Search NASA⌕ Search

SEARCH · Search NASA

Results for “Dynamic Graph Clustering”

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.

36 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Cluster-Graph Fingerprinting: A Framework for Quantitative Analysis of Machine-Learned Interatomic Model Training and Simulation Data

Machine-learned interatomic models represent a significant advancement in simulation methods, extending the predictive ability of first-principles methods to previously inaccessible length and time scales. However, the data-driven nature of these models can lead to difficult-to-detect errors that can compromise prediction accuracy. To address this challenge, we introduce a novel fingerprinting approach based on the Chebyshev Interaction Model for Efficient Simulation (ChIMES) ML-IAM graph-based descriptor. Our strategy enables efficient and statistically rigorous analysis of system configurations used in ML-IAM training and those generated by their application, e.g., in molecular dynamics simulations. We demonstrate that these fingerprints can effectively assess novelty of a configuration relative to an existing data set and determine dissimilarity among individual configurations, which are two key tasks in workflows for active learning-based ML-IAM training, data set curation, and on-the-fly uncertainty quantification.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Causal interaction in high frequency turbulence at the biosphere–atmosphere interface: Structural behavior

High-frequency (e.g., 10 Hz) eddy covariance measurements are typically used to estimate fluxes at the land–atmosphere interface at timescales of 15–60 min. These multivariate data contain information about the interdependency at high frequency between the interacting variables such as wind, humidity, temperature, and CO 2⁠ . We use data at 10 Hz from an eddy covariance instrument located at 25 m above agricultural land in the Midwestern US, which offers an opportunity to move beyond the traditional spectral analyses to explore causal dependency among variables. In this study, we quantify the structure of inter-dependencies of interacting variables at high frequency represented by a directed acyclic graph (DAG). We compare DAGs to investigate changes in structural differences in causal interactions. We then apply a distance-based classification and -means clustering approach to identify the evolution of the causal structure represented by a DAG. Our method selects an unbiased number of clusters of similar structures and characterizes the similarities and differences between them. We explore a range of dynamic behavior using data from a clear sky day and during a solar eclipse in 2017. Our results show well-defined clusters of similar causal dependencies as the system evolves. Furthermore, our approach provides a methodological framework to understand how causal dependence in turbulence manifests in high-frequency data when represented through a DAG.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Robust training of machine learning interatomic potentials with dimensionality reduction and stratified sampling

Abstract Machine learning interatomic potentials (MLIPs) enable accurate simulations of materials at scales beyond that accessible by ab initio methods and play an increasingly important role in the study and design of materials. However, MLIPs are only as accurate and robust as the data on which they are trained. Here, we present DImensionality-Reduced Encoded Clusters with sTratified (DIRECT) sampling as an approach to select a robust training set of structures from a large and complex configuration space. By applying DIRECT sampling on the Materials Project relaxation trajectories dataset with over one million structures and 89 elements, we develop an improved materials 3-body graph network (M3GNet) universal potential that extrapolates more reliably to unseen structures. We further show that molecular dynamics (MD) simulations with the M3GNet universal potential can be used instead of expensive ab initio MD to rapidly create a large configuration space for target systems. We combined this scheme with DIRECT sampling to develop a reliable moment tensor potential for titanium hydrides without the need for iterative augmentation of training structures. This work paves the way for robust high-throughput development of MLIPs across any compositional complexity.

Qi, Ji (ORCID:0000000158089931)↗
DOE OSTI2024

Molecular architecture and functional dynamics of the pre-incision complex in nucleotide excision repair

Nucleotide excision repair (NER) is vital for genome integrity. Yet, our understanding of the complex NER protein machinery remains incomplete. Combining cryo-EM and XL-MS data with AlphaFold2 predictions, we build an integrative model of the NER pre-incision complex(PInC). Here TFIIH serves as a molecular ruler, defining the DNA bubble size and precisely positioning the XPG and XPF nucleases for incision. Using simulations and graph theoretical analyses, we unveil PInC’s assembly, global motions, and partitioning into dynamic communities. Remarkably, XPG caps XPD’s DNA-binding groove and bridges both junctions of the DNA bubble, suggesting a novel coordination mechanism of PInC’s dual incision. XPA rigging interlaces XPF/ERCC1 with RPA, XPD, XPB, and 5' ssDNA, exposing XPA’s crucial role in licensing the XPF/ERCC1 incision. Mapping disease mutations onto our models reveals clustering into distinct mechanistic classes, elucidating xeroderma pigmentosum and Cockayne syndrome disease etiology.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2021

