SEARCH · Search NASA

Results for “architecture”

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 127 records · Page 7

All sources NASA NTRS DOE OSTI

DOE OSTI2026

HARMONY: Large-Scale Architecture Search for Efficient Hybrid Language Models

As large language models scale to trillions of parameters, their computational and memory requirements present critical challenges for efficient training and deployment. While Mixture of Experts (MoE) architectures enable efficient scaling through sparse parameter activation, and state-space models like Mamba offer linear-time complexity, principled methods for combining these paradigms remain undeveloped. We introduce HARMONY (Hybrid Architecture Research for Mamba, Optimized with Neural efficiencY), a multi-objective evolutionary neural architecture search framework for discovering efficient hybrid language models that integrate Transformer attention mechanisms, Mixture-of-Experts routing, and Mamba state-space components. Through large-scale distributed search using 16,384 MI250X GPUs on the Frontier supercomputer, HARMONY explores a comprehensive design space encompassing six attention variants (MHA, MQA, GQA, MLA, SWA, and Mamba-2), variable MoE configurations with both routed and shared experts, and extensive Mamba hyperparameters. Our framework discovers heterogeneous architectures that balance training performance with computational efficiency through multi-objective optimization incorporating latency penalties and fitness-based selection. Analysis of discovered architectures reveals that optimal hybrid designs favor heterogeneous component mixing rather than homogeneous patterns, with Mamba-2 and Multi-Head Latent Attention (MLA) emerging as preferred mechanisms. Discovered architectures demonstrate superior training efficiency: our best configuration achieves a final perplexity of 1.0874 with 2.38B parameters while processing 4,320 tokens/second, outperforming significantly larger manually designed models. Full-scale evaluation shows HARMONY's top architectures achieve better loss trajectories than equivalently-sized models using state-of-the-art configurations including Mixtral, Jamba, and Samba. Additionally, we demonstrate 91% weak scaling efficiency when training discovered 36B-parameter models across 1,024 GPUs. HARMONY is released as an open framework with comprehensive tools for building and training hybrid models using expert-data-pipeline parallelism, democratizing access to automated architecture design for next-generation language models.

Herron, Emily [ORNL] (ORCID:0000000273008172)↗

DOE OSTI2025

Land-Based Wind Reference Architecture

This report describes a summary and the final deliverables for the Wind Reference Architecture project funded by the Department of Energy's (DOE) Wind Energy Technology Office (WETO). The project objective was to further refine the reference architecture of the existing wind power plant reference architecture developed by Idaho National Laboratory (INL) and Sandia National Laboratories (SNL) by expanding the wind turbine generator (WTG) into a separate individual wind turbine reference architecture. Additional details regarding the wind turbine system and how it integrates with a wind power plant were researched and reflected in the reference architecture. This addition is important because it allows researchers to understand the components and devices in a wind turbine, how they function and how a wind turbine integrates into a wind power plant to be able to perform cybersecurity evaluations on the system. The communication and control systems were the focus while developing this reference architecture to understand the cybersecurity posture of a wind turbine and power plant. The information technology (IT) and operational technology (OT) systems perspective are integral in helping improve the cybersecurity posture by creating a starting point to study how wind energy can be safely designed and deployed in our power grid from an integrated systems standpoint. A holistic wind power plant and turbine reference architecture and simulation was developed and made open-sourced for further research efforts.

17 WIND ENERGY↗

NASA NTRS2004

Tools for describing the reference architecture for space data systems

This paper has briefly presented the Reference Architecture for Space Data Systems (RASDS) that is being developed by the CCSDS Systems Architecture Working Group (SAWG). The SAWG generated some sample architectures (spacecraft onboard architectures, space link architectures, cross-support architectures) using this RASDS approach, and RASDS was proven to be a powerful tool for describing and relating different space data system architectures.

communication↗

NASA NTRS2006

A Reference Architecture for Space Information Management

We describe a reference architecture for space information management systems that elegantly overcomes the rigid design of common information systems in many domains. The reference architecture consists of a set of flexible, reusable, independent models and software components that function in unison, but remain separately managed entities. The main guiding principle of the reference architecture is to separate the various models of information (e.g., data, metadata, etc.) from implemented system code, allowing each to evolve independently. System modularity, systems interoperability, and dynamic evolution of information system components are the primary benefits of the design of the architecture. The architecture requires the use of information models that are substantially more advanced than those used by the vast majority of information systems. These models are more expressive and can be more easily modularized, distributed and maintained than simpler models e.g., configuration files and data dictionaries. Our current work focuses on formalizing the architecture within a CCSDS Green Book and evaluating the architecture within the context of the C3I initiative.

