Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “model benchmarking”

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 1,765 records · Page 98

Spectroscopic Determination of the Physical Conditions in Hot Optically Thin Sources

The Astrophysics Plasma Emission Code and Database (APEC/APED), developed in part under this grant, have been upgraded to Version 1.3.0 and are now beginning to find widespread applications to X-ray spectral data from Chandra and XMM-Newton. These models represent the best theoretical data currently available and are reasonably complete below about 30 A. We are making significant progress in improving the spectral models between 30 and 90 A, by the addition of HULLAC calculations and NIST wavelengths for L-shell ions of Ne, Mg, Si, S, Ar, and Ca. While the output models are now public, the code itself is not ready for public release. We have made significant progress toward this goal through programming improvements, including optimized error codes, a major (elusive) bug fix, and additions to our testing protocols. Secondarily, we are beginning to investigate limited uses of APEC for X-ray photoionized plasma, as for opacity modeling. Stellar coronae are being used to benchmark the atomic data in APED as part of the Emission Line Project. The models appear to be in good agreement with the observations for most of the strong lines; however, we have identified significant discrepancies in the 3s/3d line ratios not only for Fe XVII, but also for Fe XVIII and XIX. The careful analysis of these high quality X-ray spectra has also shown that there are serious problems with some of the analysis techniques in widespread use. There are several manifestations of the problem, but the main problem is that line-to-continuum ratios may be seriously mismeasured. We continue to collaborate widely with atomic physicists in order to find the best atomic data, and to solicit calculations and measurements that are needed.

Brickhouse, Nancy↗

Multiscale Modeling of Structurally-Graded Materials Using Discrete Dislocation Plasticity Models and Continuum Crystal Plasticity Models

A multiscale modeling methodology that combines the predictive capability of discrete dislocation plasticity and the computational efficiency of continuum crystal plasticity is developed. Single crystal configurations of different grain sizes modeled with periodic boundary conditions are analyzed using discrete dislocation plasticity (DD) to obtain grain size-dependent stress-strain predictions. These relationships are mapped into crystal plasticity parameters to develop a multiscale DD/CP model for continuum level simulations. A polycrystal model of a structurally-graded microstructure is developed, analyzed and used as a benchmark for comparison between the multiscale DD/CP model and the DD predictions. The multiscale DD/CP model follows the DD predictions closely up to an initial peak stress and then follows a strain hardening path that is parallel but somewhat offset from the DD predictions. The difference is believed to be from a combination of the strain rate in the DD simulation and the inability of the DD/CP model to represent non-monotonic material response.

Saether, Erik↗

Joint neutron measurements on the flat-top HEU critical assembly machine - 14388

Flat-Top is a critical assembly with a spherical core (either ∼17 kg of HEU or ∼6 kg of δ-phase Pu) surrounded by a spherical 1000 kg natural uranium reflector. In 2013, joint LANL and CEA measurements were conducted to measure Flat-Top configurations with the HEU core using He-3 detector systems. Similar experiments were previously documented in which the Caliban and Godiva critical assemblies were measured. The core and reflector of Flat-Top have a glory hole in which nuclear material or non-nuclear material may be placed. Five configurations were measured and included subcritical configurations at various reactivity states as well as a configuration at delayed critical. The various reactivity states were achieved by varying the placement of the safety blocks and control rods (composed of natural uranium). A description of the assembly and the specific configurations measured will be presented. The subcritical and delayed critical measurements were analyzed using the Hage-Cifarelli formalism of the Feynman Variance-to-Mean method. This analysis can be used to calculate the multiplication of each configuration, from which the multiplication factor and reactivity are determined. These results are compared to the reactivity states based upon control rod worth curves and MCNP simulations. For this work, a new MCNP model was made to accurately represent the measured configurations. The ICSBEP handbook contains a benchmark of Flat-Top but all models are very simple and do not have the required fidelity to accurately represent the measured configurations. This work also examines the number of neutrons emitted per fission for this assembly (which has a significant contribution of induced fissions in U-238 due to the large natural uranium reflector). (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Development of Unsteady Aerodynamic and Aeroelastic Reduced-Order Models Using the FUN3D Code

Recent significant improvements to the development of CFD-based unsteady aerodynamic reduced-order models (ROMs) are implemented into the FUN3D unstructured flow solver. These improvements include the simultaneous excitation of the structural modes of the CFD-based unsteady aerodynamic system via a single CFD solution, minimization of the error between the full CFD and the ROM unsteady aero- dynamic solution, and computation of a root locus plot of the aeroelastic ROM. Results are presented for a viscous version of the two-dimensional Benchmark Active Controls Technology (BACT) model and an inviscid version of the AGARD 445.6 aeroelastic wing using the FUN3D code.