Amphiphile conformation impacts aggregate morphology and solution structure across multiple lengthscales

Although the self-assembly of amphiphiles is well-studied in aqueous solutions, much less is understood about the fundamental driving forces and structure property relationships in non-polar media. In recent work [Journal of Physical Chemistry B, 2020, 124, 10822.] the authors have studied a series of malonomide-based amphiphiles that are relevant to liquid-liquid extraction. That work demonstrated that aggregation is largely driven by local dipole-dipole interactions between molecules. Here, we build upon this observation to develop a more detailed understanding of how the balance of dipole-dipole interactions (controlled by conformation) and molecular architecture influences the morphology of the aggregates across lengthscales. Using constrained molecular dynamics about key degrees of freedom, we demonstrate that the conformation of N,N-dimethyl,N,N-dioctylhexylethoxy malonamide (DMDOHEMA) and N,N-dimethyl,N,N-dibutyltetradecyl malonamide (DMDBTDMA) has a significant impact upon self-association - where appropriate conformational sampling is essential. To quantify the aggregate morphology, several graph theoretic and persistent homology based properties are determined. The former examines the patterns of intermolecular interactions within clusters, while the latter examines the 3-dimensional spatial distribution. Based upon these analyses, we find that the morphology of aggregates, particularly at higher malonamide concentration, depends on a balance of dipole alignment and alkyl tail sterics. Dipole alignment encourages linear patterns of the intermolecular interactions within aggregates, while the alkyl tail steric interactions between the malonamide result in noticeably less linear aggregates for DMDOHEMA than DMDBTDMA. This is reflected in the spatial distribution, where more holes or voids exist between extractants within the DMDOHEMA that distribute within the solution in more of a "swiss cheese" arrangement as opposed to the more filamentous distribution of DMBDTDMA. This study links conformation and molecular structure to the morphology of amphiphile assemblies, and serves as a basis for ongoing study of multicomponent amphiphile solutions with polar and other solutes, and how these impact aggregation phenomena.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

Dynamic conformational switching underlies TFIIH function in transcription and DNA repair and impacts genetic diseases

Transcription factor IIH (TFIIH) is a protein assembly essential for transcription initiation and nucleotide excision repair (NER). Yet, understanding of the conformational switching underpinning these diverse TFIIH functions remains fragmentary. TFIIH mechanisms critically depend on two translocase subunits, XPB and XPD. To unravel their functions and regulation, we build cryo-EM based TFIIH models in transcription- and NER-competent states. Using simulations and graph-theoretical analysis methods, we reveal TFIIH’s global motions, define TFIIH partitioning into dynamic communities and show how TFIIH reshapes itself and self-regulates depending on functional context. Our study uncovers an internal regulatory mechanism that switches XPB and XPD activities making them mutually exclusive between NER and transcription initiation. By sequentially coordinating the XPB and XPD DNA-unwinding activities, the switch ensures precise DNA incision in NER. Mapping TFIIH disease mutations onto network models reveals clustering into distinct mechanistic classes, affecting translocase functions, protein interactions and interface dynamics.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

Netostat: analyzing dynamic flow patterns in high-speed networks

Understanding flow traffic patterns in networks, such as the Internet or service provider networks, is crucial to improving their design and building them robustly. However, as networks grow and become more complex, it is increasingly cumbersome and challenging to study how the many flow patterns, sizes and the continually changing source-destination pairs in the network evolve with time. Here, we present Netostat, a visualization-based network analysis tool that uses visual representation and a mathematics framework to study and capture flow patterns, using graph theoretical methods such as clustering, similarity and difference measures. Netostat generates an interactive graph of all traffic patterns in the network, to isolate key elements that can provide insights for traffic engineering. We present results for U.S. and European research networks, ESnet and GEANT, demonstrating network state changes, to identify major flow trends, potential points of failure, and bottlenecks.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

Intelligent Experiments through Real-Time AI: Fast Data Processing and Autonomous Detector Control for High-Energy Nuclear Experiments

