Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “compute in memory”

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 955 records · Page 53
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Computational Performance Bounds Prediction in Quantum Computing With Unstable Noise

Quantum computing has significantly advanced in recent years, boasting devices with hundreds of quantum bits (qubits), hinting at its potential quantum advantage over classical computing. Yet, noise in quantum devices poses significant barriers to realizing this supremacy. Understanding noise’s impact is crucial for reproducibility and application reuse; moreover, the next-generation quantum-centric supercomputing essentially requires efficient and accurate noise characterization to support system management (e.g., job scheduling), where ensuring correct functional performance (i.e., fidelity) of jobs on available quantum devices can even be higher-priority than traditional objectives. However, noise fluctuates over time, even on the same quantum device, which makes predicting the computational bounds for on-the-fly noise is vital. Noisy quantum simulation can offer insights but faces efficiency and scalability issues. Here, in this work, we propose a data-driven workflow, namely QuBound, to predict computational performance bounds. It decomposes historical performance traces to isolate noise sources and devises a novel encoder to embed circuit and noise information processed by a Long Short-Term Memory (LSTM) network. For evaluation, we compare QuBound with a state-of-the-art learning-based predictor, which only generates a single performance value instead of a bound. Experimental results show that the result of the existing approach falls outside of performance bounds, while all predictions from our QuBound with the assistance of performance decomposition better fit the bounds. Moreover, QuBound can efficiently produce practical bounds for various circuits with over 106 speedup over simulation; in addition, the range from QuBound is over 10× narrower than the state-of-the-art analytical approach.

Li, Jinyang [George Mason Univ., Fairfax, VA (Unit↗
DOE OSTI2023

Ab Initio Transport Calculations: From Normal to Superconducting Current

Applying the Bogoliubov-de Gennes equations with density-functional theory, it is possible to formulate first principles description of current-phase relationships in superconducting/normal (magnetic)/superconducting tri-layers. Such structures are the basis for the superconducting analog of Magnetoresistive random access memory devices (JMRAM). In a recent paper1 we presented results from the first attempt to formulate such a theory, applied to the Nb/Ni/Nb trilayers. In the present work we provide computational details, explaining how to construct key ingredient (scattering matrices SN ) in a framework of linear muffin-tin orbitals (LMTO).

CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSICS↗
DOE OSTI2021

APNN-TC: Accelerating Arbitrary Precision Neural Networks on Ampere GPU Tensor Cores

Over the years, accelerating neural networks with quantization has been widely studied. Unfortunately, prior efforts with diverse precisions (e.g., 1-bit weights and 2-bit activations) are usually restricted by limited precision support on GPUs (e.g., int1 and int4). To break such restrictions, we introduce the first Arbitrary Precision Neural Network framework (APNN-TC) to fully exploit quantization benefits on Ampere GPU Tensor Cores. Specifically, APNN-TC first incorporates a novel emulation algorithm to support arbitrary short bit-width computation with int1 compute primitives and XOR/AND Boolean operations. Second, APNN-TC integrates arbitrary precision layer designs to efficiently map our emulation algorithm to Tensor Cores with novel batching strategies and specialized memory organization. Third, APNN-TC embodies a novel arbitrary precision NN design to minimize memory access across layers and further improve performance. Extensive evaluations show that APNN-TC can achieve significant speedup over CUTLASS kernels and various NN models, such as ResNet and VGG.

