SEARCH · Search NASA

Results for “Auto Tuning”

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 37 records · Page 2

All sources NASA NTRS DOE OSTI

DOE OSTI2019

A Pattern Based Algorithmic Autotuner for Graph Processing on GPUs

This paper presents Gswitch, a pattern-based algorithmic autotuning system that dynamically switches to the suitable optimization variants with negligible overhead. Its novelty is a small set of algorithmic patterns that enables configurable assembling of algorithm variants. The fast transition of Gswitch is based on a machine learning model trained from 644 real graphs from the network repository. In addition, Gswitch provides succinct programming interface which hides all low-level tuning details. We evaluate Gswitch for typical graph algorithms (BFS, CC, PR, SSSP, and BC) on Nvidia Kepler and Pascal GPUs. The results show that Gswitch runs up to 10× faster than the best configuration of the state-of-the-art programmable GPU-based graph processing libraries on ten representative graphs. Gswitch wins on 92.4% cases of 644 graph data which is the largest dataset evaluation reported to date.

HPC, graph algorithms, GPUs, auto tuning↗

DOE OSTI2019

Adaptive relay autotuning under static and non-static disturbances with application to friction stir welding

When a traditional relay feedback test is biased, asymmetry is induced, erroneous system identification occurs, and the test may stop oscillating. Adaptive methods which restore symmetry to the relay test are developed, and additional methods are presented to detect and recover from a non-oscillation condition. This is demonstrated both in simulation and in a process with significant transients and disturbances - friction stir welding (FSW). In testing, it is shown that the adaptive method is able to restore symmetry even in the presence of severe and changing disturbances, including after the weld plunge where temperature gradients change rapidly. Finally, these improvements enable system identification of systems such as FSW in a more robust and accurate manner.

42 ENGINEERING↗

DOE OSTI2023

Flexible and Architecture Aware Kalman Filter Code Generation

Explore the source record for details and available documents.

43 PARTICLE ACCELERATORS↗

DOE OSTI2022

Automatic Generation of High-Performance Convolution Kernels on ARM CPUs for Deep Learning

In this work, we present FastConv, a template-based code auto-generation open source library that can automatically generate high-performance deep learning convolution kernels of arbitrary matrices/tensors shapes. FastConv is based on the Winograd algorithm, which is reportedly the highest performing algorithm for the time-consuming convolution layers of convolutional neural networks. ARM CPUs cover a wide range designs and specifications, from embedded devices to HPC-grade CPUs. The leads to the dilemma of how to consistently optimize Winograd-based convolution solvers for convolution layers of different shapes. FastConv addresses this problem by using templates to auto-generate multiple shapes of tuned kernels variants suitable for skinny tall matrices. As a performance portable library, FastConv transparently searches for the best combination of kernel shapes, cache tiles, scheduling of loop orders, packing strategies, access patterns, and online/offline computations. Auto-tuning is used to search the parameter configuration space for the best performance for a given target architecture and problem size. The experiments with layer-wise evaluation on the VGG--16 model confirms a 1.25x performance gains is got by tuning the Winograd library. Integrated comparison results shows 1.02x to 1.40x, 1.14x to 2.17x, and 1.22x and 2.48x speedup is achieved over NNPACK, Arm NN, and FeatherCNN on the Kunpeng 920 beside few cases. Furthermore, problem size performance portability experiments with various convolution shapes shows that FastConv achieves 1.2x to 1.7x speedup and 2x to 22x speedup over NNPACK and ARM NN inference engine using Winograd on Kunpeng 920 . CPU performance portability evaluation on the VGG--16 show an average speedup over NNPACK of 1.42x, 1.21x, 1.26x, 1.37x, 2.26x, and 11.02x is observed on Kunpeng 920, Snapdragon 835, 855, 888, Apple M1, and AWS Graviton2, respectively.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2025