Silva, Walter A.↗

Fast reactor 3D core and burnup analysis using VESTA

Burnup analyses using the VESTA code have been performed on a MOX-fuelled fast reactor model as specified by an IAEA computational benchmark. VESTA is a relatively new code that has been used for burnup credit calculations and thermal reactor models, but not typically for fast reactor applications. The detailed input and results of the IAEA benchmark provides an opportunity to gauge the use of VESTA in a fast reactor application. VESTA employs an ultra-fine multi-group binning approach that accelerates Monte Carlo burnup calculations. Using VESTA to compute the end of cycle (EOC) power fractions by enrichment zone showed agreement with the published values within 5%. When comparing the ultra-fine multi-group binning approach to the tally-based approach, EOC isotopic masses also agree within 5%. Using the ultra-fine multi-group binning approach, we obtain a wall-time speedup factor of 35 when compared to the tally-based approach for computing a k{sub eff} eigenvalue with burnup problem. The authors conclude the use of VESTA's ultra-fine multi-group binning approach with Monte Carlo transport performs accurate depletion calculations for this fast reactor benchmark. (authors)

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Introduction to the Bioproduct Transition Dynamics Model

The U.S. Department of Energy, Bioenergy Technologies Office has a broad understanding of different conversion processes that produce bioproducts (chemicals derived from biomass feedstocks) and the market attributes of those processes. However, bioproducts have in many cases proven to be difficult to scale up to commercial production and bring to market, and there is currently a need for greater understanding around the possible successful scenarios for advancing the bioproducts industry. Insight into the myriad factors that impact the success of individual bioproducts and the growth of the bioproducts industry will enable bioproduct stakeholders to evaluate strategies for development and investment that could achieve the greatest impact. An understanding of these factors will also enable the identification of synergies between the bioproducts and biofuels industries that occur through shared learning and co-production. The Bioproduct Transition Dynamics (BTD) model explores questions such as the following: what are the factors that separate a successful, commercially produced bioproduct like succinic acid from one that never progresses beyond bench-scale lab research? Can these factors be influenced by stakeholders, and to what extent? The BTD model uses system dynamics to capture the impacts of investor decision-making, bioproduct techno-economics, and market factors during the early stages of bioproduct development. Key components of the BTD model include techno-economic benchmarks that inform the investor decision-making process, failures and setbacks in development and the resulting need for additional work, and the impacts of effective or ineffective management during each development stage. These factors and their interactions are tracked through bench-scale laboratory research, piloting, demoing, and the construction and operation of the first commercial-scale plant. Bioproduct development is linked to the biofuels industry through a shared-learning model, enabling the benefit to biofuels from bioproduct technology and feedstock supply chain development to be quantified. A wide variety of bioproducts can be analyzed using the BTD model, including bioproducts that have been developed through economic or policy-driven mechanisms, and those with either niche (low volume) or scalable (high volume) markets. The BTD model also has the potential to apply to technology development processes outside the bioproducts industry; background research performed for this project has shown that the early stages of technology development are similar, and impacted by similar factors, regardless of the exact industry. This presentation will cover the BTD model logic, input data, validation process, and results of a sensitivity analysis based on techno-economic data for succinic acid. The sensitivity analysis was performed to identify factors with the largest impact on the eventual success or failure of a bioproduct. Results indicate that management effectiveness, which affects how efficiently money spent is converted into technological advancements, is the single most important factor in determining whether a bioproduct reaches commercial-scale production. The capacity of the first commercial plant and the type of biomass feedstock used, as well as the presence of government support in the form of cost-sharing, were also identified as significant factors.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

A System Dynamics Model of Early-Stage Transition Dynamics in the Bioproducts Industry

