Search NASA⌕ Search

SEARCH · Search NASA

Results for “Execution Models”

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 415 records · Page 23
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Large-scale parametric modeling of spent nuclear fuel dynamics in the 30 cm package drop scenario

Packages used to transport spent nuclear fuel (SNF) are required by the U.S. Code of Federal Regulations 10 CFR 71.71 to demonstrate satisfactory performance during a drop scenario. While the CFR is meant to ensure safe package function, it does not evaluate survival of the SNF within. The U.S. Department of Energy Spent Fuel and Waste Science and Technology program is working on closing the knowledge gap related to the response of SNF to external mechanical loads, including the hypothetical 30 cm package drop scenario in the CFR. In support of this effort, LS-DYNA finite element simulations were developed by Pacific Northwest National Laboratory (PNNL) to model generic drop scenarios at both the package and fuel assembly level. The models were validated against one-third scale package and full scale fuel assembly drop test data and were exercised to predict fuel cladding strains in a narrow range of model configurations. This work describes a large-scale parametric study conducted by PNNL using the previously developed and validated PWR finite element model, with the addition of a new generic BWR assembly model. The motivation for the parametric study was to characterize the broad range of SNF responses in the 30 cm package drop scenario. This was accomplished by varying the drop orientation, fuel assembly type (17x17 PWR and 10x10 BWR), burnup, cladding temperature, spacer grid buckling load, package mass, impact limiter stiffness, and mechanical gap conditions within the basket. A MATLAB framework was developed to automate LS-DYNA model generation and execution on PNNL institutional computing resources. In total, over 2000 simulations were performed. For each simulation, the SNF response was quantified in terms of permanent grid deformation, fuel rod contact pressure, and strains within the fuel rods, guide tubes, and water rods. The results provide valuable insight into the range of responses that could be reasonably expected from SNF in the hypothetical drop scenario, as well as the sensitivity to each input parameter. The results of this parametric study are a key component of the testing and modeling strategy the Spent Fuel and Waste Science and Technology program is using to close the external loads knowledge gap.

Kadooka, Kevin↗
DOE OSTI2023

A Unification Framework for Euclidean and Hyperbolic Graph Neural Networks

Hyperbolic neural networks have recently gained significant attention due to their promising results on several graph problems including node classification and link prediction. The primary reason for this success is the effectiveness of hyperbolic space in capturing the inherent hierarchy of graph datasets. However, they are limited in terms of generalization, scalability, and have inferior performance when applied to non-hierarchical datasets. In this paper, we take a completely different perspective for modeling hyperbolic networks and answer the following question: is an Euclidean model able to approximate a function or behavior in the hyperbolic space? Extending the universal approximation theory developed for Euclidean models, We draw an analogy from the hyperbolic components to the Euclidean counterparts and conclude that, in order to capture hierarchical features, it is possible to generalize hyperbolic models to be a special case of Euclidean models with the proposed Pseudo-Poincaré technique. We applied our non-linear hyperbolic normalization to the current state-of-the-art homogeneous and multi-relational graph networks and demonstrate significant improvements in performance compared to both Euclidean and hyperbolic counterparts. The primary impact of this work lies in its ability to capture hierarchical features in the Euclidean space, and thus, can replace hyperbolic networks without any loss in performance metrics while simultaneously leveraging the power of Euclidean networks such as interpretability and efficient execution of various model components.

