Search NASA⌕ Search

SEARCH · Search NASA

Results for “kernel method”

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 1,315 records · Page 73

UNVEILING THE NATURE OF THE UNIDENTIFIED GAMMA-RAY SOURCES. V. ANALYSIS OF THE RADIO CANDIDATES WITH THE KERNEL DENSITY ESTIMATION

Nearly one-third of the γ-ray sources detected by Fermi are still unidentified, despite significant recent progress in this area. However, all of the γ-ray extragalactic sources associated in the second Fermi-LAT catalog have a radio counterpart. Motivated by this observational evidence, we investigate all the radio sources of the major radio surveys that lie within the positional uncertainty region of the unidentified γ-ray sources (UGSs) at a 95% level of confidence. First, we search for their infrared counterparts in the all-sky survey performed by the Wide-field Infrared Survey Explorer (WISE) and then we analyze their IR colors in comparison with those of the known γ-ray blazars. We propose a new approach, on the basis of a two-dimensional kernel density estimation technique in the single [3.4] – [4.6] – [12] μm WISE color-color plot, replacing the constraint imposed in our previous investigations on the detection at 22 μm of each potential IR counterpart of the UGSs with associated radio emission. The main goal of this analysis is to find distant γ-ray blazar candidates that, being too faint at 22 μm, are not detected by WISE and thus are not selected by our purely IR-based methods. We find 55 UGSs that likely correspond to radio sources with blazar-like IR signatures. An additional 11 UGSs that have blazar-like IR colors have been found within the sample of sources found with deep recent Australia Telescope Compact Array observations.

79 ASTRONOMY AND ASTROPHYSICS↗

Quantification of arterial plaque and lumen density with MDCT

Purpose: This study aimed to derive a mathematical correction function in order to normalize the CT number measurements for small volume arterial plaque and small vessel mimicking objects, imaged with multidetector CT (MDCT). Methods: A commercially available calcium plaque phantom (QRM GmbH, Moehrendorf, Germany) and a custom built cardiovascular phantom were scanned with 320 and 64 MDCT scanners. The calcium hydroxyapatite plaque phantom contained objects 0.5-5.0 mm in diameter with known CT attenuation nominal values ranging 50-800 HU. The cardiovascular phantom contained vessel mimicking objects 1.0-5.0 mm in diameter with different contrast media. Both phantoms were scanned using clinical protocols for CT angiography and images were reconstructed with different filter kernels. The measured CT number (HU) and diameter of each object were analyzed on three clinical postprocessing workstations. From the resultant data, a mathematical formula was derived based on absorption function exp(-{mu}{sup *}d) to demonstrate the relation between measured CT numbers and object diameters. Results: The percentage reduction in measured CT number (HU) for the group of selected filter kernels, apparent during CT angiography, is dependent only on the object size (plaque or vessel diameter). The derived formula of the form 1-c{sup *}exp(-a{sup *}d{sup b}) showed reduction in CT number for objects between 0.5 and 5 mm in diameter, with asymptote reaching background noise for small objects with diameters nearing the CT in-plane resolution (0.35 mm). No reduction was observed for the objects with diameters equal or larger than 5 mm. Conclusions: A clear mathematical relationship exists between object diameter and reduction in measured CT number in HU. This function is independent of exposure parameters and inherent attenuation properties of the objects studied. Future developments include the incorporation of this mathematical model function into quantification software in order to automatically generate a true assessment of measured CT number (HU) corresponding to plaque physical density {rho} (g/cm{sup 3}). This is a significant development for the accurate, noninvasive classification of noncalcified arterial plaque.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Agentic AI vs ML-Based Autotuning: A Comparative Study for Loop Reordering Optimization

High Performance Computing (HPC) applications rely heavily on code optimizations to achieve good performance on modern CPU and GPU architectures. Traditional Machine Learning auto-tuning approaches have demonstrated success in exploring high-dimensional spaces, but they often require expensive compile-run evaluations and lack adaptability for large HPC applications. The recent advances in Large Language Models (LLMs) and Agentic AI systems raise intriguing questions about the potential of these approaches to address specific optimization methodologies. This work aims to answer an essential question for the HPC community: “How Agentic AI Systems Compare to Traditional ML Autotuning Techniques?” To address this question, we present a comparative analysis between a traditional ML-based optimization approach and an Agentic AI system, evaluating their respective capabilities and limitations for loop-level optimization. In addition, we introduced a new Agentic AI system named LoopGen-AI using three different Large Language Models: GPT-4.1, Claude 4.0, and Gemini 2.5. A key finding is that LoopGen-AI achieves competitive per-formance with only a few program runs, the reasoning logs from the agents revealed that their decisions rely heavily on the combination of semantic understanding of the target kernel with dynamic feedback from the environment, highlighting a promising new dimension in performance tuning. In contrast, ML-based autotuners focus on statistical exploration, and require orders of magnitude more runs to reach peak performance. Additionally, our analysis shows that prompt engineering, particularly using Persona + Context Manager patterns, significantly impacts the effectiveness of Agentic AI. Our results indicate that while Agentic AI systems are not yet a complete replacement for ML-based autotuners, it can effectively complement traditional methods.

