S.J. Turner and M.Q. Xu. Performance evaluation of the bounded time warp algorithm. In 6th Workshop on Parallel and Distributed Simulation (PADS'92), pages 117--126, 1992. 10  Home/Search   Document Not in Database   Summary   Related Articles   Check

....no longer be used by the rollback synchronization mechanism. The Time Warp protocol terminates the execution of all LPs once GVT has reached a predefined simulation endtime. 4. 2 Time Warp Performance The question on the Time Warp protocol performance has often been raised in the literature [12] [22], 1] 4] General performance rules, however, cannot be formulated [10] due to a very high degree of interweaving of influencing factors. The rollback mechanism is known to be prone to inefficient behavior in situations where event occurrences are highly dispersed in space and time. Such ....

Stephen Turner and M. Xu. Performance Evaluation of the Bounded Time Warp Algorithm. In Proceedings of the 6 Workshop on Parallel and Distributed Simulation, pages 117--128. SCS, January 1992.

....influence the behavior of Time Warp, e.g. induce rollback if the simulation of internal events has progressed further into the future than the time stamp carried by the message etc. The question on the Time Warp protocol performance has often been raised in the literature [Fuji 90b] Gupt 91] Turn 92] Fuji 93] Akyi 93] Das 93] Das 94b] General performance rules cannot be formulated [Fuji 90b] since performance due to a very high degree of interweaving of influencing factors cannot be sufficiently characterized by models. Even full implementations of the Time Warp protocol ....

S. Turner and M. Xu. "Performance Evaluation of the Bounded Time Warp Algorithm". In: Proceedings of the 6 Workshop on Parallel and Distributed Simulation, pp. 117--128, SCS, January 1992.

....it is also possible to limit the length of rollback chains in order to control the cascading of rollbacks. Several aggressive windowing algorithms have been proposed: Moving Time Window [24] and Window based throttling [25] Unified Distributed Simulation System [26] 18 Bounded Time Warp [27], Aggressive Global Windowing Algorithm [28] and Breathing Time Warp [29] These algorithms apply the same basic principle described above. The major di#erences lie in how the time window is determined. Notable among these are UDSS and Breathing Time Warp. The latter uses a fixed size window ....

S. Turner and M. Xu, "Performance evaluation of the bounded Time Warp algorithm, " in Proceedings of the 6th Workshop on Parallel and Distributed Simulation,

....on local state and adaptive based upon global state. In the non adaptive approach we rst nd the use of time windows, which constrain the events executed at an LP to lie within a time window having GVT as its lower edge. This limits both the frequency and length of rollbacks. Bounded Time Warp [2] and Moving Time Windows [3] are examples of this technique. A clear problem with this approach is determining the size of the time window. The breathing Time Bucket al..gorithm [33] makes use of local rollback only, i.e. it does not send anti messages. Breathing Time Warp combines the Breathing ....

Turner, S. and Xu, M.. Performance Evaluation of the Bounded Time Warp Algorithm, Proceedings of the SCS Multi-conference on Parallel and Distributed Simulation, volume 22, pages 117-126, 1992.

....of the time window (as observed by Reiher, Wieland, and Jefferson [74] A narrow time window will reduce rollbacks but at the cost of reduced parallelism. A large time window can potentially exploit more parallelism but at the risk of more rollbacks. A similar idea is explored by Turner and Xu [83] in the bounded Time Warp (BTW) protocol, where no events are processed beyond a bound in simulation time until all processes have reached that bound. Several methods for implementing the bounded time windows optimization in Time Warp simulators are available in the literature [6, 29, 30, 58, 74, ....

....by Turner and Xu [83] in the bounded Time Warp (BTW) protocol, where no events are processed beyond a bound in simulation time until all processes have reached that bound. Several methods for implementing the bounded time windows optimization in Time Warp simulators are available in the literature [6, 29, 30, 58, 74, 80, 83]. Limiting the optimism in Time Warp through the use of bounded time windows was originally envisaged to overcome rollback overheads. But it can be seen as a form of scheduling to optimize the execution behavior of Time Warp logical processes. In addition, limiting the number of rollbacks ....

Turner, S. J., and Xu, M. Q. Performance evaluation of the bounded Time Warp algorithm. Proceedings of the SCS Multiconference on Parallel and Distributed Simulation 24, 3 (1992), 117--126.

....that is blocked on such a channel on the runnable list, and continue. Hierarchical approaches to synchronization have been considered in other contexts. In [108] the authors introduced a scheme that combines a conservative time window synchronization with the Time Warp. This approach and others ([117, 122]) combine conservatism and optimism in an e ort to overcome the limitations of either one of them used alone. Time management for HLA [38] can also be viewed as a hierarchical 1 We do need to apply ow control here. 21 approach. A hierarchical scheduling algorithm for shared memory is proposed ....

