SEARCH · Search NASA

Results for “execution model”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,351 records · Page 75

All sources NASA NTRS DOE OSTI

DOE OSTI2017

Lawrence Berkeley National Laboratory 2016 Annual Financial Report

FY2016 was a year of significant change and progress at Berkeley Lab. In March, Laboratory Director Michael Witherell assumed his new role when former Lab Director Paul Alivisatos became Vice Chancellor for Research at UC Berkeley. Dr. Witherell has solidified the Lab’s strategy, with a focus on long term science and technology priorities. Large-scale science efforts continued to expand at the Lab, including the Dark Energy Spectroscopic Instrument now heading towards construction, and the LUX-ZEPLIN dark matter detector to be built underground in South Dakota. Another proposed project, the Advanced Light Source-Upgrade, was given preliminary approval and will be the Lab’s largest scientific investment in years. Construction of the Integrative Genomics Building began, and will bring together researchers from the Lab’s Joint Genome Institute, now based in Walnut Creek, and the Systems Biology Knowledgebase (K-Base) under one roof. Investment in the Lab’s infrastructure also continues, informed by the Lab’s Infrastructure Strategic Plan. Another important focus is on developing the next generation of scientists with the talent and diversity needed to sustain Berkeley Lab’s scientific leadership and mission contributions to DOE and the Nation. Berkeley Lab received $897.5M in new FY2016 funding, a 12.5% increase over FY2015, for both programmatic and infrastructure activities. While the Laboratory experienced a substantial increase in funding, it was accompanied by only a modest increase in spending, as areas of growth were partially offset by the completion of several major efforts in FY2015. FY2016 costs were $826.9M, an increase of 1.9% over FY2015. Similar to the prior year, the indirect-funded Operations units worked with generally flat budgets to yield more funding for strategic needs. A key challenge for Berkeley Lab continues to be achieving the best balance to fund essential investments, deliver highly effective operational mission support and remain cost-competitive. Through a comprehensive approach to prioritize competing needs, the Lab ended the year in a favorable financial position. The Office of the Chief Financial Officer (OCFO) played a key role in providing analysis and decision support to Executive Leadership, enabling the Lab to enhance its financial management strategies. In FY2016, the OCFO updated its analytic approaches and models to enhance long term financial projections under various funding and investment scenarios, and to assess total cost of ownership for major proposed investments. These improvements provided the new Lab Director and Senior Leadership with more comprehensive information and analytic support for planning and prioritization efforts. Within the OCFO, we focused on core operations and key initiatives defined in our OCFO Strategic Roadmap. The Lab’s Financial System transitioned from stabilization to optimization, with a focus on expanding the financial reporting capabilities considerably. We completed implementation of the eCommerce platform, achieving a notable outcome for the Lab in close partnership with DOE’s Office of Science. In other accomplishments, we launched a financial literacy program to enable Lab managers and staff to understand and execute their financial management and stewardship responsibilities more effectively; made substantial progress in enhancing our Field Finance model that provides financial support to client divisions and areas; developed a business process governance model to define OCFO business processes, clarify roles, and strengthen service delivery; and implemented a Partners in Leadership training program to build leadership capacity among our staff. We completed a ‘refresh’ of our Strategic Roadmap, which now defines our priorities for FY2017-FY2019. As a part of this effort, we made a subtle but important change to the OCFO mission statement to call out the Lab’s research and stewardship mission to sustain the Lab’s science and technology capabilities now and into the future. Berkeley Lab’s FY2016 progress on all fronts - scientific, operations, and financial management – position the Lab to continue bringing science solutions to the world as we charge into the 21st Century.

Source record↗

NASA NTRS1980

Interactive display of molecular models using a microcomputer system

A simple, microcomputer-based, interactive graphics display system has been developed for the presentation of perspective views of wire frame molecular models. The display system is based on a TERAK 8510a graphics computer system with a display unit consisting of microprocessor, television display and keyboard subsystems. The operating system includes a screen editor, file manager, PASCAL and BASIC compilers and command options for linking and executing programs. The graphics program, written in USCD PASCAL, involves the centering of the coordinate system, the transformation of centered model coordinates into homogeneous coordinates, the construction of a viewing transformation matrix to operate on the coordinates, clipping invisible points, perspective transformation and scaling to screen coordinates; commands available include ZOOM, ROTATE, RESET, and CHANGEVIEW. Data file structure was chosen to minimize the amount of disk storage space. Despite the inherent slowness of the system, its low cost and flexibility suggests general applicability.

