Digital Commons Network^™

Automating Output Size And Reuse Metrics In A Repository-Based Computer-Aided Software Engineering (Case) Environment, R. D. Banker, Robert J. Kauffman, C. Wright, D. Zweig

Research Collection School Of Computing and Information Systems

Measurement of software development productivity is needed in order to control software costs, but it is discouragingly labor-intensive and expensive. Computer-aided software engineering (CASE) technologies-especially repository-based, integrated CASE-have the potential to support the automation of this measurement. We discuss the conceptual basis for the development of automated analyzers for function point and software reuse measurement for object-based CASE. Both analyzers take advantage of the existence of a representation of the application system that is stored within an object repository, and that contains the necessary information about the application system. We also discuss metrics for software reuse measurement, including reuse leverage, …

Dyadic Wavelet Features For Isolated Word Speaker Dependent Speech Recognition, Stephen Ainge

Theses and Dissertations

This research examines the use of dyadic wavelet features for the recognition of speaker dependent isolated word speech. The features were generated using three different wavelet filters-Daubechies 4 coefficient (Db4), Daubechies 20 coefficient (Db20) and a 31 coefficient cubic spline and three different window lengths-15ms, 8ms and 4ms. The accuracy of the standard and over-sampled dyadic wavelet methods were compared. The over-sampled dyadic wavelet method using the Db4 scaling function, with a maximum accuracy of 65.5, was found to be the most accurate of the wavelet methods tested. The accuracy of this over-sampled dyadic Db4 wavelet method was compared to …

Clustering Techniques In Speaker Recognition, Douglas N. Prescott

Theses and Dissertations

This thesis presents a comparison based on identification rate, of three clustering techniques applied to cepstral features for speaker identification. LBG vector quantization as developed by Linde, Buzo and Gray; is used to provide benchmark performance for comparison with Fuzzy clustering (based on the unsupervised fuzzy partition-optimal number of classes, UFP-ONC algorithm by Gath and Geva) and an Artificial Neural Network, the Multilayer Perceptron. Cepstral features from the TIMIT, King and AFIT93 corpus speaker databases are used to produce speaker-identification classifiers using each of the clustering algorithms. The experiment reported evaluates the speaker identification performance using the 20-dimensional cepstral features …

A Reduction Semantics For Array Expressions:The Psi Compiler, L. Mullin, S. Thibault

Computer Science Technical Reports

No abstract provided.

Formal Verification Of Distributed Deadlock Detection Algorithm Using A Time-Dependent Proof Technique, Pei-Yu Li, Bruce Mcmillin

Computer Science Technical Reports

A large number of published distributed deadlock detection/resolution algorithms are found to be incorrect because they have used informal approaches to prove the correctness of their algorithms. In this paper, we present a formal approach for the correctness proof and give an example of the proof. In this proposed approach, a formal model of distributed deadlock is presented with a local-time deadlock specification for correctness verification. With the formal model, we have an insight into the definition of deadlock in local views which is used to show the existence of a real deadlock. A rigorous proof to show the equivalence …

Conjugating Polynomials On Finite Rings, M. Insall, L. Mullin, R. Wilkerson

Computer Science Technical Reports

No abstract provided.

Design And Implementation Of Ipis : An X-Window Based Image Processing Interactive System, Eduardo Morales

Theses

Most of image processing systems are based on command line functions or can only display one image at a time. This is a serious inconvenience for those who need an interactive system session or want to compare two images processed by different techniques at the same time.

The system was designed with these problems in mind. It is able to display the processed image right after an operation and to display several images simultaneously, making it simple to compare techniques. The system was also created with the purpose to be used in an academic environment. Its structured design makes it …

A Pipeline Implementation Of Lu-Decomposition On A Hypercube, Scott Thibault, Lenore Mullin

Computer Science Technical Reports

This paper presents a method of performing LU-Decomposition on a hypercube. The algorithm is constructed similar to the way a pipeline would be used in hardware or on a systolic array processor. Using this method eliminates the need to broadcast data to all the processors at each iteration of the algorithm. Communication takes place at each iteration but since we are using a pipeline model, the communication is between each processor and the next processor in the pipeline. On the hypercube this communication can be done in parallel by using the Gray code ordering to embed a ld mesh (pipeline) …

Formal Methods For Portable, Scalable, Scheduling, Routing And Communication Protocol, Lenore M. R. Mullin, Scott A. Thibault, Daria R. Dooling, Erik A. Sandberg

Computer Science Technical Reports

The PRAM model has been shown to be an optimal design for emulating both loos and tightly coupled multiprocessors for unit time operations. Extensions to this design employ software pipelining on a network of homogeneous workstations. Partitioning array data structures and pipelining groups of partitions to processors can minimize latency and bottlenecking on distributted message passing multiprocessing architectures. Our previous paper developed a general message passing design that was conjectured to port to a CMS and scale to more processors connected via TCPIP protocol. This paper presents the results of of our ported and scaled designs. We have indeed developed …

Effective Data Parallel Computation Using The Psi Calculus, L. R. Mullin, M. A. Jenkins

Computer Science Technical Reports

No abstract provided.