Coupling magnons to an opto-electronic parametric oscillator

Hybrid magnonic systems have emerged as versatile modular components for quantum signal transduction and sensing applications owing to their capability of connecting distinct quantum platforms. To date, the majority of the magnonic systems have been explored in a local, near-field scheme, due to the close proximity required for realizing a strong coupling between magnons and other excitations. This constraint greatly limits the applicability of magnons in developing remotely coupled, distributed quantum network systems. On the contrary, opto-electronic architectures hosting self-sustained oscillations have been a unique platform for long-haul signal transmission and processing. Here, we integrated an opto-electronic oscillator with a magnonic oscillator consisting of a microwave waveguide and a Y 3 Fe 5 O 12 (YIG) sphere, and demonstrated strong and coherent coupling between YIG’s magnon modes and the opto-electronic oscillator’s characteristic photon modes—revealing the hallmark anti-crossing gap in the measured spectrum. In particular, the photon mode is produced on-demand via a nonlinear, parametric process as stipulated by an external seed pump. Both the internal cavity phase and the external pump phase can be precisely tuned to stabilize either degenerate or nondegenerate auto-oscillations. Our result lays out a new, hybrid platform for investigating the long-distance coupling and nonlinearity in coherent magnonic phenomena, which may be found useful in constructing the future “distributed hybrid magnonic systems.”

36 MATERIALS SCIENCE↗

DOE OSTI2022

Vertex finding in neutrino-nucleus interaction: a model architecture comparison

We compare different neural network architectures for machine learning algorithms designed to identify the neutrino interaction vertex position in the MINERvA detector. The architectures developed and optimized by hand are compared with the architectures developed in an automated way using the package “Multi-node Evolutionary Neural Networks for Deep Learning” (MENNDL), developed at Oak Ridge National Laboratory. While the domain-expert hand-tuned network was the best performer, the differences were negligible and the auto-generated networks performed as well. There is always a trade-off between human, and computer resources for network optimization and this work suggests that automated optimization, assuming resources are available, provides a compelling way to save significant expert time.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

DOE OSTI2024

autoGEMM: Pushing the Limits of Irregular Matrix Multiplication on Arm Architectures

This paper presents an open-source library that pushes the limits of performance portability for irregular General Matrix Multiplication (GEMM) on the widely-used Arm architectures. Our library, autoGEMM, is designed to support a wide range of Arm processors: from edge devices to HPC-grade CPUs. autoGEMM generates optimized kernels for various hardware configurations by auto-combining fragments of autogenerated micro-kernels that employ hand-written optimizations to maximize computational efficiency. We optimize the kernel pipeline by tuning the register reuse and the data load/store overlapping. In addition, we use a dynamic tiling scheme to generate balanced tile shapes. Finally, we position autoGEMM on top of the TVM framework where our dynamic tiling scheme prunes the search space for TVM to identify the optimal combination of parameters for code optimization. Evaluations on five different classes of Arm chips demonstrate the advantages of autoGEMM. For small matrices, autoGEMM achieves 98% of peak and up to 2.0x speedup over state-of-the-art libraries such as LIBXSMM and LibShalom. For irregular matrices (i.e. tall skinny and long rectangles), autoGEMM is 1.3-2.0x faster than widely-used libraries such as OpenBLAS and Eigen. autoGEMM is publicly available at: https://github.com/wudu98/autoGEMM.

Wu, Du↗

DOE OSTI2018

Automatic tuning of the Community Atmospheric Model (CAM5) by using short-term hindcasts with an improved downhill simplex optimization method

Traditional trial-and-error tuning of uncertain parameters in global atmospheric general circulation models (GCMs) is time consuming and subjective. This study explores the feasibility of automatic optimization of GCM parameters for fast physics by using short-term hindcasts. An automatic workflow is described and applied to the Community Atmospheric Model (CAM5) to optimize several parameters in its cloud and convective parameterizations. We show that the auto-optimization leads to 10 % reduction of the overall bias in CAM5, which is already a well-calibrated model, based on a predefined metric that includes precipitation, temperature, humidity, and longwave/shortwave cloud forcing. The computational cost of the entire optimization procedure is about equivalent to a single 12-year atmospheric model simulation. The tuning reduces the large underestimation in the CAM5 longwave cloud forcing by decreasing the threshold relative humidity and the sedimentation velocity of ice crystals in the cloud schemes; it reduces the overestimation of precipitation by increasing the adjustment time in the convection scheme. The physical processes behind the tuned model performance for each targeted field are discussed. Limitations of the automatic tuning are described, including the slight deterioration in some targeted fields that reflect the structural errors of the model. It is pointed out that automatic tuning can be a viable supplement to process-oriented model evaluations and improvement.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2018

Automatic tuning of the Community Atmospheric Model CAM5.3 by using short-term hindcasts with an improved downhill simplex optimization method

