Showing papers on "Performance prediction published in 1995"

Computer generation of correlated non-Gaussian radar clutter

Abstract: We develop computer simulation procedures which enable us to generate any correlated non-Gaussian radar clutter that can be modeled as a spherically invariant random process (SIRP). In most cases, when the clutter is a correlated non-Gaussian random process, performance of the optimal radar signal processor cannot be evaluated analytically. Therefore, in order to evaluate such processors, there is a need for efficient computer simulation of the clutter. We present two canonical simulation procedures for the generation of correlated non-Gaussian clutter. A new approach for the goodness-of-fit test is proposed in order to assess the performance of the simulation procedure. >

Analysis and simulation of the high-speed torque performance of brushless DC motor drives

Abstract: A simulation model for the performance prediction of brushless DC drives is used to investigate the high-speed torque performance for a motor switched by a PWM inverter. The solution is given for both 120 degrees and 180 degrees conduction angles of inverter operation, with provision for selecting the advance commutation angle. The influence of the more important drive parameters on high-speed performance is described. The paper also presents two simplified analytical approaches for obtaining the optimum phase-advance angle in the high-speed ranges. The merit of these methods, compared with trial-and-error based on simulation, is described.

Performance analysis of an oil-injected screw compressor and its application

Abstract: Flow and heat transfer characteristics required for computer simulation of screw compressors are experimentally obtained and used in performance prediction, where the working medium is air. The heat transfer coefficient is determined from the experimental relation between the volumetric efficiency and the inlet temperature. Flow coefficients are obtained from the efficiency-clearance curves. Applying those coefficients to the performance simulation, good agreement is obtained between the tested and calculated performance of three different prototype air compressors. A new rotor profile aimed at higher performance is designed based on the simulation results. The tested performance of the new profile compressor is much higher than that of a conventional compressor, as predicted by the simulation. The new profile compressor has been applied to a commercial series of packaged air screw compressors.

Development and Application of a Multistage Navier-Stokes Solver: Part I — Multistage Modeling Using Bodyforces and Deterministic Stresses

Abstract: A novel multistage compressor performance analysis method based on the three-dimensional Reynolds averaged Navier-Stokes equations is presented in this paper. This approach is a “continuous interface plane approach” where deterministic stresses are used to ensure continuous physical properties across interface planes. The average unsteady effects due to neighboring blades and/or vanes are approximated using deterministic stresses along with the application of bodyforces. Bodyforces are used to account for the “potential” interaction between closely coupled (staged) rows. Deterministic stresses account for the “average” wake blockage and mixing effects both axially and radially. The attempt here is to implement an approximate technique for incorporating periodic unsteady flow physics that provides for a robust multistage design procedure incorporating reasonable computational efficiency. The present paper gives the theoretical development of the stress/bodyforce models incorporated in the code, and demonstrates the usefulness of these models in practical compressor applications. Compressor performance prediction capability is then established through a rigorous code/model validation effort using the power of networked workstations. The numerical results are compared with experimental data in terms of one-dimensional performance parameters such as total pressure ratio and circumferentially averaged radial profiles deemed critical to compressor design. This methodology allows the designer to design from hub to tip with a high level of confidence in the procedure.Copyright © 1995 by ASME

Design, construction and testing of explosive-driven helical generators

Abstract: This paper describes the development of a very efficient computer model for the design and performance prediction of explosive-driven helical generators. The model is based on simple theoretical considerations. Validation of the model is achieved by comparing the theoretical and measured performances of existing both high- and low-energy generators. It is shown that, although the basic model predicts accurately the load current history of high-energy generators, a somewhat more elaborate model is needed for low-energy devices. The model has been used in the design of a simple 1 MJ generator with an eight-section stator coil, intended for use as a current source in an investigation of high-current conditioning systems. A description is given of the construction and testing of this device. Experimental results are in accordance with predictions from the design code and establish that, when primed with 40 kJ at 50 kA from a capacitor bank and using 15 kg of high explosive, the generator is capable of delivering an output of 1 MJ at 7 MA to a coaxial load.

System and method for modeling the flow performance features of an object

Abstract: The method and apparatus includes a neural network for generating a model of an object in a wind tunnel from performance data on the object. The network is trained from test input signals (e.g., leading edge flap position, trailing edge flap position, angle of attack, and other geometric configurations, and power settings) and test output signals (e.g., lift, drag, pitching moment, or other performance features). In one embodiment, the neural network training method employs a modified Levenberg-Marquardt optimization technique. The model can be generated "real time" as wind tunnel testing proceeds. Once trained, the model is used to estimate performance features associated with the aircraft given geometric configuration and/or power setting input. The invention can also be applied in other similar static flow modeling applications in aerodynamics, hydrodynamics, fluid dynamics, and other such disciplines. For example, the static testing of cars, sails, and foils, propellers, keels, rudders, turbines, fins, and the like, in a wind tunnel, water trough, or other flowing medium.

