Search NASA⌕ Search

SEARCH · Search NASA

Results for “Computational Performance”

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 1,783 records · Page 99

Frequency-domain computing using nonlinear acoustic-wave device on lithium niobate

Abstract Multiply-accumulation are crucial computing operations in signal processing, numerical simulations, and machine learning. In recent years, optical analog approaches have demonstrated higher computing performance and better power efficiency than their digital counterparts. However, analog computing chips usually need large areas and complex structures for parallel computing, as a single device element only executes one computing operation at a single time. Here, we demonstrate frequency-domain computing using the nonlinear acoustic-wave devices on lithium niobate, featuring a normalized external second-harmonic generation conversion efficiency of ~ 5.7 × 10-4 W-1. The second-order sum-frequency nonlinear process of lithium niobate enables multiplication of inputs encoded in the frequency domain. Compared to the analog schemes, our device features a notably simpler design, and nanofabrication requires only one lift-off. Using a single acoustic-wave device within an area of 0.03 mm2, we can simultaneously conduct over 130,000 multiply-accumulation operations. Our acoustic-wave device shows applications in real and complex vector convolutions and image processing. This demonstration sets the stage for experimental realizations into frequency-domain integrated nonlinear acoustic computing systems, potentially shaping future developments in acoustic neural networks and quantum computing.

chai, mingzhao (ORCID:0009000466226341)↗

Towards Efficient N - x Contingency Selection Using Group Betweenness Centrality

The goal of N -x contingency selection is to pick a subset of critical cases to assess their potential for causing a severe crippling of an electric power grid. Even for a moderate-sized system there can be an overwhelming number of contingency cases that need to be studied. The number grows exponentially with x. This combinatorial explosion renders any exhaustive search strategy computationally infeasible, even for small to medium sized systems. We propose a novel method for N- x selection for x >1 using group betweenness centrality and show that the amount of computation can be decoupled from the problem size, thus making the contingency analysis for large systems with x > 1 computationally feasible. Consequently, it may be that N - x (for x > 1) contingency selection can be effectively deployed despite the combinatorial explosion of the number of possible N - x contingencies. Our strategy replaces computation with storage of intermediate results and therefore relies on high performance computing resources.

Contingency analysis, Contingency selection, group↗

Agentic AI vs ML-Based Autotuning: A Comparative Study for Loop Reordering Optimization

High Performance Computing (HPC) applications rely heavily on code optimizations to achieve good performance on modern CPU and GPU architectures. Traditional Machine Learning auto-tuning approaches have demonstrated success in exploring high-dimensional spaces, but they often require expensive compile-run evaluations and lack adaptability for large HPC applications. The recent advances in Large Language Models (LLMs) and Agentic AI systems raise intriguing questions about the potential of these approaches to address specific optimization methodologies. This work aims to answer an essential question for the HPC community: “How Agentic AI Systems Compare to Traditional ML Autotuning Techniques?” To address this question, we present a comparative analysis between a traditional ML-based optimization approach and an Agentic AI system, evaluating their respective capabilities and limitations for loop-level optimization. In addition, we introduced a new Agentic AI system named LoopGen-AI using three different Large Language Models: GPT-4.1, Claude 4.0, and Gemini 2.5. A key finding is that LoopGen-AI achieves competitive per-formance with only a few program runs, the reasoning logs from the agents revealed that their decisions rely heavily on the combination of semantic understanding of the target kernel with dynamic feedback from the environment, highlighting a promising new dimension in performance tuning. In contrast, ML-based autotuners focus on statistical exploration, and require orders of magnitude more runs to reach peak performance. Additionally, our analysis shows that prompt engineering, particularly using Persona + Context Manager patterns, significantly impacts the effectiveness of Agentic AI. Our results indicate that while Agentic AI systems are not yet a complete replacement for ML-based autotuners, it can effectively complement traditional methods.

