Search NASA⌕ Search

SEARCH · Search NASA

Results for “Reconfigurable Computing”

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 145 records · Page 8

A new approach to reliability estimation for reconfigurable systems

A new approach to computing the probability of failure for process-control systems that achieve reliability by redundancy and reconfiguration is described. This approach has three distinguishing characteristics. It is based on information about the components and the system that can be observed; it requires a minimum of information; and it fills the gaps by locating the redundant unit misbehavior pattern that produces the largest probability of system failure.

White, Allan L.↗

Scalable FPGA Accelerator for Deep Convolutional Neural Networks with Stochastic Streaming

Here, FPGA-based heterogeneous computing platform, due to its extreme logic reconfigurability, emerges to be a strong contender as computing fabric in modern AI. As a result, various FPGA-based accelerators for deep CNN—the key driver of modern AI—have been proposed due to their advantages of high performance, reconfigurability, and fast development round, etc. In general, the consensus among researchers is that, although FPGA-based accelerator can achieve much higher energy efficiency, its raw computing performance lags behind when compared with GPUs with similar logic density. In this paper, we develop an alternative methodology to efficiently implement CNNs with FPGAs that outperform GPUs in terms of both power consumption and performance. Our key idea is to design a scalable hardware architecture and circuit design for large-scale CNNs that leverages a stochastic-based computing principle. Specifically, there are three major performance advantages. First, all key components of our deep learning CNN are designed and implemented to compute stochastically, thus achieving excellent computing performance and energy efficiency. Second, because our proposed CNN architecture enables a stream-mode computing, all of its stages can process even the partial results from preceding stages, therefore not incurring unnecessary latency due to data dependency. Finally, our FPGA-based deep CNN also provides a superior hardware scalability when compared with conventional FPGA implementations by reducing the bandwidth requirement between layers. The results show that our proposed CNN architecture significantly outperforms all previous FPGA-based deep CNN implementation approaches. It achieves 1.58x more GOPS, 6.42x more GOPS/Slice, and 10.92x more GOPS/W when compared with state-of-the-art CNN architecture. The top-5 accuracy of stochastic VGG-16 CNN is 86.77 percent with 18.91 fps frame rate.

97 MATHEMATICS AND COMPUTING↗

Closed-form solution of decomposable stochastic models

Markov and semi-Markov processes are increasingly being used in the modeling of complex reconfigurable systems (fault tolerant computers). The estimation of the reliability (or some measure of performance) of the system reduces to solving the process for its state probabilities. Such a model may exhibit numerous states and complicated transition distributions, contributing to an expensive and numerically delicate solution procedure. Thus, when a system exhibits a decomposition property, either structurally (autonomous subsystems), or behaviorally (component failure versus reconfiguration), it is desirable to exploit this decomposition in the reliability calculation. In interesting cases there can be failure states which arise from non-failure states of the subsystems. Equations are presented which allow the computation of failure probabilities of the total (combined) model without requiring a complete solution of the combined model. This material is presented within the context of closed-form functional representation of probabilities as utilized in the Symbolic Hierarchical Automated Reliability and Performance Evaluator (SHARPE) tool. The techniques adopted enable one to compute such probability functions for a much wider class of systems at a reduced computational cost. Several examples show how the method is used, especially in enhancing the versatility of the SHARPE tool.

Sjogren, Jon A.↗

Upper and lower bounds for semi-Markov reliability models of reconfigurable systems

This paper determines the information required about system recovery to compute the reliability of a class of reconfigurable systems. Upper and lower bounds are derived for these systems. The class consists of those systems that satisfy five assumptions: the components fail independently at a low constant rate, fault occurrence and system reconfiguration are independent processes, the reliability model is semi-Markov, the recovery functions which describe system configuration have small means and variances, and the system is well designed. The bounds are easy to compute, and examples are included.