Accelerated testing of space mechanisms

Abstract: This report contains a review of various existing life prediction techniques used for a wide range of space mechanisms. Life prediction techniques utilized in other non-space fields such as turbine engine design are also reviewed for applicability to many space mechanism issues. The development of new concepts on how various tribological processes are involved in the life of the complex mechanisms used for space applications are examined. A 'roadmap' for the complete implementation of a tribological prediction approach for complex mechanical systems including standard procedures for test planning, analytical models for life prediction and experimental verification of the life prediction and accelerated testing techniques are discussed. A plan is presented to demonstrate a method for predicting the life and/or performance of a selected space mechanism mechanical component.

Design geometry and design/off-design performance computer codes for compressors and turbines

Abstract: This report summarizes some NASA Lewis (i.e., government owned) computer codes capable of being used for airbreathing propulsion system studies to determine the design geometry and to predict the design/off-design performance of compressors and turbines. These are not CFD codes; velocity-diagram energy and continuity computations are performed fore and aft of the blade rows using meanline, spanline, or streamline analyses. Losses are provided by empirical methods. Both axial-flow and radial-flow configurations are included.

Multivariate statistical techniques for parallel performance prediction

Abstract: Performance prediction can play an important role in improving the efficiency of multicomputers in executing scalable parallel applications. An accurate model of program execution time must include detailed algorithmic and architectural characterizations. The exact values for critical model parameters such as message latency and cache miss penalty can often be difficult to determine. This research uses multivariate data analysis to estimate the values of these coefficients in an analytical model. Representing the coefficients as random variables with a specified mean and variance improves the utility of a performance model. Confidence intervals for predicted execution time can be generated using the standard error values for model parameters. Improvements in the model can also be made by investigating the cause of large variance values for a particular architecture. >

A parallelism analyzer for conservative parallel simulation

Abstract: Most small-scale simulation applications are implemented by sequential simulation techniques. As the problem size increases, however, sequential techniques may be unable to manage the time complexity of the simulation applications adequately. It is natural to consider re-implementing the corresponding large-scale simulations using parallel techniques, which have been reported to be successful in reducing the time complexity for several examples. However, parallel simulation may not be effective for every application. Since the implementation of parallel simulation for an application is usually very expensive, it is required to investigate the performance of parallel simulation for a particular application before re-implementing the simulation. The Chandy-Misra parallel, discrete-event simulation paradigm has been utilized in many large-scale simulation experiments, and several significant extensions have been based on it. Hence the Chandy-Misra protocol is adopted here as a basic model of parallel simulation to which our performance prediction techniques are applied. For an existing sequential simulation program based on the process interaction model, this paper proposes a technique for evaluating Chandy-Misra parallel simulation without actually implementing the parallel program. The idea is to insert parallelism analysis code into the sequential simulation program. When the modified sequential program is executed, the time complexity of the parallel simulation based on the Chandy-Misra protocol is computed. Our technique has been used to determine whether a giant Signaling System 7 simulation (sequential implementation) should be re-implemented using the parallel simulation approach. >

Coverage and performance prediction of DGPS systems employing radiobeacon transmissions.

Abstract: The NAVSTAR-Clobal Positioning System (CPS) is a state-of-the-art, satellite-based, world-wide, navigation system which was declared operational this year. With GPS, the position of a user can be determined to within 100 m; by employing differential techniques this uncertainty can be reduced to 10 m or less. In real time, this requires a Differential GPS (DGPS) system to measure and communicate correction information to the user. This research considers nGPS systems employing modified marine radiobeacons and receivers, examining the factors affecting their coverage and performance. Existing techniques of coverage prediction have been examined and found to be inadequate. Improved methods of coverage and performance prediction are developed which take into account several new factors. The groundwave and skywave field strengths of a beacon are calculated, taking into account detailed information about the ground conductivity. A new method of determining own-skywave fading has been derived and the effect at night determined to be significant. Levels of interference from other beacons on the co- and adjacent channels, via groundwave and skywave, are shown to be dominant coverage limiting factors in the European environment. Atmospheric noise and its effect on receiver performance is evaluated and forms the basis of performance predictions. These new techniques have been built into a computer model which automatically evaluates coverage and performance contours for DGPS radiobeacons and which is being used world-wide in the planning and evaluation of these systems.