Traditional trial-and-error tuning of uncertain parameters in global atmospheric General Circulation Models (GCM) is time consuming and subjective. This study explores the feasibility of automatic optimization of GCM parameters for fast physics by using short-term hindcasts. An automatic workflow is described and applied to the Community Atmospheric Model (CAM5) to optimize several parameters in its cloud and convective parameterizations. We show that the auto-optimization leads to 10 % reduction of the overall bias in CAM5, which is already a well calibrated model, based on a pre-defined metric that includes precipitation, temperature, humidity, and longwave/shortwave cloud forcing. The computational cost of the entire optimization procedure is about equivalent to about a single 12-year atmospheric model simulation. The tuning reduces the large underestimation in the CAM5 longwave cloud forcing by decreasing the threshold relative humidity and the sedimentation velocity of ice crystals in the cloud schemes; it reduces the overestimation of precipitation by increasing the adjustment time in the convection scheme. The physical processes behind the tuned model performance for each targeted field are discussed. Limitations of the automatic tuning are described, including the slight deterioration in some targeted fields that reflect the structural errors of the model. It is pointed out that automatic tuning can be a viable supplement to process-oriented model evaluations and improvement.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2020

Photodetachment spectroscopy and resonant photoelectron imaging of the 2-naphthoxide anion via dipole-bound excited states

We report a photodetachment spectroscopy and high-resolution resonant photoelectron imaging study of cryogenically-cooled 2-naphthoxide anions (C 10 H 7 O - ). The photodetachment spectrum revealed a dipole-bound state (DBS) 202(4) cm -1 below the detachment threshold and thirty-eight resonances corresponding to the vibrational levels of the DBS. By tuning the detachment laser to these above-threshold resonances, we obtained thirty-eight resonantly-enhanced photoelectron spectra, which were highly non-Franck-Condon as a result of mode-selective vibrational auto-detachment from the DBS. The resonances were assigned by comparing the resonant and non-resonant photoelectron spectra, assisted by the computed vibrational frequencies. Specifically, vibrational features with low Franck-Condon factors or from Franck-Condon-forbidden vibrational modes were significantly enhanced in the resonant photoelectron spectra, resulting in much richer spectroscopic information. The electron affinity of the 2-naphthoxy radical was measured accurately to be 19,387(4) cm -1 or 2.4037(5) eV. In addition, a total of seventeen vibrational frequencies were obtained for the 2-naphthoxy radical. In particular, seven Franck-Condon-forbidden out-of-plane bending modes, including the two lowest frequency modes (ν 48 at 102 cm -1 and ν47 at 171 cm -1 ), were observed, demonstrating the advantages of combining photodetachment spectroscopy and resonant photoelectron spectroscopy in obtaining vibrational information for polar radical species via DBS.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2023

Implementation and test of an automated control hunting fault correction algorithm in a fault detection and diagnostics tool

Control hunting due to improper proportional–integral–derivative (PID) parameters in the building automation system (BAS) is one of the most common faults identified in commercial buildings. It can cause suboptimal performance and early failure of heating, ventilation, and air conditioning (HVAC) equipment. Commercial fault detection and diagnostics (FDD) software represents one of the fastest growing market segments in smart building technologies in the United States. Implementation of PID retuning procedures as an auto-correction algorithm and integration into FDD software has the potential to mitigate control hunting across a heterogeneous portfolio of buildings with different BAS in a scalable way. This paper presents the development, implementation, and field testing of an automated control hunting fault correction algorithm based on lambda tuning open-loop rules. The algorithm was developed in a commercial FDD software and successfully tested among nine variable air volume boxes in an office building in the United States. The paper shows the feasibility of using FDD tools to automatically correct control hunting faults, discusses scalability considerations, and proposes a path forward for the HVAC industry and academia to further improve this technology.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

DOE OSTI2023

Design and Operation of a Multi-Bed Catalytic Micro-Reactor for the Study of Co-Processing of Bio-Oils with VGO