S. J. Turner and M. Q. Xu. Performance evaluation of the Bounded Time Warp algorithm. Proceedings of the 6th Workshop on Parallel and Distributed Simulation (PADS'92), pages 117-26, 1992.

....over other synchronous algorithms, it causes a lot of rollback and memory space overhead. In order to obtain better performance, many techniques are proposed by researchers who believe that the TW algorithm with low overhead will be the best: adaptive process scheduling [79] bounding window [19, 91, 103], synchronization granularities [98] memory management [2, 24, 39, 44, 66, 67] and partitioning and load balancing [19, 57, 60, 69, 96] In the Time Warp algorithm, logical processes may advance too far ahead and produce a significant rollback cost. Limiting optimism may be a solution to reduce ....

....Limiting optimism may be a solution to reduce the rollback cost. The first method is to employ a time window to prevent incorrect 22 computation from propagating too far ahead. However, using a fixed window can either cause deterioration or provide improvement depending on the specific cases [103]. To reduce the rollback cost and the queue length in an event list, the Moving Time Window (MTW) scheme is applied to optimistic simulations in [91] It uses a global time window to limit the maximum difference between the local virtual times of communicating processes. In other words, the global ....

[Article contains additional citation context not shown here]

S. Turner and M. Xu, "Performance evaluation of the bounded Time Warp algorithm," Proc. of the 6th Workshop on Parallel and Distributed Simulation, pp. 117-128, 1992.

.... There are other throttling schemes that have been suggested in the literature with different controls, e.g. width of the time window in window based schemes [21] 28] 29] 30] resynchronization intervals with probabilistic synchronization [22] bound value in the bounded Time Warp algorithm [31], among others. They were developed primarily to limit the amount of rolled back computation in Time Warp. However, the controls used in the above mechanisms to limit optimism have only an indirect relationship to the amount of optimism provided by the system, and hence it may be difficult to ....

S.J. Turner and M.Q. Xu, "Performance Evaluation of the Bounded Time Warp Algorithm," Proc. SCS Multiconf. Parallel and Distributed Simulation, vol. 24, no. 3, pp. 117-126, Jan. 1992.

....but to create a separate external thread and to return a message which indicates that a deliberation process is under way. Thus, the transition function and the overall simulation can proceed. However, as in other synchronization protocols, e.g. Moving Time Windows [11] and Bounded Time Warp [13], barrier synchronizations are introduced to prevent the simulation from proceeding too far ahead compared to the external processes still running. To prevent cascading rollbacks over several simulation steps the simulation ensures at each step that it is safe to proceed. At each step the ....

S. Turner and M. Xu. Performance Evaluation of the Bounded Time Warp Algorithm. In Proc. of the 6th Workshop on Parallel and Distributed Simulation, pages 117--126, 1992.

....lengths of rollback chains and therefore control the cascading of rollbacks. Several aggressive windowing algorithms have been proposed: Moving Time Windows [SoBW88] Filtered Rollback [LuWS89] Window based throttling [ReJe89] Unified Distributed Simulation System [McAf90] Bounded Time Warp [TuXu92], Aggressive 18 Global Windowing Algorithm [Dick93] and Breathing Time Warp [Stei93] Most of these algorithms follow the same basic principle described above. The differences are in the methods used to determine the time window. Notable among these are UDS and Breathing Time Warp. The former ....

....and distributed systems. It is well known that good load balancing requires good estimation of tunable parameters 141 such as thresholds and maximum number of transfer hops. Accordingly, we observe that all existing adaptive protocols admit tunable parameters of some sort (the time window bound in [TuXu92], N 1 and N 2 in [Stei93] cluster size in [RaAT93] blocking window size in [BaHo90] the scaling factor c i in [HaTr94] and M f and M p in [DaFu94] Clearly, there is a need to invent a scheme by which adaptive protocols may tune the values of such parameters dynamically. The Adaptive Time ....

Turner, S.J. and Xu, M.Q., "Performance evaluation of the Bounded Time Warp algorithm", Proceedings of the 6th Workshop on Parallel and Distributed Simulation, January 1992, 117-126.

....simulators, coordinated at a high level through message passing. Hierarchical approaches to synchronization have been considered in other contexts. In [5] the authors introduced a scheme that combines a conservative time window synchronization with the Time Warp. This approach and others ( [7, 8]) combine conservatism and optimism in an effort to overcome the limitations of either one of them used alone. Time management for HLA [3] can also be viewed as a hierarchical approach. A hierarchical scheduling algorithm for shared memory is proposed in [9] Our work is distinguished by its focus ....

