Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “sparse”

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 1,405 records · Page 78
All sourcesNASA NTRSDOE OSTI
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 OSTI2023

Implementing and Benchmarking the Locally Competitive Algorithm on the Loihi 2 Neuromorphic Processor

Neuromorphic processors have garnered considerable interest in recent years for their potential in enabling energy-efficient and high-speed computing. The Locally Competative Algorithm (LCA) has been utilized for power efficient sparse coding on neuromophic processors, including the first Loihi processor \cite{appletospikes, loihi1}. With the Loihi 2 processor enabling custom neuron models and graded spike communication, more complex implementations of LCA are possible \cite{loihi2}. We present a new implementation of LCA designed for the Loihi 2 processor and perform an initial set of benchmarks comparing it to LCA on CPU and GPU devices. In these experiments LCA on Loihi 2 is faster and orders of magnitude more efficient, while maintaining similar reconstruction quality. We find this performance improvement increases as the LCA parameters are tuned towards greater representation sparsity. Our study highlights the potential of neuromorphic processors, particularly Loihi 2, in enabling intelligent,autonomous, real-time processing on small robots, satellite where there are strict SWaP (small, lightweighr, and low-power) requirement. By demonstrating the superior performance of LCA on Loihi 2 compared to conventional computing device, our study suggests that Loihi 2 could be a valuable tool in advancing these types of applications. Overall, our study highlights the potential of neuromorphic processors for efficient and accurate data processing on resource-constrained devices.

Parpart, Gavin G.↗
DOE OSTI2023

Optimizing Irregular Communication with Neighborhood Collectives and Locality-Aware Parallelism

Irregular communication often limits both the performance and scalability of parallel applications. Typically, applications individually implement irregular communication as point-to-point, and any optimizations are integrated directly into the application. As a result, these optimizations lack portability. It is difficult to optimize point-to-point messages within MPI, as the interface for single messages provides no information on the collection of all communication to be performed. However, the persistent neighbor collective API, released in the MPI 4 standard, provides an interface for portable optimizations of irregular communication within MPI libraries. This paper presents methods for implementing existing optimizations for irregular communication within neighborhood collectives, analyzes the impact of replacing point-to-point communication in existing codebases such as Hypre BoomerAMG with neighborhood collectives, and finally shows up to a 1.38x speedup on sparse matrix-vector multiplication communication within a BoomerAMG solve through the use of our optimized neighbor collectives. Here, the authors analyze three implementations of persistent neighborhood collectives for Alltoallv: an unoptimized wrapper of standard point-to-point communication, and two locality-aware aggregating methods. The second locality-aware implementation exposes an non-standard interface to perform additional optimization, and the authors present the additional 0.07x speedup from the extended interface. All optimizations are available in an open-source codebase, MPI Advance, which sits on top of MPI, allowing for optimizations to be added into existing codebases regardless of the system MPI install.

AMG↗
DOE OSTI2025

Scaling the memory wall using mixed-precision - HPG-MxP on an exascale-class machine

Mixed-precision algorithms have been proposed as a way for scientific computing to benefit from some of the gains seen for AI on recent high performance computing (HPC) platforms. A few applications dominated by dense matrix operations have seen substantial speedups by utilizing low precision formats such as FP16. However, a majority of scientific simulation applications are memory bandwidth limited. Beyond preliminary studies, the practical gain from using mixed-precision algorithms on a given high-performance computing (HPC) system is largely unclear. The High Performance GMRES Mixed Precision (HPG-MxP) benchmark has been proposed to measure the useful performance of a HPC system on sparse matrix-based mixed-precision applications. In this work, we present an implementation of the HPG-MxP benchmark for an exascale system and describe our algorithm enhancements. We show for the first time a speedup of 1.6x using a combination of double- and single-precision keeping the same residual level on modern GPU-based supercomputers.

