|
Muhammad E. Shaaban
Associate Professor
Computer Engineering, RIT
|
|
 Education
Ph.D. in Computer Engineering, University of Southern California,
May, 1994.
Thesis: Mapping Methodologies for
Heterogeneous Computing
Advisor:
Professor Viktor Prasanna.
M.S. in Electrical Engineering,
University of Petroleum and Minerals, Dhahran, Saudi Arabia, 1985.
B.S. in Electrical Engineering,
University of Petroleum and Minerals, 1983.
Course Work
 arallel Processing, Parallel Image Computing, Design of Parallel Algorithms,
Advanced Computer Architecture, Advanced Computer Arithmetic,
Dataflow Architectures, Interconnection Networks, Artificial Intelligence,
Operating Systems, Advanced Digital Systems Design.
Ph.D. Thesis Summary
 apping parallel
algorithms having diverse computational requirements,
such as those found in image processing tasks, onto a single architecture
may adversely affect algorithm performance. For such applications,
Heterogeneous Computing (HC) concurrently utilizing a suite of heterogeneous
computers has the potential of increased performance and cost effectiveness.
This thesis concentrates on fundamental design issues in HC. This includes
devising a programming paradigm (Hierarchical Cluster-M),
and mapping strategies that meet the requirements of HC.
One such mapping algorithm developed is suitable for mapping image processing
tasks onto a two-machine heterogeneous suite.
Research Interests
 igh Performance computing and advanced computer architecture, including:
Parallel algorithm design and mapping;
heterogeneous computing;
massively parallel computer architectures;
image computing.
Professional History
Assistant Professor, Department of Computer Engineering,
Rochester Institute of Technology, Rochester, New York, since December 1996.
Part-time Faculty, Department of Computer Engineering and Computer
Science, California State University, Long Beach, August 1996 to December 1996.
Postdoctoral Research Associate, Department of Computer and
Information Sciences , New Jersey Institute of Technology,
June 1994 to June 1995.
Research/Teaching Assistant,
Department of Electrical Engineering-Systems ,
University of Southern California,
January 1988 to May 1994.
Computer Room Operations Supervisor, University Computing Services,
University of Southern California, May 1987 to December 1993.
Research/Teaching Assistant, Department of Electrical Engineering,
University of Petroleum and Minerals, Fall 1983 to Fall 1984.
Projects
-
Design of Heterogeneous Algorithms for Image Understanding
Architecture:
A set of heterogeneous algorithms for computer vision tasks were
designed for the Image Understanding Architecture.
Problems studied include Hugh Transform Line detection,
finding geometric properties of images, and high level
image understanding tasks such as object matching.
-
Optical Techniques for Parallel Image Computing:
A study of efficient optical organizations for a set
of preprocessing tasks such as texture analysis, histogramming,
edge detection, dilation and contraction.
-
Partitioning of Image Processing Tasks on Heterogeneous Computer
Systems:
A fully polynomial time approximation scheme was devised
to partition chain or tree structured image processing tasks
onto a two processor heterogeneous system.
-
Design Issues for Heterogeneous Supercomputing:
issues and problems arising in using heterogeneous environments.
Preliminary solutions to a set of problems in heterogeneous
supercomputing are presented. Suitability of heterogeneous
environments for image Understanding is studied.
-
Heterogeneous Optimal Selection Theory (HOST):
a mathematical programming formulation was devised for choosing
the most appropriate suite of heterogeneous machines for a given
HOST incorporates heterogeneous parallelism embedded in the tasks
and it reflects the cost associated using various fine grain
mapping strategies.
-
Cluster-M Parallel Programming Paradigm:
A novel programming model which provides an environment
for efficiently designing highly parallel machine independent
software is proposed. Hierarchical Cluster-M with extensions
to accommodate heterogeneous systems is also developed.
-
Cluster-M based Mapping Methodologies:
Utilizing Cluster-M several efficient mapping strategies
were devised for a number of test applications.
Portions of this work were conducted as part of my postdoctoral
work at the New Jersey Institute of Technology.
