Digital Commons Network^™

The Use Of The National Information Infrastructure And High Performance Computers In Industry, Geoffrey C. Fox, Wojtek Furmanski

Northeast Parallel Architecture Center

We divide potential NII (National Information Infrastructure) services into five broad areas: Collaboration and televirtuality; InfoVISiON (Information, Video, Imagery, and Simulation on Demand), and digital libraries; commerce; metacomputing; WebTop productivity services. The latter denotes the broad suite of tools we expect to be offered on the Web in a general environment we term WebWindows. We review current and future World Wide Web technologies, which could underlie these services. In particular, we suggest an integration framework WebWork for High Performance (parallel and distributed) computing and the NII. We point out that pervasive WebWork and WebWindows technologies will enable, facilitate and substantially …

Cluster Computing Review, Mark Baker, Geoffrey C. Fox, Hon W. Yau

Northeast Parallel Architecture Center

In the past decade there has been a dramatic shift from mainframe or ‘host−centric’ computing to a distributed ‘client−server’ approach. In the next few years this trend is likely to continue with further shifts towards ‘network−centric’ computing becoming apparent. All these trends were set in motion by the invention of the mass−reproducible microprocessor by Ted Hoff of Intel some twenty−odd years ago. The present generation of RISC microprocessors are now more than a match for mainframes in terms of cost and performance. The long−foreseen day when collections of RISC microprocessors assembled together as a parallel computer could out perform the …

Exploiting High Performance Fortran For Computational Fluid Dynamics, Volume 919, Ken Hawick, Geoffrey C. Fox

Northeast Parallel Architecture Center

We discuss the High Performance Fortran data parallel programming language as an aid to software engineering and as a tool for exploiting High Performance Computing systems for computational uid dynamics applications. We discuss the use of intrinsic functions, data distribution directives and explicitly parallel constructs to optimize performance by minimizing communications requirements in a portable manner. In particular we use an implicit method such as the ADI algorithm to illustrate the major issues. We focus on regular mesh problems, since these can be efficiently represented by the existing HPF definition, but also discuss issues arising from the use of irregular …

High Performance Distributed Computing, Geoffrey C. Fox

Northeast Parallel Architecture Center

High Performance Distributed Computing (HPDC) is driven by the rapid advance of two related technologies -- those underlying computing and communications, respectively. These technology pushes are linked to application pulls, which vary from the use of a cluster of some 20 workstations simulating fluid flow around an aircraft, to the complex linkage of several hundred million advanced PCs around the globe to deliver and receive multimedia information. The review of base technologies and exemplar applications is followed by a brief discussion of software models for HPDC, which are illustrated by two extremes -- PVM and the conjectured future World Wide …

Basic Issues And Current Status Of Parallel Computing -- 1995, Geoffrey C. Fox

Northeast Parallel Architecture Center

The best enterprises have both a compelling need pulling them forward and an innovative technological solution pushing them on. In high-performance computing, we have the need for increased computational power in many applications and the inevitable long-term solution is massive parallelism. In the short term, the relation between pull and push may seem unclear as novel algorithms and software are needed to support parallel computing. However, eventually parallelism will be present in all computers -- including those in your children's video game, your personal computer or workstation, and the central supercomputer.

Effects Of Technology Mapping On Fault Detection Coverage In Reprogrammable Fpgas, Kevin A. Kwiat, Warren Debany, Salim Hariri

Electrical Engineering and Computer Science - All Scholarship

Although Field-Programmable Gate Arrays (FPGAs) are tested by their manufacturers prior to shipment, they are still susceptible to failures in the field. In this paper, test vectors generated for the emulated (i.e., mission) circuit are fault simulated on two different models: the original view of the circuit, and the design as it is mapped to the FPGA's logic cells. Faults in the cells and in the programming logic are considered. Experiments show that this commonly-used approach fails to detect most of the faults in the FPGA.