Efficient Run-Time Assurance In Distributed Systems Through Selection Of Executable Assertions, Martina Schollmeyer, Bruce Mcmillin

Computer Science Technical Reports

Run-time assurance of a distributed system can be obtained by comparing, at run-time, the behavior of the program with the expected behavior described in the program's specification. Executable assertions, embedded into the program code, can determine when there are discrepancies, due to processor failures, between actual and expected behavior. Thus, there is no global monitoring scheme but processes will check each other.

A non-faulty process will always perform correct computation. It can detect errors in other processes after receiving information from them and checking it against expected values by using executable assertions. In order to efficiently check programs at run-time, …

A Data Parallel Algorithm For Solving The Region Growing Problem On The Connection Machine, Nawal Copty, Sanjay Ranka, Geoffrey C. Fox, Ravi V. Shankar

College of Engineering and Computer Science - Former Departments, Centers, Institutes and Projects

Region growing is a general technique for image segmentation, where image characteristics are used to group adjacent pixels together to form regions. This paper presents a parallel algorithm for solving the region growing problem based on the split and merge approach, and uses it to test and compare various parallel architectures and programming models. The implementations were done on the Connection Machine, models CM-2 and CM-5, in the data parallel and message passing programming models. Randomization was introduced in breaking ties during merging to increase the degree of parallelism, and only one and two-dimensional arrays of data were used in …

The Complexity Of Local Stratification, Peter Cholak, Howard A. Blair

College of Engineering and Computer Science - Former Departments, Centers, Institutes and Projects

The class of locally stratified logic programs is shown to be Π 1 1-complete by the construction of a reducibility of the class of infinitely branching nondeterministic finite register machines.

Genetic Algorithms For Graph Partitioning And Incremental Graph Partitioning, Harpal Maini, Kishan Mehrotra, Chilukuri K. Mohan, Sanjay Ranka

College of Engineering and Computer Science - Former Departments, Centers, Institutes and Projects

Partitioning graphs into equally large groups of nodes, minimizing the number of edges between different groups, is an extremely important problem in parallel computing. This paper presents genetic algorithms for suboptimal graph partitioning, with new crossover operators (KNUX, DKNUX) that lead to orders of magnitude improvement over traditional genetic operators in solution quality and speed. Our method can improve on good solutions previously obtained by using other algorithms or graph theoretic heuristics in minimizing the total communication cost or the worst case cost of communication for a single processor. We also extend our algorithm to Incremental Graph Partitioning problems, in …

The Transportation Primitive, Ravi V. Shankar, Khaled A. Alsabti, Sanjay Ranka

College of Engineering and Computer Science - Former Departments, Centers, Institutes and Projects

This paper presents algorithms for implementing the transportation primitive on a distributed memory parallel architecture. The transportation primitive performs many-to-many personalized communication with bounded incoming and outgoing traffic. We present a two-stage deterministic algorithm that decomposes the communication with possibly high variance in message size into two communication stages with low message size variance. If the maximum outgoing or incoming traffic at any processor is t, transportation can be done in 2t¯ time (+ lower order terms) when t O(p 2 + pø=¯) (¯ is the inverse of the data transfer rate, ø is the startup overhead). If the maximum …

Parallel Incremental Graph Partitioning Using Linear Programming, Chao Wei Ou, Sanjay Ranka

College of Engineering and Computer Science - Former Departments, Centers, Institutes and Projects

Partitioning graphs into equally large groups of nodes while minimizing the number of edges between different groups is an extremely important problem in parallel computing. For instance, efficiently parallelizing several scientific and engineering applications requires the partitioning of data or tasks among processors such that the computational load on each node is roughly the same, while communication is minimized. Obtaining exact solutions is computationally intractable, since graph-partitioning is an NP-complete. For a large class of irregular and adaptive data parallel applications (such as adaptive meshes), the computational structure changes from one phase to another in an incremental fashion. In incremental …

Random Data Accesses On A Coarse-Grained Parallel Machine Ii. One-To-Many And Many-To-One Mappings, Ravi V. Shankar, Sanjay Ranka

College of Engineering and Computer Science - Former Departments, Centers, Institutes and Projects