Rosas, Miguel Romero↗

Real-space inversion and super-resolution of ultrafast scattering

Ultrafast scattering using x-rays or electrons is an emerging method to obtain structure dynamics at the atomic lengthscales and timescales. However, directly resolving in real-space atomic motions is inherently limited by the finite detector range and the probe energy. So, as a result, the time-resolved signal interpretation is mostly done in reciprocal space and relies on modeling and simulations of specific structures and processes. Here, we introduce a model-free approach to directly resolve scattering signals in real space, surpassing the diffraction limit, using scattering kernels and signal priors that naturally arise from the measurement constraints. We demonstrate the approach on simulated and experimental data, recover multiple atomic motions at sub-angstrom resolutions, and discuss the recovery accuracy and resolution limits versus signal fidelity. The approach offers a robust path to obtain high-resolution real-space information of atomic-scale structure dynamics using current time-resolved x-ray or electron scattering sources.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Protein Kinase Classification with 2866 Hidden Markov Models and One Support Vector Machine

The main application considered in this paper is predicting true kinases from randomly permuted kinases that share the same length and amino acid distributions as the true kinases. Numerous methods already exist for this classification task, such as HMMs, motif-matchers, and sequence comparison algorithms. We build on some of these efforts by creating a vector from the output of thousands of structurally based HMMs, created offline with Pfam-A seed alignments using SAM-T99, which then must be combined into an overall classification for the protein. Then we use a Support Vector Machine for classifying this large ensemble Pfam-Vector, with a polynomial and chisquared kernel. In particular, the chi-squared kernel SVM performs better than the HMMs and better than the BLAST pairwise comparisons, when predicting true from false kinases in some respects, but no one algorithm is best for all purposes or in all instances so we consider the particular strengths and weaknesses of each.

Weber, Ryan↗

Effects of Fuel Composition on EGR Dilution Tolerance in Spark Ignited Engines

Fuel-specific differences in exhaust gas recirculation (EGR) dilution tolerance are studied in a modern, direct-injection single-cylinder research engine. A total of 6 model fuel blends are examined at a constant research octane number (RON) of 95 using n-heptane, iso-octane, toluene, and ethanol. Laminar flame speeds for these mixtures, which were calculated two different methods (an energy fraction mixing rule and a detailed kinetic simulation), spanned a range of about 6 cm/s. A constant fueling nominal load of 350 kPa IMEPg at 2000 rpm was operated with varying CA50 from 8-20 CAD aTDCf, and with EGR increasing until a COV of IMEP of 5% is reached. The results illustrate that flame speed affects EGR dilution tolerance; fuels with increased flame speeds increase EGR tolerance. Specifically, flame speed correlates most closely to the initial flame kernel growth, measured as the time of ignition to 5% mass fraction burned. The effect of the latent heat of vaporization on the flame speed is taken into account for the ethanol-containing fuels. At a 30 vol% blend level, the increased enthalpy of vaporization of ethanol compared to conventional hydrocarbons can decrease the temperature at the time of ignition by a maximum of 15 C, which can account for up to a 3.5 cm/s decrease in flame speed. The ethanol-containing fuels, however, still exhibit a flame speed advantage, and a dilution tolerance advantage over the slower flame-speed fuels. The fuel-specific differences in dilution tolerance are significant at the condition examined, allowing for a 50% relative increase in EGR (4% absolute difference in EGR) at a constant COV of IMEP of 3%.

dilution tolerance↗

Dedicated breast CT: Fibroglandular volume measurements in a diagnostic population