Rosas, Miguel Romero↗

Dielectric-Lined High-Gradient Accelerator Structure (Final Report)

Rectangular particle accelerator structures with internal planar dielectric elements have been studied, with a view towards devising structures with lower surface fields for a given accelerating field, as compared with structures without dielectrics. Success with this concept is expected to allow operation at higher accelerating gradients than otherwise on account of reduced breakdown probabilities. The project involves studies of RF breakdown on amorphous dielectrics in test cavities that could enable high-gradient structures to be built for a future multi-TeV collider. The aim is to determine what the limits are for RF fields at the surfaces of selected dielectrics, and the resulting acceleration gradient that could be achieved in a working structure. The dielectric of principal interest in this study is artificial CVD diamond, on account of its advertised high breakdown field (~2 GV/m for dc), low loss tangent, and high thermal conductivity. Experimental studies at mm-wavelengths on materials and structures for achieving high acceleration gradient were based on the availability of the 34.3 GHz third-harmonic magnicon amplifier developed by Omega-P, and installed at the Yale University Beam Physics Laboratory. Peak power from the magnicon was measured to be about 20 MW in 0.5 μs pulses, with a gain of 54 dB. Experiments for studying RF high-field effects on CVD diamond samples failed to show any evidence after more than 10 5 RF pulses of RF breakdown up to a tangential surface field strength of 153 MV/m; studies at higher fields were not possible due to a degradation in magnicon performance. A rebuild of the tube is underway at this writing. Computed performance for a dielectric-loaded rectangular accelerator structure (DLA) shows highly competitive properties, as compared with an existing all-metal structure. For example, comparisons were made of a DLA structure having two planar CVD diamond elements with a all-metal CERN structure HDS operating at 30 GHz. It was shown that the ratio of maximum surface electric field to accelerating field at the metal wall is only 0.35-0.4 for DLA, much smaller than the value 2.2 for HDS; and the ratio of surface magnetic field to accelerating field is 3.0 mA/V for DLA, compared with 3.45 mA/V for HDS. These values bode well for DLA in helping to avoid breakdown and to reducing pulsed surface heating and fatigue. The shunt impedance is found to be 160-175 MΩ/m for DLA, as compared to 99 MΩ/m for HDS. Conclusions are reached from this project that CVD diamond appears promising as a dielectric with a high threshold for RF breakdown, and that rectangular accelerator structures can be devised using planar CVD diamond elements that could be operated at higher acceleration gradients with low probability of RF breakdown, as compared with corresponding all-metallic structures.

36 MATERIALS SCIENCE↗

Ceramic Coatings for Clad (The C 3 Project): Advanced Accident-Tolerant Ceramic Coatings for Zr-Alloy Cladding

The goal of this NEUP-IRP project is to develop a fuel concept based on an advanced ceramic coating for Zr-alloy cladding. The coated cladding must exhibit demonstrably improved performance compared to conventional Zr-alloy clad in the following respects: During normal service, the ceramic coating should decrease cladding oxidation and hydrogen pickup (the latter leads to hydriding and embrittlement). During a reactor transient (e.g., a loss of coolant accident), the ceramic coating must minimize or at least significantly delay oxidation of the Zr-alloy cladding, thus reducing the amount of hydrogen generated and the oxygen ingress into the cladding. The specific objectives of this project are as follows: To produce durable ceramic coatings on Zr-alloy clad using two possible routes: (i) MAX phase ceramic coatings or similar nitride or carbide coatings; and (ii) graded interface architecture (multilayer) ceramic coatings, using, for instance, an oxide such as yttria-stabilized zirconia (YSZ) as the outer protective layer. To characterize the structural and physical properties of the coated clad samples produced in 1. above, especially the corrosion properties under simulated normal and transient reactor operating conditions. To perform computational analyses to assess the effects of such coatings on fuel performance and reactor neutronics, and to perform fuel cycle analyses to assess the economic viability of modifying conventional Zr-alloy cladding with ceramic coatings. This project meets a number of the goals outlined in the NEUP-IRP call for proposals, including: Improve the fuel/cladding system through innovative designs (e.g. coatings/liners for zirconium-based cladding) Reduce or eliminate hydrogen generation Increase resistance to bulk steam oxidation Achievement of our goals and objectives, as defined above, will lead to safer light-water reactor (LWR) nuclear fuel assemblies, due to improved cladding properties and built-in accident resistance, as well as the possibilities for enhanced fuel/clad system performance and longevity.