Khatir, Mehrdad↗
DOE OSTI2021

A Hybrid Energy System Workflow for Energy Portfolio Optimization

This manuscript develops a workflow, driven by data analytics algorithms, to support the optimization of the economic performance of an Integrated Energy System. The goal is to determine the optimum mix of capacities from a set of different energy producers (e.g., nuclear, gas, wind and solar). A stochastic-based optimizer is employed, based on Gaussian Process Modeling, which requires numerous samples for its training. Each sample represents a time series describing the demand, load, or other operational and economic profiles for various types of energy producers. These samples are synthetically generated using a reduced order modeling algorithm that reads a limited set of historical data, such as demand and load data from past years. Numerous data analysis methods are employed to construct the reduced order models, including, for example, the Auto Regressive Moving Average, Fourier series decomposition, and the peak detection algorithm. All these algorithms are designed to detrend the data and extract features that can be employed to generate synthetic time histories that preserve the statistical properties of the original limited historical data. The optimization cost function is based on an economic model that assesses the effective cost of energy based on two figures of merit: the specific cash flow stream for each energy producer and the total Net Present Value. An initial guess for the optimal capacities is obtained using the screening curve method. The results of the Gaussian Process model-based optimization are assessed using an exhaustive Monte Carlo search, with the results indicating reasonable optimization results. The workflow has been implemented inside the Idaho National Laboratory’s Risk Analysis and Virtual Environment (RAVEN) framework. The main contribution of this study addresses several challenges in the current optimization methods of the energy portfolios in IES: First, the feasibility of generating the synthetic time series of the periodic peak data; Second, the computational burden of the conventional stochastic optimization of the energy portfolio, associated with the need for repeated executions of system models; Third, the inadequacies of previous studies in terms of the comparisons of the impact of the economic parameters. The proposed workflow can provide a scientifically defendable strategy to support decision-making in the electricity market and to help energy distributors develop a better understanding of the performance of integrated energy systems.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2016

High Performance Teams Enable Mission Success for SRNL - 16363

The accomplishment of the DOE EM mission of cleaning up the radioactive waste generated as a byproduct of the development and maintenance of nuclear materials over the last 60 years is a daunting task. SRNL specializes in working directly with the 'owners' of the cleanup challenges to identify the desired outcome and develop a solution that can be put into practice. 'We put science to work' is not just a 'tag line' rather it is a concise statement of our values and approach to applying science and engineering to the cleanup objectives. The management of radioactive waste poses multi-decade challenges to DOE EM and the Nation. The ability to harness the depth of knowledge, experience, expertise and passion for safely and effectively solving the technical challenges has reached a critical junction. Perhaps too late to be a natural and organic transition the industry finds itself critically devoid of resources to effectively and efficiently keep focused on the highest order task: cleanup. When project managers plan the resources necessary to perform the planned scope they typically are focused on financial, human, and industrial resources that would be beneficial to project execution. The critical shortage we will be evaluating is experienced and skilled leaders, mentors and coaches in the workforce. The production of educated scientists and engineers is a primary focus at all levels of educational institutions with Science Technology Engineering and Mathematics (STEM) focused curricula beginning in elementary school. The need for leaders who are skilled in inspiring organizational creativity, adept in harnessing and focusing individual energy while effectively communicating to the work force and the clientele equally compelling messages, has never been greater, or their absence in the workforce more ominous. Managers trained in the classical management theory (insert references here) are aging out of the workforce. In many cases these managers have not retained close connection to advancements in their primary field of study but rather have succumbed to the administrivia burdens that typify highly regulated work. The current technically trained workforce while interested in the 'perks' of management (i.e. visibility, pay increases, advancing status) are driven to keep close connection to their technical 'roots' as a fall-back plan should they need to change their career path either by choice or necessity. The challenge is how to achieve the productive hum of a finely tuned organizational model that blends executive vision and drive with technical leadership and excellence in execution. If any of these three tenets of a high performance team falters the ability to recover erodes the performance of the team and begins to weaken the cultural foundation essential to sustaining high performance work teams. At SRNL, recognition of the need to adapt the management dynamics and the business model and to harness the energy of the workforce to sustain - and in fact increase - the level of performance is a driving force in the transformation of the organization. Examples of changes in organizations structure, business systems and employee engagement are discussed as they relate to the impact on accomplishing the cleanup mission and fostering the next generation of great leaders and innovators. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
NASA NTRS1991