N-MAP: a virtual processor discrete event simulation tool for performance prediction in the CAPSE environment

Abstract: The CAPSE (Computer Aided Parallel Software Engineering) environment aims to assist a performance oriented parallel program development approach by integrating tools for performance prediction in the design phase, analytical or simulation based performance analysis in the detailed specification and coding phase, and finally monitoring in the testing and correction phase. In this work, the N-MAP tool as part of the CAPSE environment is presented. N-MAP covers the crucial aspect of performance prediction to support a performance oriented, incremental development process of parallel applications such that implementation design choices can be investigated far ahead of the full coding of the application. Methodologically, N-MAP in an automatic parse and translate step generates a simulation program from a skeletal SPMD program, with which the programmer expresses just the constituent and performance critical program parts, subject to an incremental refinement. The simulated execution of the SPMD skeleton supports a variety of performance studies. We demonstrate the use and performance of the N-MAP tool by developing a linear system solver for the CM-5. >

PerPreT - A Performance Prediction Tool for Massive Parallel Sysytems

Abstract: Today's massively parallel machines are typically message passing systems consisting of hundreds or thousands of processors. Implementing parallel applications efficiently in this environment is a challenging task. The Performance Prediction Tool (PerPreT) presented in this paper is useful for system designers and application developers. The system designers can use the tool to examine the effects of changes of architectural parameters on parallel applications (e.g., reduction of setup time, increase of link bandwidth, faster execution units). Application developers are interested in a fast evaluation of different parallelization strategies of their codes. PerPreT uses a relatively simple analytical model to predict speedup, execution time, computation time, and communication time for a parametrized application. Especially for large numbers of processors, PerPreT's analytical model is preferable to traditional models (e.g., Markov based approaches such as queueing and Petri net models). The applications are modelled through parameterized formulae for communication and computation. The parameters used by PerPreT include the problem size and the number of processors used to execute the program. The target systems are described by architectural parameters (e.g., setup times for communication, link bandwidth, and sustained computing performance per node).

Symbolic performance prediction of scalable parallel programs

Abstract: Recent advances in the power of parallel computers have made them attractive for solving large computational problems. Scalable parallel programs are particularly well suited to Massively Parallel Processing (MPP) machines since the number of computations can be increased to match the available number of processors. Performance tuning can be particularly difficult for these applications since it must often be performed with a smaller problem size than that targeted for eventual execution. This research develops a performance prediction methodology that addresses this problem through symbolic analysis of program source code. Algebraic manipulations can then be performed on the resulting analytical model to determine performance for scaled up applications on different hardware architectures. >

An integrated approach for criticality prediction

Abstract: The paper provides insight in techniques for criticality prediction as they are applied within the development of Alcatel 1000 S12 switching software. The primary goal is to identify critical components and to make failure predictions as early as possible during the life cycle and hence reduce managerial risk combined with too early or too late release. The approach is integrated in the development process and starts with complexity based criticality prediction of modules. Modules identified as overly complex are given additional tests or review efforts. Release time prediction and field performance prediction are both based on tailored ENHPP reliability models. For the complete approach of criticality prediction, recent data from the development of a switching system with around 2 MLOC is provided. The switching system is currently in operational use, thus allowing for validation and tuning of the prediction models.

Load balancing technique for parallel search with statistical model

Abstract: A parallel performance enhancement technique using statistical model to increase parallel system efficiency is discussed in this paper. This idea is applied to solving the constraint satisfaction problem (CSP) on a multi-processor shared memory system. The combination of reducing standard deviation of the execution time in each processor and increasing the correlation coefficient between every two processors' execution times proves to be a very effective method in balancing load without resorting to dynamic approaches. Significant improvement in system efficiency over a traditional divide-and-conquer approach is shown in solving CSP in parallel. A near-optimal (100%) average load balance result is obtained. The underlying theory can also be easily applied to other problems and can be further extended to provide system performance prediction and evaluation. >

Theoretical analysis and performance prediction of tracking in clutter with strongest neighbor filters

Abstract: A simple and commonly used method for tracking in clutter is the so-called strongest neighbor filter (SNF), which uses the "strongest neighbor" measurement, that is, the one with the strongest intensity (amplitude) in the neighborhood of the predicted target measurement, as if it were the true one. The purpose of this paper is two-fold. First, the following theoretical results of tracking in clutter with SNF are derived: the a priori probabilities of data association events and the one-step prediction of the matrix mean square error conditioned on these events. Secondly, a technique for prediction without recourse to expensive Monte Carlo simulation of the performance of SNF is presented. This technique can quantify the dynamic process of tracking divergence as well as the steady state performance. The technique is a new development along the line of the recently developed general approach prediction of algorithms with both continuous and discrete uncertainties.