Purpose: To determine the mean and range of volumetric glandular fraction (VGF) of the breast in a diagnostic population using a high-resolution flat-panel cone-beam dedicated breast CT system. This information is important for Monte Carlo-based estimation of normalized glandular dose coefficients and for investigating the dependence of VGF on breast dimensions, race, and pathology. Methods: Image data from a clinical trial investigating the role of dedicated breast CT that enrolled 150 women were retrospectively analyzed to determine the VGF. The study was conducted in adherence to a protocol approved by the institutional human subjects review boards and written informed consent was obtained from all study participants. All participants in the study were assigned BI-RADS{sup Registered-Sign} 4 or 5 as per the American College of Radiology assessment categories after standard diagnostic work-up and underwent dedicated breast CT exam prior to biopsy. A Gaussian-kernel based fuzzy c-means algorithm was used to partition the breast CT images into adipose and fibroglandular tissue after segmenting the skin. Upon determination of the accuracy of the algorithm with a phantom, it was applied to 137 breast CT volumes from 136 women. VGF was determined for each breast and the mean and range were determined. Pathology results with classification as benign, malignant, and hyperplasia were available for 132 women, and were used to investigate if the distributions of VGF varied with pathology. Results: The algorithm was accurate to within {+-}1.9% in determining the volume of an irregular shaped phantom. The study mean ({+-} inter-breast SD) for the VGF was 0.172 {+-} 0.142 (range: 0.012-0.719). VGF was found to be negatively correlated with age, breast dimensions (chest-wall to nipple length, pectoralis to nipple length, and effective diameter at chest-wall), and total breast volume, and positively correlated with fibroglandular volume. Based on pathology, pairwise statistical analysis (Mann-Whitney test) indicated that at the 0.05 significance level, there was no significant difference in distributions of VGF without adjustment for age between malignant and nonmalignant breasts (p= 0.41). Pairwise comparisons of the distributions of VGF in increasing order of mammographic breast density indicated all comparisons were statistically significant (p < 0.002). Conclusions: This study used a different clinical prototype breast CT system than that in previous studies to image subjects from a different geographical region, and used a different algorithm for analysis of image data. The mean VGF estimated from this study is within the range reported in previous studies, indicating that the choice of 50% glandular weight fraction to represent an average breast for Monte Carlo-based estimation of normalized glandular dose coefficients in mammography needs revising. In the study, the distributions of VGF did not differ significantly with pathology.

60 APPLIED LIFE SCIENCES↗

CAST: Contraction Algorithm for Symmetric Tensors

Tensor contractions represent the most compute-intensive core kernels in ab initio computational quantum chemistry and nuclear physics. Symmetries in these tensor contractions makes them difficult to load balance and scale to large distributed systems. In this paper, we develop an efficient and scalable algorithm to contract symmetric tensors. We introduce a novel approach that avoids data redistribution in contracting symmetric tensors while also avoiding redundant storage and maintaining load balance. We present experimental results on two parallel supercomputers for several symmetric contractions that appear in the CCSD quantum chemistry method. We also present a novel approach to tensor redistribution that can take advantage of parallel hyperplanes when the initial distribution has replicated dimensions, and use collective broadcast when the final distribution has replicated dimensions, making the algorithm very efficient.

distributed memory↗

WE-AB-207B-07: Dose Cloud: Generating “Big Data” for Radiation Therapy Treatment Plan Optimization Research

Purpose: To provide a tool to generate large sets of realistic virtual patient geometries and beamlet doses for treatment optimization research. This tool enables countless studies exploring the fundamental interplay between patient geometry, objective functions, weight selections, and achievable dose distributions for various algorithms and modalities. Methods: Generating realistic virtual patient geometries requires a small set of real patient data. We developed a normalized patient shape model (PSM) which captures organ and target contours in a correspondence-preserving manner. Using PSM-processed data, we perform principal component analysis (PCA) to extract major modes of variation from the population. These PCA modes can be shared without exposing patient information. The modes are re-combined with different weights to produce sets of realistic virtual patient contours. Because virtual patients lack imaging information, we developed a shape-based dose calculation (SBD) relying on the assumption that the region inside the body contour is water. SBD utilizes a 2D fluence-convolved scatter kernel, derived from Monte Carlo simulations, and can compute both full dose for a given set of fluence maps, or produce a dose matrix (dose per fluence pixel) for many modalities. Combining the shape model with SBD provides the data needed for treatment plan optimization research. Results: We used PSM to capture organ and target contours for 96 prostate cases, extracted the first 20 PCA modes, and generated 2048 virtual patient shapes by randomly sampling mode scores. Nearly half of the shapes were thrown out for failing anatomical checks, the remaining 1124 were used in computing dose matrices via SBD and a standard 7-beam protocol. As a proof of concept, and to generate data for later study, we performed fluence map optimization emphasizing PTV coverage. Conclusions: We successfully developed and tested a tool for creating customizable sets of virtual patients suitable for large-scale radiation therapy optimization research.