information systems↗

NASA NTRS2011

Space Missions Trade Space Generation and Assessment Using JPL Rapid Mission Architecture (RMA) Team Approach

The JPL Rapid Mission Architecture (RMA) capability is a novel collaborative team-based approach to generate new mission architectures, explore broad trade space options, and conduct architecture-level analyses. RMA studies address feasibility and identify best candidates to proceed to further detailed design studies. Development of RMA first began at JPL in 2007 and has evolved to address the need for rapid, effective early mission architectural development and trade space exploration as a precursor to traditional point design evaluations. The RMA approach integrates a small team of architecture-level experts (typically 6-10 people) to generate and explore a wide-ranging trade space of mission architectures driven by the mission science (or technology) objectives. Group brainstorming and trade space analyses are conducted at a higher level of assessment across multiple mission architectures and systems to enable rapid assessment of a set of diverse, innovative concepts. This paper describes the overall JPL RMA team, process, and high-level approach. Some illustrative results from previous JPL RMA studies are discussed.

real time systems↗

NASA NTRS2015

Fault Management Architectures and the Challenges of Providing Software Assurance

The satellite systems Fault Management (FM) is focused on safety, the preservation of assets, and maintaining the desired functionality of the system. How FM is implemented varies among missions. Common to most is system complexity due to a need to establish a multi-dimensional structure across hardware, software and operations. This structure is necessary to identify and respond to system faults, mitigate technical risks and ensure operational continuity. These architecture, implementation and software assurance efforts increase with mission complexity. Because FM is a systems engineering discipline with a distributed implementation, providing efficient and effective verification and validation (VV) is challenging. A breakout session at the 2012 NASA Independent Verification Validation (IVV) Annual Workshop titled VV of Fault Management: Challenges and Successes exposed these issues in terms of VV for a representative set of architectures. NASA's IVV is funded by NASA's Software Assurance Research Program (SARP) in partnership with NASA's Jet Propulsion Laboratory (JPL) to extend the work performed at the Workshop session. NASA IVV will extract FM architectures across the IVV portfolio and evaluate the data set for robustness, assess visibility for validation and test, and define software assurance methods that could be applied to the various architectures and designs. This work focuses efforts on FM architectures from critical and complex projects within NASA. The identification of particular FM architectures, visibility, and associated VVIVV techniques provides a data set that can enable higher assurance that a satellite system will adequately detect and respond to adverse conditions. Ultimately, results from this activity will be incorporated into the NASA Fault Management Handbook providing dissemination across NASA, other agencies and the satellite community. This paper discusses the approach taken to perform the evaluations and preliminary findings from the research including identification of FM architectures, visibility observations, and methods utilized for VVIVV.

Fault Management↗

NASA NTRS2017

Enabling Future Science and Human Exploration with NASA's Next Generation near Earth and Deep Space Communications and Navigation Architecture

The National Aeronautics and Space Administration (NASA) is studying alternatives for the United States space communications architecture through the 2040 timeframe. This architecture provides communication and navigation services to both human exploration and science missions throughout the solar system. Several of NASA's key space assets are approaching their end of design life and major systems are in need of replacement. The changes envisioned in the relay satellite architecture and capabilities around both Earth and Mars are significant undertakings and occur only once or twice each generation, and therefore is referred to as NASA's next generation space communications architecture. NASA's next generation architecture will benefit from technology and services developed over recent years. These innovations will provide missions with new operations concepts, increased performance, and new business and operating models. Advancements in optical communications will enable high-speed data channels and the use of new and more complex science instruments. Modern multiple beam/multiple access technologies such as those employed on commercial high throughput satellites will enable enhanced capabilities for on-demand service, and with new protocols will help provide Internet-like connectivity for cooperative spacecraft to improve data return and coordinate joint mission objectives. On-board processing with autonomous and cognitive networking will play larger roles to help manage system complexity. Spacecraft and ground systems will coordinate among themselves to establish communications, negotiate link connectivity, and learn to share spectrum to optimize resource allocation. Spacecraft will autonomously navigate, plan trajectories, and handle off-nominal events. NASA intends to leverage the ever-expanding capabilities of the satellite communications industry and foster its continued growth. NASA's technology development will complement and extend commercial capabilities to meet unique space environment requirements and to provide capabilities that are beyond the commercial marketplace. The progress of the communications industry, including the emerging global space internet segment and its planned constellations of 100's of satellites offer additional opportunities for new capability and mission concepts. The opportunities and challenges of a future space architecture require an optimal solution encompassing a global perspective. The concepts and technologies intentionally define an architecture that applies not only to NASA, but to other U.S. government agencies, international space and government agencies, and domestic and international industries to advance the openness, interoperability, and affordability of space communications. Cooperation among the worlds space agencies, their capabilities, standards, operations, and interoperability are key to advancing humankind's understand of the universe and extending human presence into the solar system.