The aim of this project is to develop software and hardware for fast real-time data processing and autonomous detector control and calibration for the sPHENIX and the future EIC experiments. Below summarizes Georgia Tech team efforts in the past year: 1. We developed a real-time clustering algorithm and FPGA-based pipeline architecture for processing fired pixel data from ALPIDE sensors in sPHENIX experiments. Our Columnar Clustering Co-Design introduces a hardware-aware, stream-friendly approach that segments pixel data by column pairs using a Column Pair Clustering (CPC) strategy, followed by Cluster Stitching to merge adjacent subclusters. Implemented in Vitis HLS, the pipeline comprises five stages—read-in, subclustering, stitching, analysis, and write-out—connected by tagged HLS streams with custom end-of-event signaling for robust synchronization. We designed a pipelined dataflow model optimized for throughput, low latency, and minimal buffering, enabling scalable clustering across events of arbitrary size. Our system maintains spatial precision via center-of-mass and shape key extraction and efficiently handles edge cases such as fragmented or nested clusters. Compared against DBSCAN in both software and hardware, our approach demonstrates competitive performance under FPGA constraints. 2. We also conducted a comprehensive algorithm-to-hardware co-design of connected component analysis tailored for sPHENIX experiments, focusing on real-time, low-latency processing using FPGAs and High-Level Synthesis (HLS). Starting from a Python-based particle tracking pipeline, the team translated the core logic—graph traversal via DFS and Union-Find—into an HLS-compatible C++ model, replacing dynamic memory and recursion with static arrays and pipelined control flow. The final design includes a fully streamed and dataflow-compatible Union-Find kernel optimized across five iterations, incorporating loop pipelining, array partitioning, AXI/FIFO interface tuning, and function flattening. Experimental results show up to 14.8× speedup over the CPU baseline, reducing per-graph latency to 1.58 μs and demonstrating strong resource efficiency with only ~7k LUTs and zero BRAM usage. The design maintains functional correctness against the Python reference using a Python-based C-simulation framework and Mean Squared Error metrics. This work validates the potential of HLS-driven FPGA designs for edge-level HEP data acquisition, laying a scalable foundation for future integration with real-time detector pipelines and multi-graph processing systems.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2023

CEGANN: Crystal Edge Graph Attention Neural Network for multiscale classification of materials environment

Abstract We introduce Crystal Edge Graph Attention Neural Network (CEGANN) workflow that uses graph attention-based architecture to learn unique feature representations and perform classification of materials across multiple scales (from atomic to mesoscale) and diverse classes ranging from metals, oxides, non-metals to hierarchical materials such as zeolites and semi-ordered mesophases. CEGANN can classify based on a global, structure-level representation such as space group and dimensionality (e.g., bulk, 2D, clusters, etc.). Using representative materials such as polycrystals and zeolites, we demonstrate its transferability in performing local atom-level classification tasks, such as grain boundary identification and other heterointerfaces. CEGANN classifies in (thermal) noisy dynamical environments as demonstrated for representative zeolite nucleation and growth from an amorphous mixture. Finally, we use CEGANN to classify multicomponent systems with thermal noise and compositional diversity. Overall, our approach is material agnostic and allows for multiscale feature classification ranging from atomic-scale crystals to heterointerfaces to microscale grain boundaries.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Templates for Risk Informed Assurance with Curvature Embeddings (TRACE)

We investigate recovery of geometric structure from networks embedded in manifolds with spatially varying curvature, extending the constant-curvature framework of Lubold et al. (2023). Our work supports cascade risk assessment in critical infrastructure through the Templates for Risk-informed Assurance with Curvature Embeddings (TRACE) framework. Simulations on a bi-modal Gaussian surface show that constant-curvature methods yield weighted averages shaped by clique patterns, while hierarchical clustering identifies distinct regimes. Localized estimation, however, reveals boundary contamination in transitional regions. To address heterogeneity, we develop distance metrics for graphs with edge and node features, proving their metric validity, and validate them via deterministic graph generation from canonical tilings. We further propose a diffusion-based anomaly detection approach that treats networks as glued manifolds, using curvature discontinuities to detect structural anomalies. Employing the carré-du-champ operator and scalar curvature, we achieve robust anomaly discrimination, demonstrated on the Singapore Water Treatment (SWaT) dataset with joint network-traffic and sensor features. Integration with TRACE reveals how curvature shapes cascade dynamics: positive curvature impedes, while negative curvature accelerates propagation. This geometric perspective provides interpretable risk metrics and visualization tools for critical infrastructure managers. While full validation remains ongoing, our contributions establish a rigorous foundation for geometric analysis of network resilience and cascade vulnerability.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Causal interaction in high frequency turbulence at the biosphere–atmosphere interface: Structure–function coupling

