporal clustering was applied to down-scaled general circulation model results and data for the State of Alaska at 4 km2 resolution to define multiple sets of ecoregions across two decadal time periods. Maps of ecoregions for the present (2000-2009) and future (2090-2099) were produced, showing how combinations of 37 characteristics are distributed and how they may shift in the future. Representative sampling locations are identified on present and future ecoregion maps. A representativeness metric was developed, and representativeness maps for eight candidate sampling locations were produced. This metric was used to characterize the environmental similarity of each site. This analysis provides model-inspired insights into optimal sampling strategies, offers a framework for up-scaling measurements, and provides a down-scaling approach for integration of models and measurements. These techniques can be applied at different spatial and temporal scales to meet the needs of individual measurement campaigns. This dataset contains one zipped file, one .txt file, and one .sh file.The Next-Generation Ecosystem Experiments: Arctic (NGEE Arctic), was a research effort to reduce uncertainty in Earth System Models by developing a predictive understanding of carbon-rich Arctic ecosystems and feedbacks to climate. NGEE Arctic was supported by the Department of Energy's Office of Biological and Environmental Research.The NGEE Arctic project had two field research sites: 1) located within the Arctic polygonal tundra coastal region on the Barrow Environmental Observatory (BEO) and the North Slope near Utqiagvik (Barrow), Alaska and 2) multiple areas on the discontinuous permafrost region of the Seward Peninsula north of Nome, Alaska.Through observations, experiments, and synthesis with existing datasets, NGEE Arctic provided an enhanced knowledge base for multi-scale modeling and contributed to improved process representation at global pan-Arctic scales within the Department of Energy's Earth system Model (the Energy Exascale Earth System Model, or E3SM), and specifically within the E3SM Land Model component (ELM).

54 ENVIRONMENTAL SCIENCES↗

Effects of aerosols on the dynamics and microphysics of squall lines simulated by spectral bin and bulk parameterization schemes

A new spectral bin microphysical scheme (SBM) was implemented into the WRF (referred to as Fast-SBM), which uses a smaller number of size distribution functions than the previous (standard) version of SBM (referred to as Exact-SBM). It was shown that both the Fast SBM and Exact SBM reproduced the typical structure of an idealized squall line quite realistically. The schemes simulated similar dynamical and microphysical structures, and there was excellent agreement in the simulated precipitation amounts between the schemes under a very wide range of aerosol conditions in which initial condensation nuclei concentration varied from 100 cm-3 to 3000 cm-3. Moreover, the Fast-SBM uses about 40% of the computing power of the exact-SBM, allowing it to be used for “real-time” simulations over limited domains. The results of SBM simulations have been compared with a modified version of the Thompson bulk parameterization scheme within the same dynamical framework (2D WRF). The main extension of the bulk scheme was the implementation of the process of drop nucleation, so that drop concentration is no longer prescribed a priori, but rather calculated depending on the prescribed aerosol concentration. This scheme is referred to as DROP scheme. A large set of sensitivity studies have been performed, in which the sensitivity of results (microphysical parameters and precipitation) to aerosol concentration, droplet nucleation above cloud base, etc. has been compared with the results from the SBM. It is shown that the SBM scheme produces more realistic dynamical and microphysical structure of the squall line. The addition of the DROP scheme did relatively little to change the underlying results of the bulk scheme, and unlike the SBM simulations, that show different precipitation sensitivities to aerosol concentrations in relatively clean and polluted environments, the DROP scheme simulates large monotonic decrease in precipitation with increasing aerosol concentrations.

Squall line, aerosol effects on precipitation, spe↗

Deep learning Hamiltonians from disordered image data in quantum materials

The capabilities of image probe experiments are rapidly expanding, providing new information about quantum materials on unprecedented length- and timescales. Many such materials feature inhomogeneous electronic properties with intricate pattern formation on the observable surface. This rich spatial structure contains information about interactions, dimensionality, and disorder—a spatial encoding of the Hamiltonian driving the pattern formation. Image recognition techniques from machine learning are an excellent tool for interpreting information encoded in the spatial relationships in such images. Here, we develop a deep learning framework for using the rich information available in these spatial correlations in order to discover the underlying Hamiltonian driving the patterns. We first vet the method on a known case, scanning near-field optical microscopy on a thin film of V⁢O 2 . We then apply our trained convolutional neural network architecture to new optical microscope images of a different V⁢O 2 film as it goes through the metal-insulator transition. We find that a two-dimensional Hamiltonian with both interactions and random field disorder is required to explain the intricate, fractal intertwining of metal and insulator domains during the transition. This detailed knowledge about the underlying Hamiltonian paves the way for using the model to control the pattern formation via, e.g., tailored hysteresis protocols. Finally, we also introduce a distribution-based confidence measure on the results of a multilabel classifier, which does not rely on adversarial training. In addition, we propose a machine-learning-based criterion for diagnosing a physical system's proximity to criticality.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Data Security Defense: Modeling and Detection of Synchrophasor Data Spoofing Attack for Grid Edge

Data security and cyberattack have become critical issues in the distributed power system where adversaries can swap the source information of sensors or even spoof and alter measurements. However, the cyber security of the power system is challenged by the unpredictability and stealth of the spoofing attacks. Here, to protect the data security at the grid edge, this paper developed a synchrophasor data spoofing attack detection framework based on the time-frequency feature extraction techniques including the short-time Fourier transform (STFT) and object detection network for real-time synchrophasor data categorization and spoofing attack localization. The proposed approach outperforms earlier work in terms of spoofing attack detection and offers a vital localization function employing distributed synchrophasor sensors.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Achieving High Performance With TCP Over 40 GbE on NUMA Architectures for CMS Data Acquisition

TCP and the socket abstraction have barely changed over the last two decades, but at the network layer there has been a giant leap from a few megabits to 100 gigabits in bandwidth. At the same time, CPU architectures have evolved into the multicore era and applications are expected to make full use of all available resources. Applications in the data acquisition domain based on the standard socket library running in a Non-Uniform Memory Access (NUMA) architecture are unable to reach full efficiency and scalability without the software being adequately aware about the IRQ (Interrupt Request), CPU and memory affinities. During the first long shutdown of LHC, the CMS DAQ system is going to be upgraded for operation from 2015 onwards and a new software component has been designed and developed in the CMS online framework for transferring data with sockets. This software attempts to wrap the low-level socket library to ease higher-level programming with an API based on an asynchronous event driven model similar to the DAT uDAPL API. It is an event-based application with NUMA optimizations, that allows for a high throughput of data across a large distributed system. This paper describes the architecture, the technologies involved and the performance measurements of the software in the context of the CMS distributed event building.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