Feng, Boyuan↗
DOE OSTI2023

CrossLink: General Overview [Slides]

Problem: Traditional mesh generation approaches are labor intensive and have limited robustness when applied to parametric design exploration and optimization of complex geometries. While automatic mesh generation approaches exist, they tend to generate tetrahedral or mixed-hybrid meshes which are generally unsuitable for physics applications with strong shock waves, thin boundary layers, and strong gradients. In addition, simulations sizes in the billions of cells are becoming more common with traditional mesh generation methods quickly reaching scalability limits. Solution: CrossLink offers a topology-based mesh generation approach with unstructured block-filling methods and a scalable mesh generation engine. In addition, CrossLink incorporates a python-based API for seamless workflow integration and robust repeatability of the geometry handling and mesh generation process. This makes it ideal for parametric design study and optimization of complex geometries. Finally, future versions of CrossLink will offer a parametric mesh capability that optimizes a high-order mesh and enables reconstruction of the final mesh in memory by the physics solver.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

CrossLink: Advancements in Scalable Unstructured Mesh Generation [Slides]

Traditional mesh generation approaches are labor intensive and have limited robustness when applied to parametric design exploration and optimization of complex geometries. While automatic mesh generation approaches exist, they tend to generate tetrahedral or mixed-hybrid meshes which are generally unsuitable for physics applications with strong shock waves, thin boundary layers, and strong gradients. In addition, simulations sizes in the billions of cells are becoming more common with traditional mesh generation methods quickly reaching scalability limits. CrossLink offers a topology-based mesh generation approach with unstructured block-filling methods and a scalable mesh generation engine. In addition, CrossLink incorporates a python based API for seamless workflow integration and robust repeatability of the geometry handling and mesh generation process. This makes it ideal for parametric design study and optimization of complex geometries. Finally, future versions of CrossLink will offer a parametric mesh capability that optimizes a high-order mesh and enables reconstruction of the final mesh in memory by the physics solver.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Application of Dimensionality Reduction in Machine Learning Modeling of CO2 Storage

In this study, we developed deep learning models that are capable of predicting spatio-temporal outputs of CO2 saturation, pressure, and brine production in a 3D saline storage reservoir over 30 years of continuous CO2 injection and a 50-year post-injection timeframe. To improve computational efficiency and maintain performance accuracy, the model framework involves ensembling multi-layer autoencoder networks that provide dimensionality reduction of geologic inputs with fully connected long short-term memory (LSTM) neural networks that generate time-series prediction. This study was presented as poster at the 2022 Carbon Management Project Review Meeting held in Pittsburgh, PA (August 15 – 19, 2022).

Bello, Kolawole↗
DOE OSTI2023

Method and apparatus for front end gather/scatter memory coalescing

A system for processing gather and scatter instructions can implement a front-end subsystem, a back-end subsystem, or both. The front-end subsystem includes a prediction unit configured to determine a predicted quantity of coalesced memory access operations required by an instruction. A decode unit converts the instruction into a plurality of access operations based on the predicted quantity, and transmits the plurality of access operations and an indication of the predicted quantity to an issue queue. The back-end subsystem includes a load-store unit that receives a plurality of access operations corresponding to an instruction, determines a subset of the plurality of access operations that can be coalesced, and forms a coalesced memory access operation from the subset. A queue stores multiple memory addresses for a given load-store entry to provide for execution of coalesced memory accesses.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Device-Centric Ransomware Detection using Machine Learning-Based Memory Forensics for Smart Inverters

Ransomware attacks are the fastest-growing form of cyberattacks worldwide. Recently, ransomware attacks have targeted industrial control systems (ICSs), including power grids. Lessons learned from recent incidents in ICSs show that ransomware groups can deliver ransomware into not only the organization’s control servers, but also the operational technology (OT) devices such as smart inverters and smart grid devices. This paper proposes a machine learning (ML)- based memory forensics method enabling the detection of ransomware binaries stored in the memory of a commercial smart inverter. Device firmware binary files are extracted from a Serial Peripheral Interface (SPI) flash memory, and samples of both benign and ransomware binaries are generated by a binary manipulation method and a real-world ransomware encryption, separately. A deep transfer learning (DTL) method is used to retrain a convolutional neural network (CNN)-based ransomware detection algorithm using the generated samples. The experimental result validates that the proposed ML-based memory forensics method can accurately detect ransomware files.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Machine-learning based model reduction for partial differential equations

