Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “parallel machines”

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 1,045 records · Page 58
All sourcesNASA NTRSDOE OSTI
NASA NTRSDate not supplied

SPHINX: An SEP Model Validation Infrastructure developed through Community Challenges and the SEP Scoreboards

Solar Energetic Particle (SEP) events are interesting from a scientific perspective as they are the product of a broad set of physical processes from the corona out through the extent of the heliosphere, and provide insight into processes of particle acceleration and transport that are widely applicable in astrophysics. From the operations perspective, SEP events pose a radiation hazard for aviation, electronics in space, and human space exploration, in particular for missions outside of the Earth’s protective magnetosphere including to the Moon and Mars (Whitman et al 2022). For these reasons, SEP modelers have developed a rich and diverse set of models with a wide variety of aims. Some models probe the basic physics at the heart of particle acceleration and transport. Others produce fast statistical forecasts or employ disruptive new techniques like Machine Learning with the goal to assist end users in making operational decisions. To enable a consistent and quantitative understanding of SEP model performance, a generalized, automated validation infrastructure, called SPHINX, is being developed at NASA SRAG in close collaboration with NASA CCMC, NASA M2M, NOAA SWPC, and BIRA-IASB. This infrastructure has been built up through a multi-year community challenge. Starting in 2018 at the SHINE workshop, an effort was launched through SHINE, ISWAT, and ESWW to encourage quantitative, comprehensive, and consistent validation of SEP models. This effort has defined a set of challenge SEP events with the aim of generating quantitative comparisons between forecasts and observations and a set of challenge “non-events” to assess false alarms. In 2023, these challenge lists have been extended to statistically significant numbers with a prescribed set of rules for producing forecasts and supported through the dedicated SEPVAL working meetings. The participation of the research community has allowed the infrastructure to validate all the types of outputs being produced by SEP models. In parallel, the SPHINX code is being applied to real time forecasts submitted to the SEP Scoreboards, ensuring that the validation infrastructure can interpret forecasts produced in an operational scenario and provide metrics meaningful for operations. Upon completion, SPHINX and its interactive user interface, SPHINX-Web, will be made available for public use.

space weather↗
NASA NTRS1993

A summary of image segmentation techniques

Machine vision systems are often considered to be composed of two subsystems: low-level vision and high-level vision. Low level vision consists primarily of image processing operations performed on the input image to produce another image with more favorable characteristics. These operations may yield images with reduced noise or cause certain features of the image to be emphasized (such as edges). High-level vision includes object recognition and, at the highest level, scene interpretation. The bridge between these two subsystems is the segmentation system. Through segmentation, the enhanced input image is mapped into a description involving regions with common features which can be used by the higher level vision tasks. There is no theory on image segmentation. Instead, image segmentation techniques are basically ad hoc and differ mostly in the way they emphasize one or more of the desired properties of an ideal segmenter and in the way they balance and compromise one desired property against another. These techniques can be categorized in a number of different groups including local vs. global, parallel vs. sequential, contextual vs. noncontextual, interactive vs. automatic. In this paper, we categorize the schemes into three main groups: pixel-based, edge-based, and region-based. Pixel-based segmentation schemes classify pixels based solely on their gray levels. Edge-based schemes first detect local discontinuities (edges) and then use that information to separate the image into regions. Finally, region-based schemes start with a seed pixel (or group of pixels) and then grow or split the seed until the original image is composed of only homogeneous regions. Because there are a number of survey papers available, we will not discuss all segmentation schemes. Rather than a survey, we take the approach of a detailed overview. We focus only on the more common approaches in order to give the reader a flavor for the variety of techniques available yet present enough details to facilitate implementation and experimentation.