The U.S. Department of Energy, Bioenergy Technologies Office has a broad understanding of different conversion processes that produce bioproducts (chemicals derived from biomass feedstocks) and the techno-economic attributes of those processes. However, bioproducts have in many cases proven to be difficult to scale up to commercial production, and there is currently a need for greater understanding around the possible successful scenarios for advancing the bioproducts industry. Insight into the myriad factors that impact the success of individual bioproducts and the growth of the bioproducts industry will enable bioproduct stakeholders to evaluate strategies for development and investment that could achieve the greatest impact. An understanding of these factors will also enable the identification of synergies between the bioproducts and biofuels industries that occur through shared learning and co-production. The Bioproduct Transition Dynamics (BTD) model explores questions such as the following: what are the factors that separate a successful, commercially produced bioproduct like succinic acid from one that never progresses beyond bench-scale lab research? Can these factors be influenced by stakeholders, and to what extent? The BTD model uses system dynamics to capture the impacts of investor decision-making, bioproduct techno-economics, and end use factors during the early stages of bioproduct development. Key components of the BTD model include techno-economic benchmarks that inform the investor decision-making process, failures and setbacks in development and the resulting need for additional work, and the impacts of effective or ineffective management during each development stage. These factors and their interactions are tracked through bench-scale laboratory research, piloting, demoing, and the construction and operation of the first commercial-scale plant. Bioproduct development is linked to the biofuels industry through a shared-learning model, enabling the benefit to biofuels from bioproduct technology and feedstock supply chain development to be quantified. A wide variety of bioproducts can be analyzed using the BTD model, including bioproducts that have been developed through economic or policy-driven mechanisms, and those with either niche (low volume) or scalable (high volume) demand. The BTD model also has the potential to apply to technology development processes outside the bioproducts industry; background research performed for this project has shown that the early stages of technology development are similar, and impacted by similar factors, regardless of the exact industry. This presentation will cover the BTD model logic, input data, validation process, and results of a sensitivity analysis based on techno-economic data for succinic acid. The sensitivity analysis was performed to identify factors with the largest impact on the eventual success or failure of a bioproduct. Results indicate that management effectiveness, which affects how efficiently money spent is converted into technological advancements, is the single most important factor in determining whether a bioproduct reaches commercial-scale production. The capacity of the first commercial plant and the type of biomass feedstock used, as well as the presence of government support in the form of cost-sharing, were also identified as significant factors.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

GMctdospp: Description and validation of a CT dose calculation system

Purpose: To develop a Monte Carlo (MC)-based computed tomography (CT) dose estimation method with a graphical user interface with options to define almost arbitrary simulation scenarios, to make calculations sufficiently fast for comfortable handling, and to make the software free of charge for general availability to the scientific community. Methods: A framework called GMctdospp was developed to calculate phantom and patient doses with the MC method based on the EGSnrc system. A CT scanner was modeled for testing and was adapted to half-value layer, beam-shaping filter, z-profile, and tube-current modulation (TCM). To validate the implemented variance reduction techniques, depth-dose and cross-profile calculations of a static beam were compared against DOSXYZnrc/EGSnrc. Measurements for beam energies of 80 and 120 kVp at several positions of a CT dose-index (CTDI) standard phantom were compared against calculations of the created CT model. Finally, the efficiency of the adapted code was benchmarked against EGSnrc defaults. Results: The CT scanner could be modeled accurately. The developed TCM scheme was confirmed by the dose measurement. A comparison of calculations to DOSXYZnrc showed no systematic differences. Measurements in a CTDI phantom could be reproduced within 2% average, with a maximal difference of about 6%. Efficiency improvements of about six orders of magnitude were observed for larger organ structures of a chest-examination protocol in a voxelized phantom. In these cases, simulations took 25 s to achieve a statistical uncertainty of ∼0.5%. Conclusions: A fast dose-calculation system for phantoms and patients in a CT examination was developed, successfully validated, and benchmarked. Influences of scan protocols, protection method, and other issues can be easily examined with the developed framework.

60 APPLIED LIFE SCIENCES↗

Solvent Dynamics Are Critical to Understanding Carbon Dioxide Dissolution and Hydration in Water