We develop a novel synergistic approach between model reduction and machine learning. The specific goal of this project is to aid in the construction of reduced order models for basis functions that are custom-made to represent the solution of partial differential equations. Partial differential equations (PDEs) are one of the main mathematical tools for describing physical phenomena. However, due to either efficiency or necessity, for many real-world problems, we are interested in constructing reduced order models (ROMs) which focus only on the explicit computation of subsets of the active spatio-temporal scales in the problem, while treating the interaction with the rest of the scales approximately. The task of accurate representation of such interactions (usually called memory terms) constitutes a vast area of research known as model reduction. PI Stinis has significant expertise in the construction of ROMs for complex systems. In addition, in recent work with the project key participant Qadeer, they have utilized machine learning to acquire custom-made basis functions (CBFs) to expand the solutions of PDEs. In the proposed work, we will merge the two concepts by constructing ROMs for subsets of the CBFs needed to represent the solution of a PDE. Specifically, we will use the Mori-Zwanzig model reduction formalism to construct ROMs for subsets of CBFs for nonlinear PDEs of various complexity, as well as investigate the usage of CBFs in the spectral vanishing viscosity method for problems that can form shocks in finite time. The outcome of the research is aimed to be proof-of-concept about a novel synergistic approach between model reduction and machine learning, thus advancing the field of scientific machine learning. Such a capability will benefit the efficient modeling of physical systems appearing in various areas of interest to the DOE.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

$\mathrm{SageNet}$: Fast Neural Network Emulation of the Stiff-amplified Gravitational Waves from Inflation

Accurate modeling of the inflationary gravitational waves (GWs) requires time-consuming, iterative numerical integrations of differential equations to take into account their backreaction on the expansion history. To improve computational efficiency while preserving accuracy, we present the Stiff-amplified Gravitational-wave Emulator Network (SageNet), a deep learning framework designed to replace conventional numerical solvers (code available at https://github.com/YifangLuo/SageNet). SageNet employs a long short-term memory architecture to emulate the present-day energy density spectrum of the inflationary GWs with possible stiff amplification, Ω GW (f). Trained on a data set of 25,689 numerically generated solutions, SageNet allows accurate reconstructions of Ω GW (f) and generalizes well to a wide range of cosmological parameters; 90.9% of the test emulations with randomly distributed parameters exhibit errors of under 4%. In addition, SageNet demonstrates its ability to learn and reproduce the artificial, adaptive sampling patterns in numerical calculations, which implement denser sampling of frequencies around changes in spectral indices in Ω GW (f). The dual capability of learning both physical and artificial features of the numerical GW spectra establishes SageNet as a robust alternative to exact numerical methods. Finally, our benchmark tests show that SageNet reduces the computation time from tens of seconds to milliseconds, achieving a speedup of ∼10 4 times over standard CPU-based numerical solvers with the potential for further acceleration on GPU hardware. These capabilities make SageNet a powerful tool for accelerating Bayesian inference procedures for extended cosmological models. In a broad sense, the SageNet framework offers a fast, accurate, and generalizable solution to modeling cosmological observables whose theoretical predictions demand costly differential equation solvers.