Spirkovska, Lilly↗
DOE OSTI2021

Artificial Intelligence for Multiphysics Nuclear Design Optimization with Additive Manufacturing

The geometric flexibility of additively manufactured metals and ceramics generates a very large and open design space that requires advanced modeling and simulation tools for physics simulations and the rigorous definition of design problems. This effort deploys artificial intelligence (AI) and machine learning (ML) algorithms to understand the design space, evaluate potential designs, and more efficiently generate optimized results. The Transformational Challenge Reactor (TCR) program is leveraging advances in several scientific areas—including materials, manufacturing, sensors and control systems, data analytics, and high-fidelity modeling and simulation—to accelerate the design, manufacturing, qualification, and deployment of advanced nuclear energy systems. Through a manufacturing-informed design approach, the TCR program seeks to integrate digital data for rapid nuclear innovation; accelerate the adoption of advances in manufacturing, materials, and computational sciences for nuclear applications; and dramatically reduce deployment costs and timelines for new nuclear reactor technologies. This report documents efforts under the TCR program to leverage advanced modeling and simulation techniques driven by AI/ML algorithms on high-performance computing (HPC) systems to yield more optimized TCR core designs. A multiphysics ML surrogate model was developed to run on the HPC architectures. The surrogate model is trained on high-fidelity simulation data of coupled neutronics and thermofluidics and is used to quickly evaluate thousands of candidate core designs in parallel, which drives the evolution of the cooling channel shapes to minimize temperature peaking and material stress. Outcomes from these activities provide design information and feedback into the core design efforts.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2022

Throughput Analytics of Cloud Networks

A network of virtual machines at cloud server sites connected over virtual IO connections is a flexible, easily deployable, and cost-effective alternative to a physical network infrastructure with dedicated servers connected over leased fiber lines. We study the throughput performance of such a cloud network by collecting measurements over Google Cloud infrastructure spanning multiple continents. To study its ideal performance and impact of packet losses, we utilize its emulation using dedicated servers and connection hardware emulation devices. We compare the measurements over the cloud network to those over its emulation on a testbed. We examine the throughput profiles of both networks as a function of the round trip time and their utilization-concavity coefficients, estimated using measurements for common TCP versions. The throughput profile's concave-convex shape and its coefficient are critical indicators of the network performance, qualitatively and quantitatively, respectively. The results indicate their overall agreement between the production cloud network and its emulation using dedicated connections, and a near optimal throughput performance of the former except for a few under-performing connections. Also, the number of parallel flows is found to be a dominant factor in optimizing the throughput across various conditions and TCP versions.

Phanekham, Derek↗
DOE OSTI2024

Evaluation of DED and LPBF Fe-based Alloys Process Application Envelopes based on Performance, Process Economics, Supply Chain Risks, and Reactor-specific Targeted Components

The U.S. Department of Energy (DOE), Office of Nuclear Energy (NE), Advanced Materials and Manufacturing Technologies (AMMT) program aims to develop extreme-environment materials solutions for use in the deployment of advanced nuclear reactors and the sustainment of the current fleet. To achieve this objective, a combination of experiment, a computational tool, and machine learning (ML) for the design of materials is adopted for the maturation of materials for nuclear technology. Through advanced manufacturing techniques such as laser powder bed fusion (LPBF) and laser powder direct energy deposition (LP-DED), components with complex geometries can be fabricated with reduced time and effort. Such advanced manufacturing methods can also provide the opportunity to improve materials performance through optimized microstructures and mechanical properties. However, existing engineering alloys are not always well suited for fabrication with additive manufacturing (AM), as their compositions have been tuned to optimize fabrication via conventional methods. Thus, similar alloys with modified compositions that are better suited for AM can be studied for improved performance. Over the past three years, the AMMT teams from Argonne National Laboratory (ANL) and Pacific Northwest National Laboratory (PNNL) studied various known Fe-based alloys by evaluating their initial printability using LPBF, and an AMMT-developed down-selection and decision matrix reduced the number of alloys to be studied from six to three in fiscal year (FY) 2024. Additionally, in FY 2024, for parallel evaluation, these three alloys were studied using LPDED. While LPBF is better for small- to medium-sized components with high detail and internal features, LP-DED combines a material feed system to place the powder onto the exact spot where the laser will melt the material. This AM method can be easily scaled to extremely large components and provides high build rate speeds compared to those of conventional LPBF systems. Additionally, DED is a better choice for complex geometries and compositional gradients.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
NASA NTRS2017

