SEARCH · Search NASA

Results for “architecture optimization”

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 739 records · Page 41

All sources NASA NTRS DOE OSTI

NASA NTRSDate not supplied

Implicit Thermochemical Nonequilibrium Flow Simulations on Unstructured Grids using GPUs

Thermochemical nonequilibrium flow simulation capabilities have been previously implemented, verified, and validated for central processing unit (CPU) systems in NASA’s unstructured-grid computational fluid dynamics solver FUN3D. Many exascale-class high-performance computing systems will rely on graphics processing unit (GPU) architectures for high throughput and energy efficiency; thus, CPU-based scientific computing software unable to effectively utilize these systems must be updated. In this work, we present a CUDA C++ implementation of FUN3D’s thermochemical nonequilibrium flow simulation capabilities targeting NVIDIA Tesla GPUs. An overview of the porting and optimization strategy is described and performance comparisons with other recent architectures are presented. Scaling to thousands of GPUs is demonstrated, yielding computational performance equivalent to that of several million CPU cores. The implementation enables efficient, high-fidelity, scale-resolving simulations of thermochemical nonequilibrium flows for many applications including atmospheric entry, hypersonics, and combustion.

GPU↗

NASA NTRSDate not supplied

Use NASA Earthdata in the Cloud

With the impending arrival of new, high-data-volume Earth observation missions, NASA’s ability to effectively ingest, process, and archive large amounts of data requires the most cost-effective, flexible, and scalable data-management architectures and technologies. To meet these demands, NASA's Earth Science Data Systems (ESDS) Program is implementing a strategic vision to develop and operate multiple components of NASA's Earth Observing System Data and Information System (EOSDIS) in a commercial cloud environment. As more earth observing data is moved into the cloud, NASA EOSDIS wants to share our reasoning for moving data into the commercial cloud, as well as demonstrate how end-users can take advantage of both data in the cloud and cloud-deployed services. This presentation explains the enterprise reasons for moving data into the cloud, specifically the scalability of cloud systems, the flexibility to pivot to the most optimal language/system/architecture, and the reduced redundancy of a common architecture. Then we primarily focus on the user-driven reasons for moving to the cloud. Equal access to high performance computing helps all parties, especially users that don't have access to their own on-premise computing infrastructure. Rapid deployment allows users to spin up their own systems without needing the extensive platform previously required for large scale development. Cost effectiveness allows users a wide breadth of computing options, from super computer ability to small scale development, which can help underserved communities. Finally we link to internal resources that users can utilize to begin their cloud based development journey without having to be an expert in the field.

Nicholas Doty↗

DOE OSTI2021

Agent-Based Distributed Energy Resources for Supporting Intelligence at the Grid Edge

This article proposes a novel multi-agent framework that can link various forms of resources and power electronic systems into distributed energy resources (DER). The proposed multiagent architecture can also integrate DERs to a central controller for optimization and control to support the grid. To demonstrate the flexibility of this novel framework, the developed agent system is applied to a set of end-use systems. Furthermore, the agent framework is validated in hardware using controller-hardware-in-the-loop simulation platform.

24 POWER TRANSMISSION AND DISTRIBUTION↗

DOE OSTI2025

Enabling Efficient Sparse Computations using Linear Algebra Aware Compilers

This project developed the LAPIS compiler framework, built on the Multilevel Intermediate Representation (MLIR), to optimize sparse linear algebra operations and support performance portability across diverse architectures. The main innovation of LAPIS is the Kokkos dialect, which allows for lowering codes from a high productivity language to different architectures in an elegant way. The dialect also allows the conversion of lower-level MLIR code to C++ Kokkos code, facilitating the integration of scientific machine learning (SciML) models into applications. To extend LAPIS for distributed memory architectures, a new partition dialect was created to manage the distribution of sparse tensors and express communication patterns for sparse linear algebra operations. This dialect also supports the distributed execution of operators and includes algorithmic optimizations to minimize communication to improve performance. The project also demonstrates that MLIR can enable effective linear algebra-level optimizations, improving performance on different GPUs for both sparse and dense linear algebra kernels. Key applications of LAPIS include sparse linear algebra and graph kernels, TenSQL, a relational database management solution built on GraphBLAS, and the development of subgraph isomorphism and monomorphism kernels, showcasing performance portability. In summary, the LAPIS framework supports productivity, performance, portability, and distributed memory execution, while also enabling linear algebra-level optimizations that are challenging in traditional programming languages, with successful applications ranging from simple sparse linear algebra to complex graph kernels.