Communications↗

NASA NTRS2017

Enabling Future Science and Human Exploration with NASA's Next Generation Near Earth and Deep Space Communications and Navigation Architecture

The National Aeronautics and Space Administration (NASA) is studying alternatives for the United States space communications architecture through the 2040 timeframe. This architecture provides communication and navigation services to both human exploration and science missions throughout the solar system. Several of NASA's key space assets are approaching their end of design life and major systems are in need of replacement. The changes envisioned in the relay satellite architecture and capabilities around both Earth and Mars are significant undertakings and occur only once or twice each generation, and therefore is referred to as NASA's next generation space communications architecture. NASA's next generation architecture will benefit from technology and services developed over recent years. These innovations will provide missions with new operations concepts, increased performance, and new business and operating models. Advancements in optical communications will enable high-speed data channels and the use of new and more complex science instruments. Modern multiple beam/multiple access technologies such as those employed on commercial high throughput satellites will enable enhanced capabilities for on-demand service, and with new protocols will help provide Internet-like connectivity for cooperative spacecraft to improve data return and coordinate joint mission objectives. On-board processing with autonomous and cognitive networking will play larger roles to help manage system complexity. Spacecraft and ground systems will coordinate among themselves to establish communications, negotiate link connectivity, and learn to share spectrum to optimize resource allocation. Spacecraft will autonomously navigate, plan trajectories, and handle off-nominal events. NASA intends to leverage the ever-expanding capabilities of the satellite communications industry and foster its continued growth. NASA's technology development will complement and extend commercial capabilities to meet unique space environment requirements and to provide capabilities that are beyond the commercial marketplace. The progress of the communications industry, including the emerging global space internet segment and its planned constellations of 100's of satellites offer additional opportunities for new capability and mission concepts. The opportunities and challenges of a future space architecture require an optimal solution encompassing a global perspective. The concepts and technologies intentionally define an architecture that applies not only to NASA, but to other U.S. government agencies, international space and government agencies, and domestic and international industries to advance the openness, interoperability, and affordability of space communications. Cooperation among the worlds space agencies, their capabilities, standards, operations, and interoperability are key to advancing humankinds understand of the universe and extending human presence into the solar system.

deep space↗

NASA NTRS2018

NASA GSFC Avionics Architectures and Future Directions

NASA Goddard Spaceflight Center (GSFC) employs standard avionics hardware and software architectures. Hardware architectures include GMSA (Goddard Modular Smallsat Architecture), MUSTANG (Modular Unified Space Technology Avionics for Next Generation), SpaceCube. The spaceflight software architecture is based on cFS (Core Flight System). Future driving requirements for avionics architectures include increase sensor data rates, increased onboard processing, autonomous applications, and distributed space missions. Architectural concepts that can meet these requirements include the use of the High Performance Spaceflight Computing (HPSC) Chiplet, employing hybrid computing architectures, and increasing network bandwidths.