Engine Icing Data - An Analytics Approach

Engine icing researchers at the NASA Glenn Research Center use the Escort data acquisition system in the Propulsion Systems Laboratory (PSL) to generate and collect a tremendous amount of data every day. Currently these researchers spend countless hours processing and formatting their data, selecting important variables, and plotting relationships between variables, all by hand, generally analyzing data in a spreadsheet-style program (such as Microsoft Excel). Though spreadsheet-style analysis is familiar and intuitive to many, processing data in spreadsheets is often unreproducible and small mistakes are easily overlooked. Spreadsheet-style analysis is also time inefficient. The same formatting, processing, and plotting procedure has to be repeated for every dataset, which leads to researchers performing the same tedious data munging process over and over instead of making discoveries within their data. This paper documents a data analysis tool written in Python hosted in a Jupyter notebook that vastly simplifies the analysis process. From the file path of any folder containing time series datasets, this tool batch loads every dataset in the folder, processes the datasets in parallel, and ingests them into a widget where users can search for and interactively plot subsets of columns in a number of ways with a click of a button, easily and intuitively comparing their data and discovering interesting dynamics. Furthermore, comparing variables across data sets and integrating video data (while extremely difficult with spreadsheet-style programs) is quite simplified in this tool. This tool has also gathered interest outside the engine icing branch, and will be used by researchers across NASA Glenn Research Center. This project exemplifies the enormous benefit of automating data processing, analysis, and visualization, and will help researchers move from raw data to insight in a much smaller time frame.

Engine Icing↗
DOE OSTI2023

Measurements and modeling of type-I and type-II ELMs heat flux to the DIII-D divertor

Type-I and type-II edge-localized-modes (ELMs) heat flux profiles measured at the DIII-D divertor feature a peak in the vicinity of the strike-point and a plateau in the scrape-off-layer (SOL), which extends to the first wall. The plateau is present in attached and detached divertors and it is found to originate with plasma bursts upstream in the SOL. The integrated ELM heat flux is distributed at ~65% in the peak and ~35% in this plateau. The parallel loss model, currently used at ITER to predict power loads to the walls, is benchmarked using these results in the primary and secondary divertors with unprecedented constraints using experimental input data for ELM size, radial velocity, energy, electron temperature and density, heat flux footprints and number of filaments. The model can reproduce the experimental near-SOL peak within ~20%, but cannot match the SOL plateau. Employing a two-component approach for the ELM radial velocity, as guided by intermittent data, the full radial heat flux profile can be well matched. The ELM-averaged radial velocity at the separatrix, which explains profile widening, increases from ~0.2 km s –1 in attached to ~0.8 km s –1 in detached scenarios, as the ELM filaments' path becomes electrically disconnected from the sheath at the target. The results presented here indicate filaments fragmentation as a possible mechanism for ELM transport to the far-SOL and provide evidence on the beneficial role of detachment to mitigate ELM flux in the divertor far-SOL. However, these findings imply that wall regions far from the strike points in future machines should be designed to withstand significant heat flux, even for small-ELM regimes.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2021

A High Performance Sparse Tensor Algebra Compiler in MLIR

Sparse tensor algebra is widely used in many applications, including scientific computing, machine learning, and data analytics. The performance of sparse tensor algebra kernels strongly depends on the intrinsic characteristics of the input tensors, hence many storage formats are designed for tensors to achieve optimal performance for particular applications/architectures, which makes it challenging to implement and optimize every tensor operation of interest on a given architecture. We propose a tensor algebra domain-specific language (DSL) and compiler framework to automatically generate kernels for mixed sparse-dense tensor algebra operations. The proposed DSL provides high-level programming abstractions that resemble the familiar Einstein notation to represent tensor algebra operations. The compiler introduces a new Sparse Tensor Algebra dialect built on top of LLVM's extensible MLIR compiler infrastructure for efficient code generation while covering a wide range of tensor storage formats. Our compiler also leverages input-dependent code optimization to enhance data locality for better performance. Our results show that the performance of automatically generated kernels outperforms the state-of-the-art sparse tensor algebra compiler, with up to 20.92x, 6.39x, and 13.9x performance improvement over state-of-the-art tensor algebra compilers, for parallel SpMV, SpMM, and TTM, respectively.