97 MATHEMATICS AND COMPUTING↗

NASA NTRS1992

Optical systems integrated modeling

An integrated modeling capability that provides the tools by which entire optical systems and instruments can be simulated and optimized is a key technology development, applicable to all mission classes, especially astrophysics. Many of the future missions require optical systems that are physically much larger than anything flown before and yet must retain the characteristic sub-micron diffraction limited wavefront accuracy of their smaller precursors. It is no longer feasible to follow the path of 'cut and test' development; the sheer scale of these systems precludes many of the older techniques that rely upon ground evaluation of full size engineering units. The ability to accurately model (by computer) and optimize the entire flight system's integrated structural, thermal, and dynamic characteristics is essential. Two distinct integrated modeling capabilities are required. These are an initial design capability and a detailed design and optimization system. The content of an initial design package is shown. It would be a modular, workstation based code which allows preliminary integrated system analysis and trade studies to be carried out quickly by a single engineer or a small design team. A simple concept for a detailed design and optimization system is shown. This is a linkage of interface architecture that allows efficient interchange of information between existing large specialized optical, control, thermal, and structural design codes. The computing environment would be a network of large mainframe machines and its users would be project level design teams. More advanced concepts for detailed design systems would support interaction between modules and automated optimization of the entire system. Technology assessment and development plans for integrated package for initial design, interface development for detailed optimization, validation, and modeling research are presented.

Shannon, Robert R.↗

DOE OSTI2021

Evaluating Spatial Accelerator Architectures with Tiled Matrix-Matrix Multiplication.

There is a growing interest in custom spatial accelerators for machine learning applications. These accelerators employ a spatial array of processing elements (PEs) interacting via custom buffer hierarchies and networks-on-chip. The efficiency of these accelerators comes from employing optimized dataflow (i.e., spatial/temporal partitioning of data across the PEs and fine-grained scheduling) strategies to optimize data reuse. The focus of this work is to evaluate these accelerator architectures using a tiled general matrix-matrix multiplication (GEMM) kernel. To do so, we develop a framework that finds optimized mappings (dataflow and tile sizes) for a tiled GEMM for a given spatial accelerator and workload combination, leveraging an analytical cost model for runtime and energy. Finally, our evaluations over five spatial accelerators demonstrate that the tiled GEMM mappings systematically generated by our framework achieve high performance on various GEMM workloads and accelerators.

42 ENGINEERING↗

DOE OSTI2021

Evaluating Spatial Accelerator Architectures with Tiled Matrix-Matrix Multiplication

There is a growing interest in custom spatial accelerators for machine learning applications. These accelerators employ a spatial array of processing elements (PEs) interacting via custom buffer hierarchies and networks-on-chip. The efficiency of these accelerators comes from employing optimized dataflow (i.e., spatial/temporal partitioning of data across the PEs and fine-grained scheduling) strategies to optimize data reuse. The focus of this work is to evaluate these accelerator architectures using a tiled general matrix-matrix multiplication (GEMM) kernel. To do so, we develop a framework that finds optimized mappings (dataflow and tile sizes) for a tiled GEMM for a given spatial accelerator and workload combination, leveraging an analytical cost model for runtime and energy. Our evaluations over five spatial accelerators demonstrate that the tiled GEMM mappings systematically generated by our framework achieve high performance on various GEMM workloads and accelerators.

43 PARTICLE ACCELERATORS↗

DOE OSTI2017

CMS readiness for multi-core workload scheduling

In the present run of the LHC, CMS data reconstruction and simulation algorithms benefit greatly from being executed as multiple threads running on several processor cores. The complexity of the Run 2 events requires parallelization of the code to reduce the memory-per- core footprint constraining serial execution programs, thus optimizing the exploitation of present multi-core processor architectures. The allocation of computing resources for multi-core tasks, however, becomes a complex problem in itself. The CMS workload submission infrastructure employs multi-slot partitionable pilots, built on HTCondor and GlideinWMS native features, to enable scheduling of single and multi-core jobs simultaneously. This provides a solution for the scheduling problem in a uniform way across grid sites running a diversity of gateways to compute resources and batch system technologies. This paper presents this strategy and the tools on which it has been implemented. The experience of managing multi-core resources at the Tier-0 and Tier-1 sites during 2015, along with the deployment phase to Tier-2 sites during early 2016 is reported. The process of performance monitoring and optimization to achieve efficient and flexible use of the resources is also described.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2023

XploreNAS : Explore Adversarially Robust and Hardware-efficient Neural Architectures for Non-ideal Xbars

