Search NASA⌕ Search

SEARCH · Search NASA

Results for “faster”

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 847 records · Page 47

Randomized Projection for Rank-Revealing Matrix Factorizations and Low-Rank Approximations

Rank-revealing matrix decompositions provide an essential tool in spectral analysis of matrices, including the Singular Value Decomposition (SVD) and related low-rank approximation techniques. QR with Column Pivoting (QRCP) is usually suitable for these purposes, but it can be much slower than the unpivoted QR algorithm. For large matrices, the difference in performance is due to increased communication between the processor and slow memory, which QRCP needs in order to choose pivots during decomposition. Our main algorithm, Randomized QR with Column Pivoting (RQRCP), uses randomized projection to make pivot decisions from a much smaller sample matrix, which we can construct to reside in a faster level of memory than the original matrix. This technique may be understood as trading vastly reduced communication for a controlled increase in uncertainty during the decision process. Furthermore, for rank-revealing purposes, the selection mechanism in RQRCP produces results that are the same quality as the standard algorithm, but with performance near that of unpivoted QR (often an order of magnitude faster for large matrices). Additionally, we also propose two formulas that facilitate further performance improvements. The first efficiently updates sample matrices to avoid computing new randomized projections. The second avoids large trailing updates during the decomposition in truncated low-rank approximations. Our truncated version of RQRCP also provides a key initial step in our truncated SVD approximation, TUXV. These advances open up a new performance domain for large matrix factorizations that will support efficient problem-solving techniques for challenging applications in science, engineering, and data analysis.

97 MATHEMATICS AND COMPUTING↗

New procedure for evaluation of U(3) coupling and recoupling coefficients

A simple method to calculate Wigner coupling coefficients and Racah recoupling coefficients for U(3) in two group–subgroup chains is presented. While the canonical U(3) coupling and recoupling coefficients are applicable to any system that respects U(3) symmetry, the U(3) coupling coefficients are more specific to nuclear structure studies. This new procedure precludes the use of binomial coefficients and alternating sums which were used in the 1973 formulation of Draayer and Akiyama, and in so doing provides a faster and more accurate determination of any and all required results. The resolution of the outer multiplicity is based on the null space concept of the U(3) generators proposed by Alex et al., whereas the inner multiplicity in the angular momentum subgroup chain is obtained from the dimension of the null space of the SO(3) raising operator. It is anticipated that a C++ library will ultimately be available for determining generic coupling and recoupling coefficients associated with both the canonical and the physical group–subgroup chains of U(3).

Cross-Coupling Reaction↗

Scaling Deep Learning on GPU and Knights Landing clusters

The speed of deep neural networks training has become a big bottleneck of deep learning research and development. For example, training GoogleNet by ImageNet dataset on one Nvidia K20 GPU needs 21 days. To speed up the training process, the current deep learning systems heavily rely on the hardware accelerators. However, these accelerators have limited on-chip memory compared with CPUs. To handle large datasets, they need to fetch data from either CPU memory or remote processors. We use both self-hosted Intel Knights Landing (KNL) clusters and multi-GPU clusters as our target platforms. From an algorithm aspect, current distributed machine learning systems are mainly designed for cloud systems. These methods are asynchronous because of the slow network and high fault-tolerance requirement on cloud systems. We focus on Elastic Averaging SGD (EASGD) to design algorithms for HPC clusters. Original EASGD used round-robin method for communication and updating. The communication is ordered by the machine rank ID, which is inefficient on HPC clusters. First, we redesign four efficient algorithms for HPC systems to improve EASGD's poor scaling on clusters. Async EASGD, Async MEASGD, and Hogwild EASGD are faster \textcolor{black}{than} their existing counterparts (Async SGD, Async MSGD, and Hogwild SGD, resp.) in all the comparisons. Finally, we design Sync EASGD, which ties for the best performance among all the methods while being deterministic. In addition to the algorithmic improvements, we use some system-algorithm codesign techniques to scale up the algorithms. By reducing the percentage of communication from 87% to 14%, our Sync EASGD achieves 5.3x speedup over original EASGD on the same platform. We get 91.5% weak scaling efficiency on 4253 KNL cores, which is higher than the state-of-the-art implementation.