Astronomy data modeling↗
DOE OSTI2021

Improvement of Numerical Methods in Pronghorn

During the fiscal year of 2021, the finite-volume method (FVM) was deployed in Pronghorn to provide improved efficiency, stability, and accuracy for coarse-mesh, thermal-hydraulics problems. While the main goals of fiscal year 2021 were met, several issues emerged from the early deployment of the finite-volume method in Pronghorn. These issues were: The compressible and incompressible formulations that were implemented are inadequate for many nuclear reactor flow problems; Omission of terms accounting for the porosity and Darcy-Forchheimer body force discontinuities in the Rhie-Chow interpolation lead to oscillations in pressure and velocity at these discontinuities; The FVM lacks a correction for non-orthogonal grids for computing accurate pressure gradient leading to loss of accuracy in regions with skewed elements; The FVM currently uses a monolithic solver. Monolithic solvers have issues dealing with the saddle-point nature of the discretized fluid equations. The result are bad convergence if direct factorization is not used and large memory consumption when direct factorization is used. These four issues are addressed in this report. In particular, we report the completion of the following task: Implementation of a weakly compressible formulation in the MOOSE Navier-Stokes module; Implementation of Moukalled’s method for including body forces in the RCI. Additionally, we identified the need to smooth the porosity using Moukalled’s face-cell smoothing operator; Implementation of a non-orthogonal correction for the Green-Gauss gradient computation; Preliminary implementation of a SIMPLE segregated solver.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Progress and Challenges in Self-Healing Cementitious Materials

Concrete is the most used man-made material today consisting of a mixture of aggregates (sand, gravel, crashed rock) and paste. The paste component, made up of cement and water, cracks with time. Despite this shortcoming, research towards next-generation cement and concrete has only taken a boost in the past 10-15 years. In particular, self-healing cementitious materials is a research area that has attracted a great deal of attention. A number of reported novel formulations have demonstrated an increase in mechanical and chemical stability with respect to conventional Portland cement, through the addition of inorganic, organic, and even biological additives. This review reports on the latest advances in cement research related to the synthesis of self-healing cements applied in concrete structures. These include geopolymers, engineered cementitious materials (ECC), bacterial cement composites, microencapsulated self-healing materials, shape memory alloys (SMA), and polymer-cement composites. This work describes the performance of each cementitious material and the mechanism responsible for self-healing, including a section on the latest computational modeling studies towards this end. A detailed understanding of these various self-healing cement technologies, including their strengths and weaknesses, is critical at the time of determining the type of structures that these materials could be used for.

Fernandez, Carlos A.↗
DOE OSTI2021

Asynchronous distributed-memory task-parallel algorithm for compressible flows on unstructured 3D Eulerian grids

Here, we discuss the implementation of a finite element method, used to numerically solve the Euler equations of compressible flows, using an asynchronous runtime system (RTS). The algorithm is implemented for distributed-memory machines, using stationary unstructured 3D meshes, combining data-, and task-parallelism on top of the Charm++ RTS. Charm++’s execution model is asynchronous by default, allowing arbitrary overlap of computation and communication. Task-parallelism allows scheduling parts of an algorithm independently of, or dependent on, each other. Built-in automatic load balancing enables continuous redistribution of computational load by migration of work units based on real-time CPU load measurement. The RTS also features automatic checkpointing, fault tolerance, resilience against hardware failure, and supports power-, and energy-aware computation. We demonstrate scalability up to 25 x 10 9 cells at $\mathscr{O}$10 4 compute cores and the benefits of automatic load balancing for irregular workloads. The full source code with documentation is available at https://quinoacomputing.org.

42 ENGINEERING↗
DOE OSTI2025

Convergence study of wakefield simulations with GdfidL and ECHO3D

The interaction of charged particle beams with vacuum chamber components gives rise to electromagnetic wakefields, whose frequency-domain representation is known as beam coupling impedance. Geometric impedance arising from discontinuities and transitions in the vacuum chamber is the focus of this study. Minimizing this impedance is essential to mitigate adverse collective effects in modern storage rings operating with high-intensity particle beams. Accurate and reliable impedance simulations is a key factor of the vacuum chamber design. This paper presents the results of a convergence study of two widely used electromagnetic solvers, GdfidL and ECHO3D, applied to key vacuum-chamber components of the National Synchrotron Light Source II (NSLS-II) storage ring. Detailed comparisons are performed for several geometries, including flange absorbers, RF bellows, button-type beam position monitors, and an in-vacuum undulator (IVU). The results show notable differences in convergence and computational efficiency between the two codes. While GdfidL provides highly resolved results and serves as a common benchmark tool, ECHO3D yields consistent results with coarser meshes, significantly reducing simulation time and memory demands. Simulations with a full-geometry IVU model demonstrate that simplified taper-transition models can miss important impedance contributions. In conclusion, these findings provide practical guidelines for efficient and accurate impedance modeling to optimize design of vacuum chamber components for accelerators.

