Search NASA⌕ Search

SEARCH · Search NASA

Results for “Performance Model”

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 1,693 records · Page 94

A Sparse Tensor Benchmark Suite for CPUs and GPUs

Tensor computations present significant performance chal- lenges that impact a wide spectrum of applications ranging from machine learning, healthcare analytics, social network analysis, data mining to quantum chemistry and signal processing. Efforts to improve the perfor- mance of tensor computations include exploring data layout, execution scheduling, and parallelism in common tensor kernels. This work presents a benchmark suite for arbitrary-order sparse tensor kernels using state- of-the-art tensor formats: coordinate (COO) and hierarchical coordinate (HiCOO) on CPUs and GPUs. It presents a set of reference tensor kernel implementations that are compatible with real-world tensors and power law tensors extended from synthetic graph generation techniques. We also propose Roofline performance models for these kernels to provide insights of computer platforms from sparse tensor view. This benchmark suite along with the synthetic tensor generator is publicly available.

Li, Jiajia↗

Deffe: A Data-efficient Framework for Performance Characterization in Domain-Specific Computing

Predicting workload performance is a crucial task for many architecture and system research. In this paper we present De￿e, a framework to estimate the workload performances under varying architectural configurations. The infrastructure component of Deffe is based on scalable and easy-to-use open-source software components. By casting the performance modeling as transfer learning tasks, the modeling component of Deffe can leverage the learned knowledge on one workload, and “transfer” it to a new workload. Extensive experimental results show that the method can achieve superior testing accuracy with an effective reduction of 32-80x in terms of the amount of required training data.

Liu, Frank↗

DFSynthesizer: Dataflow-based Synthesis of Spiking Neural Networks to Neuromorphic Hardware

Spiking Neural Networks (SNNs) are an emerging computation model that uses event-driven activation and bio-inspired learning algorithms. SNN-based machine learning programs are typically executed on tile-based neuromorphic hardware platforms, where each tile consists of a computation unit called a crossbar, which maps neurons and synapses of the program. However, synthesizing such programs on an off-the-shelf neuromorphic hardware is challenging. This is because of the inherent resource and latency limitations of the hardware, which impact both model performance, e.g., accuracy, and hardware performance, e.g., throughput. We propose DFSynthesizer, an end-to-end framework for synthesizing SNN-based machine learning programs to neuromorphic hardware. The proposed framework works in four steps. First, it analyzes a machine learning program and generates SNN workload using representative data. Second, it partitions the SNN workload and generates clusters that fit on crossbars of the target neuromorphic hardware. Third, it exploits the rich semantics of the Synchronous Dataflow Graph (SDFG) to represent a clustered SNN program, allowing for performance analysis in terms of key hardware constraints such as number of crossbars, dimension of each crossbar, buffer space on tiles, and tile communication bandwidth. Finally, it uses a novel scheduling algorithm to execute clusters on crossbars of the hardware, guaranteeing hardware performance. We evaluate DFSynthesizer with 10 commonly used machine learning programs. Our results demonstrate that DFSynthesizer provides a much tighter performance guarantee compared to current mapping approaches.

Computer Science↗

Tournament-Based Pretraining to Accelerate Federated Learning

Advances in hardware, proliferation of compute at the edge, and data creation at unprecedented scales have made federated learning (FL) necessary for the next leap forward in pervasive machine learning. For privacy and network reasons, large volumes of data remain stranded on endpoints located in geographically austere (or at least austere network-wise) locations. However, challenges exist to the effective use of these data. To solve the system and functional level challenges, we present an three novel variants of a serverless federated learning framework. We also present tournament-based pretraining, which we demonstrate significantly improves model performance in some experiments. Overall, these extensions to FL and our novel training method enable greater focus on science rather than ML development.