IC performance prediction system

Abstract: This paper presents a methodology for predicting the electrical performance of integrated circuits prior to packaging and final test. A simulation-based approach is used to build response surface prediction models. The work focuses on determining the optimal set of test measurements required for on-line product-based control, package selection, binning and die selection for MCMs.

System performance prediction for personal communication systems

Abstract: A new location reliability formula, based on a path loss model appropriate for microcellular communication, is derived. Also an improved interpolation formula, for the calculation of log-normal shadow fading loss is presented.

Performance prediction of large MIMD systems for parallel neural network simulations

Abstract: In this paper, we present a performance prediction model for indicating the performance range of MIMD parallel processor systems for neural network simulations. The model expresses the total execution time of a simulation as a function of the execution times of a small number of kernel functions, which have to be measured on only one processor and one physical communication link. The functions depend on the type of neural network, its geometry, decomposition and the connection structure of the MIMD machine. Using the model, the execution time, speedup, scalability and efficiency of large MIMD systems can be predicted. The model is validated quantitatively by applying it to two popular neural networks, backpropagation and the Kohonen self-organizing feature map, decomposed on a GCel-512, a 512 transputer system. Measurements are taken from network simulations decomposed via dataset and network decomposition techniques. Agreement of the model with the measurements is within 1–14%. Estimates are given for the performances that can be expected for the new T9000 transputer systems. The presented method can also be used for other application areas such as image processing.

Thermal sensitivity analysis for the 119 PBGA-a framework for rapid prototyping

Abstract: New package prototyping is often a sequential process where the chip and system parameters are specified first, then the package design is initiated. A reduction in overall cycle time can be affected if the events occur simultaneously. This study proposes a methodology for addressing this issue. The method is outlined in the context of prototyping the 119 plastic ball grid array (PBGA) package thermal performance. The parameters influencing performance are system, device, or package based. Ranges for the "yet-to-be-fixed" parameters are determined and factorial analyses are used to yield approximate linear models with interactions for package performance. Once the device and system parameters are "fixed", the linear equations are solved simultaneously with junction and board temperature constraints to yield a design options map for package layout. The prototyping sequence for the PBGA results in substrate thermal conductivity, mother board thermal conductivity, mother board load, and heat sink attachment as the set of "variable" parameters-with other parameters being "fixed". The design options map gives the minimum substrate thermal conductivity needed to meet the thermal performance specification for a particular set of parameters. The substrate specification is further related to physical attributes required of the package in terms of thermal vias, thermal bumps and metal layers. These results are generically applicable to the PBGA family. A 119 PBGA package enclosing a 2 W chip requires a minimum substrate thermal conductivity of 0.03 W/cm-/spl deg/C to meet the junction temperature constraint for high performance workstation environments. Results of the performance prediction are further verified by a composite finite element simulation and experimental validation with prototypes. The design options map can be recreated without any additional simulation studies in the event any change in the "fixed" parameters occurs. The methodology described allows anticipation of design options in the "dynamic" environment of prototyping, and implementation of optimized package designs to meet performance under multiple customer environments.

Computer Simulation of Radiographic Process — A Study of Complex Component and Defect Geometry

Abstract: In proof of the applicability of a testing method, a global concept of reliability of NDE should be used which is based on both the theoretical and experimental procedure demonstration. This concept includes consequently the performance demonstration of the whole testing chain, including the physical method, the device and the inspector, by using statistical methods as well as empirical data base, and the performance prediction by computer simulation. In addition to this, the computer modeling can be used to optimise testing parameters, to make feasibility analysis for special testing problems, and to support the interpretation of indications, especially for very complicated component geometry. Thus, this variety of applications makes it necessary to develop an easy and practical model to describe the process of radiographie testing with good accuracy and with a small numerical expense, which may include any component geometry and uses experimental findings. A macroscopic model for the X-ray projection process was developed where the X-ray beam and target interaction is described as a simplified transport process. The influence of the scattering effects on the imaging process can be realised with acceptable accuracy by using a global description. The efficiency and limits of the model are discussed in terms of different examples exploring the detectability of cracks with complicated geometry. These cracks are considered for different types of high performance materials: (i) composites in the aircraft and space industry sector and (ii) realistic austenitic welding from the nuclear industry sector. These examples will show the capability of the modeling tool to handle also complicated object geometries and testing conditions like double wall penetration and ellipsoid shot.

