SEARCH · Search NASA

Results for “extra”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 559 records · Page 31

All sources NASA NTRS DOE OSTI

DOE OSTI2022

MemHC: An Optimized GPU Memory Management Framework for Accelerating Many-body Correlation

The many-body correlation function is a fundamental computation kernel in modern physics computing applications, e.g., Hadron Contractions in Lattice quantum chromodynamics (QCD). This kernel is both computation and memory intensive, involving a series of tensor contractions, and thus usually runs on accelerators like GPUs. Existing optimizations on many-body correlation mainly focus on individual tensor contractions (e.g., cuBLAS libraries and others). In contrast, this work discovers a new optimization dimension for many-body correlation by exploring the optimization opportunities among tensor contractions. More specifically, it targets general GPU architectures (both NVIDIA and AMD) and optimizes many-body correlation’s memory management by exploiting a set of memory allocation and communication redundancy elimination opportunities: first, GPU memory allocation redundancy: the intermediate output frequently occurs as input in the subsequent calculations; second, CPU-GPU communication redundancy: although all tensors are allocated on both CPU and GPU, many of them are used (and reused) on the GPU side only, and thus, many CPU/GPU communications (like that in existing Unified Memory designs) are unnecessary; third, GPU oversubscription: limited GPU memory size causes oversubscription issues, and existing memory management usually results in near-reuse data eviction, thus incurring extra CPU/GPU memory communications.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2023

Brief Announcement: Communication Optimal Sparse LU Factorization for Planar Matrices

We introduce a new parallel algorithm for solving sparse LU factorization of planar matrices, which commonly arise in the finite element method for 2D PDEs. Existing scalable methods, such as the multifrontal approach with subtree-to-subcube mapping by Gupta et al. [1] and right-looking with 3D mapping by Sao et al. [2] fail to achieve optimal communication costs for these matrices. Our new algorithm combines 3D mapping and subtree-to-subcube mapping to minimize communication costs while allowing trade-offs between extra memory and reduced communication. We demonstrate that our proposed algorithm attains the communication lower bound up to a factor of O(log log n) in the memory-optimal case and up to a factor of O(log P) in the memory-independent case for an n-dimensional planar sparse matrix on P processors.

Sao, Piyush↗

DOE OSTI2024

Tailor : Altering Skip Connections for Resource-Efficient Inference

Deep neural networks use skip connections to improve training convergence. However, these skip connections are costly in hardware, requiring extra buffers and increasing on- and off-chip memory utilization and bandwidth requirements. In this article, we show that skip connections can be optimized for hardware when tackled with a hardware-software codesign approach. We argue that while a network’s skip connections are needed for the network to learn, they can later be removed or shortened to provide a more hardware-efficient implementation with minimal to no accuracy loss. We introduceTailor, a codesign tool whose hardware-aware training algorithm gradually removes or shortens a fully trained network’s skip connections to lower the hardware cost.Tailorimproves resource utilization by up to 34% for block random access memories (BRAMs), 13% for flip-flops (FFs), and 16% for look-up tables (LUTs) for on-chip, dataflow-style architectures.Tailorincreases performance by 30% and reduces memory bandwidth by 45% for a two-dimensional processing element array architecture.

Computer Science↗

DOE OSTI2024

The Standard Model from String Theory: What Have We Learned?

Amid all candidates of physics beyond the Standard Model, string theory provides a unique proposal for incorporating gauge and gravitational interactions. In string theory, a four-dimensional theory that unifies quantum mechanics and gravity is obtained automatically if one posits that the additional dimensions predicted by the theory are small and curled up—a concept known as compactification. The gauge sector of the theory is specified by the topology and geometry of the extra dimensions, and the challenge is to reproduce all of the features of the Standard Model of particle physics from them. We review the state of the art in reproducing the Standard Model from string compactifications and summarize the lessons drawn from this fascinating quest. We describe novel scenarios and mechanisms that string theory provides to address some of the Standard Model puzzles as well as the most frequent signatures of new physics that could be detected in future experiments. We then comment on recent developments that connect, in a rather unexpected way, the Standard Model with quantum gravity and that may change our field theory notion of naturalness.