Baughman, Matt↗

Carbon Support Microstructure Impact on High Current Density Transport Resistances in PEMFC Cathode

Cell voltages at high current densities (HCD) of an operating proton-exchange membrane fuel cell (PEMFC) cathode suffer from losses due to the local-O 2 and bulk-H + transport resistances in the catalyst layer. Particularly, the microstructure of high surface area carbon (HSC) support upon which both the platinum catalyst and ionomer are dispersed play a pivotal role in controlling the reactant transport to the active site in the catalyst layer. In this study, we perform a systematic analysis of the underlying microstructure of platinum-cobalt catalyst dispersed on various HSC supports in terms of their surface area and pore-size distribution. The carbon microstructure was found to strongly influence the PtCo nanoparticle dispersion, catalyst layer ionomer distribution and transport losses governing the performance at HCD. Catalyst layer electrochemical diagnostics carried out to quantify local-O 2 transport resistance and bulk-H + transport resistance in the cathode were found to be directly correlated to the micropore (<2 nm) and macropore (>8 nm) surface areas of the carbon support, respectively. Finally, a 1D-performance model has been developed to assimilate our understanding of the catalyst layer microstructure and transport resistances at HCD.

25 ENERGY STORAGE↗

Editors’ Choice—Ionomer Side Chain Length and Equivalent Weight Impact on High Current Density Transport Resistances in PEMFC Cathodes

Cell voltage at high current densities (HCD) of an operating proton-exchange membrane fuel cell (PEMFC) suffers from losses due to the local-O 2 and bulk-H + transport resistances in the cathode catalyst layer (CCL). Particularly, the interaction of perfluorosulfonic acid (PFSA) ionomer with the carbon supported platinum catalyst plays a critical role in controlling reactant transport to the active site. In this study, we perform a systematic analysis of the side chain length and equivalent weight (EW) of PFSA ionomers on the CCL transport resistances. Ex situ measurements were carried out to quantify the ionomer characteristics such as the molecular weight, proton conductivity and water uptake. Nanomorphology of ionomers cast as 60–120 nm thin-films is characterized using grazing-incidence X-ray scattering. In situ fuel cell electrochemical diagnostic measurements were carried out to quantify the reactant (H + /O 2 ) transport properties of the CCL. Ionomer EW was found to play a major role with decreasing EW yielding higher proton conductivity and water uptake that led to lower bulk-H + and local-O 2 transport resistances in the CCL. Finally, a 1D-semi-empirical performance model has been developed to quantify the impact of ionomer EW on cell voltage loss factors.

08 HYDROGEN↗

Solergy (Beta Version 1)

SOLERGY simulates the operation and power output of a user-defined solar central receiver power plant for a time period of up to one year. SOLERGY utilizes recorded or simulated weather data and plant component performance models to calculate the power flowing through each part of the solar plant. A plant control subroutine monitors these powers and determines when to operate the various plant subsystems. The original version of the code was released in May 1987, within SAND86-8060 and was widely distributed. The Beta Version 1 to be released in 2009, includes some relatively small modifications to the original code.

KOLB, GREGORY↗

Code for Axial and Crossflow Turbine Simulation

CACTUS is a code that calculates the performance and aero/hydrodynamic loads of a wind or water turbine. The turbine may be either a vertical-axis or a horizontal-axis machine, and may consist of one or more blade curved or straight blades. The performance model is based on a lifting-line free wake formulation that calculates rotor power and blade loads in the time domain. The rotor is treated as a rotating rigid body, such that structural dynamical motions are not modeled. The turbine is assumed to operate within a steady, but possibly sheared, wind or current velocity. For marine hydrokinetic energy applications, the presence of a river/tidal channel bed and water surface boundaries may be modeled.

Barone, Matthew↗

MBMS1.0: An Open-Source Code for Modeling and Simulation of Membrane-Based Dehumidification and Energy Recovery

Membrane-based dehumidification is currently being considered as a promising solution for the building application due to its low cost and very limited energy consumption. Developing a simple and efficient open-source code simulation tool is important for boosting the optimization and evaluation of such device in HVAC community. This paper reports a first-order physics based model which accounts for the fundamental heat and mass transfer of humid-air vapor at feed side to flow stream at permeate side. The current model comprises two membrane mass transfer submodels (i.e. microstructure model and performance map model); and it adopts a segment-by-segment methodology for discretizing heat and mass transfer governing equations. The model is capable of simulating both dehumidifiers and energy recovery ventilators with parallel-flow cross-flow, and counter-flow configurations. The model was validated with the measurements at appropriate device. The practices in dehumidification and energy recovery exchangers are also discussed. The model and open-source codes are expected to become a solid fundament for developing a more comprehensive and accurate membrane-based dehumidification in the future.

