Time-Event based processing, a Survey Bruce Jones Vanessa Wallace

Size: px
Start display at page:

Download "Time-Event based processing, a Survey Bruce Jones Vanessa Wallace"

Transcription

1 Time-Event based processing, a Survey Bruce Jones Vanessa Wallace 1. Introduction Event based processing is used in many computational processes, whether by intent or as a result of the task requirement. Event processing can be subdivided into two methodologies; conservative a methodology that requires all inputs to be known before further processing, and optimistic a methodology that allows for speculative processing. This paper will primarily cover the optimistic model, though it will also cover some of the advances in conservative processing Relevance When optimistic event processing is mentioned, it is generally mentioned in the context of simulation. This is not the only context for optimistic or conservative event processing. Event processing is involved in most of the programs written. When most programs are invoked, procedures called, assembly language instructions executed; this is a form of event processing. The event being the invocation. Under optimistic processing, the form it might take is; invoking a program without all of the files or data present, calling a procedure without all of the arguments present or executing an assembly language instruction without needed intermediate calculations complete. The latter is most evident in the form of pipelined microprocessors that use branch prediction, and retirement of incorrect results. 2. Synchronous conservative event processing 2.1. Overview Conservative event processing ends up being the most intuitive to design, build and run. It is based upon the concept of a logical process taking the most recent input event and applying it to the process to which it belongs. Processes are not allowed to create events that exist in an earlier virtual time than the current virtual time of the executing process (a future event can not affect a past event). Because it is the earliest event and all processing is complete at that time. Though it is the most intuitive to design, build and run with single execution threads, implementation on a multithreaded environment becomes problematic. The problem is derived from guaranteeing causality. All of the events for a given process have to be available before the process is to be allowed to run. This requires either a segment that coordinates all tasks, or a methodology that allows tasks to know when causality is guaranteed, without requiring a synchronizing task. Using a single coordinating point for events creates a bottleneck. At its worse case, this design will degrade to single threaded operation regardless of the number of processors. The conservative protocols are intuitive, except when trying to spread across multiple processors. They are easy to implement on single processor design oriented towards single processor architectures (serial processing). The processing includes queue of events, where the most recent event is processed and output events go back onto queue(s) The Chandy-Misra-Bryant algorithm The conservative algorithm by Chandy and Misra and Bryant [36] uses a distrubuted-time scheme that allows elements to execute whenever they are ready in an attempt to exploit more parallelism in logic simulations. The extra parallelism comes at the cost of more complex element evaluations and the cost of deadlocks. Once the deadlock has been detected, it is resolved by finding the minimum event time of all unconsumed events in the system. Its base methodology is simple. The algorithm is straightforward to implement with simple control and data structures. CMB performs well as long as all static channels are equally utilized. Large dispersion of events in space and time is not bothersome. In addition, memory consumption is conservative.

2 One of the disadvantages of the CMB algorithm is the high cost of maintaining time synchronicity on a distributed system. There has to be a global control point or a method of agreeing upon current/next GVT amongst distributed systems. This conservative CMB algorithm does not maximize parallelism because of single point of timing control (or arbitration which has a high message overhead). In addition, synchronization mechanism is processor blocking, as a consequence prone to deadlock situations. The processing can not easily handle output of the type where generated event times are not monotonically increasing (time delay generated events). This is because input event times are used by the LPs to generate the LVTH to know whether to continue processing events. A way around this is to generate a message to ones self, indicating that output value x is to be generated after a time delay Null Messages Deadlock in CMB algorithm is avoided by transmitting the null message, [37]which is used to announce the absence of real message. Thus, when receiving a null message, an LP knows it can not receive any real messages from the corresponding incoming link until the time stamped on the null message. Processing of a null message, therefore, just simply resets the link-lock value of the incoming link. Since deadlock in parallel simulation are usually caused by feedback loops in the simulation network, to break the deadlock, usually an enormous number of null messages must be transmitted along these loops. However, in some cases, an LP need not to be blocked if the feedback loop can be detected The Carrier-Null Message Algorithm The Carrier-Null Message algorithm is developed by Cai and Turner[37] to detect the existence of feedback loops during the course of the simulation, and to improve the lookahead of the simulation upon the discovery of the feedback loop. The feedback loop detection and the lookahead improvement make uses of a special protocol message -- carrier-null message. Similar to the CMB algorithm, carrier-null message algorithm also consists of five sequential steps: input, simulation, computation 1, output, and computaion 2. Carrier-null message works different from CBM algorithm in Simulation and Output Steps as described in Cai and Turner [37]. However, the Carrier-Null message algorithm only deals with certain feed-backward networks. More complicated strategies need to be developed in order to efficiently exploit the algorithm. (e.g. the simulation on a LP network with multi-hierarchical feedback loops) Time Windows Based on a provably efficient scheduler for multithread programming, a predictable and robust conservative algorithm was also developed by Cai and Turner [38]. The simulation involves a divide and conquer style of execution which does not require the use of null message or deadlock detection. The algorithm is therefore robust, in that it can be used for simulation applications where local lookahead is small or zero or where lookahead information is difficult to extract. Although local lookahead information can be exploited, the algorithm is robust in that it also involves a global calculation of the time up to which events may be processed safely. The algorithm is novel in a way that it incorporates both local and global information in calculating the SafeTime for each LP. As a result, good speed-up was obtained when the average parallelism is large compared to the number of processors. Cai and Turner s Dag Consistent Parallel Simulation algorithm is further improved by Lim, Low, and Turner[39] on to modify the SafeTime computation a virtual factory simulation platform. It was found that SafeTime computation can be relaxed further to potentially allow more events to execute within a super-step. There are two advantages aspects in this modified algorithm. First it is simple to incorporate lookahead values in the SafeTime computation to dramatically reduce the number of super steps needed. Second, external events from LP I are no longer constrained to generated in timestamp order. 3. Synchronous optimistic event processing.