Egan, J. T.↗

DOE OSTI2023

An automated and portable method for selecting an optimal GPU frequency

Power consumption poses a significant challenge in current and emerging graphics processing unit (GPU) enabled high-performance computing systems. In modern GPUs, dynamic voltage frequency scaling (DVFS) appears to be a reliable control to regulate power consumption and performance. However, the DVFS design space is large - hence, brute-force approaches are infeasible to select the optimal frequency. Furthermore, no single frequency can be universally optimal for applications with varying computational intensities. Thus, the application's complexity and the availability of a wide range of frequency settings are a challenge in selecting the optimal frequency configuration for a given GPU workload. To that end, this paper proposes a systematic approach that consists of three steps. The feature characterization study identifies the fine-grain GPU utilization metrics that influence the power consumption and execution time of a given workload. To understand the performance, power, and energy consumption behaviors of a workload across GPU's DVFS design space, we derived analytical power and performance models using the identified fine-grain features. Here, it is shown that the same set of GPU utilization metrics can estimate both the power consumption and execution time while being agnostic of changes to frequency and input sizes. Applying a power control with the single objective of reducing power may cause performance degradation, leading to more energy consumption. A multi-objective approach is proposed to select the optimal GPU DVFS configuration for a workload that reduces power consumption with negligible degradation in performance. The evaluation was conducted using SPEC ACCEL benchmarks and three real applications - NAMD LAMMPS, and LSTM on NVIDIA GV100, GA100, and AMD MI210 GPUs. On average, real applications showed 29.6% energy savings with a performance loss of 5.2% on GA100 and 22.6% energy savings with a performance loss of 4.7% on GV100. Moreover, the proposed models are portable to real applications, GPU architectures, and vendors, and require metric collection at only the default frequency rather than all supported DVFS configurations. Additionally, we conducted a comparison between our models and the GPU assembly instructions (PTX)-based static models. The results revealed a significant reduction in the average error rates, with a decrease from 19.7% to 3.1% for power models and from 29.4% to 5.2% for performance models.

97 MATHEMATICS AND COMPUTING↗

NASA NTRS2010

Systems Engineering Model and Training Application for Desktop Environment

Provide a graphical user interface based simulator for desktop training, operations and procedure development and system reference. This simulator allows for engineers to train and further understand the dynamics of their system from their local desktops. It allows the users to train and evaluate their system at a pace and skill level based on the user's competency and from a perspective based on the user's need. The simulator will not require any special resources to execute and should generally be available for use. The interface is based on a concept of presenting the model of the system in ways that best suits the user's application or training needs. The three levels of views are Component View, the System View (overall system), and the Console View (monitor). These views are portals into a single model, so changing the model from one view or from a model manager Graphical User Interface will be reflected on all other views.

May, Jeffrey T.↗

NASA NTRS1993

Space Shuttle Main Engine real time stability analysis

The Space Shuttle Main Engine (SSME) is a reusable, high performance, liquid rocket engine with variable thrust. The engine control system continuously monitors the engine parameters and issues propellant valve control signals in accordance with the thrust and mixture ratio commands. A real time engine simulation lab was installed at MSFC to verify flight software and to perform engine dynamic analysis. A real time engine model was developed on the AD100 computer system. This model provides sufficient fidelity on the dynamics of major engine components and yet simplified enough to be executed in real time. The hardware-in-the-loop type simulation and analysis becomes necessary as NASA is continuously improving the SSME technology, some with significant changes in the dynamics of the engine. The many issues of interfaces between new components and the engine can be better understood and be resolved prior to the firing of the engine. In this paper, the SSME real time simulation Lab at the MSFC, the SSME real time model, SSME engine and control system stability analysis, both in real time and non-real time is presented.