This paper describes deterministic communication-efficient algorithms for performing random data accesses with hot spots on a coarse-grained parallel machine. The general random access read/write operations with hot spots can be completed in Clanip (+ lower order terms) time and is optimal and scalable provided n _> O(pa+p2r/l) (n is the number of elements distributed across p processors, r is the start-up overhead and 1/It is the data transfer rate). C is a small constant between 3 and 4 for the random access write operation, slightly higher for the random access read operation. Monotonic random access reads/writes can be completed with …

Optimal Processor Assignment For A Class Of Pipelined Computations, Alok N. Choudhary, Bhagirath Narahari, David Nicol, Rahul Simha

Electrical Engineering and Computer Science - All Scholarship

The availability of large scale multitasked parallel architectures introduces the following processor assignment problem: we are given a long sequence of data sets, each of which is to undergo processing by a collection of tasks whose inter-task data dependencies form a series-parallel partial order. Each individual task is potentially parallelizable, with a known experimentally determined execution signature. Recognizing that data sets can be pipelined through the task structure, the problem is to find a "good" assignment of processors to tasks. Two objectives interest us: minimal response time per data set given a throughput requirement, and maximal throughput given a response …

Design Issues For The Parallelization Of An Optimal Interpolation Algorithm, Gregor Von Laszewski, Mike Seablom, Miloje Makivic, Peter Lyster

Northeast Parallel Architecture Center

A regionalized optimal interpolation algorithm is currently used at the NASA Goddard Data Assimilation Office (DAO) for four dimensional data assimilation. Instead of using all observations regions are defined to approximate the solution. The sequential code for the regionalized optimal interpolation is very complex and in its current form unusable for MIMD machines. This paper describes the efforts at the DAO to parallelize the existing sequential algorithm. We outline three strategies transforming the sequential algorithm gradually to MIMD machines. A major requirement for the new parallel algorithm is the portability to as many MIMD machines as possible. Therefore, the parallel …

A Generalized Expression Optimization Hook For C++ On High-Performance Architectures, David J. Edelsohn

Northeast Parallel Architecture Center

C++ has gained broad acceptance as an object-oriented evolutionary extension to the C language, but it severely constrains methods for operating on class objects by forcing all data manipulation through an interface which assumes that all basic operations can be implemented as they are written: as unary or binary operators. C++ allows great flexibility in the creation of complex data structures which can perform the same functionality as built-in types of many other languages, but unfortunately it does not allow an equivalent level of flexibility so that operators acting on those data types can achieve the same level of efficiency …

The Binational English & Spanish Telecommunications Network, Armando A. Arias Jr.

SSGS Faculty Publications and Presentations

BESTNET was established in the early 1980's, as an effort to link universities on both sides of the U.S.- Mexico border through microwave, satellite and cable television technologies. In the late 1980's BESTNET focused primarily on the development of asynchronous computer mediated learning and teaching in an internationally networked virtual environment. For the past six years (1990's) BESTNET has strengthened its binational ties and continued its "high tech" focus through the development of active or vibrant model technology which is assisting in the creation of an on-line binational university setting that is "borderless" (albeit, seamless to the user). Today, this …

Performance Modeling Of Load Balancing Algorithms Using Neural Networks, Ishfaq Ahmad, Arif Ghafoor, Kishan Mehrotra, Chilukuri K. Mohan, Sanjay Ranka

College of Engineering and Computer Science - Former Departments, Centers, Institutes and Projects

This paper presents a new approach that uses neural networks to predict the performance of a number of dynamic decentralized load balancing strategies. A distributed multicomputer system using any distributed load balancing strategy is represented by a unified analytical queuing model. A large simulation data set is used to train a neural network using the back–propagation learning algorithm based on gradient descent. The performance model using the predicted data from the neural network produces the average response time of various load balancing algorithms under various system parameters. The validation and comparison with simulation data show that the neural network is …

Scheduling Of Unstructured Communication On The Intel Ipsc/860, Jhy-Chun Wang, Sanjay Ranka

College of Engineering and Computer Science - Former Departments, Centers, Institutes and Projects

In this paper we present several algorithms for decomposing all-to-many personalized communication into a set of disjoint partial permutations. These partial permutations avoid node contention as well as link contention. We discuss the theoretical complexity of these algorithms and study their effectiveness both from the view of static scheduling and from runtime scheduling. Experimental results for our algorithms are presented on the iPSC/860.