60 APPLIED LIFE SCIENCES↗

SuperNeuro: A Fast and Scalable Simulator for Neuromorphic Computing

In many neuromorphic workflows, simulators play a vital role for important tasks such as training spiking neural networks, running neuroscience simulations, and designing, implementing, and testing neuromorphic algorithms. Currently available simulators cater to either neuroscience workflows (e.g., NEST and Brian2) or deep learning workflows (e.g., BindsNET). Problematically, the neuroscience-based simulators are slow and not very scalable, and the deep learning-based simulators do not support certain functionalities that are typical of neuromorphic workloads (e.g., synaptic delay). In this paper, we address this gap in the literature and present SuperNeuro, which is a fast and scalable simulator for neuromorphic computing capable of both homogeneous and heterogeneous simulations as well as GPU acceleration. We also present preliminary results that compare SuperNeuro to widely used neuromorphic simulators such as NEST, Brian2, and BindsNET in terms of computation times. We demonstrate that SuperNeuro can be approximately 10×--300× faster than some of the other simulators for small sparse networks. On large sparse and large dense networks, SuperNeuro can be approximately 2.2×--3.4× faster than the other simulators, respectively.

Date, Prasanna↗

PANDORA: A Parallel Dendrogram Construction Algorithm for Single Linkage Clustering on GPU

This paper introduces Pandora, a parallel algorithm for computing dendrograms, the hierarchical cluster trees for single linkage clustering (SLC). Current parallel approaches construct dendrograms by partitioning a minimum spanning tree and removing edges. However, they struggle with skewed, hard-to-parallelize real-world dendrograms. Consequently, computing dendrograms is the sequential bottleneck in HDBSCAN*[21], a popular SLC variant. Pandora uses recursive tree contraction to address this limitation. Pandora contracts nodes to construct progressively smaller trees. It computes the smallest contracted dendrogram and expands it by inserting contracted edges. This recursive strategy is highly parallel, skew-independent, work-optimal, and well-suited for GPUs and multicores. We develop a performance portable implementation of Pandora in Kokkos[31] and evaluate its performance on multicore CPUs and multi-vendor GPUs (e.g., Nvidia, AMD) for dendrogram construction in HDBSCAN*. Multithreaded Pandora is 2.2x faster than the current best-multithreaded implementation. Our GPU version achieves 6-20x speedup on AMD GPUs and 10-37x on NVIDIA GPUs over multithreaded Pandora. Pandora removes HDBSCAN*’s sequential bottleneck, greatly boosting efficiency, particularly with GPUs.

Sao, Piyush↗

An Analysis of the Impact of Particle Growth on Transport Losses in Polymer-Electrolyte Fuel Cells

Voltage cycling causes catalyst nanoparticles in polymer-electrolyte fuel cells to grow. The concomitant loss of interfacial area results in larger kinetic, transport, and possibly ohmic overpotentials. This paper uses recently published experimental data and mathematical models to investigate the evolutions of transport and ohmic resistances to platinum nanoparticles located on the surface and inside the micropores of carbon black supports. Resistance to oxygen transport rises as surface area declines primarily because the flux to each remaining larger platinum particle increases. The path lengths governing oxygen diffusion to surface and buried platinum sites also increase as nanoparticles grow. Platinum nanoparticles on the surface become relatively less favorable as voltage cycling proceeds because they grow faster than platinum in micropores. Because voltage cycling causes total interfacial area to decline, and the fraction located inside micropores to increase, nanoscale ohmic losses increase as a catalyst layer decays. The practical importance of this effect is difficult to discern because proton conductivity in carbon micropores is not well characterized.