-
Implementation Of Cluster-M:
The various components of Cluster-M were implemented and
tested to verify theoretical results using Program Composition
Notation (PCN). In addition several mapping strategies
for parallel applications were implemented to verify
mapping performance. This ongoing work was also conducted
during my postdoctoral assignment at the New Jersey Institute
of Technology with the team of Dr. Mary Eshaghian. The results of
this and previous project form the bases of two of Dr. Eshaghian's
students Ph.D. thesis.
Professional Activities
- Helped organize and referee papers submitted to the
Workshop on the Heterogeneous Computing,
at the IEEE International Parallel Processing Symposium,
1992--1994.
- Panelist of the panel session entitled
"Heterogeneous Computing: How practical is it?" held
at the 24th International Conference on Parallel Processing,
August, 1995.
Book / Magazine Contributions
-
"A Heterogeneous Model of Computation,"
with S. Chen, M. Eshaghian, and Y. Wu,
a book chapter in the edited volume
"Heterogeneous Computing," by
M. Eshaghian, Artech House, April 1996.
-
"Selected Heterogeneous Computing Applications,"
with Arif Ghafoor and Ashfaq Khokhar,
a book chapter in the edited volume
"Heterogeneous Computing," by
M. Eshaghian, Artech House, April 1996.
-
"Heterogeneous Computing: Challenges and Opportunities,"
with V. K. Prasanna, A. Khokhar, and C. Wang, IEEE Computer Magazine,
pages 18-27, June 1993.
-
"Image Computing using Optical Technology,"
with M. Eshaghian, and Sing Lee, appeared as a
chapter in the edited volume on "Parallel Architectures and
Algorithms for Image Understanding," by V. K. Prasanna-Kumar,
published by Academic Press, pages 29-58, 1991.
Journal Publications
-
"Cluster-M Programming Paradigm," with
M. Eshaghian, the International Journal of High Speed Computing,
Volume 6, Number 2, pages 287-309, 1994.
-
"Optical Techniques for Parallel Image Computing,"
with M. Eshaghian, Journal of Parallel and
Distributed Computing, Volume 21, pages 190-201, 1994.
-
"Heterogeneous Algorithms for Image Understanding Architecture,"
with M. Eshaghian, J. G. Nash, and D. B. Shu,
Journal of Parallel Algorithms and Applications,
Volume 1, pages 273-284, 1993.
Conference Publications
-
"Real-Time Implementation of JPEG Encoder/Decoder,"
with T. Czyszczon, R. Czernikowski, and K. Hsu,
Proceedings of the SPIE Conference on Input/Output and Imaging Technologies
, pages 281-292, July 1998.
-
"Partitioning of Image Processing Tasks on Heterogeneous Computer
Systems," with M. Iqbal and S. Iqbal,
Proceedings of the Third Workshop
on Heterogeneous Computing, pages 43-50, April 1994.
-
"Heterogeneous Partitioning of Chain Structured Image Processing
Tasks," with M. Iqbal,
proceedings of the Workshop on Computer
Architectures For Machine Perception, pages 302-311, December 1993.
-
"Heterogeneous Algorithms for Image Understanding Architecture,"
with M. Eshaghian, J. G. Nash, and D. B. Shu,
proceedings of the Workshop on Computer
Architectures For Machine Perception, pages 286-292, December 1993.
-
"A Selection Theory and Methodology for Heterogeneous
Supercomputing,"
with M. Eshaghian, S. Chen, and A. Khokhar,
Proceedings of the Heterogeneous Processing Workshop, pages 15-22,
April 1993.
-
"A Cluster-M based Mapping Methodology,"
with M. Eshaghian,
proceedings of the Seventh IEEE International Parallel Processing
Symposium, pages 213-221, April 1993.
-
"Heterogeneous Supercomputing: Problems and Issues,"
with A. Khokhar, V. K. Prasanna, and C. Wang,
Proceedings of the Heterogeneous Processing
Workshop, pages 3-12, March 1992.
References
Available upon request
|