S. Turner and M. Xu. Performance evaluation of the bounded time warp algorithm. 6th Workshop on Parallel and Distributed Simulation (PADS92). Proceedings of the 1992 SCS Western Simulation MultiConference on Parallel and Distributed Simulation, pages 117--26, 1992.

....approaches include [19] where authors introduce a scheme that combines a conservative time window synchronization with Time Warp. Intra cluster activities are synchronized optimistically while inter cluster synchronization is handled conservatively. This approach and similar others [21, 22] combine conservatism and optimism in an effort to overcome the limitations of either one of them used alone. 8 Conclusions Approaches to conservative synchronous can generally be categorized as synchronous or asynchronous . Synchronous approaches are simple and scalable, but vulnerable to ....

S.J. Turner and M.Q. Xu. Performance evaluation of the bounded time warp algorithm. 6th Workshop on Parallel and Distributed Simulation (PADS92). Proceedings of the 1992 SCS Western Simulation MultiConference on Parallel and Distributed Simulation, pages 117--26, 1992.

.... systems [Lomo88] training [Stei91] military systems [Wiel89] Mors90] environment systems [Ebli89] Pres90] Bagr90] Nico90] flexible manufacturing systems [Nevi90] automobile traffic modelling [Merr90] aerodynamic simulation [Norr87] telecommunication networks [Muft90] Robe92] Eick91] [Turn92], queueing networks [Reed88a] Fuji89] Fuji89a] Chan89] Ayan89] Nico93] Nico93] and Petri nets [Kuma90] Thom91] Nico91] Bacc93] Fers93] The execution of these computationally intensive simulations requires either modelling at higher levels of abstraction in order to reduce the ....

Turner, S, Xu, M., "Performance Evaluation of the bounded Time Warp Algorithm", Proceedings of the 6th Workshop on Parallel and Distributed Simulation, SCS Simulation Series, January 1992, pp. 117128.

....be simulated by the LP (to be elaborated) C. PDES Implementation Platforms PDES systems have been implemented in different parallel architectures such as BBN Butterfly [12] 13] 14] Sequent [15] 16] 17] JPL Mark III [18] Simulated Stanford Dash Multiprocessor [19] Transputers [20] [21], CM LIN AND FISHWICK: ASYNCHRONOUS PARALLEL DISCRETE EVENT SIMULATION 3 1 CM 5 [22] KSR [23] and iPSC 860 [24] PDES has also been implemented in workstations connected by a local area network [4] which is widely available in both the industrial and the academic environments. III. Personal ....

Turner, S. and Xu, M., "Performance evaluation of the bounded Time Warp algorithm", The 6th Workshop on Parallel and Distributed Simulation, 1992.

....sending an anti message for each output message, and storing input messages which might later have to be canceled. Both of these problems have been the subject of much attention in the research literature. One approach to these problems has been trying to restrain the optimism of the simulation [18] via a global synchronization barrier. In this approach [16] a window of virtual time is defined within which it is permissible to simulate events at LPs provided that they do not violate causality constraints. In probabilistic algorithms each LP determines a blocking period. The amount of time ....

Turner, S. and Xu, M., "Performance Evaluation of the Bounded Time Warp Algorithm", In Proceedings of the SCS Multi-conference on PADS, volume 22, pages 117-126, 1992.

....generate events for each other so that the cluster would be simulated all the way until the end of the simulation time without running out of events. This is not desirable, since we don t want to create almost certain causality errors. Therefore we need to explore techniques for limiting optimism [6, 60, 65, 66, 67, 73]. These techniques involve setting 25 some limit on how far ahead an entity may simulate. The bound could be based on the amount of simulation time which elapses, the amount of memory required for state saving, or some information which the algorithm has about the simulation. We may be able to ....

S. J. Turner and M. Q. Xu. Performance evaluation of the bounded time warp algorithm. In Proceedings of the SCS Multiconference on Parallel and Distributed Simulation, pages 117--126, January 1992.

.... the future could affect the past (Fujimoto, 1990) a causality error. Correct process synchronisation is, therefore, of great importance in PDES. Hence, much research has been directed at this problem, with new protocols appearing from time to time, for examples, see Nicol et al. (1992) and Turner Xu (1992). Synchronisation protocols are usually divided into those that are conservative and those that are optimistic. In conservative protocols, a process is only permitted to process an event when it is known to be safe to do so. That is, a process containing an event with time stamp t, will only ....