At the biosphere–atmosphere interface, nonlinear interdependencies among components of an ecohydrological complex system can be inferred using multivariate high frequency time series observations. Information flow among these interacting variables allows us to represent the causal dependencies in the form of a directed acyclic graph (DAG). Here, we use high frequency multivariate data at 10 Hz from an eddy covariance instrument located at 25 m above agricultural land in the Midwestern US to quantify the evolutionary dynamics of this complex system using a sequence of DAGs by examining the structural dependency of information flow and the associated functional response. We investigate whether functional differences correspond to structural differences or if there are no functional variations despite the structural differences. We base our analysis on the hypothesis that causal dependencies are instigated through information flow, and the resulting interactions sustain the dynamics and its functionality. To test our hypothesis, we build upon causal structure analysis in the companion paper to characterize the information flow in similarly clustered DAGs from 3-min non-overlapping contiguous windows in the observational data. We characterize functionality as the nature of interactions as discerned through redundant, unique, and synergistic components of information flow. Through this analysis, we find that in turbulence at the biosphere–atmosphere interface, the variables that control the dynamic character of the atmosphere as well as the thermodynamics are driven by non-local conditions, while the scalar transport associated with CO and H 2 O is mainly driven by short-term local conditions.

58 GEOSCIENCES↗
DOE OSTI2022

Opinion dynamics in financial markets via random networks

We investigate financial market dynamics by introducing a heterogeneous agent-based opinion formation model. In this work, we organize individuals in a financial market according to their trading strategy, namely, whether they are noise traders or fundamentalists. The opinion of a local majority compels the market exchanging behavior of noise traders, whereas the global behavior of the market influences the decisions of fundamentalist agents. We introduce a noise parameter, q , to represent the level of anxiety and perceived uncertainty regarding market behavior, enabling the possibility of adrift financial action. We place individuals as nodes in an Erdös-Rényi random graph, where the links represent their social interactions. At any given time, individuals assume one of two possible opinion states ±1 regarding buying or selling an asset. The model exhibits fundamental qualitative and quantitative real-world market features such as the distribution of logarithmic returns with fat tails, clustered volatility, and the long-term correlation of returns. We use Student’s t distributions to fit the histograms of logarithmic returns, showing a gradual shift from a leptokurtic to a mesokurtic regime depending on the fraction of fundamentalist agents. Furthermore, we compare our results with those concerning the distribution of the logarithmic returns of several real-world financial indices.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Impact of Water Extraction on Malonamide Aggregation: A Molecular Dynamics and Graph Theoretic Approach

Solution structure in liquid-liquid extraction affects the efficacy of separation; however, even for simplified organic phases, structural characterization and attribution of aggregation to intermolecular interactions are fundamental challenges. We investigate water uptake into organic phases for two malonamides commonly applied to actinide and lanthanide separations. Extracted water induces reorganization of the amphiphilic extractant molecules, although we find this rearrangement is not strongly manifested in small-angle X-ray scattering making it challenging to probe without methods such as atomistic simulation. Using a graph theoretic approach to define hydrogen bonded water/malonamide aggregates from molecular dynamics simulations, we find evidence of a characteristic aggregate size by water number that results from geometric accommodation of the surrounding malonamide molecules. Furthermore, this implies a degree of size selectivity inherent to these water-in-oil aggregates. Conversely, we find no evidence of a characteristic size of the aggregates with respect to their malonamide number. By defining a separate graphical representation of self-association of the amphiphilic malonamides, we quantify how water affects the local and nonlocal topology of the malonamide network, providing a basis for characterization of the structure and impact of polar solutes in increasingly complex organic phases.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

FLASH: FPGA-Accelerated Smart Switches with GCN Case Study

Some communication switches, e.g., the Mellanox SHArP and those in the IBM BlueGene clusters, are augmented to process packets at the application level with fixed-function collectives. This approach, however, lacks flexibility, which limits their applicability in diverse and dynamic workloads. Recently, a new type of programmable packet processor, which uses high-level languages, e.g., P4, has emerged as possible candidates. P4-based switches, however, fall short in certain applications, including machine learning, where capabilities not currently supported by P4 are needed. These include more complex calculation, such as sparse computation and fused multiply-accumulate, data-intensive floating point operations, data reuse, and significant memory. The problem addressed here is that such a switch augmentation needs to support: a large amount of state, significant flexible compute capability, and ease of programming, all while maintaining full functionality, including ensuring high throughput, and demonstrating utility. In this work, we propose a programmable look-aside-type accelerator that can be embedded into, or attached to, existing communication switch pipelines and that is capable of processing packets at line-rate. The proposed in-switch accelerator is based on mixing an ISA (subset of RISC-V instructions) with dataflow graphs (found in CGRAs). To augment performance, vector instructions are also supported. To facilitate usability, we have developed a complete toolchain to compile user-provided C/C++ codes to appropriate back-end instructions for configuring the accelerator. While this approach is flexible enough to support various workloads, in this paper, we consider Graph Convolutional Networks (GCNs) as a case study. Experimental results show that this approach considerably improves the performance of distributed GCN applications.

