Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “iterative refinement”

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 19 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Combining Sparse Approximate Factorizations with Mixed-precision Iterative Refinement

The standard LU factorization-based solution process for linear systems can be enhanced in speed or accuracy by employing mixed-precision iterative refinement. Most recent work has focused on dense systems. We investigate the potential of mixed-precision iterative refinement to enhance methods for sparse systems based on approximate sparse factorizations. In doing so, we first develop a new error analysis for LU- and GMRES-based iterative refinement under a general model of LU factorization that accounts for the approximation methods typically used by modern sparse solvers, such as low-rank approximations or relaxed pivoting strategies. We then provide a detailed performance analysis of both the execution time and memory consumption of different algorithms, based on a selected set of iterative refinement variants and approximate sparse factorizations. Our performance study uses the multifrontal solver MUMPS, which can exploit block low-rank factorization and static pivoting. We evaluate the performance of the algorithms on large, sparse problems coming from a variety of real-life and industrial applications showing that mixed-precision iterative refinement combined with approximate sparse factorization can lead to considerable reductions of both the time and memory consumption.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Multistage mixed precision iterative refinement

Abstract Low precision arithmetic, in particular half precision (16‐bit) floating point arithmetic, is now available in commercial hardware. Using lower precision can offer significant savings in computation and communication costs with proportional savings in energy. Motivated by this, there has been a renewed interest in mixed precision iterative refinement schemes for solving linear systems , and new variants of GMRES‐based iterative refinement have been developed. Each particular variant with a given combination of precisions leads to different condition number‐based constraints for convergence of the backward and forward errors, and each has different performance costs. The constraints for convergence given in the literature are, as an artifact of the analyses, often overly strict in practice, and thus could lead a user to select a more expensive variant when a less expensive one would have sufficed. In this work, we develop a multistage mixed precision iterative refinement solver which aims to combine existing mixed precision approaches to balance performance and accuracy and improve usability. For a user‐specified initial combination of precisions, the algorithm begins with the least expensive approach and convergence is monitored via inexpensive computations with quantities produced during the iteration. If slow convergence or divergence is detected using particular stopping criteria, the algorithm switches to use a more expensive, but more reliable variant. A novel aspect of our approach is that, unlike existing implementations, our algorithm first attempts to use “stronger” GMRES‐based solvers for the solution update before resorting to increasing the precision(s). In some scenarios, this can avoid the need to refactorize the matrix in higher precision. We perform extensive numerical experiments on a variety of random dense problems and problems from real applications which confirm the benefits of the multistage approach.

Oktay, Eda↗
DOE OSTI2025

ARCS: Agentic Retrieval-Augmented Code Synthesis with Iterative Refinement

Agentic Retrieval-Augmented Code Synthesis with Iterative RefinementIn supercomputing, efficient and optimized code generation is essential to leverage high-performance systems effectively. We have developed Agentic Retrieval-Augmented Code Synthesis (ARCS), an advanced framework for accurate, robust, and efficient code generation, completion, and translation. ARCS integrates Retrieval-Augmented Generation (RAG) with Chain-of-Thought (CoT) reasoning to systematically break down and iteratively refine complex programming tasks. An agent-based RAG mechanism retrieves relevant code snippets, while real-time execution feedback drives the synthesis of candidate solutions. This process is formalized as a state-action search tree optimization, balancing code correctness with editing efficiency. Evaluations on the Geeks4Geeks and HumanEval benchmarks demonstrate that ARCS significantly outperforms traditional prompting methods in translation and generation quality. By enabling scalable and precise code synthesis, ARCS offers transformative potential for automating and optimizing code development in supercomputing applications, enhancing computational resource utilization

Bhattarai, Manish [Los Alamos National Labs]↗
DOE OSTI2020

Mixed-precision iterative refinement using tensor cores on GPUs to accelerate solution of linear systems