Gao, Zhiming↗

PathRepHAN: Hierarchical attention networks for pathology report classification

Developers at ORNL explored how a deep learning (DL) approach based on hierarchical attention networks (HANs) can improve model performance for multiple information extraction tasks from unstructured cancer pathology reports compared to conventional methods that do not sufficiently capture syntactic and semantic contexts from free-text documents. The HAN was implemented for 2 information extraction tasks: (1) primary site, matched to 12 International Classification of Diseases for Oncology topography codes (7 breast, 5 lung primary sites), and (2) histological grade classification, matched to G1–G4. Our results demonstrate that for both information tasks, HAN performed significantly better compared to the conventional ML and DL techniques.

Ramanathan, Arvind↗

HEMLOCK

HEMLOCK (Heterogeneous Ensemble Machine Learning Open Classification Kit) is a software tool for constructing, evaluating, and applying heterogeneous ensemble data models for use in solving supervised machine learning problems. Specifically, the main class of problems targeted by HEMLOCK is the problem of multiple-class classification (also called labeling or categorization) of data with continuous or discrete features. HEMLOCK consists of various data readers, machine learning algorithms, model combination and comparison routines, evaluation methods for model performance testing, and interfaces to external, state-of-the-art machine learning software libraries.

Gilpin, Sean Andrew↗

LLEPE

LLEPE (liquid-liquid equilibrium parameter estimator) is a tool for estimating thermodynamic parameters like standard enthalpies, standard molar volumes and Pitzer coefficients, by combining Gibbs minimization with regression to multiphase equilibrium isotherm data. It also includes visualization tools to access data quality and prediction model performance.

ILOEJE, CHUKWUNWIKE↗

Data Science Infrastructure SOFTWARE

LANL science workflows generate complex data sets and ensembles of data requiring significant compute and storage resources. The Data Science Infrastructure (DSI) project focuses on data-driven approaches to make data more readily available to LANL projects. DSI workflows leverage metadata stored in data-agnostic databases, supported by an abstraction layer API to simplify searching and accessing data across simulation runs, experimental runs, filesystems and environments. The abstraction layer API allows the user to query a range of data types: raw output, processed data, configuration data, machine learning models, performance data, etc. In addition to the abstraction backend API, the DSI project is developing client-driven query APIs and UIs to support specific user workflows.

Turton, Terece↗

Dashboard for Visualizing Molecular Property Prediction Machine Learning Results

This is a dashboard for exploring the results of machine learning models for predicting molecular properties from molecular structure. It includes tools for: 1. Modifying molecules to observe the change in predicted properties 2. Exploring the relationship between molecular structure and predicted properties 3. Recommending structurally similar molecules with improved properties 4. Exploring the impact of data subsampling on model performance metrics

Xu, Audrey↗

pnnl/ruleset-checking-tool

The Ruleset Checking Tool (RCT) will be a tool for verifying implementation of a ruleset (such as ASHRAE Standard 90.1 Appendix G) in building performance modeling (BPM) tools.

McNeill, James↗

JGCRI/hydro-emulator

The hydrological emulator (HE) was built to mimic complex global hydrological models (GHMs) at a range of spatial scales. The HE was written in MATLAB and utilizes a genetic algorithm to calibrate the a, b, c, d, m parameters from the ABCD runoff algorithm to observed runoff from GHMs when forced by the same climate data. The HE employs two methods to choose from when processing basin data: lumped (by basin) and distributed (gridded). Both methods can be evaluated using the built-in Kling-Gupta efficiency as a measure of model performance when compared to the outputs from the more complex GHM. Testing shows that the HE is 7 orders of magnitude faster then the widely used Variable Infiltration Capacity (VIC) GHM.

Vernon, Chris↗

Platform Of Optimal Experiment Management

The platform of optimal experiment management, POEM, powered with automated machine learning to accelerate the discovery of optimal solutions, and automatically guide the design of experiments to be evaluated. POEM currently supports 1) random model explorations for experiment design, 2) sparse grid model explorations with Gaussian Polynomial Chaos surrogate model to accelerate experiment design ,3) time-dependent model sensitivity and uncertainty analysis to identify the importance features for experiment design, 4) model calibrations via Bayesian inference to integrate experiments to improve model performance, and 5) Bayesian optimization for optimal experimental design. In addition, POEM aims to simplify the process of experimental design for users, enabling them to analyze the data with minimal human intervention, and improving the technological output from research activities.

Wang, Congjian [Idaho National Laboratory (INL), I↗