Abstract
The behavior of n interacting processors synchronized by the "Time Warp" protocol is analyzed using a discrete state continuous time Markov chain model. The performance and dynamics of the processes are analyzed under the following assumptions: exponential task times and times-tamp increments on messages, each event message generates one new message that is sent to a randomly selected process, negligible rollback, state saving, and communication delay, unbounded message buffers, and homogeneous processors that are never idle. We determine the fraction of processed events that commit, speedup, rollback probability, expected length of rollback, the probability mass function for the number of uncommitted processed events, and the probability distribution function for the virtual time of a process. The analysis is approximate, so the results have been validated through performance measurements of a Time Warp testbed (PHOLD workload model) executing on a shared memory multiprocessor.
- 1 K.M. Chandy. and J. Misra. Asynchronous Distributed Simulation wa a Sequence of Parallel Computations. Communications of the ACM, Vol. 24(11), pp. 198-206, November 1981. Google Scholar
Digital Library
- 2 R. E. Felderman and L. Kleinrock. An Upper Bound on the improvement of Asynchronous versus Synchronous Distributed Processing. In Proceedings of the SCS Multiconference on Distributed Simulation, Vol. 22(1), pp. 131-136, January 1990.Google Scholar
- 3 R.E. Felderman. Speedup for Large Number of Processors. Personal Communication, December 1990.Google Scholar
- 4 R.E. Felderman and L. Kleinrock. Two Processor Time Warp Analysis: Some Results on a Unifying Approach. To appear in 1991 SCS Parallel and Distributed Simulation Workshop, January 1991.Google Scholar
- 5 R.M. Fujimoto. Time Warp on a Shared Memory Multiprocessor. Trans. Soc. for Comput. Simul., Vol. 6(3), pp. 211-239, July 1989. Google Scholar
Digital Library
- 6 R. M. Fujimoto. Performance of Time Warp under Synthetic Workloads. In Proceedings of the SCS Multicon- }erence on Distributed Simulation, Vol. 22(1), pp. 23- 28, January 1990.Google Scholar
- 7 R. M. Fujimoto. Parallel Discrete Event Simulation. In Communications of the ACM, Vol. 33(10), pp. 30-53, October 1990. Google Scholar
Digital Library
- 8 E. Gelenbe. Product Form Networks with Negative and Positive Customers. To appear in Journal of Applied ProbabilityGoogle Scholar
- 9 A. G. Greenberg, B. D. Lubachevsky, and I. Mitrani. Unb6undedly Parallel Simulation~ Via Recurrence Relations. In Proc. of the 1990 A CM SIGMETRICS Conf. on Measurement and Modeling of Computer Systems, Vol. 18(1), pp. 11-12, May 1990. Google Scholar
Digital Library
- 10 A. Gupta, I.F. Akyildiz, and R.M. Fujimoto. Performance Analysis of Time Warp with Multiple Homogeneous Processors. Submitted for publication.Google Scholar
- 11 D. R. Jefferson. Virtual Time. In A CM Transactions on Programming Languages and Systems, Vol. 7, pp. 404-425, 1985. Google Scholar
Digital Library
- 12 D. Jefferson and A. Witkowski. An Approach to Performance Analysis of Timestamp-driven Synchronization Mechanisms. In Proceedings of the 3~ A CM Annual Symposium on Principles of Distributed Computing, 1984. Google Scholar
Digital Library
- 13 L. Kleinrock. On Distributed Systems Performance. In Computer Networks and 1SDN Journal, Vol. 20(1-5), pp. 209-216, December 1990.Google Scholar
- 14 L. Lamport. Timez, Clocks, and the Ordering of Events in a Distributed System. In Communications of the A CM, Vol. 21, No. 7, 1978. Google Scholar
Digital Library
- 15 S. Lavenberg, R. Muntz, and B. Samadi. Performance Analysis of a Rollback Method for Distributed Simulation. In Performance '83, North-Holland Publishing Company, pp. 117-132, 1983. Google Scholar
Digital Library
- 16 Y. B. Lin, and E. D. Lazowska. OptimMity Considerations for "Time Warp" Parallel Simulation. In Proc. 1990 SCS Multiconference on Distributed Simulation, pp. 29-34, January 1990.Google Scholar
- 17 R. J. Lipton and D. W. Mizell. Time Warp vs. Chandy- Misra : A Worst-Case Comparison. In Proceedings of the SCS Multiconference on Distributed Simulation, Vol. 22(1), pp. 137-143, January 1990.Google Scholar
- 18 B. Lubachevsky, A. Shwartz, and A. Weiss. Rollback Sometimes Works .. if Filtered. In Proc. of the 1989 Winter Simulation Conf., pp. 630-639, 1989. Google Scholar
Digital Library
- 19 V. Madisetti, J. Walrand, and D. Messerschmitt. Synchronization in Message Passing Computers: Models, Algorithms, and Analysis. In Distributed Simulation 1990, pp. 35-48, 1990.Google Scholar
- 20 J. Misra. Distributed Discrete-Event Simulation. In Computing Surveys, 18(1), pp. 39-65, 1986. Google Scholar
Digital Library
- 21 D. Mitra and I. Mitrani. Analysis and Optimum Performance of Two Message-Passing Parallel Processors Synchronized by Rollback. In Performance Evaluation 7, Elsevier Science Publishers B.V. (North-Holland), pp. 111-124, 1987. Google Scholar
- 22 D. M. Nicol. Performance Bounds on Parallel Self- Initiating Discrete-Event Simulations. Technical report ICASE 90-21, March 1990.Google Scholar
- 23 B. D. Plateau and S. K. Trip~thi. Performance Analysis of Synchronization for Two Communicating Processes. In Performance Evaluation 8, Elsevier Science Publishers B.V. (North-Holland), pp. 305-320, 1988. Google Scholar
Index Terms
Performance analysis of Time Warp with homogeneous processors and exponential task times
Recommendations
Performance analysis of Time Warp with homogeneous processors and exponential task times
SIGMETRICS '91: Proceedings of the 1991 ACM SIGMETRICS conference on Measurement and modeling of computer systemsThe behavior of n interacting processors synchronized by the "Time Warp" protocol is analyzed using a discrete state continuous time Markov chain model. The performance and dynamics of the processes are analyzed under the following assumptions: ...
Performance Analysis of Time Warp with Multiple Homogeneous Processors
The behavior of n interacting processors synchronized by the Time Warp protocol is analyzed using a discrete-state, continuous-time Markov chain model. The performance and dynamics of the processes (or processors) are analyzed under the following ...
Time-warp: lightweight abort minimization in transactional memory
PPoPP '14The notion of permissiveness in Transactional Memory (TM) translates to only aborting a transaction when it cannot be accepted in any history that guarantees correctness criterion. This property is neglected by most TMs, which, in order to maximize ...






Comments