60 APPLIED LIFE SCIENCES↗

A Theoretical Framework for Calibration in Computer Models: Parametrization, Estimation and Convergence Properties

Calibration parameters in deterministic computer experiments are those attributes that cannot be measured or available in physical experiments. Here, an approach to estimate them by using data from physical experiments and computer simulations. A theoretical framework is given which allows us to study the issues of parameter identifiability and estimation. We define the L 2 -consistency for calibration as a justification for calibration methods. It is shown that a simplified version of the original KO method leads to asymptotically L 2 -inconsistent calibration. This L 2 -inconsistency can be remedied by modifying the original estimation procedure. A novel calibration method, called the L 2 calibration, is proposed and proven to be L 2 -consistent and enjoys optimal convergence rate. Furthermore a numerical example and some mathematical analysis are used to illustrate the source of the L 2 -inconsistency problem.

97 MATHEMATICS AND COMPUTING↗

Entropy-based closure for probabilistic learning on manifolds

In a recent paper, the authors proposed a general methodology for probabilistic learning on manifolds. The method was used to generate numerical samples that are statistically consistent with an existing dataset construed as a realization from a non-Gaussian random vector. The manifold structure is learned using diffusion manifolds and the statistical sample generation is accomplished using a projected Itˆo stochastic differential equation. This probabilistic learning approach has been extended to polynomial chaos representation of databases on manifolds and to probabilistic nonconvex constrained optimization with a fixed budget of function evaluations. The methodology introduces an isotropic-diffusion kernel with hyperparameter ". Currently, " is more or less arbitrarily chosen. In this paper, we propose a selection criterion for identifying an optimal value of ", based on a maximum entropy argument. The result is a comprehensive, closed, probabilistic model for characterizing data sets with hidden constraints. This entropy argument ensures that out of all possible models, this is the one that is the most uncertain beyond any specified constraints, which is selected. Applications are presented for several databases.

97 MATHEMATICS AND COMPUTING↗

Next-Cycle Optimal Fuel Control for Cycle-to-Cycle Variability Reduction in EGR-Diluted Combustion

In this simulation study, cycle-to-cycle fuel control was used to reduce CCV by injecting additional fuel in operating conditions with sporadic misfires and partial burns. An optimal control policy was proposed that utilizes 1) a physics-based model that tracks in-cylinder gas composition and 2) a one-step-ahead prediction of the combustion efficiency based on a kernel density estimator. The optimal solution, however, presents a tradeoff between the reduction in combustion CCV and the increase in fuel injection quantity required to stabilize the charge. Such a tradeoff can be ad- just by a single parameter embedded in the cost function.

Maldonado, BryanP. [Oak Ridge National Lab. (ORNL)↗

A performance model for the communication in fast multipole methods on high-performance computing platforms

Exascale systems are predicted to have approximately 1 billion cores, assuming gigahertz cores. Limitations on affordable network topologies for distributed memory systems of such massive scale bring new challenges to the currently dominant parallel programing model. Currently, there are many efforts to evaluate the hardware and software bottlenecks of exascale designs. It is therefore of interest to model application performance and to understand what changes need to be made to ensure extrapolated scalability. The fast multipole method (FMM) was originally developed for accelerating N-body problems in astrophysics and molecular dynamics but has recently been extended to a wider range of problems. Its high arithmetic intensity combined with its linear complexity and asynchronous communication patterns make it a promising algorithm for exascale systems. In this paper, we discuss the challenges for FMM on current parallel computers and future exascale architectures, with a focus on internode communication. We focus on the communication part only; the efficiency of the computational kernels are beyond the scope of the present study. We develop a performance model that considers the communication patterns of the FMM and observe a good match between our model and the actual communication time on four high-performance computing (HPC) systems, when latency, bandwidth, network topology, and multicore penalties are all taken into account. To our knowledge, this is the first formal characterization of internode communication in FMM that validates the model against actual measurements of communication time. The ultimate communication model is predictive in an absolute sense; however, on complex systems, this objective is often out of reach or of a difficulty out of proportion to its benefit when there exists a simpler model that is inexpensive and sufficient to guide coding decisions leading to improved scaling. The current model provides such guidance.