Rule-based simulation models

Procedural modeling systems, rule based modeling systems, and a method for converting a procedural model to a rule based model are described. Simulation models are used to represent real time engineering systems. A real time system can be represented by a set of equations or functions connected so that they perform in the same manner as the actual system. Most modeling system languages are based on FORTRAN or some other procedural language. Therefore, they must be enhanced with a reaction capability. Rule based systems are reactive by definition. Once the engineering system has been decomposed into a set of calculations using only basic algebraic unary operations, a knowledge network of calculations and functions can be constructed. The knowledge network required by a rule based system can be generated by a knowledge acquisition tool or a source level compiler. The compiler would take an existing model source file, a syntax template, and a symbol table and generate the knowledge network. Thus, existing procedural models can be translated and executed by a rule based system. Neural models can be provide the high capacity data manipulation required by the most complex real time models.

Nieten, Joseph L.↗
DOE OSTI2018

Continuation Report: Creep-fatigue Behavior and Damage Accumulation of a Candidate Structural Material for Concentrating Solar Thermal Receiver

At the start of this phase, six alloys were determined to be candidate materials for the concentrating solar power (CSP) thermal receiver: Alloys 617, 625, 740H, 230, 282 and Alloy X. From these, Alloys 740H and 282 were determined to be the most suitable, with both alloys exhibiting similar high temperature strength properties. Alloy 740H was selected due to its ease to work with (simpler heat treatments and welding procedures) and a larger database of existing data. Fatigue testing was started in this phase, with four fatigue tests having been completed, with additional fatigue testing planned, as well as tensile, creep and creep-fatigue tests starting in the next phase. Two fatigue tests performed at 850 °C and were completed in significantly fewer cycles than the two tests performed at 750 °C. Metallography was performed on the fatigued specimens to examine crack path and surrounding microstructure. These preliminary results suggest a significant drop in strength at elevated temperatures above 750 °C, however, more test results will be needed prior to setting design limits. Several potential model high temperature design methods were evaluated that could form the basis of our final recommended creep-fatigue design method for Alloy 740H. To do this we expanded on the receiver finite element model to develop a model representative of a potential Gen 3 design using KCl-MgCl2 salt coolant and operating with a salt inlet temperature of 550°C and salt outlet temperature of 720°C. Using the temperature fields from the model thermohydraulic analysis we determined the design life of the reference receiver assuming it was constructed from 740H seamless tubing and assuming high temperature design properties for 740H from literature data and, where required, naïve extrapolation. As experimental data is generated, these tentative design parameters will be updated to match the actual performance of 740H in elevated temperature service. The results of this design study are a rank ordering of potential design methods for two criteria: ease of use and design margin. The end goal of this project is to develop a creep-fatigue design method with adequate, but not over conservative, design margin that is straightforward to execute. Our receiver model is based on the preliminary design provided to DOE by Solar Reserve. Our analysis indicates a peak metal temperature of around 840° C for this design (comparable to Solar Reserve’s analysis, which shows a peak metal temperature of around 830° C). The Solar Reserve tube thickness is too thin to sustain creep-fatigue loading at this temperature. This means that the current trial design is not well-suited for evaluating different design methodologies, as we predict it will have a very short design life. We suggest that we modify our reference design with input from DOE in order to lower the maximum metal temperature to less than 800° C. We can then repeat these analyses and give a more meaningful comparison of different design methods before making a final selection.

14 SOLAR ENERGY↗
NASA NTRS1994

Some Problems and Solutions in Transferring Ecosystem Simulation Codes to Supercomputers