Kuo, F. Y.↗

DOE OSTI2020

OpenABLext: An automatic code generation framework for agent-based simulations on CPU-GPU-FPGA heterogeneous platforms

The execution of agent-based simulations (ABSs) on hardware accelerator devices such as graphics processing units (GPUs) has been shown to offer great performance potentials. However, in heterogeneous hardware environments, it can become increasingly difficult to find viable partitions of the simulation and provide implementations for different hardware devices. To automate this process, we present OpenABLext, an extension to OpenABL, a model specification language for ABSs. By providing a device-aware OpenCL backend, OpenABLext enables the co-execution of ABS on heterogeneous hardware platforms consisting of central processing units, GPUs, and field programmable gate arrays (FPGAs).We present a novel online dispatching method that efficiently profiles partitions of the simulation during run-time to optimize the hardware assignment while using the profiling results to advance the simulation itself. In addition, OpenABLext features automated conflict resolution based on user-specified rules, supports graph-based simulation spaces, and utilizes an efficient neighbor search algorithm. We show the improved performance of OpenABLext and demonstrate the potential of FPGAs in the context of ABS. We illustrate how co-execution can be used to further lower execution times. OpenABLext can be seen as an enabler to tap the computing power of heterogeneous hardware platforms for ABS.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2025

Simplifying activations with linear approximations in neural networks

A key step in Neural Networks is activation. Among the different types of activation functions, sigmoid, tanh, and others involve the usage of exponents for calculation. From a hardware perspective, exponential implementation implies the usage of Taylor series or repeated methods involving many addition, multiplication, and division steps, and as a result are power-hungry and consume many clock cycles. We implement a piecewise linear approximation of the sigmoid function as a replacement for standard sigmoid activation libraries. This approach provides a practical alternative by leveraging piecewise segmentation, which simplifies hardware implementation and improves computational efficiency. In this paper, we detail piecewise functions that can be implemented using linear approximations and their implications for overall model accuracy and performance gain. Our results show that for the DenseNet, ResNet, and GoogLeNet architectures, the piecewise linear approximation of the sigmoid function provides faster execution times compared to the standard TensorFlow sigmoid implementation while maintaining comparable accuracy. Specifically, for MNIST with DenseNet, accuracy reaches 99.91% (Piecewise) vs. 99.97% (Base) with up to 1.31x speedup in execution time. For CIFAR-10 with DenseNet, accuracy improves to 98.97% (Piecewise) vs. 99.40% (Base) while achieving 1.24x faster execution. Similarly, for CIFAR-100 with DenseNet, the accuracy is 97.93% (Piecewise) vs. 98.39% (Base), with a 1.18x execution time reduction. These results confirm the proposed method’s capability to efficiently process large-scale datasets and computationally demanding tasks, offering a practical means to accelerate deep learning models, including LSTMs, without compromising accuracy.

Activation function↗

NASA NTRS1991

User modeling techniques for enhanced usability of OPSMODEL operations simulation software

The PC based OPSMODEL operations software for modeling and simulation of space station crew activities supports engineering and cost analyses and operations planning. Using top-down modeling, the level of detail required in the data base can be limited to being commensurate with the results required of any particular analysis. To perform a simulation, a resource environment consisting of locations, crew definition, equipment, and consumables is first defined. Activities to be simulated are then defined as operations and scheduled as desired. These operations are defined within a 1000 level priority structure. The simulation on OPSMODEL, then, consists of the following: user defined, user scheduled operations executing within an environment of user defined resource and priority constraints. Techniques for prioritizing operations to realistically model a representative daily scenario of on-orbit space station crew activities are discussed. The large number of priority levels allows priorities to be assigned commensurate with the detail necessary for a given simulation. Several techniques for realistic modeling of day-to-day work carryover are also addressed.