Double-precision floating-point arithmetic (FP64) has been the de facto standard for engineering and scientific simulations for several decades. Problem complexity and the sheer volume of data coming from various instruments and sensors motivate researchers to mix and match various approaches to optimize compute resources, including different levels of floating-point precision. In recent years, machine learning has motivated hardware support for half-precision floating-point arithmetic. A primary challenge in high-performance computing is to leverage reduced-precision and mixed-precision hardware. We show how the FP16/FP32 Tensor Cores on NVIDIA GPUs can be exploited to accelerate the solution of linear systems of equations Ax = b without sacrificing numerical stability. The techniques we employ include multiprecision LU factorization, the preconditioned generalized minimal residual algorithm (GMRES), and scaling and auto-adaptive rounding to avoid overflow. We also show how to efficiently handle systems with multiple right-hand sides. On the NVIDIA Quadro GV100 (Volta) GPU, we achieve a 4×-5× performance increase and 5× better energy efficiency versus the standard FP64 implementation while maintaining an FP64 level of numerical stability.

GMRES↗
DOE OSTI2022

A workflow for automatic generation and efficient refinement of individual pressure-dependent networks

Manual chemical kinetic model construction often requires modelists to at least implicitly guess all possible decay pathways and their relative fluxes for all important species. We present a workflow and associated tool for automatic generation and re finement of pressure-dependent networks that should enable modelists to efficiently and comprehensively identify decay pathways and estimate respective parameters for chemical species. This tool combines the capabilities of the Reaction Mechanism Generator (RMG) software to generate possible reaction paths, determine thermochemistry, approximate rate coefficients and estimate frequencies with the capabilities of the Arkane software to use quantum chemical parameters to compute thermochemistry and pressure dependent rates. A flux-based algorithm is used to decide which isomers to add to the network. Isomers added to the network are reacted to form other channels. When enough of the flux is accounted for by the isomers and bimolecular product channels, the generation process terminates to yield a comprehensive network. Network sensitivity analysis is applied to the network to identify the most important wells and barriers that can be refined using quantum chemistry calculations. Iterative refinement can then be used to achieve accuracy. The comprehensive network can be easily reduced using energy and flux-based algorithms to the most important channels and isomers in the network.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

RISE: robust iterative surface extension for sub-nanometer X-ray mirror fabrication

Precision optics have been widely required in many advanced technological applications. X-ray mirrors, as an example, serve as the key optical components at synchrotron radiation and free electron laser facilities. They are rectangular silicon or glass substrates where a rectangular Clear Aperture (CA) needs to be polished to sub-nanometer Root Mean Squared (RMS) to keep the imaging capability of the incoming X-ray wavefront at the diffraction limit. The convolutional polishing model requires a CA to be extended with extra data, from which the dwell time is calculated via deconvolution. However, since deconvolution is very sensitive to boundary errors and noise, the existing surface extension methods can hardly fulfill the sub-nanometer requirement. On one hand, the figure errors in a CA were improperly modeled during the extension, leading to continuity issues along the boundary. On the other hand, uncorrectable high-frequency errors and noise were also extended. In this study, we propose a novel Robust Iterative Surface Extension (RISE) method that resolves these problems with a data fitting strategy. RISE models the figure errors in a CA with orthogonal polynomials and ensures that only correctable errors are fit and extended. Combined with boundary conditions, an iterative refinement of dwell time is then proposed to compensate the errors brought by the extension and deconvolution, which drastically reduces the estimated figure error residuals in a CA while the increase of total dwell time is negligible. To our best knowledge, RISE is the first data fitting-based surface extension method and is the first to optimize dwell time based on iterative extension. An experimental verification of RISE is given by fabricating two elliptic cylinders (10 mm × 80 mm CAs) starting from a sphere with a radius of curvature around 173 m using ion beam figuring. The figure errors in the two CAs greatly improved from 204.96 nm RMS and 190.28 nm RMS to 0.62 nm RMS and 0.71 nm RMS, respectively, which proves that RISE is an effective method for sub-nanometer level X-ray mirror fabrication.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Efficient Mixed-Precision Matrix Factorization of the Inverse Overlap Matrix in Electronic Structure Calculations with AI-Hardware and GPUs