07 ISOTOPE AND RADIATION SOURCES↗

Investigating the Particle Growth in Bimodal Pt/C Catalysts by In-Situ Small-Angle X-ray Scattering: Challenges in the Evaluation of Stress Test Protocol-Dependent Degradation Mechanisms

The influence of different combinations of accelerated stress test (AST) protocols simulating load-cycle and start/stop conditions of a proton exchange membrane fuel cell (PEMFC) vehicle is investigated on a bimodal Pt/C catalyst. The bimodal Pt/C catalyst, prepared by mixing two commercial catalysts, serves as a model system and consists of two distinguishable size populations. The change in mean particle size was investigated by in situ small-angle X-ray scattering (SAXS). The comparison to the reference catalysts, i.e., the two single-size population catalysts, uncovers the presence of electrochemical Ostwald ripening as a degradation mechanism in the bimodal catalyst. Increasing the harshness of the applied AST protocol combinations by faster changing between load-cycle or start/stop conditions, the particle size of the larger population of the bimodal catalyst increases faster than expected. Surprisingly, the change in mean particle size of the smaller size population indicates a smaller increase for harsher AST protocols, which might be explained by a substantial electrochemical Ostwald ripening.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Acoustic Source Localization via Time Difference of Arrival Estimation for Distributed Sensor Networks Using Tera‐Scale Optical Core Devices

For real‐time acoustic source localization applications, one of the primary challenges is the considerable growth in computational complexity associated with the emergence of ever larger, active or passive, distributed sensor networks. These sensors rely heavily on battery‐operated system components to achieve highly functional automation in signal and information processing. In order to keep communication requirements minimal, it is desirable to perform as much processing on the receiver platforms as possible. However, the complexity of the calculations needed to achieve accurate source localization increases dramatically with the size of sensor arrays, resulting in substantial growth of computational requirements that cannot be readily met with standard hardware. One option to meet this challenge builds upon the emergence of digital optical‐core devices. The objective of this work was to explore the implementation of key building block algorithms used in underwater source localization on the optical‐core digital processing platform recently introduced by Lenslet Inc. This demonstration of considerably faster signal processing capability should be of substantial significance to the design and innovation of future generations of distributed sensor networks.

Imam, Neena↗

Parallel time integration software

This package implements an optimal-scaling multigrid solver for the (non) linear systems that arise from the discretization of problems with evolutionary behavior. Typically, solution algorithms for evolution equations are based on a time-marching approach, solving sequentially for one time step after the other. Parallelism in these traditional time-integrarion techniques is limited to spatial parallelism. However, current trends in computer architectures are leading twards system with more, but not faster. processors. Therefore, faster compute speeds must come from greater parallelism. One approach to achieve parallelism in time is with multigrid, but extending classical multigrid methods for elliptic poerators to this setting is a significant achievement. In this software, we implement a non-intrusive, optimal-scaling time-parallel method based on multigrid reduction techniques. The examples in the package demonstrate optimality of our multigrid-reduction-in-time algorithm (MGRIT) for solving a variety of parabolic equations in two and three sparial dimensions. These examples can also be used to show that MGRIT can achieve significant speedup in comparison to sequential time marching on modern architectures.

Source record↗

Data Elevator

Data Elavator allows extensive optimizations, such as overlapping read and write operations, choosing I/O modes, and aligning buffer boundaries. Is as much as 4.2X faster than Cray DataWarp, the start-of-art software for burst buffer, and 4X faster than directly writing to PFS.

Wu, Kesheng↗

FasTensor (FT) v0.0.1

The FasTensor implements the native array data programming mode namely SLOPE for modern data analysis tasks. It is designed to work in parallel on supercomputer and can handle terabyte data analysis tasks. It is thousand times faster than other data analysis systems such as Apache Spark and 38% faster than TensorFlow in dealing with large scale scientific data analysis. It has wide applications in image data analysis and mesh data analysis tasks in physical simulations and other fields.