3 3.1. Overview Optimistic event scheduling was implemented in response to problems in conservative event execution when it comes to parallel execution. The greatest cost in conservative execution was the guarantee of causality. If causality was relaxed, but ensured through correction, execution of a hypothesized input value could improve execution times should the hypothesized input value be correct. In some instances, the output of the process might not change if the hypothesized input was incorrect. To ensure causality through correction, there would have to be a mechanism of recovering past states should the hypothesized value prove to be incorrect. The Time-Warp Operating System[28] was proposed upon this principle Time-Warp Time-Warp incorporates a single input queue of input events. Events are processed in lowest event timestamp first order. Output events are not buffered because there is no requirement that output events occur in time-increasing order as on the Chandy-Misra[36] conservative system. A physical process keeps track of its local virtual time, which is derived directly from the input queue s event time. Should a process discover that its local virtual time is older than the time of an event on the input event queue, the executing process knows that it has processed events out of time order and has a causality violation. Time-Warp corrects the causality violation by rolling back to an earlier state and in the process creating anti-messages to cancel any incorrect output that it generated. The rollback and the costs associated with maintaining the information to be able to rollback are the greatest costs associated with the Time-Warp optimistic event processing. Rollbacks not only invalidate physical process results, through the anti-messages, rollbacks can cause other events to also invalidate the results of other physical processes and potentially cause a cascade of rollbacks. Invalidation of processing causes the execution to occur again with new data, and potentially different results. There is also a significant amount of storage required to support the rollback mechanism. For each state that the physical process goes through, all state variant data has to be stored as well as a copy of the outbound result event and the input event that caused the transition Overview of what is going on to combat weaknesses. There have been several methods proposed and tested, to deal with Time-Warp s tendency to consume processor utilization on calculations that will later be invalidated, costs of rollback, memory consumption and network bandwidth usage between LPs. These include moving window based optimization, lazy/aggressive/adaptive cancellation, control of checkpointing intervals Reduction in rollback costs Aggressive Cancellation - Wilsey[1][3] Aggressive cancellation on rollback makes the assumption that most of the output values will change when executing back to the local virtual time before rollback was caused. The concept is to get the antimessages out to cancel the erroneous output before any other physical processes use it as an input in calculations. To do this, all anti-messages are generated when rolling back, before any re-execution of events is done. This method works fairly well when a process runs as a series of states (state machine) which use input events to transition to different states, output being generated at the new states. It doesn t work well when physical process models a single step function with little or no time lag Lazy Cancellation - Wilsey[1][3] Lazy cancellation on rollback, makes the assumption that little if any of the output will change if a new event is added to the sequence. The concept is to delay sending anti-messages until it is confirmed that the output is incorrect to avoid causing other physical processes to rollback. As a result, the process rolls back and starts executing the modified input event list. If any output events are changed in the process of re-execution, anti-messages are sent out followed by the corrected output event if any. If the result of the

4 re-execution is that there is no change in the output message, nothing is changed and the previous output message is left unchanged. If the result of the re-execution is that there is no output message for the virtual time where the previous execution produced an output message, an anti-message is sent out to cancel the erroneous message and no new message is created. While this technique is effective with nonstate machine processes, it has the potential to cause larger rollbacks on dependant processes by letting them execute further in virtual time before being corrected Adaptive Cancellation - Wilsey[1][3] Adaptive cancellation presented in the Wilsey[3][1] paper is designed to attack the problem of having to predefine cancellation strategies of a physical process at startup. To be able to statically define the optimal cancellation strategy before run-time requires perfect knowledge of the execution environment. In addition, static policies do not allow for adaptation should the execution environment change part way through the execution. It my be the situation where an aggressive cancellation may most effect at first, but during later execution, lazy cancellation may be more effective. According to Wilsey[3][1], during tests, it was found that the optimal strategy was sensitive to how the physical processes were partitioned into logical processes{partitioning scheme}. Adaptive cancellation allows a process to determine which cancellation strategy to use based upon run-time metrics. The adaptation algorithm works by calculating a hit ratio. The hit ratio is essentially the percentage of rolledback output messages where the output did not change. To prevent undue overhead, the sample set depth{filter Depth} is defined at startup. The filter depth is the number of output message comparison results to store for calculating the hit ratio. Two thresholds are defined. A2L_Threshold is the hit ratio threshold at which the algorithm will switch from aggressive to lazy cancellation{a2l}. L2A_Threshold is the threshold at which the algorithm will switch from lazy to aggressive cancellation. The value for A2L_Threshold must be greater or equal to the value for L2A_Threshold. Rapid switching between lazy and aggressive cancellation is prevented by a large filter depth{to reduce statistical variance}, infrequent invocation of the control mechanism and the creation of hysteresis in the switching values by making A2L_Threshold unequal to L2A_Threshold. Testing by Wilsey[3][1] indicated that adaptive cancellation could operate as efficiently(by measuring execution time) as an optimally chosen cancellation protocol Reduction of Memory Footprint One of the greatest consumers of resources in the Time-Warp system, is the usage of memory for rollback state information, input event storage and stored output messages. Because input event storage is needed for calculating forward after a rollback, and the copy of output messages is needed to identify any antimessages that need to be sent out, most of the development has covered methods for reducing the spaced used for copies of physical process states{check-points} Pruneback This is a mechanism that Preiss & Loucks[15] proposed to use to handle memory usage when the system was at or near the limit for available memory {memory stall}. It is not a method to minimize memory usage, but a method to make it more probable that execution will complete by freeing up used but non critical memory allocation. In comparison to Artificial Rollback and Cancelback mechanisms, it promises to allow limited memory recovery while not forcing physical processes to backup in virtual time as does Cancelback and Artificial Rollback. The Pruneback mechanism is based upon recovering needed memory from the saved state information of the physical process. If a rollback occurs to the now missing state, it ends up rolling back to a the first virtual time state before the missing state, and then coasting forward using the saved input events to recalculate that state. There are constraints to pruning back. Neither the physical processes nor the GVT state are to be pruned. Preiss & Loucks[15] left the selection of which states beyond the constraints for the Pruneback mechanism, to be implementation dependent Periodic checkpointing Periodic check-pointing was originally proposed by Jefferson[28] as a way to reduce the state saving

