Samir R. Das
Abstract
It is widely believed that the Time Warp protocol for parallel discrete event simulation is prone to two potential problems: an excessive amount of wasted, rolled back computation resulting from “rollback thrashing” behaviors, and inefficient use of memory, leading to poor performance of virtual memory and/or multiprocessor cache systems. An adaptive mechanism is proposed based on the Cancelback memory management protocol for shared-memory multiprocessors that dynamically controls the amount of memory used in the simulation in order to maximize performance. The proposed mechanism is adaptive in the sense that it monitors the execution of the Time Warp program, and using simple models, automatically adjusts the amount of memory used to reduce Time Warp overheads (fossil collection, Cancelback, the amount of rolled back computation, etc.) to a manageable level. We describe an implementation of this mechanism on a shared memory, Kendall Square Research KSR-1, multiprocessor and demonstrate its effectiveness in automatically maximizing performance while minimizing memory utilzation, for several synthetic and benchmark discrete event simulation applications. We also demonstrate the adaptive ability of the mechanism by showing that it “tracks” the time-varying nature of a communication network simulation.
- AKYILDIZ, I. F., CHEN, L., DAS, S. R., FUJIMOTO, R. M., AND SERFOZO, R.F. 1993. The effect of memory capacity on Time Warp performance. J. Paral. Dist. Comput. 18, 4 (Aug.), 411-422.]] Google Scholar
- ARVIND, K. AND SMART, C. 1992. Hierarchical parallel discrete event simulation in composite ELSA. In Proceedings of 6th Workshop on Parallel and Distributed Simulation (Los Alamitos, CA, Jan.), IEEE Computer Science Press, Los Alamitos, CA, 147-158.]]Google Scholar
- BALL, D. AND HOYT, S. 1990. The adaptive Time-Warp concurrency control algorithm. Proceedings of the SCS Multiconference on Distributed Simulation 22, 1 (Jan.), SCS, San Diego, CA, 174-177.]]Google Scholar
- BRINER, J., JR. 1991. Fast parallel simulation of digital systems. Proceedings of the Multiconference on Advances in Parallel and Distributed Simulation 23, 1 (Jan.), SCS, San Diego, CA, 71-77.]]Google Scholar
- BROWN, R. 1988. Calendar queues: A fast O(1) priority queue implementation for the simulation event set problem. Commun. ACM 31, 10 (Oct.), 1220-1227.]] Google Scholar
- CAROTHERS, C. D., FUJIMOTO, R. M., LIN, Y.-B., AND ENGLAND, P. 1994. Distributed simulation of large scale PCS networks. In Proceedings of the 2nd International Conference on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems (MAS- COTS'94) (Jan.), IEEE Computer Society Press, Los Alamitos, CA, 2-11.]] Google Scholar
- CLEARY, J., GOMES, F., UNGER, B., ZHONGE, X., AND THUDT, R. 1994. Cost of state saving and rollback. In Proceedings of the 8th Workshop on Parallel and Distributed Simulation. IEEE Computer Society Press, Los Alamitos, CA, 94-101.]] Google Scholar
- DAS, S. R., FUJIMOTO, R., PANESAR, K., ALLISON, D., AND HYBINETTE, M. 1994. GTW: A Time Warp system for shared memory multiprocessors. In 1994 Winter Simulation Conference Proceedings (Dec.), SCS, San Diego, CA, 1332-1339.]] Google Scholar
- DAS, S. R. AND FUJIMOTO, R.M. 1993. A performance study of the cancelback protocol for Time Warp. In Proceedings of the 7th Workshop on Parallel and Distributed Simulation, 23 (May), IEEE Computer Society Press, Los Alamitos, CA, 135-142.]] Google Scholar
- DAS, S. R., AND FUJIMOTO, R. M. 1997. An empirical evaluation of performance-memory trade-offs in Time Warp. IEEE Trans. Parall. Dist. 8, 2 (Feb.), 210-224.]] Google Scholar
- DICKENS, P. M., AND REYNOLDS, P. F. JR. 1990. SRADS with local rollback. Proceedings of the SCS Multiconference on Distributed Simulation 22, 1 (Jan.), SCS San Diego, CA, 161-164.]]Google Scholar
- FERSCHA, A. 1995. Probabilistic adaptive direct optimism control in Time Warp. In Proceedings of the 9th Workshop on Parallel and Distributed Simulation, IEEE Computer Society Press, Los Alamitos, CA, 120-129.]] Google Scholar
- FLEISCHMANN, J. AND WILSEY, P. A. 1995. Comparative analysis of periodic state saving techniques in Time Warp simulators. In Proceedings of the 9th Workshop on Parallel and Distributed Simulation, IEEE Computer Society Press, Los Alamitos, CA, 50-58.]] Google Scholar
- FUJIMOTO, R. M. 1989. Time Warp on a shared memory multiprocessor. Trans. SCS 6, 3 (July), 211-239.]] Google Scholar