Kashi, Aditya [ORNL] (ORCID:0000000325893792)↗
DOE OSTI2025

ORBIT-2: Scaling Exascale Vision Foundation Models for Weather and Climate Downscaling

Sparse observations and coarse-resolution climate models limit effective regional decision-making, underscoring the need for robust downscaling. However, existing AI methods struggle with generalization across variables and geographies and are constrained by the quadratic complexity of Vision Transformer (ViT) self-attention. We introduce ORBIT-2, a scalable foundation model for global, hyper-resolution climate downscaling. ORBIT-2 incorporates two key innovations: (1) Residual Slim ViT (Reslim), a lightweight architecture with residual learning and Bayesian regularization for efficient, robust prediction; and (2) TILES, a tile-wise sequence scaling algorithm that reduces self-attention complexity from quadratic to linear, enabling long-sequence processing and massive parallelism. ORBIT-2 scales to 10 billion parameters across 65,536 GPUs, achieving up to 4.1 ExaFLOPS sustained throughput and 74–98% strong scaling efficiency. It supports downscaling to 0.9 km global resolution and processes sequences up to 4.2 billion tokens. On 7 km resolution benchmarks, ORBIT-2 achieves high accuracy with R2 scores in range of 0.98–0.99 against observation data.

Wang, Xiao [ORNL] (ORCID:0000000165451943)↗
DOE OSTI2025

Intelligent Sampling of Extreme-Scale Turbulence Datasets for Accurate and Efficient Spatiotemporal Model Training

With the end of Moore’s law and Dennard scaling, efficient training increasingly requires rethinking data volume. Can we train better models with significantly less data via intelligent subsampling? To explore this, we develop SICKLE, a sparse intelligent curation framework for efficient learning, featuring a novel maximum entropy (MaxEnt) sampling approach, scalable training, and energy benchmarking. We compare MaxEnt with random and phase-space sampling on large direct numerical simulation (DNS) datasets of turbulence. Evaluating SICKLE at scale on Frontier, we show that subsampling as a preprocessing step can, in many cases, improve model accuracy and substantially lower energy consumption, with observed reductions of up to 38×.

Brewer, Wes [ORNL] (ORCID:0000000236393956)↗
DOE OSTI2025

Foundation Models for Zero-Shot Segmentation of Scientific Images without AI-Ready Data

Zero-shot and prompt-based models have excelled at visual reasoning tasks by leveraging large-scale natural image corpora, but they often fail on sparse and domain-specific scientific image data. We introduce Zenesis, a no-code interactive computer vision platform designed to reduce data readiness bottlenecks in scientific imaging workflows. Zenesis integrates lightweight multimodal adaptation for zero-shot inference on raw scientific data, human-in-the-loop refinement, and heuristic-based temporal enhancement. We validate our approach on Focused Ion Beam Scanning Electron Microscopy (FIB-SEM) datasets of catalyst-loaded membranes. Zenesis outperforms baselines, achieving an average accuracy of 0.947, Intersection over Union (IoU) of 0.858, and Dice score of 0.923 on amorphous catalyst samples; and 0.987 accuracy, 0.857 IoU, and 0.923 Dice on crystalline samples. These results represent a significant performance gain over conventional methods such as Otsu thresholding and standalone models like the Segment Anything Model (SAM). Zenesis enables effective image segmentation in domains where annotated datasets are limited, offering a scalable solution for scientific discovery.