5 overhead of the Time-Warp system. The original paper did not go any further into periodic checkpointing at that time Adaptive checkpointing. Non-Linear feedback control Finding that the Periodic checkpoint algorithm of Lin, Preis, Loucks & Lazowska[2] to be computationally complex, and having the possibility of the gains being offset by the cost of computation, Fleischmann & Wilsey[6] proposed a simpler heuristic algorithm based upon supervisory control in control theory. The heuristic tries to balance two types of overhead, the cost of saving states vs. the cost of coasting forward over unsaved states. A cost function is derived from the sum of the cost of saving states and cost of coasting forwards. The checkpoint interval is recalculated every N intervals (typically 100), if the previous step was to increase the interval, and if the cost function was observed to actually increase, the checkpoint interval will be decremented by one. If the cost function was observed to decrease, the checkpoint interval will be further incremented. The heuristic algorithm is symmetric if the previous step was to decrease the checkpoint interval. Effectively the heuristic works as a local minimum search in process control. Because the heuristic is applied at run-time, it is also capable of handling dynamic changes in rollback and checkpoint behavior Reverse Computation instead of State Saving. Considering that considerable overhead might be involved in saving state, Carothers[20] took the approach have treating the physical processes as being potentially reversable. Rolling back would be accomplished by using the present state and the saved input events to apply inverse operations to the present state to arrive at the rolled back state. In the event that the operation was not completely reversible (ie: a destruct operation) the potential of saving partial state information for the non-reversible portions of the process in the output events was used. A reversible random number generator was also created with a period of since some processes require a pseudo random calculation for event triggered probabilistic events. The reversible physical processes were written by hand, but a mechanism for compiler generating the inverse operations was diagrammed. A side effect of using the reversible approach was discovered in the form of greatly increased speed. In testing, Carothers was using an SGI SMP architecture. Using a hardware profiler on the machine to track the benefit, they found that the smaller memory footprint reduced TLB and cache misses Non checkpointing of Non-state machine elements{rollback Relaxation}[7] Wilsey, Palaniswamy & Aji[7] proposed a method by which memoryless physical processes do not have to store past state data. They separated physical processes into two categories; memoried output events are defined by both input event values and internal state values {remembered history of values, before their present value}, and memoryless process whose output events are directly determined by the values of its input events. To implement this scheme, the input events were sorted into two sets of input queues. They were sorted both by the variable they effect category and the arrival time. This allowed the physical process to determine the input values on a rollback by looking directly at all input events with a previous time to the rollback causing event and proceed forwards from that point. 4. Summary Conservative methods of processing were included in this survey for two reasons. First, conservative methods are the primary way events {data} is processed, including in parallel systems. Second, because Time-Warp or optimistic event processing was an enhancement on the original Chandy-Misra[36] algorithm. In terms of looking towards the future, there will be a greater emphasis on optimistic processing techniques. This will be in part due to electrical transmission limitations. What is meant by limitations, is that as processor clocks run faster, they become more like radio transmitters. As frequencies reach into the Gigahertz, a small trace on a motherboard, 3 inches long, becomes a full wave antenna. This trace, acting as an antenna, will bleed out signal power, be more sensitive to inductive coupling and cause the signal transmitted to be distorted (need about 3 orders of Nyquist resolution to pass a reasonable squarewave. 3 rd Nyquist is the 3 rd harmonic). This will force the usage of loosely

6 coupled parallel systems to solve problems, and conservative processing has the demonstratable tendency to limit parallel processing in its synchronization and causality verification. 5. Future directions for research in the area of Parallel Discrete Event Processing High Level Architecture The High Level Architecture (HLA) which provides the specification of a common technical architecture for use across all classes of simulation in the U.S. Department of Defense. It is currently in the process of being applied with simulation developed for analysis, training, and evaluation. Both the Object Management Group and the IEEE are incorporating HLA into industry standards for distributed simulation. It is foreseeable that researchers in parallel discrete event simulation will be faced with technology challenges which arise in the future implementation and the application of HLA. There will be more research projects that are similar to Fujimoto's project (PDAS group in Georgia Tech) to provide a time management support in both conservative and optimistic synchronization among the simulations in High Level Architecture The rise of Web-based Simulations The World-Wide-Web is revolutionizing the computer industry. Recent advances in Web technology have made the Web a viable mechanism for performing, and distributing simulation. In addition, Java has been considered by most researchers as uniquely suitable to implement advanced architectures for discrete-event simulation. We predict that web-based simulators which take advantage of the capacity of web-based simulation and techniques to improve an object-oriented environment for fast simulation will remain the focus of research in the near future. Furthermore, the integration of High Level Architectures with Web technologies will be a challenging and interesting area for future research Telecommunication and wireless networks While a large amount of Parallel Discrete Event Simulation research has focused on employing multiprocessors and multicomputers, the use of a network of workstations interconnected through Ethernet or ATM has evolved into a popular and effective platform for Parallel Discrete Event Simulation. However, performance evaluation conducted on an advanced version of Time Warp(GTW) has shown that the performance of the GTW is degraded in the network computing environments. Some work have been done on PDES systems to improve simulation performance. We predict that more work on both conservative and optimistic synchronous methods will be demanded to improve simulation performance in ATM networks, cellular networks and large-scale wireless networks Time Warp optimizations For the present, researchers are still faced with the challenges of Time Warp's optimistic synchronization. As mentioned in this survey, the disadvantages of optimistic synchronization are: the high cost in maintaining validity of results, potential to spend CPU cycles on computations that will be thrown away as a result of invalidation of input values on a roll-back, the large memory footprint caused by saving states, results, and inputs between Local Virtual Time and Global Virtual Time for each process. Needless to say, future research will continue to investigate various optimization techniques for Time Warp systems. The future direction will continue to focus on efficient GVT computation, memory management, incremental state saving, etc. BIBLIOGRAPHY Note: The entries marked by an underline are papers that we have had an opportunity to read through and use. Due to time/availability constraints, we were not able to go through all of them. These papers

7 not marked are on our search horizon. The initial intent was to use MSWord s end note capability to be able to dynamically remove and renumber the references at the end of building the paper. Unfortunately MSWord failed us on that one. Because removing the papers would be a non-value-added operation, they were left in, but not underlined. The order of the listing of the bibliography was left title first. 1. On-Line Configuration of a Time Warp Parallel Discrete Event Simulator Radharamanan Radhakrishnan, Nael Abu-Ghazaleh, Mololan Chetlur Philip A. Wilsey Dept of ECES, PO Box Cincinnati University Cincinnati OH Proceedings of the 1998 International Conference on Parallel Processing August 10-14, 1998 ISBN Selecting the checkpoint interval in time warp simulation. Y.B.Lin, B.R. Preiss, W.M.Loucks, E.D.Lazowska Proceedings of the 7 th Workshop on Parallel and Distributed Simulation (PADS), pages Society for Computer Simulation, July Feedback Control in Time Warp Synchronized Parallel Simulators Philip A. Wilsey Dept of ECES, PO Box Cincinnati University Cincinnati OH First International Workshop on Distributed Interactive Simulation and Real Time Applications, An Analytical Comparison of Periodic Check-Pointing and Incremental State Saving A. Palaniswamy, P. A. Wilsey Society for Computer Simulation{PADS 93}, July 1993, pages Adaptive Check-Pointing in Time Warp R.Ronngren and R. Ayani Proceedings of the 8 th Workshop on Parallel and Distributed Simulation{PADS 94}, pages , Society for Computer Simulation. 6. Comparative Analysis of periodic state saving techniques in time warp simulators. J. Fleischmann, P.A.Wilsey Proceedings of the 9 th Workshop on Parallel and Distributed Simulation {PADS 95}, pages 50-58, June Rollback Relaxation: A Technique for Reducing Rollback Costs in Optimistically Synchronized Parallel Simulators Philip A. Wilsey, Avinash Palaniswamy, Sandeep Aji Center for Digital Systems Engineering University of Cincinnati, Cincinnati OH Proceeding of the International Conference on Simulation and Hardware Description Languages, SHDL The adaptive Time-Warp concurrency control algorithm. D. Ball & S. Hoyt Distributed Simulation pgs , Society for Computer Simulation, Jan Adaptive bounded time windows in an optimistically synchronized simulator. Palaniswamy & P.A. Wilsey Third Great Lakes Symposium on VLSI, pages , Performance evaluation of the bounded Time Warp algorithm S.J. Turner and M. Q. Xu. Proceedings of the SCS Multiconference on Parallel and Distributed Simulation 24(3): , A performance study of the cancelback protocol for time warp S.R.Das & R.M. Fujimoto Proceedings of the 7th Workshop on Parallel and Distributed Simulation. Jul 1993 pages An adaptive memory management protocol for Time Warp parallel simulation. S.R.Das and R.M.Fujimoto "Sigmetrics", pgs , May Investigations in adaptive distributed simulation. D. O. Hamnes and A. Tripathi. Proceedings of the 8th Workshop on Parallel and Distributed Simulation PADS(94)" pages Society for Computer Simulation, July Optimal Memory Management for Time Warp Parallel simulation. Y.B. Lin & B.R.Preiss. ACM Transactions on Modeling and Computer Simulation, 1(4): , Oct 1991.