Tian, Ruiqin↗
DOE OSTI2021

Characterizing Machine Learning I/O Workloads on Leadership Scale HPC Systems

High performance computing (HPC) is no longer solely limited to traditional workloads such as simulation and modeling. With the increase in the popularity of machine learning (ML) and deep learning (DL) technologies, we are observing that an increasing number of HPC users are incorporating ML methods into their workflow and scientific discovery processes, across a wide spectrum of science domains such as biology, earth science, and physics. This gives rise to a diverse set of I/O patterns than the traditional checkpoint/restart-based HPC I/O behavior. The details of the I/O characteristics of such ML I/O workloads have not been studied extensively for large-scale leadership HPC systems. This paper aims to fill that gap by providing an in-depth analysis to gain an understanding of the I/O behavior of ML I/O workloads using darshan - an I/O characterization tool designed for lightweight tracing and profiling. We study the darshan logs of more than 23, 000 HPC ML I/O jobs over a time period of one year running on Summit - the second-fastest supercomputer in the world. This paper provides a systematic I/O characterization of ML I/O jobs running on a leadership scale supercomputer to understand how the I/O behavior differs across science domains and the scale of workloads, and analyze the usage of parallel file system and burst buffer by ML I/O workloads.

Paul, Arnab↗
NASA NTRS2010

Biasable, Balanced, Fundamental Submillimeter Monolithic Membrane Mixer

This device is a biasable, submillimeter-wave, balanced mixer fabricated using JPL s monolithic membrane process a simplified version of planar membrane technology. The primary target application is instrumentation used for analysis of atmospheric constituents, pressure, temperature, winds, and other physical and chemical properties of the atmospheres of planets and comets. Other applications include high-sensitivity gas detection and analysis. This innovation uses a balanced configuration of two diodes allowing the radio frequency (RF) signal and local oscillator (LO) inputs to be separated. This removes the need for external diplexers that are inherently narrowband, bulky, and require mechanical tuning to change frequency. Additionally, this mixer uses DC bias-ability to improve its performance and versatility. In order to solve problems relating to circuit size, the GaAs membrane process was created. As much of the circuitry as possible is fabricated on-chip, making the circuit monolithic. The remainder of the circuitry is precision-machined into a waveguide block that holds the GaAs circuit. The most critical alignments are performed using micron-scale semiconductor technology, enabling wide bandwidth and high operating frequencies. The balanced mixer gets superior performance with less than 2 mW of LO power. This can be provided by a simple two-stage multiplier chain following an amplifier at around 90 GHz. Further, the diodes are arranged so that they can be biased. Biasing pushes the diodes closer to their switching voltage, so that less LO power is required to switch the diodes on and off. In the photo, the diodes are at the right end of the circuit. The LO comes from the waveguide at the right into a reduced-height section containing the diodes. Because the diodes are in series to the LO signal, they are both turned on and off simultaneously once per LO cycle. Conversely, the RF signal is picked up from the RF waveguide by the probe at the left, and flows rightward to the diodes. Because the RF is in a quasi- TEM (suspended, microstrip-like) mode, it impinges on the diodes in an anti-parallel mode that does not couple to the waveguide mode. This isolates the LO and RF signals. This operation is similar to a cross-bar mixer used at low frequencies, except the RF signal enters through the back-short end of the waveguide rather than through the side. The RF probe also conveys the down-converted intermediate frequency (IF) signal out to an off-chip circuit board through a simple LC low-pass filter to the left as indicated. The bias is brought to the diodes through a bypass capacitor at the top.