A Fuzzy Logic-Based Foundation For Analyzing Imprecise Conflicting Requirements, J. Yen, Xiaoqing Frank Liu

Computer Science Faculty Research & Creative Works

Imprecise requirements are represented by the canonical form in test-score semantics. The concepts of feasibility, satisfiability, and specificity are formalized based on the fuzzy sets. The relationships between requirements are classified to be conflicting and cooperative. A feasible overall requirement can thus be formulated based on the tradeoff analysis of the conflicting requirements by using fuzzy multi-criteria optimization technique

Genetic Algorithms For Soft Decision Decoding Of Linear Block Codes, Harpal Maini, Kishan Mehrotra, Chilukuri K. Mohan, Sanjay Ranka

College of Engineering and Computer Science - Former Departments, Centers, Institutes and Projects

Soft-decision decoding is an NP-hard problem of great interest to developers of communication systems. We show that this problem is equivalent to the problem of optimizing Walsh polynomials. We present genetic algorithms for soft-decision decoding of binary linear block codes and compare the performance with various other decoding algorithms including the currently developed A* algorithm. Simulation results show that our algorithms achieve bit-error-probabilities as low as 0:00183 for a [104; 52] code with a low signal-to-noise ratio of 2:5 dB, exploring only 22; 400 code words, whereas the search space contains 4:5 \Theta 10 15 codewords. We define a new …

Passion Runtime Library For Parallel I/O, Rajeev Thakur, Rajesh Bordawekar, Alok Choudhary, Ravi Ponnusamy

Electrical Engineering and Computer Science - All Scholarship

We are developing a compiler and runtime support system called PASSION: Parallel And Scalable Software for Input-Output. PASSION provides software support for I/O intensive out-of-core loosely synchronous problems. This paper gives an overview of the PASSION Runtime Library and describes two of the optimizations incorporated in it, namely Data Prefetching and Data Sieving. Performance improvements provided by these optimizations on the Intel Touchstone Delta are discussed, together with an out of -core Median Filtering application.

Linear Codes And Error-Correction, Karen Brown

Presidential Scholars Theses (1990 – 2006)

The process of encoding information for transmission from one source to another is a vital process in many areas of science and technology. Whenever coded information is sent, there arises a certain possibility that an error will occur, either during transmission or in decoding. Therefore, it is imperative to develop methods to detect and correct errors in a code. The study of coding theory is a "new" area of mathematics which is relatively undeveloped.

This paper focuses on the properties of linear codes -and their corresponding methods of error-correction. To simplify the issue, only binary block codes are studied; hence …

Design Of An Application Development Toolkit For Hpf/Fortran 90d, Manish Parashar, Salim Hariri, Tomasz Haupt, Geoffrey C. Fox

Northeast Parallel Architecture Center

The development of efficient application software capable of exploiting available High Performance Computing (HPC) systems is non-trivial and is largely governed by the availability of sufficiently high-level languages, tools, and application development environments. In this paper we describe the design and operation of a toolkit for HPF/Fortran 90D application development. The toolkit incorporates the following systems: (1) ESP: An Interpretive Framework for HPF/Fortran 90D Performance Prediction; (2) ESP-i: A HPF/Fortran 90D Functional Interpreter; and (3) ESPial: An Integrated Environment for HPF/Fortran 90D Application Development & Execution. The toolkit has been implemented on the iPSC/860 hypercube system, and is supported by …

Interpreting The Performance Of Hpf/Fortran 90d, Manish Parashar, Salim Hariri, Tomasz Haupt, Geoffrey C. Fox

Northeast Parallel Architecture Center

In this paper we present a novel interpretive approach for accurate and cost-effective performance prediction in a high performance computing environment, and describe the design of a source-driven HPF/Fortran 90D performance prediction framework based on this approach. The performance prediction framework has been implemented as part of a HPF/Fortran 90D application development environment. A set of benchmarking kernels and application codes are used to validate the accuracy, utility, usability, and cost-effectiveness of the performance prediction framework. The use of the framework for selecting appropriate compiler directives and for application performance debugging is demonstrated. Keywords: Performance prediction, HPF/Fortran 90D application development, …

Run-Time And Compile-Time Support For Adaptive Irregular Problems, Shamik D. Sharma, Ravi Ponnusamy, Bongki Moon, Yuan-Shin Hwang

Northeast Parallel Architecture Center

In adaptive irregular problems the data arrays are accessed via indirection arrays, and data access patterns change during computation. Implementing such problems on distributed memory machines requires support for dynamic data partitioning, efficient preprocessing and fast data migration. This research presents efficient runtime primitives for such problems. This new set of primitives is part of the CHAOS library. It subsumes the previous PARTI library which targeted only static irregular problems. To demonstrate the efficacy of the runtime support, two real adaptive irregular applications have been parallelized using CHAOS primitives: a molecular dynamics code (CHARMM) and a particle-in-cell code (DSMC). The …