8 15. Memory Management Techniques for Time Warp on a Distributed Memory Machine Bruno R. Preiss, Wayne M. Loucks Department of Electrical and Computer Engineering University of Waterloo, Waterloo ON N2L 3G1 {Ontario, Canada} The Institute of Electrical and Computer Engineers, Optimistic Fossil Collection for Time Warp Simulation Christopher H. Young, Philip A. Wilsey Computer Architecture Design Laboratory Dept. of ECES, PO Box , Cincinnati OH Proceedings of the Hawaiian International Conference on System Sciences HICSS Optimism: Not Just For Event Execution Anymore Christopher H. Young, Radharamanan Radhakrishnan, Philip A. Wilsey Dept of ECES, University of Cincinnati, Cincinnati OH th Workshop on Parallel and Distributed Simulation, PADS Performance Benefits of Optimism in Fossil Collection Christopher H. Young, Radharamanan Radhakrishnan, Philip A. Wilsey Dept. of ECES, University of Cincinnati, Cincinnati OH Proceedings of the Hawaii International Conference on System Sciences, HICSS Dynamically Switching between Lazy and Aggressive Cancellation in a Time Warp Parallel Simulator Raghundandan Rajan, Philip A. Wilsey Center for Digital Systems Engineering Dept. of ECE, PO Box , University of Cincinnati, Cincinnati OH Proceedings of the 28 th Annual Simulation Symposium, ASS Efficient Optimistic Parallel Simulations using Reverse Computation Christopher D. Carothers, Dep. Computer Science Rennselaer Polytechnic Institute Kalyan S. Perumalla & Richard M. Fujimoto, College of Computing, Georgia Inst. Tech. <No indicated publication > 21. A Control Mechanism for Parallel Discrete Simulation L. M. Sokol, B. K. Stucky, V. S. Hwang Proceedings of the 1989 International conference on Parallel Processing, Vol 3, pp , Performance evaluation of the bounded Time Warp Algorithm S. Turner M. Xu Proceedings of the 6 th Workshop on Parallel and Distributed Simulation pp , Concurrency Preserving Partitioning(CPP) for Parallel Logic Simulation Hong. K. Kim, Jack Jean Proceedings of the 10 th Workshop on Parallel and distributed Simulation, pp98-105, Load Balancing and Work Load Minimization of Overlapping Parallel Tasks V. Krishnaswamy, G. Hasteer, P. Banerjee Proceedings of the 1997 International Conference on Parallel Processing, Aug Asynchronous Distributed Simulation via a Sequence of Parallel Computations K. M. Chandy, J. Misra Communications of the ACM, Vol 24, No 11, pgs April Parallel and Distributed Simulation of Discrete Event Systems Alois Ferscha Contributed to Handbook of Parallel and Distributed Computing, McGraw-Hill, 1995 NOTE: This one had a detailed summary of different mechanisms, though from a simulation point of view. 27. Distributed Discrete-Event Simulation J. Misra ACM Computing Surveys, 18(1):39-65, Distributed Simulation and the Time Warp Operating System David Jefferson, Brian Beckman, Fred Wieland, Leo Blume, Mike DiLorento, Phil Hontalas, Pierre Laroche, Kathy Sturdevant, Jack Tupman, Van Warren, John Wedel, Herb Younger, Steve Bellenot NOTE: Original Class article. 29. Parallel Discrete-Event Simulation of FCFS Stochastic Queuing Networks. D. M. Nicol Proceedings of the ACM/SIGPLAN APPEALS 1988, pgs Algorithms for distributed termination detection F. Mattern Distributed Computing, 2: , Distributed Deadlock Detection K. M. Chandy, J. Misra, L. M. Haas ACM Transactions on Computer Systems, 1(2): , May Bounded Lag Distributed Discrete Event Simulation. B. D. Lubachevsky Proceedings of the SCS Multiconference on Distributed Simulation 19(3) pgs SCS, Feb1988

9 33. An Algorithm for Distributed Discrete-Event Simulation The Carrier Null Message Approach Wentong Cai, Steven J. Turner Proceedings of the SCS Multiconference on Distributed Simulation Vol 22(1), pages 3-8. SCS January An Algorithm for Reducing Null-Messages of CMB Approach in Parallel Discrete Event Simulation Wenton Cai, Steven J. Turner 35. OFC: A Distributed Fossil-Collection Algorithm for Time-Warp Christopher H. Young, Nael B. Abu-Ghazaleh, Philip A. Wilsey Proceedings of the 12 th International Conference on Distributed Computing DISC An Evaluation of the Chandy-Misra-Bryant Algorithm for Digital Logic Simulation Larry Soule and Anoop Gupta, Computer Systems Laboratory, Standford University, CA Distributed Simulation, Volume 24(2), Pages , California, SCS. Simulation: a Predictable and Robust Conservative. 37. An Algorithm For Reducing Null-Message of CMB Approach in Parallel Discrete Event Simulation. Wentong Cai, Stephen J. Turner Dag Consistent Parallel Simulaition 38. Dag Consistent Parallel Simulation: a Predictable and Robust Conservative Algorithm Wentong Cai, Stephen J. Turner 39. Relaxing SafeTime Computation of a Conservative Simulation Algorithm Chu-Cheow Lim, Yoke-Hean Low, and Stephen J. Turner

An Empirical Performance Study of Connection Oriented Time Warp Parallel Simulation

An Empirical Performance Study of Connection Oriented Time Warp Parallel Simulation 230 The International Arab Journal of Information Technology, Vol. 6, No. 3, July 2009 An Empirical Performance Study of Connection Oriented Time Warp Parallel Simulation Ali Al-Humaimidi and Hussam Ramadan

More information

Lookahead Accumulation in Conservative Parallel Discrete Event Simulation.