Shape Reconstruction From Shading Using Linear Approximation, Ping Sing Tsai

Retrospective Theses and Dissertations

Shape from shading (SFS) deals with the recovery of 3D shape from a single monocular image. This problem was formally introduced by Horn in the early 1970s. Since then it has received considerable attention, and several efforts have been made to improve the shape recovery. In this thesis, we present a fast SFS algorithm, which is a purely local method and is highly parallelizable. In our approach, we first use the discrete approximations for surface gradients, p and q, using finite differences, then linearize the reflectance function in depth, Z ( x , y), instead of p and q. This …

Experimenting With The Finite Element Method In The Calculation Of Radiosity Form Factors, Donna Marie Chesteen

Retrospective Theses and Dissertations

Radiosity has been used to create some of the most photorealistic computer-generated images to date. The problem, however, is that radiosity algorithms are so computationally and memory expensive that few applications can employ them successfully. Form factor calculation is the most costly part of the process. This report describes an algorithm for using the finite element method to reduce the amount of time that is used in the form factor calculation portion of the radiosity algorithm. This technique for form factor calculation significantly reduces the number of projections done at each iteration by using shape functions to determine the distribution …

A Trainable, Single-Pass Algorithm For Column Segmentation, Son Sylwester, Sharad C. Seth

School of Computing: Conference and Workshop Papers

Column Segmentation logically precedes OCR in the document analysis process. The trainable algorithm described here, XYCUT, relies on horizontal and vertical binary profiles to produce an XY- tree representing the column structure of a page of a technical document in a single pass through the bit image. Training against ground truth adjusts a single, resolution independent, parameter using only local information and guided by an edit distance function. The algorithm correctly segments the page image for a (fairly) wide range of parameter values, although small, local and repairable errors may be made, an effect measured by a repair cost function.

Production Of Excited Beauty States In Z Decays, D. Buskulic, Manoj Thulasidas

Research Collection School Of Computing and Information Systems

A data sample of about 3.0 million hadronic Z decays collected by the ALEPH experiment at LEP in the years 1991 through 1994, is used to make an inclusive selection of B hadron events.

Inclusive Production Of Neutral Vector Mesons In Hadronic Z Decays, D. Buskulic, Manoj Thulasidas

Research Collection School Of Computing and Information Systems

Data on the inclusive production of the neutral vector mesonsρ 0(770),ω(782), K*0(892), andφ(1020) in hadronic Z decays recorded with the ALEPH detector at LEP are presented and compared to Monte Carlo model predictions. Bose-Einstein effects are found to be important in extracting a reliable value for theρ 0 production rate. An averageρ 0 multiplicity of 1.45±0.21 per event is obtained. Theω is detected via its three pion decay modeω→π + π − π 0 and has a total rate of 1.07±0.14 per event. The multiplicity of the K*0 is 0.83±0.09, whilst that of theφ is 0.122±0.009, both measured using their …

Multiple Query Optimization With Depth-First Branch-And-Bound And Dynamic Query Ordering, Ee Peng Lim, Ahmet Cosar, Jaideep Srivastava

Research Collection School Of Computing and Information Systems

In certain database applications such as deductive databases, batch query processing, and recursive query processing etc., usually a single query gets transformed into a set of closely related database queries. Also, great benefits can be obtained by executing a group of related queries all together in a single unified multi-plan instead of executing each query separately. In order to achieve this Multiple Query Optimization (MQO) identifies common task(s) (e.g. common subexpressions, joins, etc.) among a set of query plans and creates a single unified plan (multi-plan) which can be executed to obtain the required outputs for all queries at once. …

Using Multiple Statistical Prototypes To Classify Continuously Valued Data, Tony R. Martinez, Dan A. Ventura

Faculty Publications

Multiple Statistical Prototypes (MSP) is a modification of a standard minimum distance classification scheme that generates muItiple prototypes per class using a modified greedy heuristic. Empirical comparison of MSP with other well-known learning algorithms shows MSP to be a robust algorithm that uses a very simple premise to produce good generalization and achieve parsimonious hypothesis representation.

Rule Extraction: From Neural Architecture To Symbolic Representation, Gail A. Carpenter, Ah-Hwee Tan

Research Collection School Of Computing and Information Systems

This paper shows how knowledge, in the form of fuzzy rules, can be derived from a supervised learning neural network called fuzzy ARTMAP. Rule extraction proceeds in two stages: pruning, which simplifies the network structure by removing excessive recognition categories and weights; and quantization of continuous learned weights, which allows the final system state to be translated into a usable set of descriptive rules. Three benchmark studies illustrate the rule extraction methods: (1) Pima Indian diabetes diagnosis, (2) mushroom classification and (3) DNA promoter recognition. Fuzzy ARTMAP and ART-EMAP are compared with the ADAP algorithm, the k nearest neighbor system, …

Hardware Assists For High Performance Computing Using A Mathematics Of Arrays, Hardy J. Pottinger, W. Eatherton, J. Kelly, T. Schiefelbein, Lenore Mullin, R. Ziegler