- FUJIMOTO, R. M. 1990a. Parallel discrete event simulation. Commun. ACM 33, 10 (Oct), 30 -53.]] Google Scholar
- FUJIMOTO, R.M. 1990b. Performance of Time Warp under synthetic workloads. In Proceedings of the SCS Multiconference on Distributed Simulation 22, 1 (Jan.), SCS, San Diego, CA, 23-28.]]Google Scholar
- HAMNES, D. O., AND TRIPATHI, A. 1994a. Evaluation of a local adaptive protocol for distributed discrete event simulation. In Proceedings of the 1994 International Conference on Parallel Processing (Aug.), CRC Press, Boca Raton, FL, 127-134.]] Google Scholar
- HAMNES, D. O., AND TRIPATHI, A. 1994b. Investigations in adaptive distributed simulation. In Proceedings of the 8th Workshop on Parallel and Distributed Simulation, IEEE Computer Society Press, Los Alamitos, CA, 20-23.]] Google Scholar
- JEFFERSON, D.R. 1985. Virtual time. ACM Trans. Program Lang. Syst. 7, 3 (July), 404-425.]] Google Scholar
- JEFFERSON, D.R. 1990. Virtual time II: Storage management in distributed simulation. In Proceedings 9th Annual ACM Symposium on Principles of Distributed Computing (Quebec City, Que., Canada, Aug. 22-24). ACM, New York, pp. 75-90.]] Google Scholar
- JEFFERSON, D. R., AND MOTRO, A. 1986. The Time Warp mechanism for database concurrency control. In Proceedings of the 2nd International Conference on Data Engineering. IEEE Computer Society Press, Los Alamitos, CA, 141-150.]] Google Scholar
- LIN, Y.-B., AND PREISS, B.R. 1991. Optimal memory management for Time Warp parallel simulation. ACM Trans. Modeling Comput. Simul. 1, 4 (Oct.), 283-307.]] Google Scholar
- LIN, Y.-B., PREISS, B. R., LOUCKS, W. M., AND LAZOWSKA, E.D. 1993. Selecting the checkpoint interval in Time Warp simulation. In Proceedings of the 7th Workshop on Parallel and Distributed Simulation (May), IEEE Computer Society Press, Los Alamitos, CA, 3-10.]] Google Scholar
- LUBACHEVSKY, B. D., SHWARTZ, A., AND WEISS, A. 1989. Rollback sometimes works ... if filtered. In 1989 Winter Simulation Conference Proceedings (Dec.), SCS, San Diego, CA, 630-639.]] Google Scholar
- LUBACHEVSKY, B. D., SHWARTZ, A., AND WEISS, A. 1991. An analysis of rollback-based simulation. ACM Trans. Model. Comput. Simul. 1, 2 (Apr.), 154-193.]] Google Scholar
- MADISETTI, V. K., HARDAKER, D. A., AND FUJIMOTO, R. M. 1993. The MIMDIX operating system for parallel simulation and supercomputing. J. Paral. Dist. Comput. 18, 4 (Aug.), 473-483.]] Google Scholar
- MEHL, H. 1991. Speedup of conservative distributed discrete-event simulation methods by speculative computing. Proceedings of the Multiconference on Advances in Parallel and Distributed Simulation 23, 1 (Jan.), SCS, San Diego, CA, 163-166.]]Google Scholar
- NICOL, D. M., AND FUJIMOTO, R.M. 1994. Parallel simulation today. Ann. Oper. Res. 53, 249-286.]]Google Scholar
- PALANISWAMY, A. C., AND WILSEY, P.A. 1993. An analytical comparison of periodic checkpointing and incremental state saving. In Proceedings of the 7th Workshop on Parallel and Distributed Simulation (May), IEEE Computer Society Press, Los Alamitos, CA, 127-134.]] Google Scholar
- PANESAR, Z., AND FUJIMOTO, R.M. 1995. Buffer management in shared memory time warp systems. In Proceedings of the 9th Workshop on Parallel and Distributed Simulation, IEEE Computer Society Press, Los Alamitos, CA, 149-156.]] Google Scholar
- PRESLEY, M., EBLING, M., WIELAND, F., AND JEFFERSON, D.R. 1989. Benchmarking the Time Warp Operating System with a computer network simulation. Proceedings of the SCS Multiconference on Distributed Simulation 21, 2 (Mar.), SCS, San Diego, CA, 8-13.]]Google Scholar
- RAJAEI, H., AYANI, R., AND THORELLI, L.-E. 1993. The local Time Warp approach to parallel simulation. In Proceedings of the 7th Workshop on Parallel and Distributed Simulation (May), IEEE Computer Society Press, Los Alamitos, CA, 119-126.]] Google Scholar
- REIHER, P. L., WIELAND, F., AND JEFFERSON, D.R. 1989. Limitation of optimism in the Time Warp Operating System. In Proceedings of the 1989 Winter Simulation Conference (Dec.), SCS, San Diego, CA, 765-770.]] Google Scholar
- RONNGREN, R., AYANI, R., FUJIMOTO, R. M., AND DAS, S.R. 1993. Efficient implementation of event sets in Time Warp. In Proceedings of the 7th Workshop on Parallel and Distributed Simulation (May), IEEE Computer Society Press, Los Alamitos, CA, 101-108.]] Google Scholar