Lookahead Accumulation in Conservative Parallel Discrete Event Simulation. Lookahead Accumulation in Conservative Parallel Discrete Event Simulation. Jan Lemeire, Wouter Brissinck, Erik Dirkx Parallel Systems lab, Vrije Universiteit Brussel (VUB) Brussels, Belgium {jlemeire,

More information

Parallel and Distributed VHDL Simulation

Parallel and Distributed VHDL Simulation Parallel and Distributed VHDL Simulation Dragos Lungeanu Deptartment of Computer Science University of Iowa C.J. chard Shi Department of Electrical Engineering University of Washington Abstract This paper

More information

Event List Management In Distributed Simulation

Event List Management In Distributed Simulation Event List Management In Distributed Simulation Jörgen Dahl ½, Malolan Chetlur ¾, and Philip A Wilsey ½ ½ Experimental Computing Laboratory, Dept of ECECS, PO Box 20030, Cincinnati, OH 522 0030, philipwilsey@ieeeorg

More information

Rollback Overhead Reduction Methods for Time Warp Distributed Simulation

Rollback Overhead Reduction Methods for Time Warp Distributed Simulation Rollback Overhead Reduction Methods for Time Warp Distributed Simulation M.S. Balsamo and C. Manconi* Dipartimento di Matematica e Informatica, University of Udine Vial delle Scienze 108, Udine, Italy,

More information

Optimistic Parallel Simulation of TCP/IP over ATM networks

Optimistic Parallel Simulation of TCP/IP over ATM networks Optimistic Parallel Simulation of TCP/IP over ATM networks M.S. Oral Examination November 1, 2000 Ming Chong mchang@ittc.ukans.edu 1 Introduction parallel simulation ProTEuS Agenda Georgia Tech. Time Warp

More information

Other Optimistic Mechanisms, Memory Management!

Other Optimistic Mechanisms, Memory Management! Other Optimistic Mechanisms, Memory Management! Richard M. Fujimoto! Professor!! Computational Science and Engineering Division! College of Computing! Georgia Institute of Technology! Atlanta, GA 30332-0765,

More information

Parallel Discrete Event Simulation

Parallel Discrete Event Simulation Parallel Discrete Event Simulation Dr.N.Sairam & Dr.R.Seethalakshmi School of Computing, SASTRA Univeristy, Thanjavur-613401. Joint Initiative of IITs and IISc Funded by MHRD Page 1 of 8 Contents 1. Parallel

More information

χ=5 virtual time state LVT entirely saved state partially saved state χ=5 ν=2 virtual time state LVT entirely saved partially saved unsaved state

χ=5 virtual time state LVT entirely saved state partially saved state χ=5 ν=2 virtual time state LVT entirely saved partially saved unsaved state ROLLBACK-BASED PARALLEL DISCRETE EVENT SIMULATION BY USING HYBRID STATE SAVING Francesco Quaglia Dipartimento di Informatica e Sistemistica, Universita di Roma "La Sapienza" Via Salaria 113, 00198 Roma,

More information

Harnessing Parallelism in a Parallel Discrete-Event Simulation

Harnessing Parallelism in a Parallel Discrete-Event Simulation Proceedings of the IASTED International Conference Modelling and Simulation May 5 8, 1999, Philadelphia, Pennsylvania, USA Harnessing Parallelism in a Parallel Discrete-Event Simulation Chu-Cheow LIM Yoke-Hean

More information

Predicting the performance of synchronous discrete event simulation systems

Predicting the performance of synchronous discrete event simulation systems Predicting the performance of synchronous discrete event simulation systems Jinsheng Xu and Moon Jung Chung Department of Computer Science Michigan State University {xujinshe,chung}@cse.msu.edu ABSTRACT

More information

DISTRIBUTED SIMULATION SYSTEMS. Richard M. Fujimoto. College of Computing Georgia Institute of Technology Atlanta, GA 30332, U.S.A.

DISTRIBUTED SIMULATION SYSTEMS. Richard M. Fujimoto. College of Computing Georgia Institute of Technology Atlanta, GA 30332, U.S.A. Proceedings of the 2003 Winter Simulation Conference S. Chick, P. J. Sánchez, D. Ferrin, and D. J. Morrice, eds. DISTRIBUTED SIMULATION SYSTEMS Richard M. Fujimoto College of Computing Georgia Institute

More information

Eect of fan-out on the Performance of a. Single-message cancellation scheme. Atul Prakash (Contact Author) Gwo-baw Wu. Seema Jetli

Eect of fan-out on the Performance of a. Single-message cancellation scheme. Atul Prakash (Contact Author) Gwo-baw Wu. Seema Jetli Eect of fan-out on the Performance of a Single-message cancellation scheme Atul Prakash (Contact Author) Gwo-baw Wu Seema Jetli Department of Electrical Engineering and Computer Science University of Michigan,

More information

JWarp: a Java library for parallel discrete-event simulations

JWarp: a Java library for parallel discrete-event simulations CONCURRENCY: PRACTICE AND EXPERIENCE Concurrency: Pract. Exper.,Vol.10(11 13), 999 1005 (1998) JWarp: a Java library for parallel discrete-event simulations PEDRO BIZARRO,LUÍS M. SILVA AND JOÃO GABRIEL

More information

Event Reconstruction in Time Warp

Event Reconstruction in Time Warp Event Reconstruction in Time Warp Lijun Li and Carl Tropper School of Computer Science McGill University Montreal, Canada lli22, carl@cs.mcgill.ca Abstract In optimistic simulations, checkpointing techniques

More information

The Impact of Lookahead on the Performance of Conservative Distributed Simulation

The Impact of Lookahead on the Performance of Conservative Distributed Simulation The Impact of Lookahead on the Performance of Conservative Distributed Simulation Bruno R Preiss Wayne M Loucks Department of Electrical and Computer Engineering University of Waterloo, Waterloo, Ontario,

More information

A Study of the Effect of Partitioning on Parallel Simulation of Multicore Systems

A Study of the Effect of Partitioning on Parallel Simulation of Multicore Systems A Study of the Effect of Partitioning on Parallel Simulation of Multicore Systems Zhenjiang Dong, Jun Wang, George Riley, Sudhakar Yalamanchili School of Electrical and Computer Engineering Georgia Institute

More information

Chair for Network Architectures and Services Prof. Carle Department of Computer Science TU München. Parallel simulation

Chair for Network Architectures and Services Prof. Carle Department of Computer Science TU München. Parallel simulation Chair for Network Architectures and Services Prof. Carle Department of Computer Science TU München Parallel simulation Most slides/figures borrowed from Richard Fujimoto Parallel simulation: Summary/Outline

More information

Distributed Deadlock Detection for. Distributed Process Networks

Distributed Deadlock Detection for. Distributed Process Networks 0 Distributed Deadlock Detection for Distributed Process Networks Alex Olson Embedded Software Systems Abstract The distributed process network (DPN) model allows for greater scalability and performance

More information

Computing Global Virtual Time!