In recent years, a new kind of accelerated hardware has gained popularity in the artificial intelligence (AI) community which enables extremely high-performance tensor contractions in reduced precision for deep neural network calculations. In this article, we exploit Nvidia Tensor cores, a prototypical example of such AI-hardware, to develop a mixed precision approach for computing a dense matrix factorization of the inverse overlap matrix in electronic structure theory, S –1 . This factorization of S –1 , written as ZZT = S –1 , is used to transform the general matrix eigenvalue problem into a standard matrix eigenvalue problem. Here we present a mixed precision iterative refinement algorithm where Z is given recursively using matrix–matrix multiplications and can be computed with high performance on Tensor cores. To understand the performance and accuracy of Tensor cores, comparisons are made to GPU-only implementations in single and double precision. Additionally, we propose a nonparametric stopping criteria which is robust in the face of lower precision floating point operations. The algorithm is particularly useful when we have a good initial guess to Z, for example, from previous time steps in quantum-mechanical molecular dynamics simulations or from a previous iteration in a geometry optimization.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Multi–Modular Design of Stable Pore–Space–Partitioned Metal–Organic Frameworks for Gas Separation Applications

Pore space partition (PSP) is an effective materials design method for developing high-performance small-pore materials for storage and separation of gas molecules. The continued success of PSP depends on broad availability and judicious choice of pore-partition ligands and better understanding of each structural module on stability and sorption properties. Here, by using substructural bioisosteric strategy (sub-BIS), a dramatic expansion of pore-partitioned materials is targeted by using ditopic dipyridyl ligands with non-aromatic cores or extenders, as well as by expanding heterometallic clusters to uncommon nickel–vanadium and nickel–indium clusters rarely known before in porous materials. The dual-module iterative refinement of pore-partition ligands and trimers leads to remarkable enhancement of chemical stability and porosity. Here a family of 23 pore-partitioned materials synthesized from five pore-partition ligands and seven types of trimeric clusters is reported. New materials with such compositionally and structurally diverse framework modules reveal key factors that dictate stability, porosity, and gas separation properties. Among these, materials based on heterometallic vanadium–nickel trimeric clusters give rise to the highest long-term hydrolytic stability and remarkable uptake capacity for CO 2 , C 2 H 2 /C 2 H 4 /C 2 H 6 , and C 3 H 6 /C 3 H 8 hydrocarbon gases. Here, the breakthrough experiment shows the potential application of new materials for separating gas mixtures such as C 2 H 2 /CO 2 .

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2026

Mixed-precision numerics in scientific applications: survey and perspectives

The explosive demand for artificial intelligence (AI) workloads has led to a significant increase in silicon area dedicated to lower-precision computations on recent high-performance computing hardware designs. However, mixed-precision capabilities, which can achieve performance improvements of up to 8x compared to double-precision in extreme compute-intensive workloads, remain largely untapped in most scientific applications. A growing number of efforts have shown that mixed-precision algorithmic innovations can deliver superior performance without sacrificing accuracy. These developments should prompt computational scientists to seriously consider whether their scientific modeling and simulation applications could benefit from the acceleration offered by new hardware and mixed-precision algorithms. In this survey, we (1) review progress across diverse scientific domains—fluid dynamics, weather and climate, quantum chemistry, and computational genomics—that have begun adopting mixed-precision strategies; (2) examine state-of-the-art algorithmic techniques such as iterative refinement, splitting and emulation schemes, and adaptive precision solvers; (3) assess their implications for accuracy, performance, and resource utilization; and (4) survey the emerging software ecosystem that enables mixed-precision methods at scale. We conclude with perspectives and recommendations on cross-cutting opportunities, domain-specific challenges, and the role of co-design between application scientists, numerical analysts, and computer scientists. Collectively, this survey underscores that mixed-precision numerics can reshape computational science by aligning algorithms with the evolving landscape of hardware capabilities.