Many computer codes for the simulation of ecological systems have been developed in the last twenty-five years. This development took place initially on main-frame computers, then mini-computers, and more recently, on micro-computers and workstations. Recent recognition of ecosystem science as a High Performance Computing and Communications Program Grand Challenge area emphasizes supercomputers (both parallel and distributed systems) as the next set of tools for ecological simulation. Transferring ecosystem simulation codes to such systems is not a matter of simply compiling and executing existing code on the supercomputer since there are significant differences in the system architectures of sequential, scalar computers and parallel and/or vector supercomputers. To more appropriately match the application to the architecture (necessary to achieve reasonable performance), the parallelism (if it exists) of the original application must be exploited. We discuss our work in transferring a general grassland simulation model (developed on a VAX in the FORTRAN computer programming language) to a Cray Y-MP. We show the Cray shared-memory vector-architecture, and discuss our rationale for selecting the Cray. We describe porting the model to the Cray and executing and verifying a baseline version, and we discuss the changes we made to exploit the parallelism in the application and to improve code execution. As a result, the Cray executed the model 30 times faster than the VAX 11/785 and 10 times faster than a Sun 4 workstation. We achieved an additional speed-up of approximately 30 percent over the original Cray run by using the compiler's vectorizing capabilities and the machine's ability to put subroutines and functions "in-line" in the code. With the modifications, the code still runs at only about 5% of the Cray's peak speed because it makes ineffective use of the vector processing capabilities of the Cray. We conclude with a discussion and future plans.

Skiles, J. W.↗
DOE OSTI2024

MOOSE–Workbench Integration and MOOSE Meshing Capability Enhancements to Facilitate Inputs and Outputs for Multiphysics Modeling

The Multiphysics Object-Oriented Simulation Environment (MOOSE) is an open-source framework that supports many of the US Department of Energy’s (DOE’s) Nuclear Energy Advanced Modeling and Simulation (NEAMS) technical areas (TA). These TAs develop and use NEAMS physics and coupling modules in multiple ways to enable the research and development of complex physics models. In addition to the MOOSE framework, the NEAMS Workbench user interface provides a common analysis environment with user-interaction accelerators that streamline the tasks of model creation, review, execution, and output inspection. In FY 2024, objectives were realized in the MOOSE framework application development support and user-oriented improvements. Application development improvements support both developers and users with an expanded Reactor Module and Mesh System, stateful material property support for mortar contact, and customizable convergence criteria. Additionally, new user-oriented features were implemented in the MOOSE framework language server, including autocompletion snippets, definition from source and find reference navigations, and syntax overrides. Lastly, improvements were made to the input interpreter necessary to support the MOOSE language server and the NEAMS Workbench so that they can interact with syntactically incomplete user inputs. These improvements and more were intended to address stakeholder feedback and improve developer and user ability to conduct advanced nuclear energy modeling and simulation in support of DOE and industry needs.

97 - MATHEMATICS AND COMPUTING↗
DOE OSTI2024

MOOSE-Workbench integration and MOOSE meshing capability enhancements to facilitate inputs and outputs for multiphysics modeling

The Multiphysics Object-Oriented Simulation Environment (MOOSE) is an open-source framework that supports many of the US Department of Energy’s (DOE’s) Nuclear Energy Advanced Modeling and Simulation (NEAMS) technical areas (TA). These TAs develop and use NEAMS physics and coupling modules in multiple ways to enable the research and development of complex physics models. In addition to the MOOSE framework, the NEAMS Workbench user interface provides a common analysis environment with user-interaction accelerators that streamline the tasks of model creation, review, execution, and out put inspection. In FY 2024, objectives were realized in the MOOSE framework application development support and user-oriented improvements. Application development improvements support both developers and users with an expanded Reactor Module and Mesh System, stateful material property support for mortar contact, and customizable convergence criteria. Additionally, new user-oriented features were implemented in the MOOSE framework language server, including autocompletion snippets, definition from source and find reference navigations, and syntax overrides. Lastly, improvements were made to the input interpreter necessary to support the MOOSE language server and the NEAMS Workbench so that they can interact with syntactically incomplete user inputs. These improvements and more were intended to address stakeholder feedback and improve developer and user ability to conduct advanced nuclear energy modeling and simulation in support of DOE and industry needs.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2018