Siegel, Peter↗
NASA NTRS2018

Dynamic Power Convertor Development for Radioisotope Power Systems at NASA GRC

The Thermal Energy Conversion Branch at NASA Glenn Research Center (GRC) is supporting the development of high-efficiency power convertors in support of Radioisotope Power System (RPS) development. Significant progress was made towards such a system that utilized Stirling conversion machines during the 2001 to 2015 timeframe. Flight development of the Advanced Stirling Radioisotope Generator (ASRG) was cancelled in 2013 by the Department of Energy (DOE) and NASA Headquarters primarily due to budget constraints, and the Advanced Stirling Convertor (ASC) technology contract was subsequently concluded in 2015. A new chapter of technology development has recently been initiated by the NASA RPS Program. This new effort is considering all dynamic power convertor options, such as Stirling and Brayton cycles. Four convertor development contracts supporting this effort were awarded in 2017. The awarded contracts include two free-piston Stirling, one thermoacoustic Stirling, and one turbo-Brayton designs. The technology development contracts each consist of up to three phases: Design, Fabricate, and Test. As of May 2018, all contracts have completed the Design Phase, and each underwent a design review with an independent review board. Two of the contracts have been awarded the Phase 2 option for fabrication. Convertors manifesting from these development efforts will then undergo independent validation and verification at NASA facilities, which will consist of convertor performance and RPS viability demonstrations. Example tests include launch vibration simulation, performance mapping over the environmental temperature range, and static acceleration exposure. In parallel with this renewed development effort, NASA GRC is still demonstrating free-piston Stirling convertor technology using assets from previous projects. The Stirling Research Laboratory (SRL) is still operating 10 convertors from previous development projects. Four of which are flexure-bearing based, and the other six are gas-bearing based. One of the flexure-bearing convertors has accumulated over 110,000 hours of operation, and holds the current maintenance-free heat-engine run-time record. Another flexure-bearing convertor was recently manually shutdown after 105,620 hours of operation, then disassembled and inspected. This inspection produced a wealth of information about the effects of this amount of runtime on the technology's components. One of the engineering unit flexure-bearing convertors recently underwent launch simulation vibration test, a static acceleration exposure up to 20 g, and was then placed on extended operation. Amongst the gas-bearing convertors, the longest running unit has accumulated over 70,000 hours of operation. Four high-fidelity gas-bearing convertors from the ASRG project are still operating continuously, for which the longest runtime has reached 26,000 hours.

Oriti, Salvatore↗
NASA NTRS2005

A unifying framework for systems modeling, control systems design, and system operation

Current engineering practice in the analysis and design of large-scale multi-disciplinary control systems is typified by some form of decomposition- whether functional or physical or discipline-based-that enables multiple teams to work in parallel and in relative isolation. Too often, the resulting system after integration is an awkward marriage of different control and data mechanisms with poor end-to-end accountability. System of systems engineering, which faces this problem on a large scale, cries out for a unifying framework to guide analysis, design, and operation. This paper describes such a framework based on a state-, model-, and goal-based architecture for semi-autonomous control systems that guides analysis and modeling, shapes control system software design, and directly specifies operational intent. This paper illustrates the key concepts in the context of a large-scale, concurrent, globally distributed system of systems: NASA's proposed Array-based Deep Space Network.