Physics↗

DOE OSTI2020

Precautions of Using Three-Electrode Configuration to Measure Electrode Overpotential in Solid Oxide Electrochemical Cells: Insights from Finite Element Modeling

Accurate determination of electrode overpotentials is essential to assess the performance of the electrode and understand the rate-limiting steps involved. Three-electrode configuration with the use of a reference electrode at a fixed potential is a standard way to measure overpotential of a specific electrode in liquid electrochemical systems. However, application of such three-electrode configuration to solid electrochemical cells for overpotential determination is not straightforward and requires extra caution. Here we report a theoretical Finite Element Analysis on the geometrical requirements for which the reference electrode can be applied to anode- or electrolyte-supported solid oxide button cells. The modeling results suggest that the symmetry of the working and counter electrodes is the key factor determining if a reference electrode is suitable to use. For anode-supported fuel cells with asymmetrical working and counter electrodes, reference electrode cannot be used under all circumstances. To use reference electrode for overpotential measurements, electrolyte-supported cells with symmetrical semicircular-shaped electrodes are preferred. A data processing procedure has also been presented to obtain the electrode overpotential from the measured potential using the three-electrode scheme in solid oxide electrochemical cells.

25 ENERGY STORAGE↗

DOE OSTI2020

Mathematical Model for Li-S Cell with Shuttling-Induced Capacity Loss Approximation

Lithium sulfur (Li-S) batteries have the potential to outperform the current lithium ion batteries and transform the technology of the future. However, dissolution, diffusion, and shuttling of the dissolved polysulfides result in parasitic reactions and substantial capacity loss. To provide a better understanding of the shuttling process, a 1D porous electrode mathematical model has been developed in this paper. An approximation method is used to account for the shuttling-induced capacity loss by adding an extra source/sink term in the material balance equations for the species involved in the parasitic reactions. Shuttling constants used in the source terms can be determined by fitting the model predictions to the experimental measurements. Here, the results showed that by including the approximation method, the model was able to predict the active material loss and the continuous decrease of volume fractions of Li 2 S on the cathode surface. The model sheds light on the capacity loss mechanism occurring inside the cell as a result of the shuttling of polysulfides.

25 ENERGY STORAGE↗

DOE OSTI2022

Different Electrochemical Behavior of Cationic Dopamine from Anionic Ascorbic Acid and DOPAC at CNT Yarn Microelectrodes

Carbon nanotube yarn microelectrodes (CNTYMEs) have micron-scale surface crevices that momentarily trap molecules. CNTYMEs improve selectivity among cationic catecholamines because secondary reactions are enhanced, but no anions have been studied. Here, we compared fast-scan cyclic voltammetry (FSCV) of dopamine and anionic interferents 3,4 dihydroxyphenylacetic acid (DOPAC) and L-ascorbic acid (AA) at CNTYMEs and carbon fiber microelectrodes (CFMEs). At CFMEs, dopamine current decreases with increasing FSCV repetition frequency at pH 7.4, whereas DOPAC and AA have increasing currents with increasing frequency, because of less repulsion at the negative holding potential. Both DOPAC and AA have side reactions after being oxidized, which are enhanced by trapping. At pH 4, the current increases for DOPAC and AA because they are not repelled. In addition, AA has a different oxidation pathway at pH 4, and an extra peak in the CV is enhanced by trapping effects at CNTYMEs. At pH 8.5, co-detection of dopamine in the presence of DOPAC and AA is enhanced at 100 Hz frequency because of differences in secondary peaks. Thus, the trapping effects at CNTYMEs affects anions differently than cations and secondary peaks can be used to identify dopamine in mixture of AA and DOPAC with FSCV.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2022

Coulombic Efficiency and Capacity Retention are Not Universal Descriptors of Cell Aging

Capacity and coulombic efficiency are often used to assess the performance of Li-ion batteries, under the assumption that these quantities can provide direct insights about the rate of electron consumption due to growth of the solid electrolyte interphase (SEI). Here, we show that electrode properties can actually change the amount of information about aging that can be directly retrieved from capacity measurements. During cycling of full-cells, only portions of the voltage profiles of the positive and negative electrodes are accessible, leaving a reservoir of cyclable Li + stored at both electrodes. The size and availability of this reservoir depends on the shape of the voltage profiles and accessing this extra Li + can offset some of the capacity that is consumed by the SEI. Consequently, capacity and efficiency measurements can, at times, severely underestimate the rate of side reactions experienced by the cell. We show, for example, that a same rate of SEI growth would cause faster capacity fade in LiFePO 4 than in NMC cells, and that the perceived effects of aging depend on testing variables such as depth of discharge. Simply measuring capacity may be insufficient to gauge the true extent of aging endured by Li-ion batteries.

