Parallel and Distributed Computing
We have several projects involving parallel and distributed computing.
First, we are investigating the specification of properties
of parallel and distributed systems.
We are also investigating the use of formal specifications as the
basis of visualizations of parallel and distributed programs. The
visualizations can be used to facilitate the debugging and design
of programs.
Next, in the area of parallel compilers, we are extending the
set of unimodular transformations to loops to include families
of transformations, including loop blocking and coalescing, and loop
fission and fusion. The original properties of unimodular
transformations are preserved in the extended set of transformation
techniques. In addition, the extended set of transformations will
be applicable to imperfectly nested loops.
The allocation and scheduling of processors for distributed shared
memory systems is being investigated. We are examining different
methods by which system throughput can be improved by supporting
time shared execution within parallel jobs with clusters (of
workstations) without penalizing the job turn around time.
Finally, we are investigating different algorithms that
can be used for
multicast communication in massively parallel computers.
Recent Papers and Presentations
``A Top-down Approach to Visualization-based Debugging of
Parallel Programs''
(Joseph L. Sharnowski and Betty H.C. Cheng),
to appear in the poster session of
Supercomputing '94, November 1994.
``A Formal Approach to Modeling Expected Behavior in Parallel Program
Visualizations,''
(J.L. Sharnowski and B.H.C. Cheng),
Proc. of Parallel Architectures and Languages Europe (PARLE'94),
Lecture Notes in Computer Science, vol. 817,
Springer Verlag, pp. 202-213, July 1994.
``A Formally-based Expected Behavior Model for Parallel Program
Visualizations''
(Joseph L. Sharnowski and Betty H.C. Cheng),
submitted for publication,
April 1994.
``Data``A Temporal Model for Transparent Monitoring of Shared-Memory
Multiprocessors''
(David F. Robinson and Betty H.C. Cheng)
Proc. of the IEEE 17th Annual International
Computer Software and Applications Conference, pp. 388-394, November 1993.
``Data Parallel Program Visualizations from Formal Specifications''
(M.V. LaPolla, J.L. Sharnowski, B.H.C. Cheng, and K. Anderson),
Journal of Parallel and Distributed Computing, Vol. 18, No. 2, pp. 252-257, June, 1993.
``Using Visualizations to Guide Data Partitioning''
(J.L. Sharnowski and Betty H.C. Cheng)
Technical Report, Michigan State University, Department of Computer,
CPS-93-19, May 1993.
``A Transparent Monitoring Tool for Shared-Memory Multiprocessors,''
(David F. Robinson, Betty H.C. Cheng, and Richard J. Enbody),
Proc. of 16th Annual International Computer Software and
Applications Conference, pp. 227-232, September 1992.
``Mapping Formal Specifications to Parallel Program Visualizations''
(J.L. Sharnowski, M.V. LaPolla and Betty H.C. Cheng),
Proc. of Minnowbrook Workshop on Software Engineering for
Parallel Computing, pp. 29-34.
August 1992.
``Using Formal Specifications to Generate Visualizations of Data Parallelism''
(M.V. LaPolla, J.L. Sharnowski, Betty H.C. Cheng, and K. Anderson),
Michigan State University Technical Report, MSU-CPS-92-05, July 1992.
``Scalable Multicast Communication in
Massively Parallel Computers'', David F. Robinson, PhD thesis, August 1994.
``Temporal Specification Systems for
Distributed Systems'', William E. McUmber, MS Thesis, August 1993.
``Contention Free 2-D Mesh Cluster Allocation in Hypercubes''
(Stephen W. Turner, Lionel M. Ni, and Betty H.C. Cheng),
accepted to appear in IEEE Transactions on Computers.
``Time and/or Space Sharing in a Workstation Cluster Environment''
(S.W. Turner, L. Ni, and B.H.C. Cheng), to appear in the
Proc. of Supercomputing'94, November 1994.
``Contention Free 2-D Mesh Cluster Allocation in Hypercubes''
(Stephen W. Turner, Lionel M. Ni, and Betty H.C. Cheng),
Proc. of International Conference on Parallel
Processing, pp. 125-129, August 1993.
``Application of the Unimodular Approach to Loop
Fission and Loop Fusion''
(David R. Chesney and Betty H.C. Cheng)
presented
at the Scalable High Performance Computing Conference, Knoxville,
Tennessee, May 1994.
``Optimal Multicast Communication in Torus Networks''
(D.F. Robinson, P.K. McKinley, and B.H.C. Cheng),
in Proc. of International Conference on Parallel
Processing, Vol. I, pp. 131-141, August 1994.
``Efficient Collective Data Distribution in All-Port Wormhole-Routed
Hypercubes''
(D. F. Robinson, D. Judd, P. K. McKinley and B.H.C. Cheng), accepted to
appear in Journal of Parallel and Distributed Computing, 1994.
``Efficient Collective Data Distribution in All-Port Wormhole-Routed
Hypercubes''
(D. F. Robinson, D. Judd, P. K. McKinley, and B.H.C. Cheng),
Proc. of Supercomputing'93, pp. 792-803, November 1993.
``Extending the Unimodular Approach to Loop Blocking and Loop Coalescing''
(David R. Chesney and Betty H.C. Cheng),
Michigan State University Technical Report, MSU-CPS-94-25, April 1994.
``Extending the Unimodular Approach to Other Transformation Techniques''
(David R. Chesney and Betty H.C. Cheng),
Michigan State University Technical Report, MSU-CPS-93-24, September 1993.
``A Formal Approach to Automatic Source Code Translation for Parallel
Architectures''
(D.R. Chesney and Betty H.C. Cheng),
Proc. of Minnowbrook Workshop on Software Engineering for
Parallel Computing, pp. 16-27,
August 1992.