Wu, Kesheng↗

JBlulce Data Acquisition Software for Macromolecular Crystallography

JBlulce (Java Beam Line Universal Integrated Configuration Environment is a data acquisition software for macromolecular crystallography conforming user interface of the SSRL Blulce that has become a de-factor standard in the field. Besides this interface conformity, JBlulce is a unique system in terms of architecture, spec, capability and software implementation. It features only two software layers, the JBlulce clients and the EPICS servers, as compared to three layers present in Blulce and most of similar systems. This layers reduction provides a faster communication with hardware and an easier access to advanced hardware capabilities like on-the-fly scanning. Then JBlulce clients are designed to operate in parallel with the other beamline controls which streamlines the tasks performed by staff such as beamline preparation, maintenance, auditing and user assistance. Another distinction is the deployment of multiple plugins that can be written in any programming language thus involving more staff into the development. further on, JBlulce makes use of unified motion controls allowing for easy scanning and optimizing of any beamline component. Finally, the graphic interface is implemented in Java making full use of rich Java libraries and Jave IDE for debugging. to compare, Blulce user interface is implemented with aging Tcl/tk language providing very restricted capabilities. JBlulce makes full use of the industrial power and wide drivers selection of EPICS in controlling hardware; all hardware communication is routed via multiple EPICS servers residing on local area network. JBlulce also includes several EPICS State Notation servers aimed at making hardware communication more robust. Besides using EPICS for controlling hardware, JBlulce extensively uses EPICS databases for efficient communications between multiple instances of JBlulce clients and JBlulce plugins that can run in parallel on different computers. All of the above makes JBlulce one of the biggest and most sophisticated EPICS client projects to date. JBlulce configuration is stored in my SQL database which provides flexibility in tuning the system. The database is also accessible by the plugins. From the users perspective JBlulce provides all standard features of data acquisition software for macromolecular crystallography plus such unique capabilities as one click beamline energy change that may involve switching undulator harmonics, mirrors lanes and beam realignment, automated diffraction rtastering for finding small crystals and sweet spots on poorly diffracting crystals with automated scoring of raster cells by the number of reflections; data collection along a vector; automated on-the-fly fluorescent tastering, a faster and lower-irradiation compliment to the diffraction raster; fully automated fluorescence measurements for MAD that include signal optimization, fast on the fly energy scanning and automated adapting of scan range to chemical shifts; fly-scan minibeam realignment; automated loop and crystal centering, controls for sample automounter, automated screening, data collection auditing, remote access and a lot more.

Stepanov, Sergey↗

Monsoonal MCS Initiation, Rainfall, and Diurnal Gravity Waves over the Bay of Bengal: Observation and a Linear Model

Abstract Previous observational studies have indicated that mesoscale convective systems (MCSs) contribute the majority of precipitation over the Bay of Bengal (BoB) during the summer monsoon season, yet their initiation and propagation remain incompletely understood. To fill this knowledge gap, we conducted a comprehensive study using a combination of 20-yr satellite observations, MCS tracking, reanalysis data, and a theoretical linear model. Satellite observations reveal clear diurnal propagation signals of MCS initiation frequency and rainfall from the west coast of the BoB toward the central BoB, with the MCS rainfall propagating slightly slower than the MCS initiation frequency. Global reanalysis data indicate a strong association between the offshore-propagating MCS initiation frequency/rainfall and diurnal low-level wind perturbations, implying the potential role of gravity waves. To verify the hypothesis, we developed a 2D linear model that can be driven by realistic meteorological fields from reanalysis. The linear model realistically reproduces the characteristics of offshore-propagating diurnal wind perturbations. The wind perturbations, as well as the offshore propagation signals of MCS initiation frequency and rainfall, are associated with diurnal gravity waves emitted from the coastal regions, which in turn are caused by the diurnal land–sea thermal contrast. The ambient wind speed and vertical wind shear play crucial roles in modulating the timing, propagation, and amplitude of diurnal gravity waves. Using the linear model and satellite observations, we further show that the stronger monsoonal flows lead to faster offshore propagation of diurnal gravity waves, which subsequently control the offshore propagation signals of MCS initiation and rainfall. Significance Statement Rainfall over the Bay of Bengal (BoB) is primarily contributed by large and organized rainfall systems in the summer monsoon season. During this season, these systems are commonly observed over the east coast of India around midnight, the western BoB in the morning, and the central BoB in the afternoon. This eastward rainfall propagation is confirmed by observations, reanalysis data, and a theoretical model to have a strong association with atmospheric diurnal gravity waves. These waves are caused by land–sea thermal contrast and can trigger rainfall systems over the offshore regions. We also found that the diurnal gravity waves, as well as their triggered rainfall systems, can be greatly modulated by the large-scale monsoonal flows. All the above findings improve our understanding of diurnal rainfall cycle over the BoB and may contribute to the future improvement of rainfall forecast over the region.