36 MATERIALS SCIENCE↗

Building Towards Discovery: Preparing for New Physics at the LHC [Final papers]

Particle physics today is at a crossroads. The Standard Model (SM) has survived four decades of experimental scrutiny and spurred multiple landmark discoveries, culminating with the 2012 discovery of the Higgs boson at the Large Hadron Collider (LHC). Yet we know that this model is incomplete. We live in a universe dominated by matter and energy unaccounted for by the SM– and the Higgs boson itself is improbably light. It falls to the current generation of high-energy physicists to go beyond the model and enter a new era of discovery. Andeen is leading a new research group at the University of Texas at Austin (UT-Austin) that is endeavoring to reveal novel physical phenomena through the discovery of new particles with the ATLAS detector at the LHC. A key element of our experience thus far has been collaborating closely with theorists to identify both interesting new physics models. In addition we work closely with electrical engineers in detector development. Our approach is therefore unique in its integration of theory, analysis and instrumentation, and is encapsulated in a three-pronged strategy. First is the essential task of analyzing our rapidly expanding dataset. I am directing my group in searches for vector-like quarks (VLQs). Should they exist, these hypothetical particles would indicate physics scenarios beyond the SM. Experimental constraints suggest VLQs decay preferentially to a third generation quark and a Higgs, W or Z boson. Today at 13 TeV, both multi-lepton (electron and muon) and jet substructure signatures are sensitive to identifying the VLQ decay products. The large Run 2 dataset (collected through 2018) of up to 150 fb -1 will allow us to simultaneously use multi-lepton signatures with jet-substructure in searches for new physics. UT-Austin is well positioned to be a leader in this area. Prof. P. Onyisi’s postdoctoral researchers and graduate students are searching for ttH signatures in three- and four-lepton channels and we benefit from their experience. We are also aided by the availability of the Texas Advanced Computing Center, which provides us with easy access to several high performance computing (HPC) systems. My previous experience in VLQ searches propels the work today. I have published both the first investigation into the single production of top partners (signatures of composite Higgs Models) of any LHC experiment 1 as well as one of the first analyses utilizing jet-substructure techniques to search for the single production of vector-like quarks. 2 Adding to this is my group’s work with the Liquid Argon (LAr) calorimeter, the primary instrument we use in detecting electrons. Second, my group is contributing to the operations of the LAr calorimeter while preparing to install and commission new electronic hardware (the Phase 1 upgrade) to this crucial component of the ATLAS detector. I have played a leading role in the development of a critical, radiation-hard, high-speed Analog-to-Digital Convertor (ADC) that is essential inselecting data collected by the LAr calorimeter. 3 This upgrade will improve the e!ciency with which we can select events in our detector. Postdoctoral researcher N. Nikiforou (based at CERN May 2016 onward) and graduate students (one based at CERN from the start of 2018 onward, with one additional graduate students from summer 2018) are growing our already deep experience in the operation of the LAr calorimeter to prepare for installation and commissioning of the new trigger readout in 2018-2019. My group will then be positioned to be the first to exploit the potential of the improved triggers in our multi-lepton searches for new physics. Third, a suite of preparations for a major detector upgrade, scheduled to be completed in 2024, are underway. My group is building hardware that will improve the electronic readout of the ATLAS detector. I am partnering with a group in the Electrical and Computer Engineering (ECE) Department at UT-Austin (Prof. N. Sun) to employ his work at the cutting edge of ADC research. We will support an ECE grad student for the layout of this design, and collaborate with engineering groups at Nevis Laboratory and the Electrical Engineering Department (Prof. P. Kinget) at Columbia University to integrate our ADC with gain selection circuits. We will provide this critical component of the ASIC at the challenging boundary of the analog and digital signals in the LAr calorimeter readout chain. Furthermore UT-Austin will lead the testing of this device during development and production (approximately 55,000 four-channel chips). This effort compliments my work as the deliverables manager for the LAr front-end ASICs for the US ATLAS HL-LHC project and strengthens the US commitment to that upgrade.In so doing, my group at the UT-Austin will be at the forefront of efforts to probe the structure of the universe to an unmatched degree both at the current LHC and with its future upgrades. Through both the ongoing analysis of data at the ATLAS detector and improvements to the detector itself, an answer to what lies beyond the SM may come within our grasp. My background in instrumentation and analysis allows my group to establish a singular research effort on the ATLAS experiment. In addition to answering fundamental questions about the subatomic world, the technologies developed could advance the broader scientific community and result in valuable spin-offs.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