systems modeling↗
NASA NTRS2003

Characterization of Al-Cu-Mg-Ag Alloy RX226-T8 Plate

Aluminum-copper-magnesium-silver (Al-Cu-Mg-Ag) alloys that were developed for thermal stability also offer attractive ambient temperature strength-toughness combinations, and therefore, can be considered for a broad range of airframe structural applications. The current study evaluated Al-Cu-Mg-Ag alloy RX226-T8 in plate gages and compared performance with sheet gage alloys of similar composition. Uniaxial tensile properties, plane strain initiation fracture toughness, and plane stress tearing resistance of RX226-T8 were examined at ambient temperature as a function of orientation and thickness location in the plate. Properties were measured near the surface and at the mid-plane of the plate. Tensile strengths were essentially isotropic, with variations in yield and ultimate tensile strengths of less than 2% as a function of orientation and through-thickness location. However, ductility varied by more than 15% with orientation. Fracture toughness was generally higher at the mid-plane and greater for the L-T orientation, although the differences were small near the surface of the plate. Metallurgical analysis indicated that the microstructure was primarily recrystallized with weak texture and was uniform through the plate with the exception of a fine-grained layer near the surface of the plate. Scanning electron microscope analysis revealed Al-Cu-Mg second phase particles which varied in composition and were primarily located on grain boundaries parallel to the rolling direction. Fractography of toughness specimens for both plate locations and orientations revealed that fracture occurred predominantly by transgranular microvoid coalescence. Introduction High-strength, low-density Al-Cu-Mg-Ag alloys were initially developed to replace conventional 2000 (Al-Cu-Mg) and 7000 (Al-Zn-Cu-Mg) series aluminum alloys for aircraft structural applications [1]. During the High Speed Civil Transport (HSCT) program, improvements in thermal stability were demonstrated for candidate aircraft wing and fuselage skin materials through the addition of silver to Al-Cu-Mg alloys based on Al 2519 chemistry [2]. Thermal stability of the resulting Al-Cu-Mg-Ag alloys, C415-T8 and C416-T8, was due to co-precipitation of the thermally stable . (AlCu) and ' (Al2Cu) strengthening phases [1-4]. The strength and toughness behavior was investigated for these alloys produced as 0.090-inch thick rolled sheet in the T8 condition and after various thermal exposures. The mechanical properties were shown to be competitive with conventional aircraft alloys, 2519-T8 and 2618-T8 [2]. During the Integral Airframe Structure (IAS) program, advanced aluminum alloys were examined for use in an integrally stiffened airframe structure where the skin and stiffeners would be machined from plate and extruded frames would be mechanically attached (see Figure 1) [5]. Advantages of integrally stiffened structure include reduced part count, and reduced assembly times compared to conventional built-up airframe structure. The near-surface properties of a thick plate are of significance for a machined integrally stiffened airframe structure since this represents the skin location. Properties measured at the mid-plane of the plate are more representative of the stiffener web. RX226 was developed to exploit strength-toughness improvements and thermal stability benefits of Al-Cu-Mg-Ag alloys in plate gages. This study evaluated the microstructure and properties of three gages of plate produced in the T8 condition.

Lach, Cynthia L.↗
DOE OSTI2024

Enhancing high-fidelity neural network potentials through low-fidelity sampling

The efficacy of neural network potentials (NNPs) critically depends on the quality of the configurational datasets used for training. Prior research using empirical potentials has shown that well-selected liquid–solid transitional configurations of a metallic system can be translated to other metallic systems. This study demonstrates that such validated configurations can be relabeled using density functional theory (DFT) calculations, thereby enhancing the development of high-fidelity NNPs. Training strategies and sampling approaches are efficiently assessed using empirical potentials and subsequently relabeled via DFT in a highly parallelized fashion for high-fidelity NNP training. Our results reveal that relying solely on energy and force for NNP training is inadequate to prevent overfitting, highlighting the necessity of incorporating stress terms into the loss functions. To optimize training involving force and stress terms, we propose employing transfer learning to fine-tune the weights, ensuring that the potential surface is smooth for these quantities composed of energy derivatives. This approach markedly improves the accuracy of elastic constants derived from simulations in both empirical potential-based NNPs and relabeled DFT-based NNPs. Overall, this study offers significant insights into leveraging empirical potentials to expedite the development of reliable and robust NNPs at the DFT level.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Development of algorithms for augmenting and replacing conventional process control using reinforcement learning

Here, this work seeks to allow for the online operation and training of model-free reinforcement learning (RL) agents but limit the risk to system equipment and personnel. The parallel implementation of RL alongside more conventional process control (CPC) allows for the RL algorithm to learn from CPC. The past performance of both methods are assessed on a continuous basis allowing for a transition from CPC to RL and, if needed, transitioning back to CPC from RL. This allows for the RL algorithm to slowly and safely assume control of the process without significant degradation in control performance. It is shown that the RL can derive a near optimal policy even when coupled with a suboptimal CPC. It is also demonstrated that the coupled RL-CPC algorithm learns at a faster rate than traditional RL methods of exploration while the algorithm’s performance does not deteriorate below CPC, even when exposed to an unknown operating condition.