White, A. L.↗

Fault tolerant hypercube computer system architecture

A fault-tolerant multiprocessor computer system of the hypercube type comprising a hierarchy of computers of like kind which can be functionally substituted for one another as necessary is disclosed. Communication between the working nodes is via one communications network while communications between the working nodes and watch dog nodes and load balancing nodes higher in the structure is via another communications network separate from the first. A typical branch of the hierarchy reporting to a master node or host computer comprises, a plurality of first computing nodes; a first network of message conducting paths for interconnecting the first computing nodes as a hypercube. The first network provides a path for message transfer between the first computing nodes; a first watch dog node; and a second network of message connecting paths for connecting the first computing nodes to the first watch dog node independent from the first network, the second network provides an independent path for test message and reconfiguration affecting transfers between the first computing nodes and the first switch watch dog node. There is additionally, a plurality of second computing nodes; a third network of message conducting paths for interconnecting the second computing nodes as a hypercube. The third network provides a path for message transfer between the second computing nodes; a fourth network of message conducting paths for connecting the second computing nodes to the first watch dog node independent from the third network. The fourth network provides an independent path for test message and reconfiguration affecting transfers between the second computing nodes and the first watch dog node; and a first multiplexer disposed between the first watch dog node and the second and fourth networks for allowing the first watch dog node to selectively communicate with individual ones of the computing nodes through the second and fourth networks; as well as, a second watch dog node operably connected to the first multiplexer whereby the second watch dog node can selectively communicate with individual ones of the computing nodes through the second and fourth networks. The branch is completed by a first load balancing node; and a second multiplexer connected between the first load balancing node and the first and second watch dog nodes, allowing the first load balancing node to selectively communicate with the first and second watch dog nodes.

Madan, Herb S.↗

Intelligence Applied to Air Vehicles

The exponential growth in information technology has provided the potential for air vehicle capabilities that were previously unavailable to mission and vehicle designers. The increasing capabilities of computer hardware and software, including new developments such as neural networks, provide a new balance of work between humans and machines. This paper will describe several NASA projects, and review results and conclusions from ground and flight investigations where vehicle intelligence was developed and applied to aeronautical and space systems. In the first example, flight results from a neural network flight control demonstration will be reviewed. Using, a highly-modified F-15 aircraft, a NASA/Dryden experimental flight test program has demonstrated how the neural network software can correctly identify and respond to changes in aircraft stability and control characteristics. Using its on-line learning capability, the neural net software would identify that something in the vehicle has changed, then reconfigure the flight control computer system to adapt to those changes. The results of the Remote Agent software project will be presented. This capability will reduce the cost of future spacecraft operations as computers become "thinking" partners along with humans. In addition, the paper will describe the objectives and plans for the autonomous airplane program and the autonomous rotorcraft project. Technologies will also be developed.

Rosen, Robert↗

Voter-comparator-switch optimization study

The design optimization and simulation effort of the voter-comparator switch (VCS) element are described. The VCS is an adaptive voter element that performs failure detection and reconfiguration function in a computer system designed to tolerate any three single failures in a fail operational fail operational fail safe manner. Detailed logic equations were developed and the design was evaluated and refined using a logic level fault simulator.

Jurison, J.↗

An approximation formula for a class of fault-tolerant computers

An approximation formula is derived for the probability of failure for fault-tolerant process-control computers. These computers use redundancy and reconfiguration to achieve high reliability. Finite-state Markov models capture the dynamic behavior of component failure and system recovery, and the approximation formula permits an estimation of system reliability by an easy examination of the model.