PHISICS User Manual

PHISICS (Parallel and Highly Innovative Simulation for INL Code System) is a reactor physics package developed at the Idaho National Laboratory. It is composed of several modules: a nodal and semi-structured transport core solver (INSTANT), a depletion module (MRTAU), a time-dependent solver (TimeIntegrator), a cross-section interpolation and manipulation framework (MIXER), a SuperHomogenization scheme (SPH) a criticality search module (CRITICALITY), a fuel management and shuffling component (SHUFFLE) and a coupling API with RELAP5-3D. PHISICS has been built in a highly modular fashion so that different physical models can be activated or deactivated depending on the physics relevant to the particular type of analysis. This approach simplifies development as well as mid- and long-term maintenance of the code. PHISICS is entirely written in FORTRAN 95/2003. Each of its components is fully parallelized and designed to take full advantage of shared-memory computers (multiprocessor desktops and workstations) and middle to large high-performance computing systems (10 to 1000 processors). The aim of this document is to detail the input requirements for PHISICS (and all its modules) focusing on the input structure.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Revolutionizing Neuromorphic Computing for Science (Brochure on the 2024 ASCR Workshop on Neuromorphic Computing for Science)

The ASCR basic research needs for Neuromorphic Computing for Science workshop was held in September 2024. The workshop brochure and report aim to inform and draft a set of grand challenges for advancing the field of neuromorphic computing and developing proof of principle neuromorphic circuits applicable for High Performance Computer (HPC) acceleration for scientific discovery, and brainstorm ideas needed for a successful, robust, and world leading basic research program. The resulting priority research directions are: (1) Neuromorphic computing circuit primitives; (2) Connectivity, communication, and hardware integration; (3) Neuroscience-derived dynamics and algorithms; and (4) Ecosystem for scalable neuromorphic co-design. Breakthroughs in understanding, designing, and prototyping the circuitry and simulation capabilities for a truly neuromorphic computer are essential to enable progress in the field.

97 MATHEMATICS AND COMPUTING↗

Revolutionizing Neuromorphic Computing for Science (Report for the 2024 ASCR Workshop on Neuromorphic Computing for Science)

The ASCR basic research needs for Neuromorphic Computing for Science workshop was held in September 2024. The workshop brochure and report aim to inform and draft a set of grand challenges for advancing the field of neuromorphic computing and developing proof of principle neuromorphic circuits applicable for High Performance Computer (HPC) acceleration for scientific discovery, and brainstorm ideas needed for a successful, robust, and world leading basic research program. The resulting priority research directions are: (1) Neuromorphic computing circuit primitives; (2) Connectivity, communication, and hardware integration; (3) Neuroscience-derived dynamics and algorithms; and (4) Ecosystem for scalable neuromorphic co-design. Breakthroughs in understanding, designing, and prototyping the circuitry and simulation capabilities for a truly neuromorphic computer are essential to enable progress in the field.

97 MATHEMATICS AND COMPUTING↗

Position Papers for the 2024 ASCR Workshop on Analog Computing for Science

Analog computing capabilities have existed since the dawn of science, but modern techniques potentially allow for the construction, verification, and characterization of complicated analog computing systems in a wide variety of contexts, from high-performance computational accelerators to nanorobotics and synthetic biology. In short, advances in analog computing can enable the creation of physical systems with complex behaviors that meet sophisticated requirements. Meeting our nation's needs, from needs in computing and modeling, to needs for advanced materials and energy technologies, continues to motivate pursuing novel kinds of complex systems, and thus analog computing techniques, in all of these spaces.

97 MATHEMATICS AND COMPUTING↗

Feasibility of a special-purpose computer to solve the Navier-Stokes equations

Orders-of-magnitude improvements in computer performance can be realized with a parallel array of thousands of fast microprocessors. In this architecture, wiring congestion is minimized by limiting processor communication to nearest neighbors. When certain standard algorithms are applied to a viscous flow problem and existing LSI technology is used, performance estimates of this conceptual design show a dramatic decrease in computational time when compared to the CDC 7600.