30 DIRECT ENERGY CONVERSION↗
DOE OSTI2020

Implementing Software Resiliency in HPX for Extreme Scale Computing

The DOE Office of Science Exascale Computing Project (ECP) outlines the next milestones in the supercomputing domain. The target computing systems under the project will deliver 10x performance while keeping the power budget under 30 megawatts. With such large machines, the need to make applications resilient has become paramount. The benefits of adding resiliency to mission critical and scientific applications, includes the reduced cost of restarting the failed simulation both in terms of time and power. Most of the current implementation of resiliency at the software level makes use of a Coordinated Checkpoint and Restart (C/R). This technique of resiliency generates a consistent global snapshot, also called a checkpoint. Generating snapshots involves global communication and coordination and is achieved by synchronizing all running processes. The generated checkpoint is then stored in some form of persistent storage. On failure detection, the runtime initiates a global rollback to the most recent previously saved checkpoint. This involves aborting all running processes, rolling them back to the previous state and restarting them.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Broadband unidirectional visible imaging using wafer-scale nano-fabrication of multi-layer diffractive optical processors

We present a broadband and polarization-insensitive unidirectional imager that operates at the visible part of the spectrum, where image formation occurs in one direction, while in the opposite direction, it is blocked. This approach is enabled by deep learning-driven diffractive optical design with wafer-scale nano-fabrication using high-purity fused silica to ensure optical transparency and thermal stability. Our design achieves unidirectional imaging across three visible wavelengths (covering red, green, and blue parts of the spectrum), and we experimentally validated this broadband unidirectional imager by creating high-fidelity images in the forward direction and generating weak, distorted output patterns in the backward direction, in alignment with our numerical simulations. This work demonstrates wafer-scale production of diffractive optical processors, featuring 16 levels of nanoscale phase features distributed across two axially aligned diffractive layers for visible unidirectional imaging. This approach facilitates mass-scale production of ~0.5 billion nanoscale phase features per wafer, supporting high-throughput manufacturing of hundreds to thousands of multi-layer diffractive processors suitable for large apertures and parallel processing of multiple tasks. Beyond broadband unidirectional imaging in the visible spectrum, this study establishes a pathway for artificial-intelligence-enabled diffractive optics with versatile applications, signaling a new era in optical device functionality with industrial-level, massively scalable fabrication.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Unified Language Frontend for Physic-Informed AI/ML

Artificial intelligence and machine learning (AI/ML) are becoming important tools for scientific modeling and simulation as in several other fields such as image analysis and natural language processing. ML techniques can leverage the computing power available in modern systems and reduce the human effort needed to configure experiments, interpret and visualize results, draw conclusions from huge quantities of raw data, and build surrogates for physics based models. Domain scientists in fields like fluid dynamics, microelectronics and chemistry can automate many of their most difficult and repetitive tasks or improve the design times by use of the faster ML-surrogates. However, modern ML and traditional scientific highperformance computing (HPC) tend to use completely different software ecosystems. While ML frameworks like PyTorch and TensorFlow provide Python APIs, most HPC applications and libraries are written in C++. Direct interoperability between the two languages is possible but is tedious and error-prone. In this work, we show that a compiler-based approach can bridge the gap between ML frameworks and scientific software with less developer effort and better efficiency. We use the MLIR (multi-level intermediate representation) ecosystem to compile a pre-trained convolutional neural network (CNN) in PyTorch to freestanding C++ source code in the Kokkos programming model. Kokkos is a programming model widely used in HPC to write portable, shared-memory parallel code that can natively target a variety of CPU and GPU architectures. Our compiler-generated source code can be directly integrated into any Kokkosbased application with no dependencies on Python or cross-language interfaces.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