architectures↗

NASA NTRS2019

Medical Data Architecture (MDA) Project Status

The Medical Data Architecture (MDA) project supports the Exploration Medical Capability (ExMC) risk to minimize or reduce the risk of adverse health outcomes and decrements in performance due to in-flight medical capabilities on human exploration missions. To mitigate this risk, the ExMC MDA project addresses the technical limitations identified in ExMC Gap Med 07: We do not have the capability to comprehensively process medically-relevant information to support medical operations during exploration missions. This gap identifies that the current in-flight medical data management includes a combination of data collection and distribution methods that are minimally integrated with on-board medical devices and systems. Furthermore, there are a variety of data sources and methods of data collection. For an exploration mission, the seamless management of such data will enable a more medically autonomous crew than the current paradigm. The medical system requirements are being developed in parallel with the exploration mission architecture and vehicle design. ExMC has recognized that in order to make informed decisions about a medical data architecture framework, current methods for medical data management must not only be understood, but an architecture must also be identified that provides the crew with actionable insight to medical conditions. This medical data architecture will provide the necessary functionality to address the challenges of executing a self-contained medical system that approaches crew health care delivery without assistance from ground support. Hence, the products supported by current prototype development will directly inform exploration medical system requirements.In fiscal year 2018, the MDA project developed Test Bed 2, the second iteration in a series of prototypes with functionality focused on data security through role-based access control and encryption, integration with One Portal exercise software and ingestion of an ultrasound Digital Imaging and Communications in Medicine (DICOM) file and image display. Test Bed 2 advances the medical data system architecture framework by providing these functionalities in a scalable system that maintained a layered, modular design. The architecture framework uses a data services approach with role-based access to data in a customized medical record system suitable for space exploration. These functionalities were demonstrated as part of the Next Space Technologies for Exploration Partnerships (NextSTEP) ground test demonstrated at the NASA Johnson Space Center Integrated Power, Avionics and Software (iPAS) facility. Interfacing to a Core Flight Software (CFS) system, the MDA system, using Consultative Committee for Space Data Systems (CCSDS) protocol, transferred an exercise file from the simulated flight MDA system to a mirrored MDA system on the ground through the CFS system. The selection of data sources and demonstrations enabled the team to address stakeholder concerns throughout the development process. In the next iteration, the MDA team will work with stakeholders to identify additional relevant functionalities to further advance system data models, standards and principles that will inform the medical system requirements development.

medical data architecture↗

NASA NTRSDate not supplied

ECLSS Does Not Exist in a Vacuum: Integrated Analysis is Necessary to Inform System Architecture Decisions

Environmental Control and Life Support System (ECLSS) architecture selection has profound implications for mission cost and mass extending far beyond the ECLSS itself. Similarly, other mission architecture decisions – particularly involving transportation systems – can influence optimal ECLSS architectures. Loop closure influences requirements for water, oxygen, and other consumables. System maintainability and reliability influences spares mass and risk. System, consumables, and spares mass interact with transportation system architectures and propellant demands and propulsion element sizing. All these interactions with other systems must be considered when evaluating ECLSS options. Analyses that focus only on maximizing ECLSS loop closure – or minimizing ECLSS mass, or minimizing ECLSS life cycle cost – may lead to sub-optimal or even counterproductive system architecture and investment decisions at the mission level. For example, an ECLSS architecture that minimizes ECLSS lifecycle cost but results in excessively high logistics mass could lead to significantly increased transportation system costs or make interplanetary transportation infeasible. Increased loop closure could result in higher development costs and higher mass if system/spares mass increases outweigh consumables reduction. In addition, systems mass and consumables mass are not directly comparable and have different impacts on propellant requirements, as consumables mass changes over the course of the mission. This paper presents an integrated analysis examining the overall impact of different ECLSS and transportation architectures on mass for a crewed Mars mission, including the habitat and transportation systems as well as consumables, spares, and propellant. Key observations are discussed, along with opportunities for further sensitivity analysis and model development. Overall, ECLSS development activities must consider their impacts at the mission level, as part of an integrated system, rather than in isolation.

Systems Analysis↗

NASA NTRSDate not supplied

Lunar Communications and Navigation: Lunar Communications, Position, Navigation, and Time (CPNT) Architecture

The NASA Moon-to-Mars (M2M) Architecture will enable the return of humans to the Moon, establish a long-term presence there, and open more of the lunar surface to exploration than ever before. This growth of lunar activity requires robust and resilient communications, position, navigation, and time capabilities for crew safety, the command and control of spacecraft, the return of science data, and the precise telerobotic maneuvering of assets both in space and on the lunar surface. Within the M2M Architecture, the sub-architecture for Communications, Position, Navigation, and Time (CPNT) details the specific CPNT systems and infrastructure required to meet the M2M objectives. Two objectives most fundamental to the CPNT sub-architecture are: (1) develop a lunar surface, orbital, and Moon-to-Earth communications architecture that scales to support long term science, exploration, and industrial needs and (2) develop a lunar position, navigation, and timing (PNT) architecture that also scales to support long term needs. The subject presentation provides an overview of the M2M Architecture, planned mission activities, and infrastructure development underway by NASA. The presentation provides overview of the surface wireless network and the challenges associated with operating on the lunar surface.