Speedy: An Integrated Performance Extrapolation Tool for pC++ Programs

Abstract: Performance extrapolation is the process of evaluating the performance of a parallel program in a target execution environment using performance information obtained for the same program in a different environment. Performance extrapolation techniques are suited for rapid performance tuning of parallel programs, particularly when the target environment is unavailable. This paper describes one such technique that was developed for data-parallel C++ programs written in the pC++ language. In pC++, the programmer can distribute a collection of objects to various processors and can have methods invoked on those objects execute in parallel. Using performance extrapolation in the development of pC++ applications allows tuning decisions to be made in advance of detailed execution measurements. The pC++ language system includes Τ, an integrated environment for analyzing and tuning the performance of pC++ programs. This paper presents speedy, a new addition to Τ, that predicts the performance of pC++ programs on parallel machines using extrapolation techniques. Speedy applies the existing instrumentation support of Τ to capture high-level event traces of a n-thread pC++ program run on a uniprocessor machine together with trace-driven simulation to predict the performance of the program run on a target n-processor machine. We describe how speedy works and how it is integrated into Τ. We also show how speedy can be used to evaluate a pC++ program for a given target environment.

Microbenchmarking and Performance Prediction for Parallel

Abstract: Previous research on this project (in work by Saavedra and Smith) has presented performance evaluation of sequential computers. That work presented (a) measurements of machines at the source language primitive operation level; (b) analysis of standard benchmarks; prediction of run times based on separate measurements of the machines and the programs; (d) analysis of the effectiveness of compiler optimizations; and (e) measurements of the performance and design of cache memories. In this paper, we extend the earlier work to parallel computers. We describe a portable benchmarking suite and performance prediction methodology, which accurately predicts the run times of Fortran 90 programs running upon supercomputers. The benchmarking suite measures the optimization capabilities of a given Fortran 90 compiler, execution rates of abstract Fortran 90 operations, and the processing characteristics of the underlying architecture as exposed by compiler-generated code. To predict the run time of an arbitrary program, we combine our benchmark results with dynamic execution measurements, and augment the resulting prediction with simple factors which account for overhead due to architecture-specific effects, such as remote reference latencies. We measure two supercomputers: a dedicated 128-node TMC CM-5, a distributed memory multiprocessor, and a 4-node partition of a Cray YMP-C90, a tightly-integrated shared memory multiprocessor. Our measurements show that the performance of the YMP-C90 far outstrips that of the CM-5, due to the quality of the compilers available and the architectural characteristics of each machine. To validate our prediction methodology, we predict the run time of five interesting kernels on these machines; nearly all of the predicted run times are within 50-percent of actual run times, much closer than might be expected.

Compile-time Performance Prediction of Parallel Systems

Abstract: A compile-time technique is outlined that yields low-cost, analytic performance models, intended for crude scalability analysis and first-order system design. The approach extends current static techniques by accounting for any type of resource contention that may occur. In this paper we report on the accuracy of the prediction method in terms of theory, simulation experiments, as well as measurements on a distributed-memory machine. It is shown that for series-parallel computations with random resource access patterns, the average prediction error is limited well within 50 % regardless the system parameters, where traditional compile-time methods yield errors up to orders of magnitude.

Implementation of finite-element analysis in a cage induction motor electrical design program

Abstract: The use of separate reduced models for both stator and rotor, together with the familiar steady-state per-phase equivalent circuit, has facilitated the development of a finite-element based analysis for induction motors that is both accurate and fast enough for use in a large induction motor design office. The method, has been shown to give performance prediction accuracies that compare favourably with classical techniques over a range of pole numbers and ratings. The method outlined has been implemented in a performance prediction program named FEPERF, which runs on a standard workstation.

Performance prediction of scalable computing: a case study

Abstract: While computers with tens of thousands of processors have successfully delivered high performance power for solving some of the so-called "grand-challenge" applications, the notion of scalability is becoming an important metric in the evaluation of parallel machine architectures and algorithms. In this study the prediction of scalability and its application are carefully investigated. A simple formula is presented to show the relation between scalability, single processor computing power, and degradation of parallelism. A case study is conducted on a multi-ring KSR-1 shared virtual memory machine. Experimental and theoretical results show that the influence of topology variation of an architecture is predictable. Therefore, the performance of an algorithm on a sophisticated hierarchical architecture can be predicted and a good algorithm-machine combination can be selected for a given application. >