Gritton, E. C.↗

Preliminary results from the NASA general aviation demonstration advanced avionics system program

NASA's Demonstration Advanced Avionics System (DAAS) is an integrated avionics system employing microprocessor technologies, data busing, and shared electronics displays. A DAAS demonstration system has been assessed through flight testing and demonstration for potential users which includes among its functions autopiloting, navigation/flight planning, a flight warning and advisory system, performance computations, normal and emergency checklists, and a ground simulation function. Exceptional performance has been obtained from the DAAS electronic horizontal situation indicator, the autopilot, the navigator/flight planner, and the discrete address beacon system.

Hardy, G. H.↗

Performance Evaluation and Modeling Techniques for Parallel Processors

In practice, the performance evaluation of supercomputers is still substantially driven by singlepoint estimates of metrics (e.g., MFLOPS) obtained by running characteristic benchmarks or workloads. With the rapid increase in the use of time-shared multiprogramming in these systems, such measurements are clearly inadequate. This is because multiprogramming and system overhead, as well as other degradations in performance due to time varying characteristics of workloads, are not taken into account. In multiprogrammed environments, multiple jobs and users can dramatically increase the amount of system overhead and degrade the performance of the machine. Performance techniques, such as benchmarking, which characterize performance on a dedicated machine ignore this major component of true computer performance. Due to the complexity of analysis, there has been little work done in analyzing, modeling, and predicting the performance of applications in multiprogrammed environments. This is especially true for parallel processors, where the costs and benefits of multi-user workloads are exacerbated. While some may claim that the issue of multiprogramming is not a viable one in the supercomputer market, experience shows otherwise. Even in recent massively parallel machines, multiprogramming is a key component. It has even been claimed that a partial cause of the demise of the CM2 was the fact that it did not efficiently support time-sharing. In the same paper, Gordon Bell postulates that, multicomputers will evolve to multiprocessors in order to support efficient multiprogramming. Therefore, it is clear that parallel processors of the future will be required to offer the user a time-shared environment with reasonable response times for the applications. In this type of environment, the most important performance metric is the completion of response time of a given application. However, there are a few evaluation efforts addressing this issue.

Dimpsey, Robert Tod↗

Model Assisted Probability of Detection for NASA Space Missions

Model assisted probability of detection (MAPOD) uses data from simulations to improve a traditional probability of detection (POD) study. This could include extending the parameter space to reduce uncertainty or substituting experimental data with simulated data to reduce the time and cost of a POD study. In the past MAPOD was difficult due to limited computational resources, but recent innovations in simulation tools and high-performance computing have made this type of high-degree-of-freedom modeling possible, and complex structures have made it necessary. This presentation will summarize the work done by the computational nondestructive evaluation (CNDE) specialists within the Nondestructive Evaluation Sciences branch at NASA Langley Research Center (LaRC) to complete a MAPOD study for phased array ultrasound testing (PAUT) of a friction stir welding (FSW) method to be used on Space Launch System (SLS) structures. The three critical needs for a MAPOD study are a validated and verified model of the inspection technique for the structure being inspected, some experimental POD data, and an uncertainty model for both the model and the experimental data. PAUT was simulated using Extende CIVA’s UT module. The model was validated using laboratory inspection data from NASA Marshall Space Flight Center (MSFC) for a Hit/Miss POD for FSW in 2219-T87 aluminum panels representative of those used in the SLS. This model was then used to simulate flaw sizes that were originally omitted from the original POD study. The results of this new MAPOD study will be presented along with a discussion of the methods and processes used to analyze the original data, selected simulation parameters, and development of the uncertainty model used for the statistical analysis. The goal of this effort is not just to improve the POD study but to demonstrate the value of MAPOD and provide a roadmap for application of MAPOD on future projects.