Meteorology & Atmospheric Sciences↗

MJO Teleconnections over the PNA Region in Climate Models. Part II: Impacts of the MJO and Basic State

In an assessment of 29 global climate models (GCMs), Part I of this study identified biases in boreal winter MJO teleconnections in anomalous 500-hPa geopotential height over the Pacific–North America (PNA) region that are common to many models: an eastward shift, a longer persistence, and a larger amplitude. In Part II, we explore the relationships of the teleconnection metrics developed in Part I with several existing and newly developed MJO and basic state (the mean subtropical westerly jet) metrics. The MJO and basic state diagnostics indicate that the MJO is generally weaker and less coherent and propagates faster in models compared to observations. The mean subtropical jet also exhibits notable biases such as too strong amplitude, excessive eastward extension, or southward shift. The following relationships are found to be robust among the models: 1) models with a faster MJO propagation tend to produce weaker teleconnections; 2) models with a less coherent eastward MJO propagation tend to simulate more persistent MJO teleconnections; 3) models with a stronger westerly jet produce stronger and eastward shifted MJO teleconnections; 4) models with an eastward extended jet produce an eastward shift in MJO teleconnections; and 5) models with a southward shifted jet produce stronger MJO teleconnections. The results are supported by linear baroclinic model experiments. Our results suggest that the larger amplitude and eastward shift biases in GCM MJO teleconnections can be attributed to the biases in the westerly jet, and that the longer persistence bias is likely due to the lack of coherent eastward MJO propagation.

59 BASIC BIOLOGICAL SCIENCES↗

Sensitivities of the West Greenland Current to Greenland Ice Sheet Meltwater in a Mesoscale Ocean/Sea Ice Model

Abstract Meltwater from the Greenland Ice Sheet can alter the continental shelf/slope circulation and cross-shelf freshwater fluxes and limit deep convection in adjacent basins through surface freshening. We explore the impacts on the West Greenland Current and eastern Labrador Sea with different vertical distributions of the meltwater forcing. In this study, we present the results from global coupled ocean/sea ice simulations, forced with atmospheric reanalysis, that are mesoscale eddy-active (∼2–3-km horizontal spacing) and eddy-permitting (∼6–7-km horizontal spacing) in the study region. We compare the West Greenland Current in mesoscale eddy-active and eddy-permitting without meltwater to highlight the role of small-scale features. The mesoscale eddy-active configuration is then used to assess the change in the eastern Labrador Sea when meltwater is added to the surface or vertically distributed to account for mixing within fjords. In both simulations with meltwater, the West Greenland and West Greenland Coastal Currents are faster than in the simulation with no meltwater; their mean surface speeds are the highest in the vertical distribution case. In the latter case, there is enhanced baroclinic conversion at the shelf break compared to the simulation with no meltwater. When meltwater is vertically distributed, there is an increase in baroclinic conversion at the shelf break associated with increased eddy kinetic energy. In addition, in the eastern Labrador Sea, the salinity is lower and the meltwater volume is greater when meltwater is vertically distributed. Therefore, the West Greenland Current is sensitive to how meltwater is added to the ocean with implications for the freshening of the Labrador Sea. Significance Statement Our goal is to understand how the flux of freshwater across the West Greenland continental slope into the Labrador Sea is modified by meltwater from the Greenland Ice Sheet. We compare the simulations of the ocean that capture key dynamics along the West Greenland continental slope that have no meltwater, meltwater added to the ocean surface, and meltwater distributed vertically to represent the mixing within fjords. When meltwater is added, the currents along the continental slope are faster, with the greatest increase when meltwater is vertically distributed. In that case, there is enhanced freshening of the Labrador Sea because modified density gradients generate more eddies. Proper representation of the vertical structure of meltwater is important for projecting the impact of freshwater on the subpolar North Atlantic.