architecture↗

DOE OSTI2024

Approximate 𝑡-Designs in Generic Circuit Architectures

Unitary 𝑡-designs are distributions on the unitary group whose first 𝑡 moments appear maximally random. Previous work has established several upper bounds on the depths at which certain specific random quantum circuit ensembles approximate 𝑡-designs. Here we show that these bounds can be extended to any fixed architecture of Haar-random two-site gates. This is accomplished by relating the spectral gaps of such architectures to those of one-dimensional brickwork architectures. Our bound depends on the details of the architecture only via the typical number of layers needed for a block of the circuit to form a connected graph over the sites. When this quantity is bounded, the circuit forms an approximate 𝑡-design in at most linear depth. We give numerical evidence for a stronger bound that depends only on the number of connected blocks into which the architecture can be divided. We also give an implicit bound for nondeterministic architectures in terms of properties of the corresponding distribution over fixed architectures.

information scrambling↗

DOE OSTI2024

Exploring Architectural-Aware Affinity Policies in Modern HPC Runtimes

Modern commodity and High-Performance Computing (HPC) systems are evolving with complex CPU architectures. These architectures now feature higher core and NUMA domain counts and implement features such as hyperthreading. When considering significant differences in hardware configurations, library availability, and hardware-tailored system/software stacks, which could substantially vary from one system to another, performance portability is hard to achieve. Throughout the years, this trend resulted in an increasingly high burden on application developers to fine-tune their workloads for each architecture. This work explores how hardware-dependent aspects such as locality/process/thread affinity affect performance in modern CPU architectures. We focus our study on the Global Memory and Threading (GMT) distributed runtime system as a representative of Partitioned Global Address Space (PGAS) software stacks commonly adopted for productivity. In particular, to appreciate performance implications, we evaluate GMT’s thread affinity policies, and, introduce two new ones which exploit architectural awareness. Finally, we explore alternative NUMA configurations via different process bindings and perform a scalability study on three HPC clusters with varying CPU architectures and NUMA layouts. Our analysis indicates that more complex architectures are more affected by affinity and binding policies and highlights the importance of setting proper runtime configurations to achieve superior performance.

Di Dio Lavore, Ian↗

DOE OSTI2021

Explainable Neural Architecture Search (XNAS)

Code for the paper Learning Interpretable Models Through Multi-Objective Neural Architecture Search by Zachariah Carmichael, Tim Moon, and Sam Ade Jacobs. Monumental advances in deep learning have led to unprecedented achievements across a multitude of domains. While the performance of deep neural networks is indubitable, the architectural design and interpretability of such models are nontrivial. Research has been introduced to automate the design of neural network architectures through neural architecture search (NAS). Recent progress has made these methods more pragmatic by exploiting distributed computation and novel optimization algorithms. However, there is little work in optimizing architectures for interpretability. To this end, we propose a multiobjective distributed NAS framework that optimizes for both task performance and introspection. We leverage the non-dominated sorting genetic algorithm (NSGA-II) and explainable AI (XAI) techniques to reward architectures that can be better comprehended by humans. The framework is evaluated on several image classification datasets. We demonstrate that jointly optimizing for introspection ability and task error leads to more disentangled architectures that perform within tolerable error.

Carmichael, ZachariahJ↗

DOE OSTI2020

Influence of Vegetation Canopy Architecture on Dynamics of Spatiotemporal Light Distribution and its Effect on Land-Biosphere-Atmosphere Fluxes of Energy, Carbon, and Water