Electrical and Computer Engineering Faculty Research & Creative Works

Work in progress at the University of Missouri-Rolla on hardware assists for high performance computing is presented. This research consists of a novel field programmable gate array (FPGA) based reconfigurable coprocessor board (the Chameleon Coprocessor) being used to evaluate hardware architectures for speedup of array computation algorithms. These algorithms are developed using a Mathematics of Arrays (MOA). They provide a means to generate addresses for data transfers that require less data movement than more traditional algorithms. In this manner, the address generation algorithms are acting as an intelligent data prefetching mechanism or special purpose cache controller. Software implementations have been …

An Efficient Multicomputer Algorithm For The Solution Of Chemical Process Flowsheeting Equations, Fikret Ercal, Neil L. Book, S. Pait, J. J. Fielding

Computer Science Faculty Research & Creative Works

This paper presents a parallel solution method of large sparse systems of linear equations arising in the context of a chemical process flowsheeting application on a message passing multicomputer. To maximize the performance, the algorithm uses a novel matrix decomposition and solution method, called parallel two-phased LU decomposition, which schedules the concurrent tasks in a maximally overlapping manner, and at the same time, tries to minimize the interprocessor data dependencies and obtain optimal load balancing. The forward elimination step is performed concurrently with the parallel two-phased LU decomposition step and backward substitution is parallelized in a piecewise manner. Implementation results …

Using Neural Networks For Aerodynamic Parameter Modeling, Gerald E. Peterson, William E. Bond, Roger Germann, Barry Streeter, James Urnes

Computer Science Faculty Research & Creative Works

Neural networks are being developed at McDonnell Douglas Corporation to provide an onboard model of an aircraft's aerodynamics to support advanced flight control systems. These flight control systems, constructed using neural networks and advanced controllers, have the potential to reduce flight control development costs and to improve inflight performance. Neural networks are useful in this situation because they can compactly represent the data and operate in real-time

Using Taguchi''S Method Of Experimental Design To Control Errors In Layered Perceptrons, William E. Bond, Gerald E. Peterson, Daniel C. St. Clair, Stephen R. Aylward

Computer Science Faculty Research & Creative Works

A significant problem in the design and construction of an artificial neural network for function approximation is limiting the magnitude and the variance of errors when the network is used in the field. Network errors can occur when the training data does not faithfully represent the required function due to noise or low sampling rates, when the network's flexibility does not match the variability of the data, or when the input data to the resultant network is noisy. This paper reports on several experiments whose purpose was to rank the relative significance of these error sources and thereby find neural …

Using Fuzzy Set Theory: Exploring How Generalties Can Pinpoint Where To Look For The Answers, Cosmin Radu, Ralph W. Wilkerson

Computer Science Faculty Research & Creative Works

No abstract provided.

Object Interactions As First Class Objects: From Design To Implementation, Mahesh Dodani, Benjamin Kok Siew Gan, Lizette Velazquez

Research Collection School Of Computing and Information Systems

Collaborations between objects make up the dynamic behavior of OO software. These collaborations among objects require careful design and implementation. Treating the interactions as responsibilities that are integrated in the participating objects, results in tight coupling between objects. Tight coupling increases complexity and reduces reusability. Object interactions need to be first class objects from design to implementation. Our research provides a unified approach to model and implement these interactions as first class objects. During analysis and design, they are modeled using DynaSpecs. During implementation, they are coded with a new language construct called Compositions. DynaSpecs and Compositions provide a consistent …

Automatic Pcb Inspection Systems, M. Moganti, Fikret Erçal

Computer Science Faculty Research & Creative Works

There are more than 50 process steps required to fabricate a printed circuit board (PCB). To ensure quality, human operators simply inspect the work visually against prescribed standards. The decisions made by this labor intensive, and therefore costly, procedure often also involve subjective judgements. Automatic inspection systems remove the subjective aspects and provide fast, quantitative dimensional assessments. Machine vision may answer the manufacturing industry's need to improve product quality and increase productivity. The major limitation of existing inspection systems is that all the algorithms need a special hardware platform to achieve the desired real-time speeds. This makes the systems extremely …

Adaptive Resonance Theory (Art): An Introduction, Lucien G. Heins, Daniel R. Tauritz

Computer Science Faculty Research & Creative Works

No abstract provided.

