This Article 
 Bibliographic References 
 Add to: 
The Processor Working Set and its Use in Scheduling Multiprocessor Systems
May 1991 (vol. 17 no. 5)
pp. 443-453

The concept of a processor working set (PWS) as a single value parameter for characterizing the parallel program behavior is introduced. Through detailed experimental studies of different algorithms on a transputer-based multiprocessor machine, it is shown that the PWS is a robust measure for characterizing the workload of a multiprocessor system. It is shown that processor allocation strategies based on the PWS provide significantly better throughput-delay characteristics. The robustness of PWS is further demonstrated by showing that allocation policies that allocate processors more than the PWS are inferior in performance to those that never allocate more than the PWS-even at a moderately low load. Based on the results, a simple static allocation policy that allocates the PWS at low load and adaptively fragments at high load to one processor per job is proposed.

[1] S. G. Akl,Parallel Sorting Algorithms. Orlando, FL: Academic, 1985.
[2] D.L. Eager, J. Zahorjan, and E.D. Lazowska, "Speedup versus efficiency in parallel systems,"IEEE Trans. Comput., vol. 38, pp. 408-423, Mar. 1989.
[3] L. W. Dowdy, "On the partitioning of multiprocessor systems," Dept. Comput. Sci., Vanderbilt Univ., Nashville, TN, Tech. Rep., Mar. 1988.
[4] M. Foxet al., Solving Problems on Concurrent Processors, vol. 1. Englewood Cliffs, NJ: Prentice-Hall, 1988.
[5] E. Gelenbe,Multiprocessor Performance. New York: Wiley, 1989.
[6] D. Ghosal, G. Serazzi, and S.K. Tripathi, "The concept of processor working set for allocation strategies in multiprogrammed parallel systems," UMIACS, Univ. Maryland, College Park, Tech. Rep. TR-89-95, CS-TR-2323, 1989.
[7] W. D. Hillis,The Connection Machine. Cambridge, MA: MIT Press, 1985.
[8] The Transputer Family: Product Information, INMOS Group of Companies, June 1986.
[9] L. Kleinrock, "Power and deterministic rules of thumb for probabilistic problems in computer communications," inProc. Int. Conf. Communications, June 1979, pp. 43.1.1-43.1.10.
[10] M. Kumar, "Measuring parallelism in computation-intensive scientific/engineering applications,"IEEE Trans. Comput., vol. 37, pp. 1088-1098, Sept. 1988.
[11] R. B.-L. Lee, "Empirical results on the speedup, efficiency, redundancy, and quality of parallel computations," inProc. Int. Conf. Parallel Processing, Aug. 1980, pp. 91-96.
[12] M. R. Leuze, L. W. Dowdy, and K. H. Park, "Multiprogramming a distributed-memory multiprocessor,"J. Concurrency Practice, 1989.
[13] S. Majumdar, D.L. Eager, and R. Bunt, "Scheduling in multiprogrammed parallel systems,"ACM SIGMETRICS, pp. 104-113, 1988.
[14] Multimax Technical Summary, Encore Computer Corp., Marlboro, MA, May 1985.
[15] K. H. Park and L. W. Dowdy, "Dynamic partitioning of multiprocessor systems,"Int. J. Parallel Program., 1989.
[16] K. C. Sevcik, "Characterization of parallelism in applications and their use in scheduling,"ACM SIGMETRICS, pp. 171-180, 1989.
[17] The QNAP2 Reference Manual, Simulog Corp., France, 1989.

Index Terms:
scheduling; processor working set; PWS; parallel program behavior; transputer-based multiprocessor machine; processor allocation strategies; static allocation policy; multiprocessing systems; scheduling; transputers
D. Ghosal, G. Serazzi, S.K. Tripathi, "The Processor Working Set and its Use in Scheduling Multiprocessor Systems," IEEE Transactions on Software Engineering, vol. 17, no. 5, pp. 443-453, May 1991, doi:10.1109/32.90447
Usage of this product signifies your acceptance of the Terms of Use.