White, A. L.↗

The Evolution of Software and Its Impact on Complex System Design in Robotic Spacecraft Embedded Systems

The growth in computer hardware performance, coupled with reduced energy requirements, has led to a rapid expansion of the resources available to software systems, driving them towards greater logical abstraction, flexibility, and complexity. This shift in focus from compacting functionality into a limited field towards developing layered, multi-state architectures in a grand field has both driven and been driven by the history of embedded processor design in the robotic spacecraft industry.The combinatorial growth of interprocess conditions is accompanied by benefits (concurrent development, situational autonomy, and evolution of goals) and drawbacks (late integration, non-deterministic interactions, and multifaceted anomalies) in achieving mission success, as illustrated by the case of the Mars Reconnaissance Orbiter. Approaches to optimizing the benefits while mitigating the drawbacks have taken the form of the formalization of requirements, modular design practices, extensive system simulation, and spacecraft data trend analysis. The growth of hardware capability and software complexity can be expected to continue, with future directions including stackable commodity subsystems, computer-generated algorithms, runtime reconfigurable processors, and greater autonomy.

software↗

Optimum spaceborne computer system design by simulation

A deterministic simulator is described which models the Automatically Reconfigurable Modular Multiprocessor System (ARMMS), a candidate computer system for future manned and unmanned space missions. Its use as a tool to study and determine the minimum computer system configuration necessary to satisfy the on-board computational requirements of a typical mission is presented. The paper describes how the computer system configuration is determined in order to satisfy the data processing demand of the various shuttle booster subsytems. The configuration which is developed as a result of studies with the simulator is optimal with respect to the efficient use of computer system resources.

Williams, T.↗

Optimum spaceborne computer system design by simulation

A deterministic digital simulation model is described which models the Automatically Reconfigurable Modular Multiprocessor System (ARMMS), a candidate computer system for future manned and unmanned space missions. Use of the model as a tool in configuring a minimum computer system for a typical mission is demonstrated. The configuration which is developed as a result of studies with the simulator is optimal with respect to the efficient use of computer system resources, i.e., the configuration derived is a minimal one. Other considerations such as increased reliability through the use of standby spares would be taken into account in the definition of a practical system for a given mission.

Williams, T.↗

Design of a modular digital computer system, CDRL no. D001, final design plan

The engineering breadboard implementation for the CDRL no. D001 modular digital computer system developed during design of the logic system was documented. This effort followed the architecture study completed and documented previously, and was intended to verify the concepts of a fault tolerant, automatically reconfigurable, modular version of the computer system conceived during the architecture study. The system has a microprogrammed 32 bit word length, general register architecture and an instruction set consisting of a subset of the IBM System 360 instruction set plus additional fault tolerance firmware. The following areas were covered: breadboard packaging, central control element, central processing element, memory, input/output processor, and maintenance/status panel and electronics.

Easton, R. A.↗

Reconfigurability in MDO Problem Synthesis

Integrating autonomous disciplines into a problem amenable to solution presents a major challenge in realistic multidisciplinary design optimization (MDO). We propose a linguistic approach to MDO problem description, formulation, and solution we call reconfigurable multidisciplinary synthesis (REMS). With assistance from computer science techniques, REMS comprises an abstract language and a collection of processes that provide a means for dynamic reasoning about MDO problems in a range of contexts. The approach may be summarized as follows. Description of disciplinary data according to the rules of a grammar, followed by lexical analysis and compilation, yields basic computational components that can be assembled into various MDO problem formulations and solution algorithms, including hybrid strategies, with relative ease. The ability to re-use the computational components is due to the special structure of the MDO problem. The range of contexts for reasoning about MDO spans tasks from error checking and derivative computation to formulation and reformulation of optimization problem statements. In highly structured contexts, reconfigurability can mean a straightforward transformation among problem formulations with a single operation. We hope that REMS will enable experimentation with a variety of problem formulations in research environments, assist in the assembly of MDO test problems, and serve as a pre-processor in computational frameworks in production environments. This paper, Part 1 of two companion papers, discusses the fundamentals of REMS. Part 2 illustrates the methodology in more detail.