Relaxing Synchronization In Distributed Simulated Annealing, Bruce M. Mcmillin, Chul-Eui Hong

Computer Science Faculty Research & Creative Works

This paper presents a cost error measurement scheme and relaxed synchronization method, for simulated annealing on a distributed memory multicomputer, which predicts the amount of cost error that an algorithm will tolerate. An adaptive error control method is developed and implemented on an Intel iPSC/2

Ensuring The Satisfaction Of A Temporal Specification At Run-Time, Grace Tsai, Matt Insall, Bruce M. Mcmillin

Mathematics and Statistics Faculty Research & Creative Works

A responsive computing system is a hybrid of real-time, distributed and fault-tolerant systems. In such a system, severe consequences can occur if the run-time behavior does not conform to the expected behavior or specifications. In this paper, we present a formal approach to ensure satisfaction of the specifications in the operational environment as follows. First we specify behavior of the systems using Interval Temporal Logic (ITL). Next we give algorithms for trace checking of programs in such systems. Finally, we present a fully distributed run-time evaluation system which causally orders the events of the system during its execution and checks …

Comparing Traditional Statistical Models With Neural Network Models: The Case Of The Relation Of Human Performance Factors To The Outcomes Of Military Combat, William Oliver Hedgepeth

Engineering Management & Systems Engineering Theses & Dissertations

Statistics and neural networks are analytical methods used to learn about observed experience. Both the statistician and neural network researcher develop and analyze data sets, draw relevant conclusions, and validate the conclusions. They also share in the challenge of creating accurate predictions of future events with noisy data.

Both analytical methods are investigated. This is accomplished by examining the veridicality of both with real system data. The real system used in this project is a database of 400 years of historical military combat. The relationships among the variables represented in this database are recognized as being hypercomplex and nonlinear.

The …

An Ada-Like Language To Facilitate Reliable Coding Of Low Cost Embedded Systems, Michael Collins

Theses : Honours

Due to a lack of operating system (0/S) support, it is more difficult to develop programs for embedded systems than for workstations. For those developing on a low budget, the problem is often further compounded by the necessity of using inappropriate, 0/S dependent, compilers. This study attempts to ascertain those elements of a High Level Language (HLL) which are absolutely necessary and implementable to produce reliable, efficient, embedded programs without the benefit of a large budget. The study is based upon the Ada philosophy as the Ada language incorporates many desirable features for modelling real-world problems in terms of embedded …

Transplanting Manufacturing Power: Observations From Five Years At Fujitsu, Steven M. Miller

Research Collection School Of Computing and Information Systems

Within a factory, what are the underpinnings of manufacturing power? During my thirty-month residency at Fujitsu's factory in Oyama, Japan, I constantly reflected on this question. The standard view of a factory structure provided by an organization chart was inadequate for summarizing my observations.

First Measurement Of The Quark-To-Photon Fragmentation Function, D. Buskulic, Manoj Thulasidas

Research Collection School Of Computing and Information Systems

Earlier measurements at LEP of isolated hard photons in hadronic Z decays, attributed to radiation from primary quark pairs, have been extended in the ALEPH experiment to include hard photon production inside hadron jets. Events are selected where all particles combine democratically to form hadron jets, one of which contains a photon with a fractional energy z 0:7. After statistical subtraction of non-prompt photons, the quark-to-photon fragmentation function, D(z), is extracted directly from the measured 2-jet rate. By taking into account the perturbative contributions to D(z) obtained from an O(S ) QCD calculation, the unknown non-perturbative component of D(z) is …

Improved Tau Polarisation Measurement, D. Buskulic, M. Thulasidas

Research Collection School Of Computing and Information Systems

Using 22 pb−1 of data collected at LEP in 1992 on the peak of the Z resonance, the ALEPH collaboration has measured the polarisation of the tau leptons decaying intoevv¯,μvv¯evv¯,μvv¯,πν, ρν and a1 ν from their individual decay product distributions. The measurement of the tau polarisation as a function of the production polar angle yields the two parametersN τ andN e, where, in terms of the axial and vector couplingsg Al andg Vl,N l=2g Vl gAl/(g 2Vl+g2Al). This analysis follows to a large extent the methods devised for the 1990 and 1991 data but with improvements which bring a better …

Adaptive Resonance Associative Map, Ah-Hwee Tan

Research Collection School Of Computing and Information Systems

This article introduces a neural architecture termed Adaptive Resonance Associative Map (ARAM) that extends unsupervised Adaptive Resonance Theory (ART) systems for rapid, yet stable, heteroassociative learning. ARAM can be visualized as two overlapping ART networks sharing a single category field. Although ARAM is simpler in architecture than another class of supervised ART models known as ARTMAP, it produces classification performance equivalent to that of ARTMAP. As ARAM network structure and operations are symmetrical, associative recall can be performed in both directions. With maximal vigilance settings, ARAM encodes pattern pairs explicitly as cognitive chunks and thus guarantees perfect storage and recall …