25 ENERGY STORAGE↗

DOE OSTI2020

Legal Judgment Prediction Based on Multiclass Information Fusion

Legal judgment prediction (LJP), as an effective and critical application in legal assistant systems, aims to determine the judgment results according to the information based on the fact determination. In real-world scenarios, to deal with the criminal cases, judges not only take advantage of the fact description, but also consider the external information, such as the basic information of defendant and the court view. However, most existing works take the fact description as the sole input for LJP and ignore the external information. We propose a Transformer-Hierarchical-Attention-Multi-Extra (THME) Network to make full use of the information based on the fact determination. We conduct experiments on a real-world large-scale dataset of criminal cases in the civil law system. Experimental results show that our method outperforms state-of-the-art LJP methods on all judgment prediction tasks.

Zhu, Kongfan↗

DOE OSTI2022

Translytics: A Novel Approach for Runtime Selection of Database Layout Based on User’s Context

Currently, organizations have to maintain separate systems for transactions and analytics inside the company. Notably, different vendors provide the capabilities for either of these tasks that require specialized hardware or software. Data engineers are required to retrieve data from one source and transform it into another format to obtain the maximum benefit in a minimum time. Organizations strive for a competitive advantage that is achieved by fetching data from their customers and getting insights earliest for timely decision making. Present practices do not permit the view of the latest data for analytics since the first data have to be fetched from the source, transformed, and loaded to other systems to be utilized for analysis by relevant teams. This paper introduces a single system for both transactions and analytics. Our proposed solution would permit companies to seamlessly adapt our solution without the need to shift all of their data to newer systems and allow all the teams. It would grant all the teams to have a view of the latest available data without extra expertise and budget.

Tanvir, Muhammad Makhshif↗

DOE OSTI2022

PyTREES

PyTREES (Python tool for Training/Testing Robust Explainable Ensembles on Spectra) is software that implements a data-driven approach to predicting the amount of specific oxides present in materials samples of laser-induced breakdown spectroscopy (LIBS); such as from the ChemCam instrument suite onboard the NASA Curiosity rover. PyTREES is designed to input LIBS data in the format provided by the ChemCam team [1]. PyTREES then applies appropriate pre-processing to this data [2], and implements several regression methods for predicting oxides from spectra. The regression methods include: ensemble methods (random forest, extra trees, and gradient boosting regression) and blended submodels using the “double blending” technique. PyTREES additionally implements methods for quantifying the importance of features in regression model: (1) mean decrease in impurity (MDI) and (2) permutation importance to investigate the wavelengths used by the regression methods. [1] Gasda et al. (2021). Spectrochim Acta B, 181, 106223. [2] Clegg et al. (2017). Spectrochim Acta B , 129, 64–85.

Oyen, Diane↗

DOE OSTI2024

Grid Modeling and Hosting Capacity Analysis

An interface with the OpenDSS distribution grid simulator, facilitating users in calibrating and validating models. Additionally, three distinct tools have been developed: PV Hosting Capacity: This tool allows users to determine the additional amount of photovoltaic (PV) generation that can be integrated into the system without breaching operational constraints. EV Hosting Capacity: This tool focuses on identifying the capacity for incorporating extra electric vehicle (EV) load into the system without surpassing operational limitations. Project Impact Analysis Tool: This tool is designed to assess and report all potential grid violations associated with a specific project, providing valuable insights into its impact on the distribution grid.

Poudel, Shiva↗

DOE OSTI2022

Name Normalizer

Our name normalizer and enrichment micro-service address the shortcomings of standard text normalization. Initially the text input is split into tokens. This process takes into account some conventions of formatting. Any dashes present between two tokens preserves the relationship of those two tokens. Conversely a comma between two tokens ensures that the separation between the two tokens is maintained. The order of tokens is also preserved. Standard text normalization (conversion to lowercase, trimming extra white space, and canonicalization) is then applied to the tokens. Once normalized every unique and sequence of tokens is given confidence values by comparing the normalized values to publicly available data.

Bleeker, Amelia↗

DOE OSTI2024

FTTN: Feature-Targeted Testing for Numerical Properties of NVIDIA & AMD Matrix Accelerators

FTTN is a test suite to evaluate the numerical behaviors of matrix accelerators of GPUs (NVIDIA Tensor Cores and AMD Matrix Cores) in a quick and simple setting. Matrix accelerators are heavily used in today's computationally intense applications to speed up matrix multiplications. This test suite provides a comprehensive study on the numerical behaviors of these accelerators, including support for subnormals, rounding modes, extra precision bits and FMA features. Is there

Laguna Peralta, Ignacio↗

DOE OSTI2025

Triangle Method for Dense ReLU Layers [SWR-25-72]