Computer Science↗

Reducing Computation and Communication in Scientific Computing: Connecting Theory to Practice.

This report summarizes the work produced as part of a Truman Fellowship appointment and its associated LDRD project. The overall goal of the project was to develop better algorithms and implementations for key computational kernels within the field of scientific computing by designing them to be communication efficient, moving as little data as possible. The primary problem of interest was dense matrix multiplication; other computations that were addressed include sparse matrix-matrix multiplication, QR factorization, solving symmetric linear systems, and the symmetric eigendecomposition. The project also involved the study of computations at the intersection of scientific computing and data analysis, including nonnegative matrix factorization for discovering latent factors, Tucker tensor decomposition for data compression, and sampling methods for similarity search.

97 MATHEMATICS AND COMPUTING↗

The Kokkos Ecosystem [Brief]

In 2016/2017, the field of High-Performance Computing (HPC) entered a new era driven by fundamental physics challenges to produce ever more energy and cost-efficient processors. Since the convergence on the Message-Passing Interface (MPI) standard in the mid-1990s, application developers enjoyed a seemingly static view of the underlying machine — that of a distributed collection of homogeneous nodes executing in collaboration. However, after almost two decades of dominance, the sole use of MPI to derive parallelism acted as a limiter to improved future performance. While MPI is widely expected to continue to function as the basic mechanism for communication between compute nodes for the immediate future, additional parallelism is required on the computing node itself if high performance and efficiency goals are to be realized. When reviewing the architectures of the top HPC systems today, the change in paradigm is clear: the compute nodes of the leading machines in the world are either powered by many-core chips with a few dozen cores each, or use heterogeneous designs, where traditional CPUs marshal work to massively parallel compute accelerators which has as many as 200,000 processing threads in flight simultaneously. Complicating matters further for application developers, each processor vendor has its own preferred way of writing code for their architecture.The Kokkos EcoSystem was released by Sandia in 2017 to address this new era in HPC system design by providing a vendor independent performance portable programming system for scientific, engineering, and mathematical software applications written in the C++ programming language. Using Kokkos, application developers can be more productive because they will not have to create and maintain separate versions of their software for each architecture, nor will they have to be experts in each architecture's peculiar requirements. Instead, they will have a single method of programming for the diverse set of modern HPC architectures. While Kokkos started in 2011 as a programming model only, it soon became clear that complex applications needed more. It is also critical to have a portable mathematical functions and developers need tools to debug their applications, gain insight into the performance characteristics of their codes and tune algorithm performance parameters through automated processes. The Kokkos EcoSystem addresses those needs through its three main components: the Kokkos Core programming model, the Kokkos Kernels math library, and the Kokkos Tools project.

97 MATHEMATICS AND COMPUTING↗

Mixed boundary-value problems in mechanics

Definitions in the case of multiple series equations and multiple integral equations are examined. In considering the solution of a given mixed boundary value problem perhaps the simplest technique is the direct application of the method of complex potentials provided the problem admits such potentials and the domain and the boundary conditions are suitable for such an application. The direct application of complex potentials is described with the aid of examples, taking into account a problem in potential theory, the case of periodic cuts, and an elasticity problem for a nonhomogeneous plane. The reduction to singular integral equations is discussed along with the numerical solution of singular integral equations of the first kind, integral equations with generalized Cauchy kernels, and singular integral equations of the second kind.

Erdogan, F.↗

Deploy production sliding mesh capability with linear solver benchmarking.