- SOKOL, L. M., AND STUCKY, B. K. 1990. MTW: Experimental results for a constrained optimistic scheduling paradigm. In Proceedings of the SCS Multiconference on Distributed Simulation 22, 1 (Jan.), SCS, San Diego, CA, 169-173.]]Google Scholar
- SRINIVASAN, S., AND REYNOLDS, P.F. 1995. NPSI adaptive synchronization algorithms for PDES. In 1995 Winter Simulation Proceedings (Dec.), SCS, San Diego, CA, 658-665.]] Google Scholar
- STEINMAN, J. S. 1992. SPEEDES: A multiple-synchronization environment for parallel discrete event simulation. Int. J. Comput. Simul. 2, 3, 251-286.]]Google Scholar
- STEINMAN, J.S. 1993. Breathing time warp. In Proceedings of the 7th Workshop on Parallel and Distributed Simulation (May), IEEE Computer Society Press, Los Alamitos, CA, 109-118.]] Google Scholar
- TURNER, S. J., AND XU, M. Q. 1992. Performance evaluation of the bounded Time Warp algorithm. Proceedings of the SCS Multiconference on Parallel and Distributed Simulation 24, 3 (Jan.), SCS, San Diego, CA, 117-126.]]Google Scholar
- WIELAND, F., HAWLEY, L., FEINBERG, A., DILORENTO, M., BLUME, L., REIHER, P., BECKMAN, B., HONTALAS, P., BELLENOT, S., AND JEFFERSON, D.R. 1989. Distributed combat simulation and Time Warp: The model and its performance. Proceedings of the SCS Multiconference on Distributed Simulation 21, 2 (Mar.), SCS, San Diego, CA, 14-20.]]Google Scholar
Index Terms
- Adaptive memory management and optimism control in time warp
Recommendations
An adaptive memory management protocol for Time Warp parallel simulation
SIGMETRICS '94: Proceedings of the 1994 ACM SIGMETRICS conference on Measurement and modeling of computer systemsIt is widely believed that Time Warp is prone to two potential problems: an excessive amount of wasted, rolled back computation resulting from “rollback thrashing” behaviors, and inefficient use of memory, leading to poor performance of virtual memory ...
An adaptive memory management protocol for Time Warp parallel simulation
It is widely believed that Time Warp is prone to two potential problems: an excessive amount of wasted, rolled back computation resulting from “rollback thrashing” behaviors, and inefficient use of memory, leading to poor performance of virtual memory ...
Optimal memory management for time warp parallel simulation
Special issue on parallel and distributed systems performanceRecently there has been a great deal of interest in performance evalution of parallel simulation. Most work is devoted to the time complexity and assumes that the amount of memory available for parallel simulation is unlimited. This paper studies the ...
Reviews
Anthony Joseph Duben
Time Warp is an optimistic synchronization protocol in parallel simulation computations. At runtime, it detects out-of-sequence events and recovers by rolling back the calculation to properly account for the events. Time Warp has two major problems: excessive amounts of wasted, rolled back computation, and inefficient use of memory, leading to poor performance of virtual memory or cache systems. Das and Fujimoto present an adaptive memory management system that can cope with both of these problems. The solution monitors the execution of the Time Warp program and, based on the runtime data colle cted, automatically adjusts the amount of memory used to reduce the Time Warp overhead. It requires only a modest amount of memory beyond that required for sequential execution. The authors thoroughly review the Time Warp method, the nature of the problems discovered, and previous proposals to deal with them. They remark that the two problems mentioned above were considered independently in the past. Their contribution is a method by which both problems can be solved using a single approach based on the Cancelback protocol. The Cancelback protocol is presented in terms of quantitative measures of the usage of the memory buffers and estimated times of moving and processing data among them. These measures comprise the data gathered in the Time Warp program, and they enable the system to operate flexibly with widely varying workloads in a fully adaptive manner. Two test problems are used to demonstrate and illustrate the adaptive memory management method—a symmetric homogeneous workload (PHOLD) and an open asymmetric flow network based on electric power grids. A third application, a personal communication service network, is also discussed as a benchmark case. All of the computational experiments are presented in detail.
Access critical reviews of Computing literature here
Become a reviewer for Computing Reviews.
Comments