TURNER, S. & XU, M. Performance Evaluation of the Bounded Time Warp Algorithm. In: Proceedings of the 6 th Workshop on Parallel and Distributed Simulation, Vol. 24 (SCS Simulation Series, January, 1992), pp. 117-128.

....during these supersteps the processors work with an estimation of the GVT advance, we use the factor 0.75 to prevent from over estimations of the actual GVT. We also use a small n v = 10 to avoid under estimations of this value. 5 Bounded Time Warp (BTW) The Bounded Time Warp (BTW) approach [10] defines a barrier in simulation time rather than a moving time window. Given a barrier at time T , an iteration of BTW consists of processing all the events e with time e:t T . The iteration finishes when there are no events below the barrier. Then the next iteration starts by increasing the ....

S.J. Turner and M.Q. Xu. Performance evaluation of the bounded time warp algorithm. In 6th Workshop on Parallel and Distributed Simulation (PADS'92), pages 117--126, 1992.

....distribution for the timestamp increments, negligible checkpointingrollback overhead and negligible communication delays. Modifications to the classical Time Warp synchronization have been proposed to embed barrier synchronizations in order to allow optimism of event execution at a limited extent [1, 4, 7, 25, 31, 32, 35]. Among the outcoming protocols, a complete performance model has been presented in [6] for the case of Bounded Time Warp [35] In this protocol a global window is iteratively defined and, in any iteration, LPs are allowed to process only the messages with timestamp in that window. The analysis in ....

.... Time Warp synchronization have been proposed to embed barrier synchronizations in order to allow optimism of event execution at a limited extent [1, 4, 7, 25, 31, 32, 35] Among the outcoming protocols, a complete performance model has been presented in [6] for the case of Bounded Time Warp [35]. In this protocol a global window is iteratively defined and, in any iteration, LPs are allowed to process only the messages with timestamp in that window. The analysis in [6] compares performance of Bounded Time Warp to that of a conservative protocol for the case of a queuing network under ....

S.Turner and M.Xu," Performance Evaluation of the Bounded Time Warp Algorithm," Proc. 6-th Workshop on Parallel and Distributed Simulation, pp.117-126, 1992.

....[17] A factor that must therefore be considered in any analysis of TW protocols is the use of an optimism control mechanism. Although a number of mechanisms have been published, they differ mainly in how the constraints on optimism are expressed. Examples include windowing mechanisms (e.g. [18]) and memory management techniques [19] A consideration of the important factors of an optimistic protocol thus leads to eight representative strategies. In terms of the factorial design we therefore identify three factors: i) cancellation policy LC (at the levels (aggressive ....

S. Turner, M. Xu, Performance evaluation of the bounded time warp algorithm, in: Proceedings of the Sixth Workshop on Parallel and Distributed Simulation, SCS, 1992, pp. 117--128.

....of these time stamps and broadcast the new GVT to all the processors. The process of discarding unneeded LP states and output messages is called fossil collection. A primary concern in Time Warp is stability, in terms of avoiding the proliferation of cascaded rollbacks or echoing [Lubachevsky 91, Turner 92] It is clearly important to minimise the number 18 of rollbacks, and it would be undesirable for an LP to advance its state into the far future only for it to be rolled back (and potentially roll back other LPs as well) The Time Warp local scheduling mechanism is of some help here, as a ....

S. J. Turner & M. Q. Xu, "Performance evaluation of the Bounded Time Warp algorithm", Proc. 1992 SCS Multiconf. on Parallel and Distributed Simulation, 20-22 January 1992, Newport Beach, California, pp. 117-126.

No context found.

S.J. Turner and M.Q. Xu. Performance evaluation of the bounded time warp algorithm. In 6th Workshop on Parallel and Distributed Simulation (PADS'92), pages 117--126, 1992. 10

No context found.

S.J. Turner and M.Q. Xu. Performance evaluation of the bounded time warp algorithm. In 6th Workshop on Parallel and Distributed Simulation (PADS'92), pages 117--126, 1992.

No context found.

S. J. Turner and M. Q. Xu. Performance evaluation of the bounded Time Warp algorithm. Proceedings of the SCS Multiconference on Parallel and Distributed Simulation, 24(3):117--126, January 1992.

No context found.

S. Turner and M. Xu, "Performance Evaluation of the BoundedTimeWarpAlgorithm", Proc. of 6th Workshop on Parallel and Distributed Simulation, Vol. 24, pp. 117-128, SCS Simulation Series, Jan 1992. Appendix A: Normalized Discrete Probability DensityFunction We construct a normalized discrete pdf fNw (see

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC