Abstract
The new class of queuing models, called Synchronized Queuing Networks, is proposed for evaluating the performance of multiprogrammed and multitasked multiprocessor systems, where workloads consists of parallel programs of similar structure and where the scheduling discipline is first-come-first-serve.
Pathwise evolution equations are established for these networks that capture the effects of competition for processors and the precedence constraints governing tasks executions.
A general expression is deduced for the stability condition of such queuing networks under general statistical assumptions (basically the stationarity and the ergodicity of input sequences), which yields the maximum program throughput of the multiprocessor system, or equivalently, the maximum rate at which programs can be executed or submitted. The proof is based on the ergodic theory of queues.
Basic integral equations are also derived for the stationary distribution of important performance criteria such as the workload of the queues and program response times. An iterative numerical schema that converges to this solution is proposed and various upper and lower bounds on moments are derived using stochastic ordering techniques.
- 1 ANDREWS, G. R., AND SCHNEIDER, F.B. Concepts and notations for concurrent programming. ACM Comput. Surv. 15, 1 (Mar. 1983), 3-43. Google Scholar
- 2 BACCELLI, F. Two parallel queues created by arrivals with two demands: The M/G/2 symmetrical case. INRIA Rapport de Recherche, No. 426. INRIA, Valbonne, France, July 1985.Google Scholar
- 3 BACCELLI, F., AND BRI~MAUD, P. Palm probabilities and stationary queues. In Lecture Notes in Statistics, No. 41. Springer-Verlag, New York, 1987.Google Scholar
- 4 BACCELLI, F., AND LIU, Z. On the stability condition of a precedence-based queueing discipline. Adv. Appl. Prob. 21 (1989), 883-887.Google Scholar
- 5 BACCELLI, F., AND LIU, Z. On a class of stochastic evolution equations. Ann. Prob., to appear.Google Scholar
- 6 BACCELLI, F., AND MAKOWSKI, A. Simple computable bounds for the fork-join queue. In Proceedings of the Conference for Information Science Systems. Johns Hopkins Univ., Baltimore, Md., March 1985, pp. 436-441.Google Scholar
- 7 BACCELLI, F., MAKOWSKI, A., AND SHWARTZ, A. Fork-join queue and related systems with synchronization constraints: Stochastic ordering, approximations and computable bounds. J. Adv. Prob. 21 (1989), 629-660.Google Scholar
- 8 BACCELLI, F., MASSEY, W. A., AND TOWSLEY, D. Acyclic fork-join queuing networks. 3. ACM 36, 3 (July 1989), 615-642. Google Scholar
- 9 BATTAREL, G. J., AND SAVARY, H.F. Interprocess communication system of the MT35 digital exchange. Comput. Commun. Rev. 13, 2 (1983), 197-203. Google Scholar
- 10 BARLOW, R., ANO PROSCHAN, F. Statistical Theory of Reliability and Life Testing. Holt, Rinehart and Winston, New York, 1975.Google Scholar
- 11 CHEN, W. K. Applied graph theory. In Applied Mathematics and Mechanics, H. A. Lauwerier and W. T. Koiter, eds. North-Holland, New York, I971.Google Scholar
- 12 CHU, W. W., AND LEUNO, K.K. Module replication and assignment for real-time distributed processing systems. Proc. IEEE 75, 5 (May 1987), 547-562.Google Scholar
- 13 COFFMAN, E. G., AND DENNINO, P.J. Operating Systems Theory. Prentice-Hall, New York, 1972. Google Scholar
- 14 COSNARD, M., AND ROBERT, Y. Algorithmique Parallble: Une E, tude de ComplexitY. Techniques et Sciences Informatiques, Paris, France, 1987.Google Scholar
- 15 DEVROVE, L.P. Inequalities for the completion times of stochastic PERT networks. Math. Oper. Res. 4 (1980), 441-447.Google Scholar
- 16 DODIN, B. Bounding the project completion time distribution in PERT networks. Oper. Res. 33, 4 (July-Aug. 1985), 862-881.Google Scholar
- 17 DUNCAN, T., AND HUEN, W.H. Software structure of no. 5 ESS-A distributed telephone switching system. IEEE Trans. Commun. COM-30, 6 (1982), 1379-1385.Google Scholar
- 18 ELMAGHRABY, S.E. On the expected duration of PERT type networks. Manage. Sci. 13 (1967), 469-481.Google Scholar
- 19 FLATTO, L., AND HAHN, S. Two parallel queues created by arrivals with two demands. SIAM J. AppL Math 44 (1984), 1041-1053.Google Scholar
- 20 GELENBE, E., AND LIU, Z. Performance analysis approximations for parallel processing on multiprocessor systems. In IFIP Working Conference on Parallel Processing (Pisa, Italy, Apr.) North-Holland, New York, 1988, pp. 363-375.Google Scholar
- 21 GELENBE, E., MONTAGNE, E., SUROS, R., AND WOODSIDE, C.M. A performance model of blockstructured parallel programs. In Proceedings of the International Workshop on Parallel Algorithms and Architectures, M. Cosnard, et al., eds. North-Holland, New York, 1986, pp. 127-138. Google Scholar
- 22 GELENBE, E., NELSON, P., PHILIPS, T., AND TANTAWI, A. The asymptotic processing time for a model of parallel processing. In Proceedings of the National Computer Conference (Las Vegas). 1986.Google Scholar
- 23 HOARE, C. A. R. Communicating sequential processes. Commun. ACM 21, 8 (Aug. 1978), 666-677. Google Scholar
- 24 HOLLIDAY, M. A., AND VERNON, M.K. Exact performance estimates for multiprocessor memory and bus interference. IEEE Trans. Comput. C-36, I (Jan. 1987), 76-85. Google Scholar
- 25 HUANG, J.P. Modeling of software partition for distributed real-time applications. IEEE Trans. Sofiw. Eng. SE-11, l0 (1985), lll3-1 I26. Google Scholar
- 26 JONES, A. K., AND SCHWARZ, P. Experience using multiprocessor systems~A status report. ACM Comput. Surv. 12, 2 ( June 1980), 121-165. Google Scholar
- 27 KINGMAN, J. F.C. The ergodic theory of subadditive stochastic processes. J. Roy. Statist. Soc. Set. B 30 (1968), 499-510.Google Scholar
- 28 KINGMAN, j. F.C. Subadditive ergodic theory. Ann. Prob 1 (1973), 883-909.Google Scholar
- 29 KINGMAN, J. F. C. Subadditive processes. In I~cole d'Et6 de Probabilit6 de Saint-Hour, P.-L. Hennequin, ed. Lecture Notes in Mathematics, vol. 539. Springer-Verlag, New York, 1976, 165-223.Google Scholar
- 30 LAMPORT, L. Time, clocks and the ordering of events in a distributed system. Commun. ACM 21, 7 (July 1978), 558-565. Google Scholar
- 31 LIU, Z. Evaluation des performances d~applications parallrles sur des systrmes multiprocesseurs. Rapport de DEA, Univ. of Paris-Sud, Paris, France, Sept. 1986.Google Scholar
- 32 LOYNES, R.M. The stability of queues with non independent inter-arrival and service times. Proc. Cambridge Ph. Soc. 58 (1962), 497-520.Google Scholar
- 33 MARSAN, M. A., BALBO, G., CONTE, G., AND GREGORETTI, F. Modeling bus contention and memory interference in a multiprocessor system. IEEE Trans. Comput. C-32, 1 (Jan. 1983), 60-71.Google Scholar
- 34 MARSAN, M. A., AND GREGORETTI, F. Memory interference models for a multiprocessor system with a shared bus and a single external common memory. Microproc. Microprog. 7 (1981), 124-133.Google Scholar
- 35 MILUTINOVI(;, V., FORTES, J. A. B., AND JAMIESON, L.H. A multimicroprocessor architecture for real-time computation of a class of DFT algorithms. IEEE Trans. ASSP ASSP-34, 5 (1986), 1301-1309.Google Scholar
- 36 NELSON, R., AND TANTAWI, A.N. Approximate analysis of fork/join synchronization in parallel queues. IEEE Trans. Comput. C-37, 6 (1988), 739-743. Google Scholar
- 37 PYLE, I.C. The Ada Programming Language. Prentice-Hall International, London, 1981. Google Scholar
- 38 SIGMAN, K. Regeneration in queues with regenerative input. Queue. Syst., to appear.Google Scholar
- 39 STAFYLOPATIS, A., AND GELENBE, E. Delay analysis of resequencing systems with partial ordering. In Proceedings of PERFORMANCE 87, P.-J. Courtois and G. Latouche, eds. North Holland, New York, Dec. 1987, pp. 433-446. Google Scholar
- 40 STOYAN, D. Comparison Methods for Queues and Other Stochastic Models (English translation), D. J. Daley, ed. Wiley, New York, 1984.Google Scholar
- 41 TALBOT, F. B., PATTERSON, J. H., AND GEHRLEIN, W.V. A comparative evaluation of heuristic line balancing techniques. Manage. Sci. 32, 4 (Apr. 1986), 430-454. Google Scholar
- 42 TOWSLEY, D., AND YU, S.P. Bounds for two server fork-join queueing systems. Tech. Rep. TR87- 123. Dept. of Comput. Inf. Sci., Univ. Mass., Amherst, Mass., Nov. 1987. Google Scholar
- 43 TSITSIKLIS, J. N., PAPADIMITRIOU, C. H., AND HUMBLET, P. The performance of a precedencebased queuing discipline. J. ACM 33, 3 (July 1986), 593-602. Google Scholar
- 44 YALAMANCHILI, S., AND AGGARWAL, J.K. A characterization and analysis of parallel processor interconnection networks. IEEE Trans. Comput. C-36, 6 (June 1987), 680-691. Google Scholar
Index Terms
On the execution of parallel programs on multiprocessor systems—a queuing theory approach
Recommendations
A multiserver retrial queue: regenerative stability analysis
We consider a multiserver retrial GI / G / m queue with renewal input of primary customers, interarrival time with rate $\lambda=1/\mathsf{E}\tau$ , service time S , and exponential retrial times of customers blocked in the orbit. In the model, an arriving primary ...
The cyclic queue and the tandem queue
We consider a closed queueing network, consisting of two FCFS single server queues in series: a queue with general service times and a queue with exponential service times. A fixed number $$N$$ N of customers cycle through this network. We determine the joint ...
On the stationary distribution of queue lengths in a multi-class priority queueing system with customer transfers
This paper deals with a multi-class priority queueing system with customer transfers that occur only from lower priority queues to higher priority queues. Conditions for the queueing system to be stable/unstable are obtained. An auxiliary queueing ...






Comments