Computing Global Virtual Time! Computing Global Virtual Time Issues and Some Solutions Richard M. Fujimoto Professor Computational Science and Engineering Division College of Computing Georgia Institute of Technology Atlanta, GA 30332-0765,

More information

COMPILED CODE IN DISTRIBUTED LOGIC SIMULATION. Jun Wang Carl Tropper. School of Computer Science McGill University Montreal, Quebec, CANADA H3A2A6

COMPILED CODE IN DISTRIBUTED LOGIC SIMULATION. Jun Wang Carl Tropper. School of Computer Science McGill University Montreal, Quebec, CANADA H3A2A6 Proceedings of the 2006 Winter Simulation Conference L. F. Perrone, F. P. Wieland, J. Liu, B. G. Lawson, D. M. Nicol, and R. M. Fujimoto, eds. COMPILED CODE IN DISTRIBUTED LOGIC SIMULATION Jun Wang Carl

More information

Distributed Simulation of Large Computer Systems

Distributed Simulation of Large Computer Systems Distributed Simulation of Large Computer Systems Moreno Marzolla Univ. di Venezia Ca Foscari Dept. of Computer Science and INFN Padova Email: marzolla@dsi.unive.it Web: www.dsi.unive.it/ marzolla Moreno

More information

PARALLEL AND DISTRIBUTED SIMULATION. Richard M. Fujimoto. College of Computing Georgia Institute of Technology Atlanta, GA 3033, U.S.A.

PARALLEL AND DISTRIBUTED SIMULATION. Richard M. Fujimoto. College of Computing Georgia Institute of Technology Atlanta, GA 3033, U.S.A. Proceedings of the 1999 Winter Simulation Conference P. A. Farrington, H. B. Nembhard, D. T. Sturrock, and G. W. Evans, eds. PARALLEL AND DISTRIBUTED SIMULATION Richard M. Fujimoto College of Computing

More information

Parallel Logic Simulation on General Purpose Machines

Parallel Logic Simulation on General Purpose Machines Parallel Logic Simulation on General Purpose Machines Larry Soule', Tom Blank Center for Integrated Systems Stanford University Abstract Three parallel algorithms for logic simulation have been developed

More information

PARALLEL SIMULATION. Rassul Ayani Department of Teleinformatics, Computer Systems Laboratory Royal Institute of Technology (KTH) Stockholm, Sweden

PARALLEL SIMULATION. Rassul Ayani Department of Teleinformatics, Computer Systems Laboratory Royal Institute of Technology (KTH) Stockholm, Sweden PARALLEL SIMULATION Rassul Ayani Department of Teleinformatics, Computer Systems Laboratory Royal Institute of Technology (KTH) Stockholm, Sweden Abstract This tutorial surveys various approaches to executing

More information

USING GENETIC ALGORITHMS TO LIMIT THE OPTIMISM IN TIME WARP. Jun Wang Carl Tropper

USING GENETIC ALGORITHMS TO LIMIT THE OPTIMISM IN TIME WARP. Jun Wang Carl Tropper Proceedings of the 2009 Winter Simulation Conference M. D. Rossetti, R. R. Hill, B. Johansson, A. Dunkin, and R. G. Ingalls, eds. USING GENETIC ALGORITHMS TO LIMIT THE OPTIMISM IN TIME WARP Jun Wang Carl

More information

Using Timestamps to Track Causal Dependencies

Using Timestamps to Track Causal Dependencies Using Timestamps to Track Causal Dependencies J. A. David McWha Dept. of Computer Science, University of Waikato, Private Bag 315, Hamilton jadm@cs.waikato.ac.nz ABSTRACT As computer architectures speculate

More information

Implementing Sequential Consistency In Cache-Based Systems

Implementing Sequential Consistency In Cache-Based Systems To appear in the Proceedings of the 1990 International Conference on Parallel Processing Implementing Sequential Consistency In Cache-Based Systems Sarita V. Adve Mark D. Hill Computer Sciences Department

More information

Just-In-Time Cloning

Just-In-Time Cloning Just-In-Time Cloning Maria Hybinette Computer Science Department University of Georgia Athens, GA 30602-7404, USA maria@cs.uga.edu Abstract In this work we focus on a new technique for making cloning of

More information

An Approach for Federating Parallel Simulators

An Approach for Federating Parallel Simulators An Approach for Federating Parallel Simulators Steve L. Ferenci Kalyan S. Perumalla Richard M. Fujimoto College Of Computing Georgia Institute of Technology Atlanta, GA 30332-0280 {ferenci,kalyan,fujimoto}@cc.gatech.edu

More information

Parallel Logic Simulation of VLSI Systems

Parallel Logic Simulation of VLSI Systems Parallel Logic Simulation of VLSI Systems Roger D. Chamberlain Computer and Communications Research Center Department of Electrical Engineering Washington University, St. Louis, Missouri Abstract Design

More information

Analysing Probabilistically Constrained Optimism

Analysing Probabilistically Constrained Optimism Analysing Probabilistically Constrained Optimism Michael Lees and Brian Logan School of Computer Science & IT University of Nottingham UK {mhl,bsl}@cs.nott.ac.uk Dan Chen, Ton Oguara and Georgios Theodoropoulos

More information

A Methodology for Automating the Parallelization of Manufacturing Simulations

A Methodology for Automating the Parallelization of Manufacturing Simulations A Methodology for Automating the Parallelization of Manufacturing Simulations Stephen J. Turner Wentong Cai Chu Cheow Lim Yoke Hean Low Wen Jing Hsu Shell Ying Huang Gintic Institute of School of Applied

More information

Fault-tolerant Distributed-Shared-Memory on a Broadcast-based Interconnection Network

Fault-tolerant Distributed-Shared-Memory on a Broadcast-based Interconnection Network Fault-tolerant Distributed-Shared-Memory on a Broadcast-based Interconnection Network Diana Hecht 1 and Constantine Katsinis 2 1 Electrical and Computer Engineering, University of Alabama in Huntsville,

More information

Towards Adaptive Caching for Parallel and Discrete Event Simulation

Towards Adaptive Caching for Parallel and Discrete Event Simulation Towards Adaptive Caching for Parallel and Discrete Event Simulation Abhishek Chugh and Maria Hybinette Computer Science Department The University of Georgia 415 Boyd Graduate Studies Research Center Athens,

More information

Optimistic Parallel Discrete Event Simulations of Physical Systems using Reverse Computation

Optimistic Parallel Discrete Event Simulations of Physical Systems using Reverse Computation Optimistic Parallel Discrete Event Simulations of Physical Systems using Reverse Computation Yarong Tang, Kalyan S. Perumalla, Richard M. Fujimoto College of Computing, Georgia Institute of Technology

More information

Event Simulation Algorithms

Event Simulation Algorithms VLSI Design 1994, Vol. 2, No. 1, pp. 1-16 Reprints available directly from the publisher Photocopying permitted by license only (C) 1994 Gordon and Breach Science Publishers S.A. Printed in the United