The dynamic spatiotemporal distribution of light and micro climatic variables (wind speed, atmospheric CO 2 concentration, humidity, pressure, and air temperature) within the plant canopy as determined by the canopy architecture play a critical role in land-biosphere-atmosphere (LBA) photosynthetic carbon uptake, transpiratory water loss and energy partitioning. Vegetation models that include effects of sun and shade flecks, leaf temperature and energy balance partitioning between sensible and latent heat fluxes, and reduced turbulent mixing in dense vegetation canopies can accurately predict canopy performance, particularly its shaded leaves, which can constitute 80% of leaf area and contribute up to 50% of carbon update and 65% of water loss. However, current vegetation models with their bulk characterization of shaded leaves poorly represent the complex three dimensional (3D) canopy architecture are insufficient to capture the spatiotemporal dynamics of shaded leaves. Here, we propose to employ an x-ray computed tomography (XCT) imaging technique in combination with an automated, high throughput, image processing pipeline (3DArch, to be developed) to obtain an explicit 3D triangulated surface mesh of a plant canopy (Fig 1). 3DArch would provide a much needed technological jump in 3D LBA modeling where computer generated artificial 3D models suffer from lack of high resolution detailed canopy architecture parameters. The 3D canopy surface mesh output from 3DArch is coupled with an explicit 3D forward ray tracing model (fastTracer) to obtain a high resolution representation of the dynamic spatiotemporal light environment within the canopy over the course of a day (Fig 1). The results from the light attenuation model are then coupled with a process based multi-scale bio-physical canopy and leaf model of photosynthesis (PlantGro) to get fluxes of carbon, water and energy at the leaf and canopy scale, in addition to estimating whole plant carbon uptake, and water loss. Our proposed model framework would be the first of its kind to provide a breakthrough capability in the field of process-based plant modeling that bridges the key technological gap in shaded leaf modeling through explicit representation of 3D canopy architecture. Computer generated artificial 3D model canopy simulations on grasses show that the canopy light distribution does not follow an exponential behavior as assumed in most 1D vegetation models, but rather exhibits a heterogeneous behavior, with a small but significant portion of interior canopy leaves experiencing peak light levels due to sun flecks while the rest of the canopy receives substantially lower light levels. These results indicate a bias in traditional models that overestimates shaded leaf contribution towards whole canopy photosynthesis by up to 15% and underestimate transpiration by up to 12% with an associated increase in sensible heat fluxes since photosynthesis, transpiration and energy partitioning are tightly coupled processes. However, these test results need to be validated using the 3D architecture obtained from a real plant imaged in 3D using 3DArch, rather than an artificial model plant. Our proposed model framework would allow us to identify inefficiencies in canopy light distribution and prescribe optimal canopy architectures (leaf optical properties, leaf area, leaf angles, row spacing, planting density, etc.) to traditional and genetic engineering based crop breeders, for maximizing target functions such as light use efficiency, water use efficiency, drought tolerance, crop yield, crop resilience etc., with important applications in biofuel crops such as sorghum, and switchgrass. Our model can also be applied for designing optimal canopy architecture that increase surface albedo as a mitigation strategy for global warming and projections of crop yield for future food and water security under climate change.

59 BASIC BIOLOGICAL SCIENCES↗

NASA NTRS1983

Parallel architectures for iterative methods on adaptive, block structured grids

A parallel computer architecture well suited to the solution of partial differential equations in complicated geometries is proposed. Algorithms for partial differential equations contain a great deal of parallelism. But this parallelism can be difficult to exploit, particularly on complex problems. One approach to extraction of this parallelism is the use of special purpose architectures tuned to a given problem class. The architecture proposed here is tuned to boundary value problems on complex domains. An adaptive elliptic algorithm which maps effectively onto the proposed architecture is considered in detail. Two levels of parallelism are exploited by the proposed architecture. First, by making use of the freedom one has in grid generation, one can construct grids which are locally regular, permitting a one to one mapping of grids to systolic style processor arrays, at least over small regions. All local parallelism can be extracted by this approach. Second, though there may be a regular global structure to the grids constructed, there will be parallelism at this level. One approach to finding and exploiting this parallelism is to use an architecture having a number of processor clusters connected by a switching network. The use of such a network creates a highly flexible architecture which automatically configures to the problem being solved.

Gannon, D.↗

NASA NTRS1990

Multiprocessor architecture: Synthesis and evaluation

Multiprocessor computed architecture evaluation for structural computations is the focus of the research effort described. Results obtained are expected to lead to more efficient use of existing architectures and to suggest designs for new, application specific, architectures. The brief descriptions given outline a number of related efforts directed toward this purpose. The difficulty is analyzing an existing architecture or in designing a new computer architecture lies in the fact that the performance of a particular architecture, within the context of a given application, is determined by a number of factors. These include, but are not limited to, the efficiency of the computation algorithm, the programming language and support environment, the quality of the program written in the programming language, the multiplicity of the processing elements, the characteristics of the individual processing elements, the interconnection network connecting processors and non-local memories, and the shared memory organization covering the spectrum from no shared memory (all local memory) to one global access memory. These performance determiners may be loosely classified as being software or hardware related. This distinction is not clear or even appropriate in many cases. The effect of the choice of algorithm is ignored by assuming that the algorithm is specified as given. Effort directed toward the removal of the effect of the programming language and program resulted in the design of a high-level parallel programming language. Two characteristics of the fundamental structure of the architecture (memory organization and interconnection network) are examined.

Standley, Hilda M.↗