Graphics processing units↗
DOE OSTI2023

Building a DFT+U machine learning interatomic potential for uranium dioxide

Despite uranium dioxide (UO 2 ) being a widely used nuclear fuel, fuel performance models rely extensively on empirical correlations of material behavior, leveraging the historical operating experience of UO 2 . Mechanistic models that consider an atomistic understanding of the processes governing fuel performance (such as fission gas release and creep) will enable a better description of fuel behavior under non-prototypical conditions such as in new reactor concepts or for modified UO 2 fuel compositions. To this end, molecular dynamics simulation is a powerful tool for rapidly predicting physical properties of proposed fuel candidates. However, the reliability of these simulations depends largely on the accuracy of the atomic forces. Traditionally, these forces are computed using either a classical force field (FF) or density functional theory (DFT). While DFT is relatively accurate, the computational cost is burdensome, especially for f-electron elements, such as actinides. By contrast, classical FFs are computationally efficient but are less accurate. For these reasons, we report a new accurate machine learning interatomic potential (MLIP) for UO 2 that provides high-fidelity reproduction of DFT forces at a similar low cost to classical FFs. We employ an active learning approach that autonomously augments the DFT training data set to iteratively refine the MLIP. To further improve the quality of our predictions, we utilize transfer learning to retrain our MLIP to higher-accuracy DFT+U data. We validate our MLIPs by comparing predicted physical properties (e.g., thermal expansion and elastic properties) with those from existing classical FFs and DFT/DFT+U calculations, as well as with experimental data when available.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2026

Agentic traffic intelligence: Augmented human-in-the-loop scenario generation for microscopic traffic simulation

Traditional microscopic traffic simulation generation often relies on static datasets and manual design, limiting its ability to simulate complex conditions easily. This paper presents a novel framework, Agentic Traffic Intelligence, which combines human approval large language models (LLMs), the Real-Twin tool, and multi-agent systems to perform realistic microscopic traffic simulation scenario generation. The proposed framework incorporates human-in-the-loop (HIL) control, retrieval-augmented generation (RAG), and multi-agent control mechanisms. HIL mechanisms are used to guide multiple LLMs focused on attributes for microscopic simulation generation and to improve the interpretability and transparency of LLM execution for users. RAG enhances context extraction by dynamically integrating external knowledge sources for traffic scenario generation foundations. A multi-agent architecture with supervisory control coordinates the interaction of simulation components, including traffic simulators, control logic, and calibration tools. This enables the synthesis of simulation-ready scenarios that reflect dynamic demand profiles and behavior controls. Furthermore, the framework fuses multisource traffic data with unstructured context and supports iterative refinement through interactive user feedback. Validated through microscopic simulation using Simulation of Urban Mobility, the generated scenarios demonstrate high-fidelity network generation with inflow and turn movement and behavioral calibration, offering a robust and efficient tool for stress-testing and optimizing urban mobility systems.

Hierarchical multi-agent control↗
DOE OSTI2026

Agentic framework for programmatic crystal structure generation using a fine-tuned worker–supervisor large language model

Platinum group metals (PGMs) underpin many catalytic technologies but face severe supply constraints, motivating the search for alternative materials and computational methods to accelerate discovery. While atomistic simulation tools such as Pymatgen and ASE have streamlined structure manipulation, they require detailed inputs, limiting accessibility for experimentalists and slowing early-stage exploration. Here, in this study, we present an AI-driven agentic framework that orchestrates worker–supervisor large language models (LLMs). The worker translates natural-language prompts of varying abstraction into valid crystallographic structures using a compact LLM fine-tuned with low-rank adaptation on a curated text–code–CIF dataset, emphasizing energy-efficient training. Benchmarking against the baseline CodeGen-350M-mono model shows that fine-tuning reduces hallucination rates from 100% to as low as 5% and improves structural match accuracy to up to 82% for fully specified inputs. Accuracy declines with decreasing prompt detail but remains nontrivial even when only stoichiometry and space group are provided, underscoring the LLM’s capacity for crystallographic inference. The supervisor Claude LLM evaluates the outputs and triggers iterative refinement through the worker’s built-in structure manipulation capabilities (e.g., supercell scaling, strain, vacancy, and substitution operations). We further demonstrate use cases for technologically relevant catalysts, including IrO 2 , pyrochlore Pb 2 Ir 2 O 7 , Ni 2 FeO 4 , and Ni 3 Mo, where the framework generates physically consistent structures that can be refined via geometry optimization. This work introduces a low-energy, language-driven pathway for integrating human and machine intelligence in materials design, paving the way for AI-assisted synthesis planning and high-throughput screening of complex oxides.