Haghi, Pouya↗
DOE OSTI2025

Graph-Based Prediction of Spatio-Temporal Vaccine Hesitancy From Insurance Claims Data

Growing vaccine hesitancy is contributing to the decline in immunization rates for highly contagious, vaccine-preventable childhood diseases. Therefore, there has been a significant interest in understanding how hesitancy is spreading at higher spatio-temporal resolutions, enabling more targeted interventions. Motivated by this, we study the problem of prediction of vaccine hesitancy at the ZIP Code level, referred to as the VaxHesitancy problem. A significant challenge for this problem is the lack of high-resolution data that indicates hesitancy. Here, we develop a hybrid VaxHesSTL framework that combines a Graph Neural Network (GNN) and a Recurrent Neural Network (RNN) to address the VaxHesitancy problem. The GNN uses a ZIP Code-level network to capture spatial signals from neighboring areas, while the RNN models the temporal dynamics present in the data. We train and evaluate VaxHesSTL using a large dataset, namely the All-Payer Claims Databases (APCD), for Virginia, consisting of insurance claims from over five million individuals for six years. We find that an aggregated contact network or graph, developed from a detailed activity-based population network, plays an important role in the performance of VaxHesSTL, compared to graph models based solely on spatial proximity. Experiments demonstrate that VaxHesSTL outperforms a range of state-of-the-art baselines, which rely solely on historical time series data without accounting for spatial relationships. Since hesitancy data at higher spatial resolution is often unavailable or hard to get, we incorporate an active learning approach with our VaxHesSTL framework to optimize the training set without compromising the prediction performance. We find that hesitancy data for only 18% of ZIP Codes selected by active learning allows us to forecast hesitancy for all the ZIP Codes in the Virginia.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2020

ADEPT: a domain independent sequence alignment strategy for gpu architectures

Bioinformatic workflows frequently make use of automated genome assembly and protein clustering tools. At the core of most of these tools, a significant portion of execution time is spent in determining optimal local alignment between two sequences. This task is performed with the Smith-Waterman algorithm, which is a dynamic programming based method. With the advent of modern sequencing technologies and increasing size of both genome and protein databases, a need for faster Smith-Waterman implementations has emerged. Multiple SIMD strategies for the Smith-Waterman algorithm are available for CPUs. However, with the move of HPC facilities towards accelerator based architectures, a need for an efficient GPU accelerated strategy has emerged. Existing GPU based strategies have either been optimized for a specific type of characters (Nucleotides or Amino Acids) or for only a handful of application use-cases. In this paper, we present ADEPT, a new sequence alignment strategy for GPU architectures that is domain independent, supporting alignment of sequences from both genomes and proteins. Our proposed strategy uses GPU specific optimizations that do not rely on the nature of sequence. We demonstrate the feasibility of this strategy by implementing the Smith-Waterman algorithm and comparing it to similar CPU strategies as well as the fastest known GPU methods for each domain. ADEPT’s driver enables it to scale across multiple GPUs and allows easy integration into software pipelines which utilize large scale computational systems. We have shown that the ADEPT based Smith-Waterman algorithm demonstrates a peak performance of 360 GCUPS and 497 GCUPs for protein based and DNA based datasets respectively on a single GPU node (8 GPUs) of the Cori Supercomputer. Overall ADEPT shows 10x faster performance in a node-to-node comparison against a corresponding SIMD CPU implementation. ADEPT demonstrates a performance that is either comparable or better than existing GPU strategies. We demonstrated the efficacy of ADEPT in supporting existing bionformatics software pipelines by integrating ADEPT in MetaHipMer a high-performance denovo metagenome assembler and PASTIS a high-performance protein similarity graph construction pipeline. Our results show 10% and 30% boost of performance in MetaHipMer and PASTIS respectively.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2026