This software is an implementation of the methods for initializing and training neural networks to be more efficient per parameter, described more fully below and in the related publication: In theory, depth should make a ReLU network EXPONENTIALLY more efficient by enabling it to produce an exponential number of piecewise linear sections in its output. This reasoning is largely based on the work of mathematicians that have hand-constructed networks that make good use of depth. In practice however, even very deep ReLU networks that have been randomly initialized will behave identically to their shallow counterparts - missing an entire exponential dimension of efficiency. The triangle method is a first attempt at realizing the exponential potential of deep networks. Instead of randomly setting weights, we force pairs of neurons in each layer learn to build triangles (i.e. functions from [0,1] -> [0,1] that look like triangles). This is a very efficient pattern for generating lots of linear pieces because composing two triangular functions doubles the number of pieces with each composition. The triangle method is more than just a different initialization, it is a new paradigm of training. Instead of making direct updates to the matrix weights, we do an extra step of backpropagation to collect the derivatives of the loss function with respect to the shapes of the triangles, training them to tilt left or right. This process essentially holds the networks hand throughout the loss landscape and forces it to always use depth effectively by producing triangular shapes internally. This can produce several orders of magnitude of improvement on convex one-dimensional regression problems. Much more theoretical work is needed to realize its full potential beyond this context, but the implementation in this repository will still work in arbitrary numbers of dimensions. The file Triangle_Method.py is a generalized form of the method that will build each neuron its own custom 1-d convex activation function (with exponential efficiency). Example usage on one dimensional problems can be found in Example_Usage.ipynb and an example of using this in a real neural network can be found in Example_VGG16_CIFAR10.ipynb.

Milkert, Max [National Renewable Energy Laboratory↗

DOE OSTI2025

ghostbus v0.1.0

Ghostbus is a Verilog meta-scripting tool which automates memory map creation, address assignment, and bus routing much like Register Description Language (RDL) and SystemRDL, but with a few stark differences. First, the design is entirely defined/captured within the Verilog/SystemVerilog codebase (no extra second language). Second, the Verilog code remains valid even without the auto-generated code, enhancing modularity and compatibility with other tools/workflows.

Huang, Gang [Lawrence Berkeley National Laboratory↗

DOE OSTI2026

Multithreaded copy ('cp')

This is a modification to 'cp' and 'mv' commands to make them multi-threaded. Simple benchmarks showed that multi-threading could reduce the time to copy a large Linux source directory by over 2x. The 'cp' and 'mv' utilities are part of the existing Coreutils (https://www.gnu.org/software/coreutils/) software package that get installed on all Linux distros. Changes: * Add '-j|--parallel ' flags to 'cp' and 'mv'. This allows the utilities to recursively copy regular files in directories in parallel. This does NOT parallelize multiple single file copies to a destination (like 'cp file2 file2 file3 dst/'). Along with this, add in new 'CP_NUM_THREADS' and 'MV_NUM_THREADS' environment variables to set the number of threads. This can be useful when you want to enable parallelism by default in /etc/profile. The maximum number of threads is internally capped to the number of CPUs. * Add a '-j' flag to 'sort' to complement its existing '--parallel' flag. This is only done for consistency with 'cp' and 'mv'. * Add test cases for the new flags. Also, run each 'cp' and 'mv' test both in single-threaded and multithreaded modes for extra coverage.

Hutter, AnthonyJ [Lawrence Livermore National Labo↗

DOE OSTI2021

Changes in Future Precipitation Mean and Variability across Scales

Changes in precipitation variability can have large societal consequences, whether at the short time scales of flash floods or the longer time scales of multiyear droughts. Recent studies have suggested that in future climate projections, precipitation variability rises more steeply than does its mean, leading to concerns about societal impacts. Here we evaluate changes in mean precipitation over a broad range of spatial and temporal scales using a range of models from high-resolution regional simulations to millennial-scale global simulations. Results show that changes depend on the scale of aggregation and involve strong regional differences. On local scales that resolve individual rainfall events (hours and tens of kilometers), changes in precipitation distributions are complex and variances rise substantially more than means, as is required given the well-known disproportionate rise in precipitation intensity. On scales that aggregate across many events, distributional changes become simpler and variability changes smaller. At regional scale, future precipitation distributions can be largely reproduced by a simple transformation of present-day precipitation involving a multiplicative shift and a small additive term. The “extra” broadening is negatively correlated with changes in mean precipitation: in strongly “wetting” areas, distributions broaden less than expected from a simple multiplicative mean change; in “drying” areas, distributions narrow less. Precipitation variability changes are therefore of especial concern in the subtropics, which tend to dry under climate change. Outside the tropics, variability changes are similar on time scales from days to decades (i.e., show little frequency dependence). This behavior is highly robust across models, suggesting it may stem from some fundamental constraint.

54 ENVIRONMENTAL SCIENCES↗