AI agent↗
DOE OSTI2026

A score-based diffusion model approach for adaptive learning of stochastic partial differential equation solutions

In this paper, we propose a novel framework for adaptively learning the time-evolving solutions of stochastic partial differential equations (SPDEs) using score-based diffusion models within a recursive Bayesian inference setting. SPDEs play a central role in modeling complex physical systems under uncertainty, but their numerical solutions often suffer from model errors and reduced accuracy due to incomplete physical knowledge and environmental variability. To address these challenges, we encode the governing physics into the score function of a diffusion model using simulation data and incorporate observational information via a likelihood-based correction in a reverse-time stochastic differential equation. This enables adaptive learning through iterative refinement of the solution as new data becomes available. To improve computational efficiency in high-dimensional settings, we introduce the ensemble score filter, a training-free approximation of the score function designed for real-time inference. Numerical experiments on benchmark SPDEs demonstrate the accuracy and robustness of the proposed method under sparse and noisy observations.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Iterative methods in GPU-resident linear solvers for nonlinear constrained optimization

Linear solvers are major computational bottlenecks in a wide range of decision support and optimization computations. The challenges become even more pronounced on heterogeneous hardware, where traditional sparse numerical linear algebra methods are often inefficient. For example, methods for solving ill-conditioned linear systems have relied on conditional branching, which degrades performance on hardware accelerators such as graphical processing units (GPUs). To improve the efficiency of solving ill-conditioned systems, our computational strategy separates computations that are efficient on GPUs from those that need to run on traditional central processing units (CPUs). Our strategy maximizes the reuse of expensive CPU computations. Iterative methods, which thus far have not been broadly used for ill-conditioned linear systems, play an important role in our approach. In particular, we extend ideas from Arioli et al., (2007) to implement iterative refinement using inexact LU factors and flexible generalized minimal residual (FGMRES), with the aim of efficient performance on GPUs. In conclusion, we focus on solutions that are effective within broader application contexts, and discuss how early performance tests could be improved to be more predictive of the performance in a realistic environment.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Resolving the Coverage Dependence of Surface Reaction Kinetics with Machine Learning and Automated Quantum Chemistry Workflows

Microkinetic models for catalytic systems require estimation of many thermodynamic and kinetic parameters that can be calculated for isolated species and transition states using ab initio methods. However, the presence of nearby coadsorbates on the surface can dramatically alter these thermodynamic and kinetic parameters causing them to be dependent on species coverage fractions. As there are combinatorially many coadsorbed configurations on the surface, computing the coverage dependence of these parameters is far less straightforward. We present a framework for generating and applying machine learning models to predict coverage-dependent parameters for microkinetic models. Our toolkit enables automatic calculation and evaluation of coadsorbed configurations allowing us to sample 2,000 coadsorbed adsorbates and transition states (TSs) for a diverse set of 9 reactions on Cu(111), a challenging surface, with four possible coadsorbates. This dataset was then used to train subgraph isomorphic decision trees (SIDTs) to predict the stability and association energy of configurations. We were able to achieve mean absolute errors (MAEs) of 0.106 eV on adsorbates, 0.172 eV on TSs, and due to natural error cancellation in SIDTs for relative properties, 0.130 eV on reaction energies and 0.180 eV on activation barriers. In conclusion, we describe how to use these models to predict coverage-dependent corrections for adsorbates and TSs and demonstrate on H*, HO*, and O* comparing the generated SIDT model with an iteratively refined version.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