Simulations of carbon dioxide (CO 2 ) in water may aid in understanding the impact of its accumulation in aquatic environments and help advance technologies for carbon capture and utilization (via, e.g., mineralization). Quantum mechanical (QM) simulations based on static molecular models with polarizable continuum solvation poorly reproduce the energetics of CO 2 hydration to form carbonic acid in water, independent of the level of QM theory employed. Only with density-functional-theory-based molecular dynamics and rare-event sampling, followed by energy corrections based on embedded correlated wavefunction theory (in conjunction with density functional embedding theory), can a close agreement between theory and experiment be achieved. Such multilevel simulations can serve as benchmarks for simpler, less costly models, giving insight into potential errors of the latter. The strong influence of sampling/averaging over dynamical solvent configurations on the energetics stems from the difference in polarity of both the transition state and product (both polar) versus the reactant (nonpolar). When a solute undergoes a change in polarity during reaction, affecting its interaction with the solvent, careful assessment of the energetic contribution of the solvent response to this change is critical. Here, we show that static models (without structural sampling) that incorporate three explicit water molecules can yield far superior results than models with more explicit water molecules because fewer water molecules yield less configurational artifacts. Static models intelligently incorporating both explicit (molecules directly participating in the reaction) and implicit solvation, along with a proper QM theory, e.g., CCSD(T) for closed-shell systems, can close the accuracy gap between static and dynamic models.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Representing Global Soil Erosion and Sediment Flux in Earth System Models

Abstract Soil erosion produces enormous amounts of sediment, carbon, and nutrient fluxes from land to rivers, thus playing crucial roles in global biogeochemical cycles and food security. To predict soil erosion in the context of climate and land use changes, we explicitly parameterize cropland management actions (i.e., irrigation, conserved agriculture, and crop residue management) and geological factors (i.e., lithology and glacier) in the Energy Exascale Earth System Model (E3SM) soil erosion module. The erosion model is calibrated using a global‐scale regionalized parameter calibration method. The spatial variabilities of the modeled and the Revised Universal Soil Loss Equation (RUSLE) based benchmark soil erosion are consistent across vegetation, climate and soil properties. Compared with independent data, our model shows a bias reduction in 59% of the observations relative to the RUSLE‐based soil erosion, with 53% of the bias reduction exceeding 50%. This improvement is mainly due to a better representation of the topographic effect on soil erosion. Our results indicate that conserved agriculture practices have effectively reduced soil erosion in cropland by over 25% in the United States and Argentina. In contrast, irrigation has increased soil erosion in many Asian countries. For upland sediment flux, our model is consistent with the WBMsed benchmark data in inter‐basin variability but could be more skillful in simulating intra‐basin variability because it couples soil erosion and sediment flux explicitly. The developed model provides useful skills for more realistic predictions of soil erosion and river sediment dynamics under environmental changes.

58 GEOSCIENCES↗

Experimental Data from the Benchmark SuperCritical Wing Wind Tunnel Test on an Oscillating Turntable

The Benchmark SuperCritical Wing (BSCW) wind tunnel model served as a semi-blind testcase for the 2012 AIAA Aeroelastic Prediction Workshop (AePW). The BSCW was chosen as a testcase due to its geometric simplicity and flow physics complexity. The data sets examined include unforced system information and forced pitching oscillations. The aerodynamic challenges presented by this AePW testcase include a strong shock that was observed to be unsteady for even the unforced system cases, shock-induced separation and trailing edge separation. The current paper quantifies these characteristics at the AePW test condition and at a suggested benchmarking test condition. General characteristics of the model's behavior are examined for the entire available data set.

Heeg, Jennifer↗

A Statistical Model for Generating a Population of Unclassified Objects and Radiation Signatures Spanning Nuclear Threats

This report describes an approach for generating a simulated population of plausible nuclear threat radiation signatures spanning a range of variability that could be encountered by radiation detection systems. In this approach, we develop a statistical model for generating random instances of smuggled nuclear material. The model is based on physics principles and bounding cases rather than on intelligence information or actual threat device designs. For this initial stage of work, we focus on random models using fissile material and do not address scenarios using non-fissile materials. The model has several uses. It may be used as a component in a radiation detection system performance simulation to generate threat samples for injection studies. It may also be used to generate a threat population to be used for training classification algorithms. In addition, we intend to use this model to generate an unclassified 'benchmark' threat population that can be openly shared with other organizations, including vendors, for use in radiation detection systems performance studies and algorithm development and evaluation activities. We assume that a quantity of fissile material is being smuggled into the country for final assembly and that shielding may have been placed around the fissile material. In terms of radiation signature, a nuclear weapon is basically a quantity of fissile material surrounded by various layers of shielding. Thus, our model of smuggled material is expected to span the space of potential nuclear weapon signatures as well. For computational efficiency, we use a generic 1-dimensional spherical model consisting of a fissile material core surrounded by various layers of shielding. The shielding layers and their configuration are defined such that the model can represent the potential range of attenuation and scattering that might occur. The materials in each layer and the associated parameters are selected from probability distributions that span the range of possibilities. Once an object is generated, its radiation signature is calculated using a 1-dimensional deterministic transport code. Objects that do not make sense based on physics principles or other constraints are rejected. Thus, the model can be used to generate a population of spectral signatures that spans a large space, including smuggled nuclear material and nuclear weapons.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Multiscale dendritic needle network model of alloy solidification with fluid flow

We present a mathematical formulation of a multiscale model for solidification with convective flow in the liquid phase. The model is an extension of the dendritic needle network approach for crystal growth in a binary alloy. We propose a simple numerical implementation based on finite differences and step-wise approximations of parabolic dendritic branches of arbitrary orientation. Results of the two-dimensional model are verified against reference benchmark solutions for steady, unsteady, and buoyant flow, as well as steady-state dendritic growth in the diffusive regime. Simulations of equiaxed growth under forced flow yield dendrite tip velocities within 10% of quantitative phase-field results from the literature. Finally, we perform illustrative simulations of polycrystalline solidification using physical parameters for an aluminum-10wt% copper alloy. Resulting microstructures show notable differences when taking into account natural buoyancy in comparison to a purely diffusive transport regime. The resulting model opens new avenues for computationally and quantitatively investigating the influence of fluid flow and gravity-induced buoyancy upon the selection of dendritic microstructures. Further ongoing developments include an equivalent formulation for directional solidification conditions and the implementation of the model in three dimensions, which is critical for quantitative comparison to experimental measurements.

36 MATERIALS SCIENCE↗

Validation Experiments of Incipient Turbulent Separation over an Axisymmetric Afterbody

Historically, the flow physics involved with most turbulent separated flows have presented fundamental challenges in validation between experimental and numerical approaches. As recognized by the CFD Vision 2030 study commissioned by NASA, validation of Reynolds-averaged Navier-Stokes (RANS) models and other scale-resolving methods for turbulent separated flow requires the support of advanced, high-fidelity experiments designed specifically for CFD implementation. In accordance with this effort, a new test platform, referred to as the NASA Axisymmetric Afterbody, was designed to obtain detailed information of the flow field over a smooth body, undergoing adverse pressure gradient induced separation for a fixed Reynolds number, Re = 180,000. The parametric body offers a range of flow states from fully attached,incipient separation, to fully separated flow based on variable afterbody geometries. In an initial effort to evaluate RANS turbulence model capabilities, the present configuration of the axisymmetric model features a mild adverse pressure gradient over the contoured boattail section, inducing incipient turbulent separation. Experiments include steady pressure measurements to serve as preliminary comparisons to simulation studies, which examine the effect of variable grid domains and RANS turbulence models, in an effort to understand and evaluate the critical variability between solutions for the present configuration. Results indicate potential discrepancies may be due to the effect of the square tunnel test section walls, relatively large blockage ratio, and inconsistent reference parameters. Ongoing work will focus on the experimental campaigns to obtain surface flow visualizations and high-resolution, off-body flow field measurements using Stereoscopic Particle Image Velocimetry (SPIV) and Laser Doppler Velocimetry (LDV) techniques to provide a detailed benchmark dataset to aid turbulence modelers.

axisymmetric afterbody↗

Tunable thermodynamics and kinetics for hydrogen storage : nanoparticle synthesis using ordered polymertemplates.

Theory and experiment suggest nanoscale hydride particles are destabilized relative to bulk, but the origin of this effect is unclear. Both size and local environment may play a role. The overall project objective is to achieve tunable thermodynamics for hydrogen storage materials by controlling nanoparticle size, composition, and environment. Key Goals for FY09 are: (1) Demonstrate and downselect infiltration methods; (2) Measure desorption kinetics for MgH{sub 2} and NaAlH{sub 4} nanoparticles and LiBH{sub 4} thin films; (3) Benchmark DFT and atomistic nanoparticle models using Quantum Monte Carlo (QMC); and (4) Quantify effect of nanoparticle size on {Delta}H{sub d}{sup o} using MgH{sub 2} as initial example. Summary of the key results are: (1) New highly ordered nanoporous templates enable systematic probing of nanoscale effects - Nanoscale NaAlH{sub 4} particles (as small as 1.5 nm diameter) exhibit improved H{sub 2} desorption kinetics relative to bulk and Preliminary data suggest MgH{sub 2} nanoparticle formation and possibly improved desorption kinetics; (2) Benchmarking DFT against QMC reveals significant errors that are non-systematic (H{sub 2} desorption energies underpredicted by as much as 30 kJ/mol); (3) QMC predicts greatest effect of size is for extremely small particles; e.g. (MgH{sub 2}){sub n}, n {le} 6 which is much smaller than predicted by Wolfe construction approach and observed in experiments and it suggests factors other than electronic structure (e.g. surrounding chemical environment) influence stability; (4) New NanoPEGS code developed and tested for MgH{sub 2} 2particles; and (5) New mass spec tool (STMBMS) reveals key details of hydrogen desorption process.

08 HYDROGEN↗

Comprehensive 3D Multiphysics Model on Electrochemical Recovery of O 2 from Metabolic CO 2 at the International Space Station (ISS)

The International Space Station (ISS) is presently equipped with an elaborate, heavy, and high-power consuming system that recovers approximately 50% of O 2 from metabolic CO 2 as part of the atmospheric revitalization (AR) at the ISS habitat. Future long-duration missions will require a sustainable and efficient system capable of yielding a minimum of 75% O 2 recovery to reach the self-sufficiency required for long space missions beyond earth’s low orbit. A Macrofluidic Electrochemical Reactor (MFECR) technology development effort is currently underway at NASA Marshall Space Flight Center (MSFC) to not only increase significantly current O 2 recovery efficiency, improving self-sufficiency on AR at the ISS habitat and future long missions, but also reduce the complexity of the system. The authors have developed and deployed a comprehensive 3D multiphysics model that thoroughly replicates the actual configuration and fluid/material domains of the MFECR. The coupled physics in this multiphysics model include multicomponent-multiphase electrochemical-driven reactions, non-ideal mass transport mechanism, free and porous flow, heat transfer, CO 2 solubility on alkaline electrolyte, water condensation on porous medium, and DC electrical current generation along with Joule heating effect. This model is aimed to conduct quantitive benchmark on three different MFECR’s layouts, one without serpentine paths (plain) and two with serpentines leading to four and twelve paths respectively. Once experimental data is generated via a test matrix of 200 tests, the model will be validated to conduct MFECR’s process optimization and revalidate the quantitive benchmark on three different MFECR’s layouts.

Jesus A. Dominguez↗

Evaluation of OpenAI Codex for HPC Parallel Programming Models Kernel Generation

We evaluate AI-assisted generative capabilities on fundamental numerical kernels in high-performance computing (HPC), including AXPY, GEMV, GEMM, SpMV, Jacobi Stencil, and CG. We test the generated kernel codes for a variety of language-supported programming models, including (1) C++ (e.g., OpenMP [including offload], OpenACC, Kokkos, SyCL, CUDA, and HIP), (2) Fortran (e.g., OpenMP [including offload] and OpenACC), (3) Python (e.g., numpy, Numba, cuPy, and pyCUDA), and (4) Julia (e.g., Threads, CUDA.jl, AMDGPU.jl, and KernelAbstractions.jl). We use the GitHub Copilot capabilities powered by the GPT-based OpenAI Codex available in Visual Studio Code as of April 2023 to generate a vast amount of implementations given simple + + prompt variants. To quantify and compare the results, we propose a proficiency metric around the initial 10 suggestions given for each prompt. Results suggest that the OpenAI Codex outputs for C++ correlate with the adoption and maturity of programming models. For example, OpenMP and CUDA score really high, whereas HIP is still lacking. We found that prompts from either a targeted language such as Fortran or the more general purpose Python can benefit from adding code keywords, while Julia prompts perform acceptably well for its mature programming models (e.g., Threads and CUDA.jl). We expect for these benchmarks to provide a point of reference for each programming model's community. Overall, understanding the convergence of large language models, AI, and HPC is crucial due to its rapidly evolving nature and how it is redefining human-computer interactions.

Godoy, William↗

Analyticity constraints for hadron amplitudes: Going high to heal low-energy issues

Analyticity constitutes a rigid constraint on hadron scattering amplitudes. This property is used to relate models in different energy regimes. Using meson photoproduction as a benchmark, we show how to test contemporary low energy models directly against high energy data. This method pinpoints deficiencies of the models and treads a path to further improvement. In conclusion, the implementation of this technique enables one to produce more stable and reliable partial waves for future use in hadron spectroscopy and new physics searches.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