Compute In-Memory platforms such as memristive crossbars are gaining focus as they facilitate acceleration of Deep Neural Networks (DNNs) with high area and compute efficiencies. However, the intrinsic non-idealities associated with the analog nature of computing in crossbars limits the performance of the deployed DNNs. Furthermore, DNNs are shown to be vulnerable to adversarial attacks leading to severe security threats in their large-scale deployment. Thus, finding adversarially robust DNN architectures for non-ideal crossbars is critical to the safe and secure deployment of DNNs on the edge. This work proposes a two-phase algorithm-hardware co-optimization approach called XploreNAS that searches for hardware efficient and adversarially robust neural architectures for non-ideal crossbar platforms. We use the one-shot Neural Architecture Search approach to train a large Supernet with crossbar-awareness and sample adversarially robust Subnets therefrom, maintaining competitive hardware efficiency. Our experiments on crossbars with benchmark datasets (SVHN, CIFAR10, CIFAR100) show up to ~8–16% improvement in the adversarial robustness of the searched Subnets against a baseline ResNet-18 model subjected to crossbar-aware adversarial training. We benchmark our robust Subnets for Energy-Delay-Area-Products (EDAPs) using the Neurosim tool and find that with additional hardware efficiency–driven optimizations, the Subnets attain ~1.5–1.6× lower EDAPs than ResNet-18 baseline.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2019

Graph-based modeling and simulation of complex systems

We present graph-based modeling abstractions to represent cyber-physical dependencies arising in complex systems. Specifically, we propose an algebraic graph abstraction to capture physical connectivity in complex optimization models and a computing graph abstraction to capture communication connectivity in computing architectures. The proposed abstractions are scalable and are used as the backbone of a Julia-based software package that we call Plasmo.j1. Here, we show how the algebraic graph abstraction facilitates the implementation, analysis, and decomposition of optimization problems and we show how the computing graph abstraction facilitates the implementation of optimization and control algorithms and their simulation in virtual environments that involve distributed, centralized, and hierarchical computing architectures.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2023

Multidisciplinary concurrent optimization framework for multi-phase building design process

Modern day building design projects require multidisciplinary expertise from architects and engineers across various phases of the design (conceptual, preliminary, and detailed) and construction processes. The Architecture Engineering and Construction (AEC) community has recently shifted gears toward leveraging design optimization techniques to make well-informed decisions in the design of buildings. However, most of the building design optimization efforts are either multidisciplinary optimization confined to just a specific design phase (conceptual/preliminary/detailed) or single disciplinary optimization (structural/thermal/daylighting/energy) spanning across multiple phases. Complexity in changing the optimization setup as the design progresses through subsequent phases, interoperability issues between modeling and physics-based analysis tools used at later stages, and the lack of an appropriate level of design detail to get meaningful results from these sophisticated analysis tools are few challenges that limit multi-phase multidisciplinary design optimization (MDO) in the AEC field. Here this paper proposes a computational building design platform leveraging concurrent engineering techniques such as interactive problem structuring, simulation-based optimization using meta models for energy and daylighting (machine learning based) and tradespace visualization. The proposed multi-phase concurrent MDO framework is demonstrated by using it to design and optimize a sample office building for energy and daylighting objectives across multiple phases. Furthermore, limitations of the proposed framework and future avenues of research are listed.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

DOE OSTI2019

Intelligent Reservoir Generation for Liquid State Machines using Evolutionary Optimization

Neuromorphic Computing is a burgeoning field of research. Many groups are exploring hardware architectures and theoretical ideas about spiking recurrent neural networks. The overarching goal is to exploit the low power promise of these neuromorphic systems. However, it is difficult to train spiking recurrent neural networks (SRNNs) to perform tasks and make efficient use of neuromorphic hardware. Reservoir Computing is an attractive methodology because it requires no tuning of weights for the reservoir itself. Yet, to find optimal reservoirs, manual tuning of hyperparameters such as hidden neurons, synaptic density, and natural structure is still required. Because of this, researchers often have to generate and evaluate many networks, which can result in non-trivial amounts of computation. This paper employs the reservoir computing technique (specifically liquid state machines) and genetic algorithms in order to develop useful networks that can be deployed on neuromorphic hardware. We build on past work in reservoir computing and genetic algorithms to demonstrate the power of combining these two techniques and the advantage it can provide over manually tuning reservoirs for use on classification tasks. We discuss the complexities of determining whether or not to use the genetic algorithms approach for liquid state machine generation.

Reynolds, John↗

NASA NTRS1991

Optimal processor assignment for pipeline computations

The availability of large scale multitasked parallel architectures introduces the following processor assignment problem for pipelined computations. Given a set of tasks and their precedence constraints, along with their experimentally determined individual responses times for different processor sizes, find an assignment of processor to tasks. Two objectives are of interest: minimal response given a throughput requirement, and maximal throughput given a response time requirement. These assignment problems differ considerably from the classical mapping problem in which several tasks share a processor; instead, it is assumed that a large number of processors are to be assigned to a relatively small number of tasks. Efficient assignment algorithms were developed for different classes of task structures. For a p processor system and a series parallel precedence graph with n constituent tasks, an O(np2) algorithm is provided that finds the optimal assignment for the response time optimization problem; it was found that the assignment optimizing the constrained throughput in O(np2log p) time. Special cases of linear, independent, and tree graphs are also considered.