Mukherjee, Shubhabrata↗
DOE OSTI2026

ChatMPI: LLM-Driven MPI Code Generation for HPC Workloads

The Message Passing Interface (MPI) standard plays a crucial role in enabling scientific applications for parallel computing and is an essential component in high-performance computing (HPC). However, implementing MPI code manually—especially applying a proper domain decomposition and communication pattern—is a challenging and error-prone task. We present ChatMPI, an AI assistant for MPI parallelization of sequential C codes. In our analysis, we focus on testing six essential HPC workloads, which are based on Basic Linear Algebra Subprograms levels 1, 2, and 3 as well as sparse, stencil, and iterative operations. We analyze the process of creating ChatMPI by using the ChatHPC library. This lightweight large language model (LLM)–based infrastructure enables HPC experts to efficiently create and supervise trustworthy AI capabilities for critical HPC software tasks. We study the data required for training (fine-tuning) ChatMPI to generate parallel codes that not only use MPI syntax correctly but also apply HPC techniques to reduce memory communication and maximize performance by using proper work decomposition. With a relatively small training dataset composed of a few dozen prompts and fewer than 15 minutes of fine-tuning on one node equipped with two NVIDIA H100 GPUs, ChatMPI elevates trustworthiness for MPI code generation of current LLMs (e.g., Code Llama, ChatGPT-4o and ChatGPT 5). Additionally, we evaluate the performance of the MPI codes generated by ChatMPI in comparison with the ones generated by ChatGPT-4o and ChatGPT-5. The codes generated by ChatMPI provide up to a 4 × boost in performance by using better problem decomposition, communication patterns, and HPC techniques (e.g., communication avoiding).

Valero Lara, Pedro [ORNL] (ORCID:0000000214794310)↗
DOE OSTI2026

CORE-BFS: Communication-Optimized REctangular-partitioned BFS Achieving 160.845 TeraTEPS on Frontier Supercomputer

Distributed Breadth-First Search (BFS) is fundamental to many large-scale graph applications, but its performance on parallel systems is often limited by high communication overhead. This paper presents CORE-BFS, an extremely scalable GPU-based BFS implementation that introduces a unique rectangular 2D partitioning-based design for Frontier supercomputer. To further improve performance, we propose four key optimizations: (1) Rectangular 2D-partition specific data formats that use two compressed row and one compressed column status array bitmaps combined with a Double Compressed Sparse Row (DCSR) format per partition, reducing memory footprint and inter-rank traffic; (2) Adaptive frontier & communication strategy that unifies top-down and bottom-up traversal on the rectangular layout, uses lazy synchronization in top-down levels, and switches variants based on frontier size to minimize communication overhead; (3) Frontier-split degree-aware update that maps frontier vertices to thread-centric, wavefront-centric, and block-centric kernels based on their degree to improve GPU utilization and memory coalescing; (4) Row-reduction pipeline that overlaps bottom-up adjacency list processing with row-wise bitmap reduction to hide inter-rank latency. Together, these techniques increase parallelism while reducing memory and communication overhead. On the Graph500 benchmark, CORE - BFS scales up to 9,248 Frontier nodes with scale-42 graphs and reaches 160.845 TTEPS, delivering a 5.42 × speedup over our previous Frontier implementation.

Yang, Haoshen [Rutgers University]↗
DOE OSTI2020

Imaging Cycle-Induced Damage of MnO 2 Microparticles

MnO 2 has been proposed as an electrode material in electrochemical energy storage devices. However, poor cycle life, especially in aqueous electrolytes, remains a detriment to commercialization. Prior studies have suggested a number of explanations for this capacity loss; however, experiments aimed at elucidating the details of the degradation process (es) are sparse. We describe here a microtube-membrane construct that allows for electrodeposition of monodisperse MnO 2 microparticles distributed across the membrane surface, and for subsequent electrochemical cycling of these MnO 2 particles. This allowed for a detailed analysis of the effect of cycling on the MnO 2 , by simply imaging the membrane surface before and after cycling. When an aqueous electrolyte was used, gross changes in particle shape, size and morphology were observed over the course of 500 cycles. Partial dissolution occurred as well. No such changes were observed when the MnO 2 particles were cycled (up to 500 times) in a propylene carbonate electrolyte solution.