Elizabeth Gregory↗

Model Assisted Probability of Detection for NASA Space Missions

Model assisted probability of detection (MAPOD) uses data from simulations to improve a traditional probability of detection (POD) study. This could include extending the parameter space to reduce uncertainty or substituting experimental data with simulated data to reduce the time and cost of a POD study. In the past MAPOD was difficult due to limited computational resources, but recent innovations in simulation tools and high-performance computing have made this type of high-degree-of-freedom modelling possible, and complex structures have made it necessary. This presentation will summarize the work done by the computational nondestructive evaluation (CNDE) group at NASA Langley Research Center (LaRC) to complete a MAPOD study for phased array ultrasound testing (PAUT) of a friction stir welding (FSW) method to be used on Space Launch System (SLS) structures. The three critical needs for a MAPOD study are a validated and verified model of the inspection technique for the structure being inspected, some experimental POD data, and an uncertainty model for both the model and the experimental data. PAUT was simulated using Extende CIVA’s UT module. The model was validated using laboratory inspection data from NASA Marshall Space Flight Center (MSFC) for a Hit/Miss POD for FSW in 2219-T87 aluminum panels representative of those used in the SLS. This model was then used to simulate flaw sizes that were originally omitted from the original POD study. The results of this new MAPOD study will be presented along with a discussion of the methods and processes used to analyze the original data, selected simulation parameters, and development of the uncertainty model used for the statistical analysis. The goal of this effort is not just to improve the POD study but to demonstrate the value of MAPOD and provide a roadmap for application of MAPOD on future projects.

Elizabeth Gregory↗

Geomancy: Automated Performance Enhancement through Data Layout Optimization

Large distributed storage systems such as high- performance computing (HPC) systems used by national or international laboratories require sufficient performance and scale for demanding scientific workloads and must handle shifting workloads with ease. Ideally, data is placed in locations to optimize performance, but the size and complexity of large storage systems inhibit rapid effective restructuring of data layouts to maintain performance as workloads shift. To address these issues, we have developed Geomancy, a tool that models the placement of data within a distributed storage system and reacts to drops in performance. Using a combination of machine learning techniques suitable for temporal modeling, Geomancy determines when and where a bottleneck may happen due to changing workloads and suggests changes in the layout that mitigate or prevent them. Our approach to optimizing throughput offers benefits for storage systems such as avoiding potential bottlenecks and increasing overall I/O throughput from 11% to 30%.

Bel, Oceane M.↗

ATHENA: Analytical Tool for Heterogeneous Neuromorphic Architectures

The ASC program seeks to use machine learning to improve efficiencies in its stockpile stewardship mission. Moreover, there is a growing market for technologies dedicated to accelerating AI workloads. Many of these emerging architectures promise to provide savings in energy efficiency, area, and latency when compared to traditional CPUs for these types of applications — neuromorphic analog and digital technologies provide both low-power and configurable acceleration of challenging artificial intelligence (AI) algorithms. If designed into a heterogeneous system with other accelerators and conventional compute nodes, these technologies have the potential to augment the capabilities of traditional High Performance Computing (HPC) platforms [5]. This expanded computation space requires not only a new approach to physics simulation, but the ability to evaluate and analyze next-generation architectures specialized for AI/ML workloads in both traditional HPC and embedded ND applications. Developing this capability will enable ASC to understand how this hardware performs in both HPC and ND environments, improve our ability to port our applications, guide the development of computing hardware, and inform vendor interactions, leading them toward solutions that address ASC’s unique requirements.

97 MATHEMATICS AND COMPUTING↗

State-of-the-Art in Heterogeneous Computing

Node level heterogeneous architectures have become attractive during the last decade for several reasons: compared to traditional symmetric CPUs, they offer high peak performance and are energy and/or cost efficient. With the increase of fine-grained parallelism in high-performance computing, as well as the introduction of parallelism in workstations, there is an acute need for a good overview and understanding of these architectures. We give an overview of the state-of-the-art in heterogeneous computing, focusing on three commonly found architectures: the Cell Broadband Engine Architecture, graphics processing units (GPUs), and field programmable gate arrays (FPGAs).We present a review of hardware, available software tools, and an overview of state-of-the-art techniques and algorithms. Furthermore, we present a qualitative and quantitative comparison of the architectures, and give our view on the future of heterogeneous computing.

99 GENERAL AND MISCELLANEOUS↗