Nicol, David M.↗

DOE OSTI2025

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

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

Rosas, Miguel Romero↗

NASA NTRS2021

Generation-based Evolutionary Tool for the Optimization of Constellations (GenETOC)

mission design↗

NASA NTRS2021

Generation-based Evolutionary Tool for the Optimization of Constellations (GenETOC)

With the rapid growth in the capabilities of smaller satellites, satellite architectures that replace a single, extremely capable spacecraft with multiple, cheaper ones are gaining in popularity. Unfortunately, the orbit design process for constellations can be significantly more involved, especiallywhen the relative placement of the individual spacecraft within the constellation is not constrained by mission and/or science objectives. Optimizing a satellite constellation in the presence of multiple, competing objectives is a highly complex problem to which many traditional mathematical optimization methods cannot be applied and few tools exist to help mission designers search for promising candidate mission designs. The Generation-based Evolutionary Tool for the Optimization of Constellations (GenETOC) has been created to search for near-optimal constellation design options. GenETOC combines a modified version of the Non-dominated Sorting Genetic Algorithm II (NSGA II) with STK Components libraries (a 3rdparty .NET package created by Analytical Graphics Inc.) to create a framework that enables a mission designer to generate a simulation that models the design problem and obtain a family of potential, near-optimal solutions that can be investigated more in detail. GenETOC was developed in C# using the .NET framework with Windows Presentation Foundation (WPF) serving as the framework from which to create the graphical user interface (GUI). GenETOC user inputs can be categorized into three major data components: definition of the problem (areas of interest, satellite decision parameters, and sensor configurations), definition of performance objectives, and specification of the genetic algorithm (GA) parameters. In the problem definition component, the user is prompted to define the areas of interest against which the performance metrics will be computed, define the sensor parameters and attach them to specific spacecraft, select which satellite orbital parameters will be added to the decision space of the GA, and specify the range of desired values for each optimization parameter. For performance objectives, the user is presented with a list of available coverage and revisit performance based calculation options from which two metrics are chosen to serve as the objective functions that the GA will use to evaluate solutions during the optimization process. Finally, the definition of the GA parameters provides user control over the number of generations (number of optimization iterations), the population size (number of candidate constellations created in each generation), and the adaptive mutation and crossover threshold values (control parameters for how frequently each process occurs during the optimization). GenETOC has been extensively tested to verify the individual components of the optimization process. The GA has been tested against a suite of GA test problems to confirm convergence to the known two and three-dimensional Pareto fronts. The coverage and revisit performance metrics obtained in GenETOC are compared with STK desktop scenarios, confirming the constellations are being appropriately modeled within GenETOC simulations. A walkthrough of a simple, example problem is provided to illustrate the workings of GenETOC and to demonstrate the output available to the mission designer.

mission design↗

NASA NTRS1989

The role of optimization in the next generation of computer-based design tools

There is a close relationship between design optimization and the emerging new generation of computer-based tools for engineering design. With some notable exceptions, the development of these new tools has not taken full advantage of recent advances in numerical design optimization theory and practice. Recent work in the field of design process architecture has included an assessment of the impact of next-generation computer-based design tools on the design process. These results are summarized, and insights into the role of optimization in a design process based on these next-generation tools are presented. An example problem has been worked out to illustrate the application of this technique. The example problem - layout of an aircraft main landing gear - is one that is simple enough to be solved by many other techniques. Although the mathematical relationships describing the objective function and constraints for the landing gear layout problem can be written explicitly and are quite straightforward, an approximation technique has been used in the solution of this problem that can just as easily be applied to integrate supportability or producibility assessments using theory of measurement techniques into the design decision-making process.

Rogan, J. Edward↗

DOE OSTI2019

Lynx X-Ray Observatory: an overview

Lynx, one of the four strategic mission concepts under study for the 2020 Astrophysics Decadal Survey, provides leaps in capability over previous and planned x-ray missions and provides synergistic observations in the 2030s to a multitude of space- and ground-based observatories across all wavelengths. Lynx provides orders of magnitude improvement in sensitivity, on-axis subarcsecond imaging with arcsecond angular resolution over a large field of view, and high-resolution spectroscopy for point-like and extended sources in the 0.2- to 10-keV range. The Lynx architecture enables a broad range of unique and compelling science to be carried out mainly through a General Observer Program. This program is envisioned to include detecting the very first seed black holes, revealing the high-energy drivers of galaxy formation and evolution, and characterizing the mechanisms that govern stellar evolution and stellar ecosystems. The Lynx optics and science instruments are carefully designed to optimize the science capability and, when combined, form an exciting architecture that utilizes relatively mature technologies for a cost that is compatible with the projected NASA Astrophysics budget.

79 ASTRONOMY AND ASTROPHYSICS↗