This Article 
   
 Share 
   
 Bibliographic References 
   
 Add to: 
 
Digg
Furl
Spurl
Blink
Simpy
Google
Del.icio.us
Y!MyWeb
 
 Search 
   
Optimal Dynamic Remapping of Data Parallel Computations
February 1990 (vol. 39 no. 2)
pp. 206-219

A large class of data parallel computations is characterized by a sequence of phases, with phase changes occurring unpredictably. Dynamic remapping of the workload to processors may be required to maintain good performance. The problem considered, for which the utility of remapping and the future behavior of the workload are uncertain, arises when phases exhibit stable execution requirements during a given phase, but requirements change radically between phases. For these situations, a workload assignment generated for one phase may hinder performance during the next phase. This problem is treated formally for a probabilistic model of computation with at most two phases. The authors address the fundamental problem of balancing the expected remapping performance gain against the delay cost, and they derive the optimal remapping decision policy. The promise of the approach is shown by application to multiprocessor implementations of an adaptive gridding fluid dynamics program and to a battlefield simulation program.

[1] I. Babuska, Ed.,Adaptive Computational Methods for Partial Differential Equations. Philadelphia, PA: SIAM, 1983.
[2] M. Berger and S. H. Bokhari, "A partitioning strategy for nonuniform problems on multiprocessors,"IEEE Trans. Comput., vol. C-36, pp. 570-580, May 1987.
[3] M. J. Berger and J. Oliger, "Adaptive mesh refinement for hyperbolic partial differential equations,"J. Computat. Phys., vol. 53, pp. 484-512, 1984.
[4] S. Bokhari, "Partitioning problems in parallel, pipelined, and distributed computing,"IEEE Trans. Comput., vol. 37, no. 1, pp. 48-57, Jan. 1988.
[5] D. L. Book, Ed.,Finite-Difference Techniques for Vectorized Fluid Dynamics Calculations. New York: Springer-Verlag, 1981.
[6] W. W. Chu, L. J. Holloway, M. Lan, and K. Efe, "Task allocation in distributed data processing,"IEEE Comput. Mag., vol. 13, no. 11, pp. 57-69, Nov. 1980.
[7] W. W. Chu and K. K. Leung, "Module replication and assignment for real-time distributed processing systems,"Proc. IEEE, May 1987.
[8] Y. Chow and W. Kowhler, "Models for dynamic load balancing in a heterogeneous multiple processor system,"IEEE Trans. Comput., vol. C-28, no. 5, pp. 354-361, May 1979.
[9] T. C. Chou and J. A. Abraham, "Load balancing in distributed systems,"IEEE Trans. Software Eng., vol. SE-8, no. 4, pp. 401-412, July 1982.
[10] D. Eager, E. Lazowska, and J. Zahorjan, "Adaptive load sharing in homogeneous distributed systems,"IEEE Trans. Software Eng., vol. SE-12, no. 5, pp. 662-675, May 1986.
[11] G. J. Foschini, "On heavy traffic diffusion analysis and dynamic routing in packet switched networks," inComputer Performance, K. M. Chandy and M. Reiser Eds. New York: North-Holland, 1977.
[12] G. Fox et al.,Solving Problems on Concurrent Processors, Vol. 1, Prentice Hall, Englewood Cliffs, N.J., 1988.
[13] J. B. Gilmer, "Documentation, state-space reconciliation version of the zipscreen prototype simulation," Tech. Rep., BDM Corp., 1986.
[14] D. Gusfield, "Parametric combinatorial computing and a problem of module distribution,"J. ACM, no. 3, pp. 551-563, 1983.
[15] P. R. Ma, E. Y. S. Lee, and M. Tsuchiya, "A task allocation model for distributed computing systems,"IEEE Trans. Comput., vol. C-31, no. 1, pp. 41-47, Jan. 1982.
[16] N. Matelan, "The Flex/32 Multicomputer," inProc. 12th Int. Symp. Comput. Architecture. Los Alamitos, CA: Computer Society Press, June 1985, pp. 209-213.
[17] D. I. Moldovan and J. A. B. Fortes, "Partitioning and mapping algorithms into fixed size systolic arrays,"IEEE Trans. Comput., vol. C-35, pp. 1-12, Jan. 1986.
[18] L. M. Ni, C. Xu, and T. B. Gendreau, "A distributed drafting algorithm for load balancing,"IEEE Trans. Software Eng., vol. SE-11, no. 10, pp. 1153-1161, Oct. 1985.
[19] D. M. Nicol, "Mapping a battlefield simulation onto parallel messagepassing architectures," inProc. 1988 SCS Conf. Distrib. Simulation, San Diego, CA, 1988, pp. 141-146.
[20] D. M. Nicol and P. F. Reynolds, Jr., "Dynamic remapping decisions in multi-phase parallel computations," ICASE Rep. 86-48, Sept. 1986.
[21] D. M. Nicol and J. H. Saltz, "Dynamic remapping of parallel computations with varying resource demands,"IEEE Trans. Comput., vol. 37, no. 9, pp. 1073-1087, Sept. 1988.
[22] D. M. Nicol and J. Townsend, "Accurate modeling of parallel scientific computations," inProc. 1989 SIGMETRICS Conf., May 1989, Berkeley, CA, pp. 165-170.
[23] D. M. Nicol and F. H. Willard, "Problem size, parallel architecture, and optimal speedup,"J. Parallel Distributed Comput., vol. 5, 1988.
[24] C. C. Price and U. W. Pooch, "Search techniques for a nonlinear multiprocessor scheduling problem,"Naval Res. Logistics Quarterly, vol. 29, no. 2, pp. 213-233, June 1982.
[25] A. Rapoport, W. E. Stein, and G. J. Burkheimer,Response Models for Detection of Change. Boston, MA: Reidel, 1979.
[26] S. Ross,Applied Probability Models with Optimization Applications. San Francisco, CA: Holden-Day, 1970.
[27] S. A. Schmitt,An Elementary Introduction to Bayesian Statistics. Reading, MA: Addison-Wesley, 1969.
[28] J. A. Stankovic, "An application of Bayesian decision theory to decentralized control of job scheduling,"IEEE Trans. Comput., vol. C-34, no. 2, pp. 117-130, Feb. 1985.
[29] J. A. Stankovic, K. Ramamritham, and S. Cheng, "Evaluation of a flexible task scheduling algorithm for distributed hard real-time systems,"IEEE Trans. Comput., vol. C-34, no. 12, pp. 1130-1143, Dec. 1985.
[30] D. Towsley, "Queueing network models with state-dependent routing,"J. ACM, vol. 27, pp. 323-337, Apr. 1980.
[31] A. N. Tantawi and D. Towsley, "Optimal static load balancing in distributed computer systems,"J. ACM, pp. 445-465, Apr. 1985.

Index Terms:
optimal dynamic remapping; data parallel computations; performance; workload assignment; probabilistic model; multiprocessor implementations; adaptive gridding fluid dynamics program; battlefield simulation program; military computing; parallel processing; performance evaluation; physics computing.
Citation:
D.M. Nicol, P.F. Reynolds, Jr., "Optimal Dynamic Remapping of Data Parallel Computations," IEEE Transactions on Computers, vol. 39, no. 2, pp. 206-219, Feb. 1990, doi:10.1109/12.45206
Usage of this product signifies your acceptance of the Terms of Use.