More information

Computing Global Virtual Time in Shared- Memory Multiprocessors

Computing Global Virtual Time in Shared- Memory Multiprocessors Computing Global Virtual Time in Shared- Memory Multiprocessors RICHARD M. FUJIMOTO and MARIA HYBINETTE Georgia Institute of Technology Global virtual time (GVT) is used in the Time Warp synchronization

More information

PARALLEL QUEUING NETWORK SIMULATION WITH LOOKBACK- BASED PROTOCOLS

PARALLEL QUEUING NETWORK SIMULATION WITH LOOKBACK- BASED PROTOCOLS PARALLEL QUEUING NETWORK SIMULATION WITH LOOKBACK- BASED PROTOCOLS Gilbert G. Chen and Boleslaw K. Szymanski Department of Computer Science, Rensselaer Polytechnic Institute 110 Eighth Street, Troy, NY

More information

PARALLEL EXECUTION OF A SEQUENTIAL NETWORK SIMULATOR

PARALLEL EXECUTION OF A SEQUENTIAL NETWORK SIMULATOR Proceedings of the 2000 Winter Simulation Conference J. A. Joines, R. R. Barton, K. Kang, and P. A. Fishwick, eds. PARALLEL EXECUTION OF A SEQUENTIAL NETWORK SIMULATOR Kevin G. Jones Division of Computer

More information

Tutorial Parallel & Distributed Simulation Systems and the High Level Architecture

Tutorial Parallel & Distributed Simulation Systems and the High Level Architecture Tutorial Parallel & Distributed Simulation Systems and the High Level Architecture Kalyan Perumalla, Ph.D. Research Faculty Member College of Computing & MSREC Georgia Institute of Technology Atlanta,

More information

FAST CELL LEVEL ATM NETWORK SIMULATION

FAST CELL LEVEL ATM NETWORK SIMULATION Proceedings of the 22 Winter Simulation Conference E. Yücesan, C.-H. Chen, J. L. Snowdon, and J. M. Charnes, eds. FAST CELL LEVEL ATM NETWORK SIMULATION Xiao Zhong-e Rob Simmonds Brian Unger Dept. Computer

More information

RTI Performance on Shared Memory and Message Passing Architectures

RTI Performance on Shared Memory and Message Passing Architectures RTI Performance on Shared Memory and Message Passing Architectures Steve L. Ferenci Richard Fujimoto, PhD College Of Computing Georgia Institute of Technology Atlanta, GA 3332-28 {ferenci,fujimoto}@cc.gatech.edu

More information

PARALLEL SIMULATION OF SCALE-FREE NETWORKS

PARALLEL SIMULATION OF SCALE-FREE NETWORKS PARALLEL SIMULATION OF SCALE-FREE NETWORKS A Dissertation Presented to The Academic Faculty by Vy Thuy Nguyen In Partial Fulfillment of the Requirements for the Degree Master of Science in the School of

More information

IMPROVING PERFORMANCE OF PARALLEL SIMULATION KERNEL FOR WIRELESS NETWORK SIMULATIONS

IMPROVING PERFORMANCE OF PARALLEL SIMULATION KERNEL FOR WIRELESS NETWORK SIMULATIONS IMPROVING PERFORMANCE OF PARALLEL SIMULATION KERNEL FOR WIRELESS NETWORK SIMULATIONS M. Thoppian, S. Venkatesan, H. Vu, R. Prakash, N. Mittal Department of Computer Science, The University of Texas at

More information

Timestamp Representations for Virtual Sequences

Timestamp Representations for Virtual Sequences Timestamp Representations for Virtual equences John G. Cleary, J. A. David McWha, Murray Pearson Dept of Computer cience, University of Waikato, Private Bag 305, Hamilton, New Zealand. {jcleary, jadm,

More information

6LPXODWLRQÃRIÃWKHÃ&RPPXQLFDWLRQÃ7LPHÃIRUÃDÃ6SDFH7LPH $GDSWLYHÃ3URFHVVLQJÃ$OJRULWKPÃRQÃDÃ3DUDOOHOÃ(PEHGGHG 6\VWHP

6LPXODWLRQÃRIÃWKHÃ&RPPXQLFDWLRQÃ7LPHÃIRUÃDÃ6SDFH7LPH $GDSWLYHÃ3URFHVVLQJÃ$OJRULWKPÃRQÃDÃ3DUDOOHOÃ(PEHGGHG 6\VWHP LPXODWLRQÃRIÃWKHÃ&RPPXQLFDWLRQÃLPHÃIRUÃDÃSDFHLPH $GDSWLYHÃURFHVVLQJÃ$OJRULWKPÃRQÃDÃDUDOOHOÃ(PEHGGHG \VWHP Jack M. West and John K. Antonio Department of Computer Science, P.O. Box, Texas Tech University,

More information

Proceedings of the 2017 Winter Simulation Conference W. K. V. Chan, A. D'Ambrogio, G. Zacharewicz, N. Mustafee, G. Wainer, and E. Page, eds.

Proceedings of the 2017 Winter Simulation Conference W. K. V. Chan, A. D'Ambrogio, G. Zacharewicz, N. Mustafee, G. Wainer, and E. Page, eds. Proceedings of the 2017 Winter Simulation Conference W. K. V. Chan, A. D'Ambrogio, G. Zacharewicz, N. Mustafee, G. Wainer, and E. Page, eds. ABSTRACT VIRTUAL TIME III: UNIFICATION OF CONSERVATIVE AND OPTIMISTIC

More information

CSCI 4717 Computer Architecture

CSCI 4717 Computer Architecture CSCI 4717/5717 Computer Architecture Topic: Symmetric Multiprocessors & Clusters Reading: Stallings, Sections 18.1 through 18.4 Classifications of Parallel Processing M. Flynn classified types of parallel

More information

MANUFACTURING SIMULATION USING BSP TIME WARP WITH VARIABLE NUMBERS OF PROCESSORS

MANUFACTURING SIMULATION USING BSP TIME WARP WITH VARIABLE NUMBERS OF PROCESSORS MANUFACTURING SIMULATION USING BSP TIME WARP WITH VARIABLE NUMBERS OF PROCESSORS Malcolm Yoke Hean Low Programming Research Group, Computing Laboratory, University of Oxford Wolfson Building, Parks Road,

More information

DISTRIBUTED SELF-SIMULATION OF HOLONIC MANUFACTURING SYSTEMS

DISTRIBUTED SELF-SIMULATION OF HOLONIC MANUFACTURING SYSTEMS DISTRIBUTED SELF-SIMULATION OF HOLONIC MANUFACTURING SYSTEMS Naoki Imasaki I, Ambalavanar Tharumarajah 2, Shinsuke Tamura 3 J Toshiba Corporation, Japan, naoki.imasaki@toshiba.co.jp 2 CSIRO Manufacturing

More information

SIMULATIONS. PACE Lab, Rockwell Collins, IMPROVING PERFORMANCE OF PARALLEL SIMULATION KERNEL FOR WIRELESS NETWORK

SIMULATIONS. PACE Lab, Rockwell Collins, IMPROVING PERFORMANCE OF PARALLEL SIMULATION KERNEL FOR WIRELESS NETWORK IMPROVING PERFORMANCE OF PARALLEL SIMULATION KERNEL FOR WIRELESS NETWORK SIMULATIONS M. Thoppian, S. Venkatesan, H. Vu, R. Prakash, N. Mittal Department of Computer Science, The University of Texas at

More information

SPECULATIVE MULTITHREADED ARCHITECTURES

SPECULATIVE MULTITHREADED ARCHITECTURES 2 SPECULATIVE MULTITHREADED ARCHITECTURES In this Chapter, the execution model of the speculative multithreading paradigm is presented. This execution model is based on the identification of pairs of instructions

More information

Optimistic Distributed Simulation Based on Transitive Dependency. Tracking. Dept. of Computer Sci. AT&T Labs-Research Dept. of Elect. & Comp.