Davis, William T.↗

DOE OSTI2021

PARIS: Predicting application resilience using machine learning

The traditional method to study application resilience to errors in HPC applications uses fault injection (FI), a time-consuming approach. Furthermore, while analytical models have been built to overcome the inefficiencies of FI, they lack accuracy. In this paper, we present PARIS, a machine-learning method to predict application resilience that avoids the time-consuming process of random FI and provides higher prediction accuracy than analytical models. PARIS captures the implicit relationship between application characteristics and application resilience, which is difficult to capture using most analytical models. We overcome many technical challenges for feature construction, extraction, and selection to use machine learning in our prediction approach. Our evaluation on 16 HPC benchmarks shows that PARIS achieves high prediction accuracy. PARIS is up to 450x faster than random FI (49x on average). Compared to the state-of-the-art analytical model, PARIS is at least 63% better in terms of accuracy and has comparable execution time on average.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2017

Integrating prediction, provenance, and optimization into high energy workflows

We propose a novel approach for efficient execution of workflows on distributed resources. The key components of this framework include: performance modeling to quantitatively predict workflow component behavior; optimization-based scheduling such as choosing an optimal subset of resources to meet demand and assignment of tasks to resources; distributed I/O optimizations such as prefetching; and provenance methods for collecting performance data. In preliminary results, these techniques improve throughput on a small Belle II workflow by 20%.

Schram, M.↗

DOE OSTI2025

Digital quantum simulation of cavity quantum electrodynamics: insights from superconducting and trapped ion quantum testbeds

We explore the potential for hybrid development of quantum hardware where currently available quantum computers simulate open cavity quantum electrodynamical (CQED) systems for applications in optical quantum communication, simulation and computing. Our simulations make use of a recent quantum algorithm that maps the dynamics of a singly excited open Tavis–Cummings model containing N atoms coupled to a lossy cavity. We report the results of executing this algorithm on two noisy intermediate-scale quantum computers: a superconducting processor and a trapped ion processor, to simulate the population dynamics of an open CQED system featuring N = 3 atoms. By applying technology-specific transpilation and error mitigation techniques, we minimize the impact of gate errors, noise, and decoherence in each hardware platform, obtaining results which agree closely with the exact solution of the system. These results can be used as a recipe for efficient and platform-specific quantum simulation of cavity–emitter systems on contemporary and future quantum computers.

cavity QED↗

DOE OSTI2016

Arion

Arion is a library and tool set that enables researchers to holistically define test system models. To define a complex system for testing an algorithm or control requires expertise across multiple domains. Simulating a complex system requires the integration of multiple simulators and test hardware, each with their own specification languages and concepts. This requires extensive set of knowledge and capabilities. Arion was developed to alleviate this challenge. Arion is a library of Java libraries that abstracts the concepts from supported simulators into a cohesive model language that allows someone to build models to their needed level of fidelity and expertise. Arion is also a software tool that translates the users model back into the specification languages of the simulators and test hardware needed for execution.

Fisher, Andy↗

DOE OSTI2019

Acceleration of tensor-product operations for high-order finite element methods

This article is devoted to graphics processing unit (GPU) kernel optimization and performance analysis of three tensor-product operations arising in finite element methods. We provide a mathematical background to these operations and implementation details. Achieving close to peak performance for these operators requires extensive optimization because of the operators’ properties: low arithmetic intensity, tiered structure, and the need to store intermediate results during the kernel execution. We give a guided overview of optimization strategies and we present a performance model that allows us to compare the efficacy of these optimizations against an empirically calibrated roofline.

Świrydowicz, Kasia↗

DOE OSTI2015

eXascale PRogramming Environment and System Software (XPRESS)

Exascale systems, with a thousand times the compute capacity of today’s leading edge petascale computers, are expected to emerge during the next decade. Their software systems will need to facilitate the exploitation of exceptional amounts of concurrency in applications, and ensure that jobs continue to run despite the occurrence of system failures and other kinds of hard and soft errors. Adapting computations at runtime to cope with changes in the execution environment, as well as to improve power and performance characteristics, is likely to become the norm. As a result, considerable innovation is required to develop system support to meet the needs of future computing platforms. The XPRESS project aims to develop and prototype a revolutionary software system for extreme-scale computing for both exascale and strongscaled problems. The XPRESS collaborative research project will advance the state-of-the-art in high performance computing and enable exascale computing for current and future DOE mission-critical applications and supporting systems. The goals of the XPRESS research project are to: A. enable exascale performance capability for DOE applications, both current and future, B. develop and deliver a practical computing system software X-stack, OpenX, for future practical DOE exascale computing systems, and C. provide programming methods and environments for effective means of expressing application and system software for portable exascale system execution.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2021