Wind applications require the ability to simulate rotating blades. To support this use-case, a novel design-order sliding mesh algorithm has been developed and deployed. The hybrid method combines the control volume finite element methodology (CVFEM) with concepts found within a discontinuous Galerkin (DG) finite element method (FEM) to manage a sliding mesh. The method has been demonstrated to be design-order for the tested polynomial basis (P=1 and P=2) and has been deployed to provide production simulation capability for a Vestas V27 (225 kW) wind turbine. Other stationary and canonical rotating ow simulations are also presented. As the majority of wind-energy applications are driving extensive usage of hybrid meshes, a foundational study that outlines near-wall numerical behavior for a variety of element topologies is presented. Results indicate that the proposed nonlinear stabilization operator (NSO) is an effective stabilization methodology to control Gibbs phenomena at large cell Peclet numbers. The study also provides practical mesh resolution guidelines for future analysis efforts. Application-driven performance and algorithmic improvements have been carried out to increase robustness of the scheme on hybrid production wind energy meshes. Specifically, the Kokkos-based Nalu Kernel construct outlined in the FY17/Q4 ExaWind milestone has been transitioned to the hybrid mesh regime. This code base is exercised within a full V27 production run. Simulation timings for parallel search and custom ghosting are presented. As the low-Mach application space requires implicit matrix solves, the cost of matrix reinitialization has been evaluated on a variety of production meshes. Results indicate that at low element counts, i.e., fewer than 100 million elements, matrix graph initialization and preconditioner setup times are small. However, as mesh sizes increase, e.g., 500 million elements, simulation time associated with \setup-up" costs can increase to nearly 50% of overall simulation time when using the full Tpetra solver stack and nearly 35% when using a mixed Tpetra- Hypre-based solver stack. The report also highlights the project achievement of surpassing the 1 billion element mesh scale for a production V27 hybrid mesh. A detailed timing breakdown is presented that again suggests work to be done in the setup events associated with the linear system. In order to mitigate these initialization costs, several application paths have been explored, all of which are designed to reduce the frequency of matrix reinitialization. Methods such as removing Jacobian entries on the dynamic matrix columns (in concert with increased inner equation iterations), and lagging of Jacobian entries have reduced setup times at the cost of numerical stability. Artificially increasing, or bloating, the matrix stencil to ensure that full Jacobians are included is developed with results suggesting that this methodology is useful in decreasing reinitialization events without loss of matrix contributions. With the above foundational advances in computational capability, the project is well positioned to begin scientific inquiry on a variety of wind-farm physics such as turbine/turbine wake interactions.

97 MATHEMATICS AND COMPUTING↗

Deploy production sliding mesh capability with linear solver benchmarking (ECP Milestone Report, Ver. 1.0)

Wind applications require the ability to simulate rotating blades. To support this use-case, a novel design-order sliding mesh algorithm has been developed and deployed. The hybrid method combines the control volume finite element methodology (CVFEM) with concepts found within a discontinuous Galerkin (DG) finite element method (FEM) to manage a sliding mesh. The method has been demonstrated to be design-order for the tested polynomial basis (P=1 and P=2) and has been deployed to provide production simulation capability for a Vestas V27 (225 kW) wind turbine. Other stationary and canonical rotating flow simulations are also presented. As the majority of wind-energy applications are driving extensive usage of hybrid meshes, a foundational study that outlines near-wall numerical behavior for a variety of element topologies is presented. Results indicate that the proposed nonlinear stabilization operator (NSO) is an effective stabilization methodology to control Gibbs phenomena at large cell Peclet numbers. The study also provides practical mesh resolution guidelines for future analysis efforts. Application-driven performance and algorithmic improvements have been carried out to increase robustness of the scheme on hybrid production wind energy meshes. Specifically, the Kokkos-based Nalu Kernel construct outlined in the FY17/Q4 ExaWind milestone has been transitioned to the hybrid mesh regime. This code base is exercised within a full V27 production run. Simulation timings for parallel search and custom ghosting are presented. As the low-Mach application space requires implicit matrix solves, the cost of matrix reinitialization has been evaluated on a variety of production meshes. Results indicate that at low element counts, i.e., fewer than 100 million elements, matrix graph initialization and preconditioner setup times are small. However, as mesh sizes increase, e.g., 500 million elements, simulation time associated with "setup-up" costs can increase to nearly 50% of overall simulation time when using the full Tpetra solver stack and nearly 35% when using a mixed Tpetra- Hypre-based solver stack. The report also highlights the project achievement of surpassing the 1 billion element mesh scale for a production V27 hybrid mesh. A detailed timing breakdown is presented that again suggests work to be done in the setup events associated with the linear system. In order to mitigate these initialization costs, several application paths have been explored, all of which are designed to reduce the frequency of matrix reinitialization. Methods such as removing Jacobian entries on the dynamic matrix columns (in concert with increased inner equation iterations), and lagging of Jacobian entries have reduced setup times at the cost of numerical stability. Artificially increasing, or bloating, the matrix stencil to ensure that full Jacobians are included is developed with results suggesting that this methodology is useful in decreasing reinitialization events without loss of matrix contributions. With the above foundational advances in computational capability, the project is well positioned to begin scientific inquiry on a variety of wind-farm physics such as turbine/turbine wake interactions.

Domino, Stefan P.↗