36 MATERIALS SCIENCE↗
DOE OSTI2023

TomocuPy – efficient GPU-based tomographic reconstruction with asynchronous data processing

Fast 3D data analysis and steering of a tomographic experiment by changing environmental conditions or acquisition parameters require fast, close to real-time, 3D reconstruction of large data volumes. Here a performance-optimized TomocuPy package is presented as a GPU alternative to the commonly used central processing unit (CPU) based TomoPy package for tomographic reconstruction. TomocuPy utilizes modern hardware capabilities to organize a 3D asynchronous reconstruction involving parallel read/write operations with storage drives, CPU–GPU data transfers, and GPU computations. In the asynchronous reconstruction, all the operations are timely overlapped to almost fully hide all data management time. Since most cameras work with less than 16-bit digital output, the memory usage and processing speed are furthermore optimized by using 16-bit floating-point arithmetic. As a result, 3D reconstruction with TomocuPy became 20–30 times faster than its multi-threaded CPU equivalent. Full reconstruction (including read/write operations and methods initialization) of a 2048 3 tomographic volume takes less than 7 s on a single Nvidia Tesla A100 and PCIe 4.0 NVMe SSD, and scales almost linearly increasing the data size. To simplify operation at synchrotron beamlines, TomocuPy provides an easy-to-use command-line interface. Efficacy of the package was demonstrated during a tomographic experiment on gas-hydrate formation in porous samples, where a steering option was implemented as a lens-changing mechanism for zooming to regions of interest.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2025

Distributed Cross-Channel Hierarchical Aggregation for Foundation Models

Vision-based scientific foundation models hold significant promise for advancing scientific discovery and innovation. This potential stems from their ability to aggregate images from diverse sources—such as varying physical groundings or data acquisition systems—and to learn spatio-temporal correlations using transformer architectures. However, tokenizing and aggregating images can be compute-intensive, a challenge not fully addressed by current distributed methods. In this work, we introduce the Distributed Cross-Channel Hierarchical Aggregation (D-CHAG) approach designed for datasets with a large number of channels across image modalities. Our method is compatible with any model-parallel strategy and any type of vision transformer architecture, significantly improving computational efficiency. We evaluated D-CHAG on hyperspectral imaging and weather forecasting tasks. When integrated with tensor parallelism and model sharding, our approach achieved up to a 75% reduction in memory usage and more than doubled sustained throughput on up to 1,024 AMD GPUs on the Frontier Supercomputer.

Tsaris, Aristeidis (aris) [ORNL] (ORCID:0000000277↗
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 OSTI2026

Arbitrary Polynomial Separations in Trainable Quantum Machine Learning

Recent theoretical results in quantum machine learning have demonstrated a general trade-off between the expressive power of quantum neural networks (QNNs) and their trainability; as a corollary of these results, practical exponential separations in expressive power over classical machine learning models are believed to be infeasible as such QNNs take a time to train that is exponential in the model size. We here circumvent these negative results by constructing a hierarchy of efficiently trainable QNNs that exhibit unconditionally provable, polynomial memory separations of arbitrary constant degree over classical neural networks—including state-of-the-art models, such as Transformers—in performing a classical sequence modeling task. This construction is also computationally efficient, as each unit cell of the introduced class of QNNs only has constant gate complexity. We show that contextuality—informally, a quantitative notion of semantic ambiguity—is the source of the expressivity separation, suggesting that other learning tasks with this property may be a natural setting for the use of quantum learning algorithms.

Anschuetz, Eric R. [California Institute of Techno↗