An industry wide shift from fossil-based fuel to renewable fuel sources including biomass, municipal waste, and plastics will require new process monitoring methods to minimize transitional risks including off specification product formation and catalyst deactivation. This project aims to provide a machine learning based process monitoring tool composed of online, slipstream mass spectra for use in biomass refineries and co-processing in existing refineries allowing operators to monitor product qualities and adjust process conditions accordingly. In order to maximize the robustness of the tool, large volumes of data must be collected to fine tune model parameters which consists of both micro and pilot scale mass spectral data. Micro-scale data is collected with a multi-tube micro-reactor housing up to six catalysts in horizontal beds, coupled with a molecular beam mass spectrometer. A pyrolizer equipped with an auto-sampler streamlines the micro-scale data collection process. This type of pyrolizer/micro-reactor configuration does not exist on the market, and therefore had to be created for the purposes of this project. The design and commissioning of this reactor will be presented in detail. This reactor set-up is highly flexible and increases throughput of analysis. For catalyst testing, each bed can be individually selected simply by turning valves. For catalyst reduction and regeneration, simultaneous flow through all six beds is used. The reproducibility of the system was first assessed with whole biomass pyrolysis along with pyrolysis of calibration standards. Initial work on this system evaluated two FCC catalysts, equilibrium catalyst (E-cat), and a proprietary catalyst from Johnson Matthey specifically design for co-processing of bio-oil with vacuum gas oil (VGO). This work used model compounds and VGO which illuminated differences in products produced by the catalysts.

biomass↗

DOE OSTI2025

JuTrack: A Julia package for auto-differentiable accelerator modeling and particle tracking

Efficient accelerator modeling and particle tracking are key for the design and configuration of modern particle accelerators. In this work, we present JuTrack, a nested accelerator modeling package developed in the Julia programming language and enhanced with compiler-level automatic differentiation (AD). With the aid of AD, JuTrack enables rapid derivative calculations in accelerator modeling, facilitating sensitivity analyses and optimization tasks. Here we demonstrate the effectiveness of AD-derived derivatives through several practical applications, including sensitivity analysis of space-charge-induced emittance growth, nonlinear beam dynamics analysis for a synchrotron light source, and lattice parameter tuning of the future Electron-Ion Collider (EIC). Through the incorporation of automatic differentiation, this package opens up new possibilities for accelerator physicists in beam physics studies and accelerator design optimization.

43 PARTICLE ACCELERATORS↗

DOE OSTI2022

Autodifferentiable Ensemble Kalman Filters

Data assimilation is concerned with sequentially estimating a temporally evolving state. This task, which arises in a wide range of scientific and engineering applications, is particularly challenging when the state is high-dimensional and the state-space dynamics are unknown. This paper introduces a machine learning framework for learning dynamical systems in data assimilation. Here, our auto-differentiable ensemble Kalman filters (AD-EnKFs) blend ensemble Kalman filters for state recovery with machine learning tools for learning the dynamics. In doing so, AD-EnKFs leverage the ability of ensemble Kalman filters to scale to high-dimensional states and the power of automatic differentiation to train high-dimensional surrogate models for the dynamics. Numerical results using the Lorenz-96 model show that AD-EnKFs outperform existing methods that use expectation-maximization or particle filters to merge data assimilation and machine learning. In addition, AD-EnKFs are easy to implement and require minimal tuning.

autodifferentiation↗

DOE OSTI2025

Predicting Initial Trans-Membrane Pressure for Optimized Operations in UF Unit Using Random Forest

With the growing scarcity of freshwater, innovative process design mechanisms like Reverse Osmosis (RO) are increasingly gaining attention among water treatment utilities to address the rising demand. Ensuring reliable water production necessitates efficient resource utilization, minimizing downtime in (ultra-filtration) UF systems. Recent advancements in machine learning (ML) have enabled the development of accurate data-driven models for Model Predictive Control (MPC), often requiring minimal prior knowledge of underlying physical processes. In this study, we present predictive regression models based on Random Forest (RF) and Auto-Regressive (AR) approaches to forecast the initial Trans-Membrane Pressure (TMP) for each filtration cycle in data generated by Direct Potable Reuse (DPR) systems. The proposed RF-based model demonstrates superior performance compared to baseline methods, including historical mean, Last Observation Carried Forward (LOCF), and naïve AR models, across various forecasting horizons in terms of root mean square error (RMSE) metric. To evaluate how different classes of process variables contribute to TMP dynamics over time, we examine the feature importance of independent covariates across multiple forecast horizons. This analysis provides insight into the temporal relevance of operational and sensor-derived features, guiding control and monitoring strategies. Additionally, the impact of hyperparameter tuning on TMP prediction performance is studied for both direct and recursive RF modelling approaches across increasing forecast horizons. Accurate prediction of initial TMP is critical for optimizing RO operations, as it enables the development of robust modelling frameworks by accurately estimating membrane fouling trends, thereby enhancing process efficiency and long-term reliability. The demonstrated efficacy of the RF-based approach highlights its potential as a tool for real-time decision-making in water treatment systems, paving the way for advanced process optimization and sustainable water resource management.