Alexandrov, Natalia M.↗

Reconfigurability in MDO Problem Synthesis

Integrating autonomous disciplines into a problem amenable to solution presents a major challenge in realistic multidisciplinary design optimization (MDO). We propose a linguistic approach to MDO problem description, formulation, and solution we call reconfigurable multidisciplinary synthesis (REMS). With assistance from computer science techniques, REMS comprises an abstract language and a collection of processes that provide a means for dynamic reasoning about MDO problems in a range of contexts. The approach may be summarized as follows. Description of disciplinary data according to the rules of a grammar, followed by lexical analysis and compilation, yields basic computational components that can be assembled into various MDO problem formulations and solution algorithms, including hybrid strategies, with relative ease. The ability to re-use the computational components is due to the special structure of the MDO problem. The range of contexts for reasoning about MDO spans tasks from error checking and derivative computation to formulation and reformulation of optimization problem statements. In highly structured contexts, reconfigurability can mean a straightforward transformation among problem formulations with a single operation. We hope that REMS will enable experimentation with a variety of problem formulations in research environments, assist in the assembly of MDO test problems, and serve as a pre-processor in computational frameworks in production environments. Part 1 of two companion papers, discusses the fundamentals of REMS. This paper, Part 2 illustrates the methodology in more detail.

Alexandrov, Natalia M.↗

Load Forecasting Based Distribution System Network Reconfiguration -- A Distributed Data-Driven Approach

In this paper, a short-term load forecasting approach based network reconfiguration is proposed in a parallel manner. Specifically, a support vector regression (SVR) based short-term load forecasting approach is designed to provide an accurate load prediction and benefit the network reconfiguration. Because of the nonconvexity of the three-phase balanced optimal power flow, a second-order cone program (SOCP) based approach is used to relax the optimal power flow problem. Then, the alternating direction method of multipliers (ADMM) is used to compute the optimal power flow in distributed manner. Considering the limited number of the switches and the increasing computation capability, the proposed network reconfiguration is solved in a parallel way. The numerical results demonstrate the feasible and effectiveness of the proposed approach.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Evaluating Proxy Apps with FPGAs

Field-programmable gate arrays (FPGAs) are becoming more promising for high performance computing (HPC). Unlike fixed hardware specialization, reconfigurability allows FPGAs to be customized for a wide variety of HPC workloads. Our ultimate goal is to understand the viability of FPGAs as accelerators for some of the key HPC codes at LLNL and identify application characteristics that can benefit from the reconfigurability of FPGAs. More specifically, this LDRD project conducted a detailed performance study of FPGAs using sparse matrix computations and identified key optimization techniques to effectively use FPGA hardware resources.

97 MATHEMATICS AND COMPUTING↗

Evolution of the Hubble Space Telescope Safing Systems

The Hubble Space Telescope (HST) was launched on April 24 1990, with an expected lifespan of 15 years. Central to the spacecraft design was the concept of a series of on-orbit shuttle servicing missions permitting astronauts to replace failed equipment, update the scientific instruments and keep the HST at the forefront of astronomical discoveries. One key to the success of the Hubble mission has been the robust Safing systems designed to monitor the performance of the observatory and to react to keep the spacecraft safe in the event of equipment anomaly. The spacecraft Safing System consists of a range of software tests in the primary flight computer that evaluate the performance of mission critical hardware, safe modes that are activated when the primary control mode is deemed inadequate for protecting the vehicle, and special actions that the computer can take to autonomously reconfigure critical hardware. The HST Safing System was structured to autonomously detect electrical power system, data management system, and pointing control system malfunctions and to configure the vehicle to ensure safe operation without ground intervention for up to 72 hours. There is also a dedicated safe mode computer that constantly monitors a keep-alive signal from the primary computer. If this signal stops, the safe mode computer shuts down the primary computer and takes over control of the vehicle, putting it into a safe, low-power configuration. The HST Safing system has continued to evolve as equipment has aged, as new hardware has been installed on the vehicle, and as the operation modes have matured during the mission. Along with the continual refinement of the limits used in the safing tests, several new tests have been added to the monitoring system, and new safe modes have been added to the flight software. This paper will focus on the evolution of the HST Safing System and Safing tests, and the importance of this evolution to prolonging the science operations of the telescope.

Pepe, Joyce↗