Creating an MPAS Ocean Shallow Water Core in Julia [Slides]

An MPAS model was created in Julia to test its potential for scientific HPC. Tradeoff between execution speed and development speed: Development languages (e.g. Python) are easy, but slow and production languages (e.g. C) are hard, but fast. Julia aims to be the best of both.

58 GEOSCIENCES↗

DOE OSTI2024

Best of Both Worlds: Integrating Slurm with Kubernetes in a Kubernetes Native Way

We present K-Foundry, a framework that enables the integration of Simple Linux Utility for Resource Management (SLURM) with Kubernetes (K8s) via a Kubernetes-like Control Plane (KCP). Our implementation seamlessly enables a unified communication and scheduling layer for a fleet of multiple diverse computing platforms. While SLURM and K8s traditionally support distinct scheduling models, they have recently started aligning their objectives by moving towards a more converged execution environment.

Beltre, Angel Manuel↗

DOE OSTI2017

VERA 3.6 Release

CASL has developed a formalized process to release the Virtual Environment for Reactor Applications (VERA) to non-CASL users through the Radiation Shielding Information Computational Center (RSICC). This milestone represents a major release of VERA with updates that significantly reduce computational resources required and introduce new capabilities related to fuel performance integration. This release will also be used by CASL Test Stands and training activities. VERA 3.6 represents an incremental release, with improvements to core component capabilities, significant performance improvements, and updated documentation. This release followed a significantly revised release process, with formal documentation, testing, readiness reviews, and release authorization. The major tasks that were completed as part of this milestone were: 1) Revision of the VERA RSICC Release Process and Procedures; 2) Identification of the VERA release components; 3) Creation of a Release Candidate; 4) Independent evaluation of the Release Candidate by the Advanced Modeling Applications (AMA) Focus Area; 5) A Release Candidate Readiness Review; 6) Development and execution of a System Test Plan; 7) Updates of manuals and other required documentation; 8) Export Control reviews; 9) A Release Readiness Review; and 10) Delivery of the release package to RSICC. Synopses of the work done to complete each of these major tasks are summarized in Section 2 below. Supporting details are provided in the PHI Kanban ticket #4424 Support VERA 3.6 Release Activities and its child tickets.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

DOE OSTI2019

A Robust and Efficient Deep Learning Method for Dynamical Mass Measurements of Galaxy Clusters

We demonstrate the ability of convolutional neural networks (CNNs) to mitigate systematics in the virial scaling relation and produce dynamical mass estimates of galaxy clusters with remarkably low bias and scatter. We present two models, CNN 1D and CNN 2D , which leverage this deep learning tool to infer cluster masses from distributions of member galaxy dynamics. Our first model, CNN 1D , infers cluster mass directly from the distribution of member galaxy line-of-sight velocities. Our second model, CNN 2D , extends the input space ofCNN 1D to learn on the joint distribution of galaxy line-of-sight velocities and projected radial distances. We train each model as a regression over cluster mass using a labeled catalog of realistic mock cluster observations generated from the MultiDark simulation and UniverseMachine catalog. We then evaluate the performance of each model on an independent set of mock observations selected from the same simulated catalog. The CNN models produce cluster mass predictions with lognormal residuals of scatter as low as 0.132 dex, greater than a factor of 2 improvement over the classical M–σ power-law estimator. Furthermore, the CNN model reduces prediction scatter relative to similar machinelearning approaches by up to 17% while executing in drastically shorter training and evaluation times (by a factor of 30) and producing considerably more robust mass predictions (improving prediction stability under variations in galaxy sampling rate by 30%).

79 ASTRONOMY AND ASTROPHYSICS↗