Performing fully parallel constraint logic programming on a quantum annealer

A quantum annealer exploits quantum effects to solve a particular type of optimization problem. The advantage of this specialized hardware is that it effectively considers all possible solutions in parallel, thereby potentially outperforming classical computing systems. However, despite quantum annealers having recently become commercially available, there are relatively few high-level programming models that target these devices. In this article, we show how to compile a subset of Prolog enhanced with support for constraint logic programming into a two-local Ising-model Hamiltonian suitable for execution on a quantum annealer. In particular, we describe the series of transformations one can apply to convert constraint logic programs expressed in Prolog into an executable form that bears virtually no resemblance to a classical machine model yet that evaluates the specified constraints in a fully parallel manner. We evaluate our efforts on a 1,095-qubit D-Wave 2X quantum annealer and describe the approach's associated capabilities and shortcomings.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS2013

Generalized Ocean Color Inversion Model for Retrieving Marine Inherent Optical Properties

Ocean color measured from satellites provides daily, global estimates of marine inherent optical properties (IOPs). Semi-analytical algorithms (SAAs) provide one mechanism for inverting the color of the water observed by the satellite into IOPs. While numerous SAAs exist, most are similarly constructed and few are appropriately parameterized for all water masses for all seasons. To initiate community-wide discussion of these limitations, NASA organized two workshops that deconstructed SAAs to identify similarities and uniqueness and to progress toward consensus on a unified SAA. This effort resulted in the development of the generalized IOP (GIOP) model software that allows for the construction of different SAAs at runtime by selection from an assortment of model parameterizations. As such, GIOP permits isolation and evaluation of specific modeling assumptions, construction of SAAs, development of regionally tuned SAAs, and execution of ensemble inversion modeling. Working groups associated with the workshops proposed a preliminary default configuration for GIOP (GIOP-DC), with alternative model parameterizations and features defined for subsequent evaluation. In this paper, we: (1) describe the theoretical basis of GIOP; (2) present GIOP-DC and verify its comparable performance to other popular SAAs using both in situ and synthetic data sets; and, (3) quantify the sensitivities of their output to their parameterization. We use the latter to develop a hierarchical sensitivity of SAAs to various model parameterizations, to identify components of SAAs that merit focus in future research, and to provide material for discussion on algorithm uncertainties and future ensemble applications.

Werdell, P. Jeremy↗
NASA NTRS2010

Improving the Representation of Snow Crystal Properties with a Single-Moment Mircophysics Scheme

Single-moment microphysics schemes are utilized in an increasing number of applications and are widely available within numerical modeling packages, often executed in near real-time to aid in the issuance of weather forecasts and advisories. In order to simulate cloud microphysical and precipitation processes, a number of assumptions are made within these schemes. Snow crystals are often assumed to be spherical and of uniform density, and their size distribution intercept may be fixed to simplify calculation of the remaining parameters. Recently, the Canadian CloudSat/CALIPSO Validation Project (C3VP) provided aircraft observations of snow crystal size distributions and environmental state variables, sampling widespread snowfall associated with a passing extratropical cyclone on 22 January 2007. Aircraft instrumentation was supplemented by comparable surface estimations and sampling by two radars: the C-band, dual-polarimetric radar in King City, Ontario and the NASA CloudSat 94 GHz Cloud Profiling Radar. As radar systems respond to both hydrometeor mass and size distribution, they provide value when assessing the accuracy of cloud characteristics as simulated by a forecast model. However, simulation of the 94 GHz radar signal requires special attention, as radar backscatter is sensitive to the assumed crystal shape. Observations obtained during the 22 January 2007 event are used to validate assumptions of density and size distribution within the NASA Goddard six-class single-moment microphysics scheme. Two high resolution forecasts are performed on a 9-3-1 km grid, with C3VP-based alternative parameterizations incorporated and examined for improvement. In order to apply the CloudSat 94 GHz radar to model validation, the single scattering characteristics of various crystal types are used and demonstrate that the assumption of Mie spheres is insufficient for representing CloudSat reflectivity derived from winter precipitation. Furthermore, snow density and size distribution characteristics are allowed to vary with height, based upon direct aircraft estimates obtained from C3VP data. These combinations improve the representation of modeled clouds versus their radar-observed counterparts, based on profiles and vertical distributions of reflectivity. These meteorological events are commonplace within the mid-latitude cold season and present a challenge to operational forecasters. This study focuses on one event, likely representative of others during the winter season, and aims to improve the representation of snow for use in future operational forecasts.