Mukherjee, Subrata [ORNL] (ORCID:0000000309930338)↗

DOE OSTI2025

Harnessing distributed GPU computing for generalizable graph convolutional networks in power grid reliability assessments

Although machine learning (ML) has emerged as a powerful tool for rapidly assessing grid contingencies, prior studies have largely considered a static grid topology in their analyses. This limits their application, since they need to be re-trained for every new topology. Here, this paper explores the development of generalizable graph convolutional network (GCN) models by pre-training them across a range of grid topologies and contingency types. We found that a GCN model with auto-regressive moving average (ARMA) layers with a line graph representation of the grid offered the best predictive performance in predicting voltage magnitudes (VM) and voltage angles (VA). We introduced the concept of phantom nodes to consider disparate grid topologies with a varying number of nodes and lines. For pre-training the GCN ARMA model across a variety of topologies, distributed graphics processing unit (GPU) computing afforded us significant training scalability. The predictive performance of this model on grid topologies that were part of the training data is substantially better than the direct current (DC) approximation. Although direct application of the pre-trained model to topologies that are not part of the grid is not particularly satisfactory, fine-tuning with small amounts of data from a specific topology of interest significantly improves predictive performance. In general, this paper highlights the feasibility of training large-scale GNN models to assess the reliability of power grids by considering a wide variety of grid topologies and contingency types. With the advent of foundational models in ML and the exponential increase in GPU computing clusters, generalizable ML models will significantly enhance how utilities manage power systems and make decisions in real-time or near-real-time.

24 - POWER TRANSMISSION AND DISTRIBUTION↗

DOE OSTI2023

Transient Efficiency, Flexibility, and Reliability Optimization of Coal-Fired Power Plants - Final Report

This program developed an advanced model-based monitoring and model-predictive control algorithms for a coal fired power plant (CFPP), and deployed these algorithms in a real-time platform to demonstrate performance benefits for transient flexibility and plant operation efficiency. More specifically, the objectives were successfully achieved through a combination of (i) developing a high-fidelity transient plant model in Apros, which was used as a high-fidelity plant simulation between $100-50\% TMCR$ where TMCR denotes the turbine maximum continuous rating, i.e., baseload, (ii) developing a very fast physics-based reduced-order model (ROM) of the plant, which ran more than $100\times$ faster than real-time, enabling its use as real-time embedded model for model-based estimation (MBE) and model predictive control (MPC) (iii) implementing a real-time MBE based on ROM using a robust unscented Kalman filter (UKF) to continuously tune the ROM to match the measurements from high-fidelity Apros plant model despite significant plant-model mismatch, and thus, obtain a Digital Twin of the plant (iv) designing and implementing a real-time MPC with dual objectives of transient plant load tracking with high ramp rates and minimizing coal consumption, i.e., improving plant efficiency in the baseload-partload operation range of $100-50\% TMCR$. Each key element above was developed and tested individually, and has been reported in corresponding Topical Reports. Finally, all the individual elements were integrated in an overall closed-loop system, that was successfully tested in desktop Simulink test harness simulations with ROM or high-fidelity model as the plant. Thereafter, the Simulink implementation was used to auto-generate C-code and deploy as real-time Docker microservice containers in Linux, and validate that they can run in real-time in the hardware-in-the loop (HIL) setup and produce the same results as in Simulink. The results of the integrated simulation tests in Simulink as well as the real-time HIL deployment are documented in this final report, showing good load tracking for load ramps at $3-4\%/min$ ramp rates, and achieving up to $5.5\%$ reduction in coal relative to baseline operation at $50\% TMCR$ load. The desktop and HIL simulations show successful performance of the overall model based estimation and control solution and achieve the key objectives of the program for flexible, efficient and reliable operation of subcritical coal fired power plants.

20 FOSSIL-FUELED POWER PLANTS↗

DOE OSTI2023

Transient Efficiency, Flexibility, and Reliability Optimization of Coal-Fired Power Plants - Final Program Review

20 FOSSIL-FUELED POWER PLANTS↗