36 MATERIALS SCIENCE↗
DOE OSTI2021

Interfacial Effects on Transport Coefficient Measurements in Li-ion Battery Electrolytes

Development of Li + -containing electrolytes with improved transport properties requires reliable, reproducible, and ideally low volume techniques to rigorously understand ion-transport with varying composition. Precisely measuring the complete set of transport coefficients in liquid electrolytes under battery-relevant operating conditions is difficult and the reliability of these methods are sparsely described in electrolyte transport literature. In this work, we apply a potentiostatic polarization-based transport characterization approach typically used for polymer electrolytes to liquid electrolyte systems in an attempt to fully measure all transport coefficients (conductivity, total salt diffusion coefficient, thermodynamic factor and transference number) for the model system of LiPF 6 in an ethylene carbonate—ethyl methyl carbonate (EC:EMC) mixture. Using systematic timescale and statistical analyses, we find that transport coefficients measured using potentiostatic polarization of Li-Li symmetric cells exhibit strong correlation to Li electrode interfacial resistance, indicating that such methods are probing both bulk and interfacial phenomena. This reveals a major roadblock in characterizing electrolyte systems where the interfacial resistance is significantly larger than ohmic electrolyte resistance. As a result, we find that methods that rely on potentiostatic Li metal stripping/plating do not readily result in reliable liquid electrolyte transport coefficients, unlike similar methods for solid polymer electrolytes, where interfacial resistances are typically smaller than electrolyte resistances at the elevated temperatures typically of interest for such electrolytes.

25 ENERGY STORAGE↗
DOE OSTI2021

Electrochemistry-Induced Restructuring of Tin-Doped Indium Oxide Nanocrystal Films of Relevance to CO 2 Reduction

The electrochemical reduction of CO 2 into fuels using renewable electricity presents an opportunity to utilize captured CO 2 . Electrocatalyst development has been a primary focus of research in this area. This is especially true at the nanoscale, where researchers have focused on understanding nanostructure-property relationships. However, electrocatalyst structure may evolve during operation. Indium- and tin-based oxides have been widely studied as electrocatalysts for CO 2 reduction to formate, but evolution of these catalysts during operation is not well-characterized. Here, we report the evolution of nanoscale structure of precise tin-doped indium oxide nanocrystals under CO 2 reduction conditions. We show that sparse monolayer nanocrystal films desorb from the electrode upon charging, but thicker nanocrystal films remain, likely due to an increased number of physical contacts. Upon applying a cathodic voltage of –1.0 V vs RHE or greater, the original 10-nm diameter nanocrystals are no longer visible, and instead form a larger microstructural network. Elemental analysis suggests the network is an oxygen-deficient indium-tin metal alloy. We hypothesize that this morphological evolution is the result of nanocrystal sintering due to oxide reduction. Furthermore, these data provide insights into the morphological evolution of tin-doped indium oxide nanocrystal electrocatalysts under reducing conditions and highlight the importance of post-electrochemical structural characterization of electrocatalysts.

25 ENERGY STORAGE↗
DOE OSTI2021

The updated ITPA global H-mode confinement database: description and analysis

The multi-machine ITPA Global H-mode Confinement Database has been upgraded with new data from JET with the ITER-like wall and ASDEX Upgrade with the full tungsten wall. This paper describes the new database and presents results of regression analysis to estimate the global energy confinement scaling in H-mode plasmas using a standard power law. Various subsets of the database are considered, focusing on type of wall and divertor materials, confinement regime (all H-modes, ELMy H or ELM-free) and ITER-like constraints. Apart from ordinary least squares, two other, robust regression techniques are applied, which take into account uncertainty on all variables. Regression on data from individual devices shows that, generally, the confinement dependence on density and the power degradation are weakest in the fully metallic devices. Using the multi-machine scalings, predictions are made of the confinement time in a standard ELMy H-mode scenario in ITER. The uncertainty on the scaling parameters is discussed with a view to practically useful error bars on the parameters and predictions. One of the derived scalings for ELMy H-modes on an ITER-like subset is studied in particular and compared to the IPB98(y,2) confinement scaling in engineering and dimensionless form. Transformation of this new scaling from engineering variables to dimensionless quantities is shown to result in large error bars on the dimensionless scaling. Regression analysis in the space of dimensionless variables is therefore proposed as an alternative, yielding acceptable estimates for the dimensionless scaling. The new scaling, which is dimensionally correct within the uncertainties, suggests that some dependencies of confinement in the multi- machine database can be reconciled with parameter scans in individual devices. This includes vanishingly small dependence of confinement on line-averaged density and normalized plasma pressure (β), as well as a noticeable, positive dependence on effective atomic mass and plasma triangularity. Extrapolation of this scaling to ITER yields a somewhat lower confinement time compared to the IPB98(y, 2) prediction, possibly related to the considerably weaker dependence on major radius in the new scaling (slightly above linear). Further studies are needed to compare more flexible regression models with the power law used here. In addition, data from more devices concerning possible ‘hidden variables’ could help to determine their influence on confinement, while adding data in sparsely populated areas of the parameter space may contribute to further disentangling some of the global confinement dependencies in tokamak plasmas.

Database↗
DOE OSTI2020

UQ Toolkit v 2.0

The Uncertainty Quantification (UQ) Toolkit is a software library for the characterizaton and propagation of uncertainties in computational models. For the characterization of uncertainties, Bayesian inference tools are provided to infer uncertain model parameters, as well as Bayesian compressive sensing methods for discovering sparse representations of high-dimensional input-output response surfaces, and also Karhunen-Loève expansions for representing stochastic processes. Uncertain parameters are treated as random variables and represented with Polynomial Chaos expansions (PCEs). The library implements several spectral basis function types (e.g. Hermite basis functions in terms of Gaussian random variables or Legendre basis functions in terms of uniform random variables) that can be used to represent random variables with PCEs. For propagation of uncertainty, tools are provided to propagate PCEs that describe the input uncertainty through the computational model using either intrusive methods (Galerkin projection of equations onto basis functions) or non-intrusive methods (perform deterministic operation at sampled values of the random values and project the obtained results onto basis functions).

Safta, Cosmin↗
DOE OSTI2021

Zero-Order Reaction Kinetics v3.0

Zero-RK is a software package that simulates chemically reacting systems using sparse, preconditioned, adaptive matrix methods to achieve orders-of-magnitude reduction in simulation time while maintaining accurate results.

Mcnenly, MatthewJ.↗
DOE OSTI2021

MuyGPyS

MuyGPs is a GP estimation method that affords fast hyperparameter optimization by way of performing leave-one-out cross-validation. MuyGPs achieves best-in-class speed and scalability by limiting inference to the information contained in k nearest neighborhoods for prediction locations for both hyperparameter optimization and tuning. This feature affords the optimization of hyperparameters by way of leave-one-out cross-validation, as opposed to the more expensive loglikelihood evaluations requires by similar sparse methods

Priest, BenjaminW.↗
DOE OSTI2021

pyDRESCALk

Modern data scientists are tasked to analyze ever-growing data sets with increasingly complex relationships. Tensor decompositions have come to play a central role in identifying underlying latent structures in higher-order data. The problem of fitting tensor models to different distributions is complicated by the combinations of size, dimensionality, and sparsity present in real world data. The situation demands efficient algorithms designed for shared-memory and distributed systems. This work will present new research that tackles these challenges on several different fronts, leveraging optimizations in numerical algorithms and sparse tensor representations in heterogeneous high performance computing environments.

Bhattarai, Manish↗
DOE OSTI2022

PIO Tools

This is a set of utilities for reading and manipulating PIO files written in python and C++. These files are intended to be building blocks for other scripts / programs that will use them to do great things. The PIO format itself is described in LA-UR-05-7425 at page 102 and embodied in code LA-CC-05-052. In addition to reading the raw PIO files, the bundled utilities will expand variable that use a compressed sparse row notation

Swaminarayan, Sriram↗