Molthan, Andrew L.↗
DOE OSTI2023

Hydropower Value Drivers

Conventional hydro resources generate the majority of their value by providing energy under most conditions, but the relative fraction of value generated by providing ancillary services and capacity increases with increasing penetration of resources with zero fuel costs. Pumped storage hydropower resources generate the majority of their value by providing capacity under most conditions, but the relative fraction of value generated by providing energy increases with increasing penetration of resources with zero fuel costs. The total value of conventional hydropower generally decreases in systems with increasing penetration of resources with zero fuel costs; this is largely due to the associated decrease in average energy prices. The total value of pumped storage generally increases in systems with increasing penetration of resources with zero fuel costs, largely due to opportunities to operate in pumping mode when energy prices are low or even negative. Energy storage representation must be enhanced to ensure that models accurately capture system value streams for these resources. Current power system models have a limited ability to capture the price dynamics of ancillary services, and it is still challenging to assess the role and magnitude of ancillary service value streams in future systems. Power systems are currently in a state of rapid and dramatic evolution due to a number of different factors, including the increasing penetration of variable renewable energy (VRE) sources, such as wind and solar, and battery energy storage systems (BESS). This evolution will change the way power systems are fundamentally planned and operated. Some of these changes may be incremental, while others may be more significant, but the result will likely be parallel evolution in the definition and requirement of different grid services and therefore a subsequent shift in their relative values. This report presents a framework developed to identify such system value drivers and quantify their relative impact on several different value streams, with a specific focus on implications for conventional hydropower and pumped storage hydropower (PSH) resources. This value drivers framework (VDF) encompasses five core analytical steps: 1. Identify potential drivers and develop scenarios, 2. Execute production cost models, 3. Calibrate prices, 4. Optimize hydropower operations, and 5. Quantify value drivers.

13 HYDRO ENERGY↗
DOE OSTI2014

Performance and Architecture Lab Modeling Tool

Analytical application performance models are critical for diagnosing performance-limiting resources, optimizing systems, and designing machines. Creating models, however, is difficult. Furthermore, models are frequently expressed in forms that are hard to distribute and validate. The Performance and Architecture Lab Modeling tool, or Palm, is a modeling tool designed to make application modeling easier. Palm provides a source code modeling annotation language. Not only does the modeling language divide the modeling task into sub problems, it formally links an application's source code with its model. This link is important because a model's purpose is to capture application behavior. Furthermore, this link makes it possible to define rules for generating models according to source code organization. Palm generates hierarchical models according to well-defined rules. Given an application, a set of annotations, and a representative execution environment, Palm will generate the same model. A generated model is a an executable program whose constituent parts directly correspond to the modeled application. Palm generates models by combining top-down (human-provided) semantic insight with bottom-up static and dynamic analysis. A model's hierarchy is defined by static and dynamic source code structure. Because Palm coordinates models and source code, Palm's models are 'first-class' and reproducible. Palm automates common modeling tasks. For instance, Palm incorporates measurements to focus attention, represent constant behavior, and validate models. Palm's workflow is as follows. The workflow's input is source code annotated with Palm modeling annotations. The most important annotation models an instance of a block of code. Given annotated source code, the Palm Compiler produces executables and the Palm Monitor collects a representative performance profile. The Palm Generator synthesizes a model based on the static and dynamic mapping of annotations to program behavior. The model -- an executable program -- is a hierarchical composition of annotation functions, synthesized functions, statistics for runtime values, and performance measurements.