Deciphering the Solvation Structure of Aqueous ZnCl 2 Solutions from X-ray Absorption Spectra Using the Interpretable Graph Neural Network

Machine learning (ML) provides powerful pathways for predicting spectroscopic observables from atomic structures, but its broader impact depends on making model predictions interpretable in terms of physical and chemical principles. Here, we introduce a physics-guided graph neural network (GNN) model that predicts Zn K-edge X-ray spectroscopy (XAS) spectra of aqueous ZnCl 2 solutions. Training data are generated from ab initio XAS calculations on molecular dynamics snapshots obtained using a machine learning interatomic potential. The GNN reproduces experimental spectra across concentrations from dilute (<0.1 m) to highly concentrated (30 m, “water-in-salt”) regimes and scales efficiently to large, disordered liquid systems beyond the reach of conventional ab initio approaches. Gradient-based attribution analysis reveals that the model learns physically meaningful structure-spectrum relationships. Ligand-specific attributions reflect orbital hybridization patterns and the origin of the excitations derived from the density functional theory. Bond-length attributions recover spectral shifts consistent with multiple-scattering theory. Finally, this work bridges data-driven prediction with electronic-structure theory, establishing a general paradigm for interpretable ML that links atomic structure, electronic structure, and spectroscopic observables.

25 ENERGY STORAGE↗
DOE OSTI2020

Compare linear-system solver and preconditioner stacks with emphasis on GPU performance and propose phase-2 NGP solver development pathway

The goal of the ExaWind project is to enable predictive simulations of wind farms comprised of many megawatt-scale turbines situated in complex terrain. Predictive simulations will require computational fluid dynamics (CFD) simulations for which the mesh resolves the geometry of the turbines and captures the rotation and large deflections of blades. Whereas such simulations for a single turbine are arguably petascale class, multi-turbine wind farm simulations will require exascale-class resources. The primary physics codes in the ExaWind project are Nalu-Wind, which is an unstructured-grid solver for the acoustically incompressible Navier-Stokes equations, and OpenFAST, which is a whole-turbine simulation code. The Nalu-Wind model consists of the mass-continuity Poisson-type equation for pressure and a momentum equation for the velocity. For such modeling approaches, simulation times are dominated by linear-system setup and solution for the continuity and momentum systems. For the ExaWind challenge problem, the moving meshes greatly affect overall solver costs as reinitialization of matrices and recomputation of preconditioners is required at every time step. In this report we evaluated GPU-performance baselines for the linear solvers in the Trilinos and hypre solver stacks using two representative Nalu-Wind simulations: an atmospheric boundary layer precursor simulation on a structured mesh, and a fixed-wing simulation using unstructured overset meshes. Both strong-scaling and weak-scaling experiments were conducted on the OLCF supercomputer Summit and similar proxy clusters. We focused on the performance of multi-threaded Gauss-Seidel and two-stage Gauss-Seidel that are extensions of classical Gauss-Seidel; of one-reduce GMRES, a communication-reducing variant of the Krylov GMRES; and algebraic multigrid methods that incorporate the afore-mentioned methods. The team has established that AMG methods are capable of solving linear systems arising from the fixed-wing overset meshes on CPU, a critical intermediate result for ExaWind FY20 Q3 and Q4 milestones. For the fixed-wing strong-scaling study (model with 3M grid-points), the team identified that Nalu-Wind simulations with the new Trilinos and hypre solvers scale to modest GPU counts, maintaining above 70% efficiency up to 6 GPUs. However, there still remain significant bottlenecks to performance: matrix assembly (hypre), AMG setup (hypre and Trilinos) In the weak-scaling experiments (going from 0.4M to 211M gridpoints), it's shown that the solver apply phases are faster on GPUs, but that Nalu-Wind simulation times grow, primarily due to the multigrid-setup process. Finally, based on the report outcomes, we propose a linear solver path-forward for the remainder of the ExaWind project. Near term, the NREL team will continue their work on GPU-based linear-system assembly. They will also investigate how the use of alternatives to the NVIDIA UVM (unified virtual memory) paradigm affects performance. Longer term, the NREL team will evaluate algorithmic performance on other types of accelerators and merge their improvements back to the main hypre repository branch. Near term, the Trilinos team will address performance bottlenecks identified in this milestone, such as implementing a GPU-based segregated momentum solve and reusing matrix graphs across linear-system assembly phases. Longer term, the Trilinos team will do detailed analysis and optimization of multigrid setup.

17 WIND ENERGY↗