58 GEOSCIENCES↗

Scalability of high-performance PDE solvers

Performance tests and analyses are critical to effective high-performance computing software development and are central components in the design and implementation of computational algorithms for achieving faster simulations on existing and future computing architectures for large-scale application problems. In this article, we explore performance and space-time trade-offs for important compute-intensive kernels of large-scale numerical solvers for partial differential equations (PDEs) that govern a wide range of physical applications. We consider a sequence of PDE-motivated bake-off problems designed to establish best practices for efficient high-order simulations across a variety of codes and platforms. We measure peak performance (degrees of freedom per second) on a fixed number of nodes and identify effective code optimization strategies for each architecture. In addition to peak performance, we identify the minimum time to solution at 80% parallel efficiency. The performance analysis is based on spectral and p-type finite elements but is equally applicable to a broad spectrum of numerical PDE discretizations, including finite difference, finite volume, and h-type finite elements.

97 MATHEMATICS AND COMPUTING↗

A new model for the biodegradation kinetics of oil droplets: application to the Deepwater Horizon oil spill in the Gulf of Mexico

Oil biodegradation by native bacteria is one of the most important natural processes that can attenuate the environmental impacts of marine oil spills. Existing models for oil biodegradation kinetics are mostly for dissolved oil. This work developed a new mathematical model for the biodegradation of oil droplets and applied the model to estimate the time scale for oil biodegradation under conditions relevant to the Deepwater Horizon oil spill in the Gulf of Mexico. In the model, oil is composed of droplets of various sizes following the gamma function distribution. Each oil droplet shrinks during the microbe-mediated degradation at the oil-water interface. Using our developed model, we find that the degradation of oil droplets typically goes through two stages. The first stage is characterized by microbial activity unlimited by oil-water interface with higher biodegradation rates than that of the dissolved oil. The second stage is governed by the availability of the oil-water interface, which results in much slower rates than that of soluble oil. As a result, compared to that of the dissolved oil, the degradation of oil droplets typically starts faster and then quickly slows down, ultimately reaching a smaller percentage of degraded oil in longer time. The availability of the water-oil interface plays a key role in determining the rates and extent of degradation. We find that several parameters control biodegradation rates, including size distribution of oil droplets, initial microbial concentrations, initial oil concentration and composition. Under conditions relevant to the Deepwater Horizon spill, we find that the size distribution of oil droplets (mean and coefficient of variance) is the most important parameter because it determines the availability of the oil-water interface. Smaller oil droplets with larger variance leads to faster and larger extent of degradation. The developed model will be useful for evaluating transport and fate of spilled oil, different remediation strategies, and risk assessment.

02 PETROLEUM↗

GenAp: a distributed SQL interface for genomic data

Abstract Background The impressively low cost and improved quality of genome sequencing provides to researchers of genetic diseases, such as cancer, a powerful tool to better understand the underlying genetic mechanisms of those diseases and treat them with effective targeted therapies. Thus, a number of projects today sequence the DNA of large patient populations each of which produces at least hundreds of terra-bytes of data. Now the challenge is to provide the produced data on demand to interested parties. Results In this paper, we show that the response to this challenge is a modified version of Spark SQL, a distributed SQL execution engine, that handles efficiently joins that use genomic intervals as keys. With this modification, Spark SQL serves such joins more than 50× faster than its existing brute force approach and 8× faster than similar distributed implementations. Thus, Spark SQL can replace existing practices to retrieve genomic data and, as we show, allow users to reduce the number of lines of software code that needs to be developed to query such data by an order of magnitude.

Kozanitis, Christos↗