Tallent, Nathan↗
DOE OSTI2019

Heterogeneous Hierarchical Workflow Composition

Workflow systems promise scientists an automated end-to-end path from hypothesis to discovery. However, expecting any single workflow system to deliver such a wide range of capabilities is impractical. A more practical solution is to compose the end-to-end workflow from more than one system. With this goal in mind, the integration of task-based and in situ workflows is explored, where the result is a hierarchical heterogeneous workflow composed of subworkflows, with different levels of the hierarchy using different programming, execution, and data models. Materials science use cases demonstrate the advantages of such heterogeneous hierarchical workflow composition.

distributed workflows↗
DOE OSTI2016

Modeling a Million-Node Slim Fly Network Using Parallel Discrete-Event Simulation

As supercomputers close in on exascale performance, the increased number of processors and processing power translates to an increased demand on the underlying network interconnect. The Slim Fly network topology, a new lowdiameter and low-latency interconnection network, is gaining interest as one possible solution for next-generation supercomputing interconnect systems. In this paper, we present a high-fidelity Slim Fly it-level model leveraging the Rensselaer Optimistic Simulation System (ROSS) and Co-Design of Exascale Storage (CODES) frameworks. We validate our Slim Fly model with the Kathareios et al. Slim Fly model results provided at moderately sized network scales. We further scale the model size up to n unprecedented 1 million compute nodes; and through visualization of network simulation metrics such as link bandwidth, packet latency, and port occupancy, we get an insight into the network behavior at the million-node scale. We also show linear strong scaling of the Slim Fly model on an Intel cluster achieving a peak event rate of 36 million events per second using 128 MPI tasks to process 7 billion events. Detailed analysis of the underlying discrete-event simulation performance shows that a million-node Slim Fly model simulation can execute in 198 seconds on the Intel cluster.

interconnection networks↗
DOE OSTI2025

Lowering and Runtime Support for Fortran’s Multi-Image Parallel Features using LLVM Flang, PRIF, and Caffeine

This paper provides an overview of the multi-image parallel features in Fortran 2023 and their implementation in the LLVM flang compiler and the Caffeine parallel runtime library. The features of interest support a Single-Program, Multiple-Data (SPMD) programming model based on executing multiple “images”, each of which is a program instance. The features also support a Partitioned Global Address Space (PGAS) in the form of “coarray” distributed data structures. The paper discusses the lowering of multi-image features to the Parallel Runtime Interface for Fortran (PRIF) and the implementation of PRIF in the Caffeine parallel runtime library. This paper also provides an early view into the design of a new multi-image dialect of the LLVM Multi-Level Intermediate Representation (MLIR). We describe validation and testing of the resulting software stack, and demonstrate that performance compares favorably to another open-source compiler and runtime library: GNU Compiler Collection (GCC) gfortran and OpenCoarrays, respectively.

Bonachea, Dan↗
DOE OSTI2023

Abisko: Deep codesign of an architecture for spiking neural networks using novel neuromorphic materials

The Abisko project aims to develop an energy-efficient spiking neural network (SNN) computing architecture and software system capable of autonomous learning and operation. The SNN architecture explores novel neuromorphic devices that are based on resistive-switching materials, such as memristors and electrochemical RAM. Equally important, Abisko uses a deep codesign approach to pursue this goal by engaging experts from across the entire range of disciplines: materials, devices and circuits, architectures and integration, software, and algorithms. Here, the key objectives of our Abisko project are threefold. First, we are designing an energy-optimized high-performance neuromorphic accelerator based on SNNs. This architecture is being designed as a chiplet that can be deployed in contemporary computer architectures and we are investigating novel neuromorphic materials to improve its design. Second, we are concurrently developing a productive software stack for the neuromorphic accelerator that will also be portable to other architectures, such as field-programmable gate arrays and GPUs. Third, we are creating a new deep codesign methodology and framework for developing clear interfaces, requirements, and metrics between each level of abstraction to enable the system design to be explored and implemented interchangeably with execution, measurement, a model, or simulation. As a motivating application for this codesign effort, we target the use of SNNs for an analog event detector for a high-energy physics sensor.

97 MATHEMATICS AND COMPUTING↗