Optimistic Distributed Simulation Based on Transitive Dependency. Tracking. Dept. of Computer Sci. AT&T Labs-Research Dept. of Elect. & Comp. Optimistic Distributed Simulation Based on Transitive Dependency Tracking Om P. Damani Yi-Min Wang Vijay K. Garg Dept. of Computer Sci. AT&T Labs-Research Dept. of Elect. & Comp. Eng Uni. of Texas at Austin

More information

MDRIP: A Hybrid Approach to. Parallelisation of Discrete Event Simulation

MDRIP: A Hybrid Approach to. Parallelisation of Discrete Event Simulation MDRIP: A Hybrid Approach to Parallelisation of Discrete Event Simulation A thesis submitted in partial fulfilment of the requirements for the Degree of Master of Science in the University of Canterbury

More information

ANALYSIS OF TIME WARP ON A 32,768 PROCESSOR IBM BLUE GENE/L SUPERCOMPUTER

ANALYSIS OF TIME WARP ON A 32,768 PROCESSOR IBM BLUE GENE/L SUPERCOMPUTER ANALYSIS OF TIME WARP ON A 32,768 PROCESSOR IBM BLUE GENE/L SUPERCOMPUTER Akintayo O. Holder and Christopher D. Carothers Department of Computer Science Rensselaer Polytechnic Institute Troy, NY 12180,

More information

2 TEST: A Tracer for Extracting Speculative Threads

2 TEST: A Tracer for Extracting Speculative Threads EE392C: Advanced Topics in Computer Architecture Lecture #11 Polymorphic Processors Stanford University Handout Date??? On-line Profiling Techniques Lecture #11: Tuesday, 6 May 2003 Lecturer: Shivnath

More information

AGGRESSIVENESS/RISK EFFECTS BASED SCHEDULING IN TIME WARP

AGGRESSIVENESS/RISK EFFECTS BASED SCHEDULING IN TIME WARP Proceedings of the 2000 Winter Simulation Conference J. A. Joines, R. R. Barton, K. Kang, and P. A. Fishwick, eds. AGGRESSIVENESS/RISK EFFECTS BASED SCHEDULING IN TIME WARP Vittorio Cortellessa Computer

More information

An Eæcient Conditional-knowledge based. Optimistic Simulation Scheme. Atul Prakash. June 29, 1991.

An Eæcient Conditional-knowledge based. Optimistic Simulation Scheme. Atul Prakash.   June 29, 1991. An Eæcient Conditional-knowledge based Optimistic Simulation Scheme Atul Prakash Rajalakshmi Subramanian Department of Electrical Engineering and Computer Science University of Michigan, Ann Arbor, MI

More information

Design of Parallel Algorithms. Course Introduction

Design of Parallel Algorithms. Course Introduction + Design of Parallel Algorithms Course Introduction + CSE 4163/6163 Parallel Algorithm Analysis & Design! Course Web Site: http://www.cse.msstate.edu/~luke/courses/fl17/cse4163! Instructor: Ed Luke! Office:

More information

Principles of Parallel Algorithm Design: Concurrency and Mapping

Principles of Parallel Algorithm Design: Concurrency and Mapping Principles of Parallel Algorithm Design: Concurrency and Mapping John Mellor-Crummey Department of Computer Science Rice University johnmc@rice.edu COMP 422/534 Lecture 3 28 August 2018 Last Thursday Introduction

More information

Distributed Simulations. Atul Prakash. simulation algorithms can be found in ë5ë. In

Distributed Simulations. Atul Prakash.   simulation algorithms can be found in ë5ë. In Filter: An Algorithm for Reducing Cascaded Rollbacks in Optimistic Distributed Simulations Atul Prakash Rajalakshmi Subramanian Department of Electrical Engineering and Computer Science University of Michigan,

More information

Petri Nets ~------~ R-ES-O---N-A-N-C-E-I--se-p-te-m--be-r Applications.

Petri Nets ~------~ R-ES-O---N-A-N-C-E-I--se-p-te-m--be-r Applications. Petri Nets 2. Applications Y Narahari Y Narahari is currently an Associate Professor of Computer Science and Automation at the Indian Institute of Science, Bangalore. His research interests are broadly

More information

An Improved Data Flow Architecture for Logic

An Improved Data Flow Architecture for Logic VLSI Design 1994, Vol. 2, No. 3, pp. 259-265 Reprints available directly from the publisher Photocopying permitted by license only (C) 1994 Gordon and Breach Science Publishers S.A. Printed in the United

More information

Introduction. Analytic simulation. Time Management

Introduction. Analytic simulation. Time Management XMSF Workshop Monterey, CA Position Paper Kalyan S. Perumalla, Ph.D. College of Computing, Georgia Tech http://www.cc.gatech.edu/~kalyan August 19, 2002 Introduction The following are some of the authors

More information

Consistent Logical Checkpointing. Nitin H. Vaidya. Texas A&M University. Phone: Fax:

Consistent Logical Checkpointing. Nitin H. Vaidya. Texas A&M University. Phone: Fax: Consistent Logical Checkpointing Nitin H. Vaidya Department of Computer Science Texas A&M University College Station, TX 77843-3112 hone: 409-845-0512 Fax: 409-847-8578 E-mail: vaidya@cs.tamu.edu Technical

More information

FAULT TOLERANT SYSTEMS

FAULT TOLERANT SYSTEMS FAULT TOLERANT SYSTEMS http://www.ecs.umass.edu/ece/koren/faulttolerantsystems Part 17 - Checkpointing II Chapter 6 - Checkpointing Part.17.1 Coordinated Checkpointing Uncoordinated checkpointing may lead

More information

Understanding The Behavior of Simultaneous Multithreaded and Multiprocessor Architectures

Understanding The Behavior of Simultaneous Multithreaded and Multiprocessor Architectures Understanding The Behavior of Simultaneous Multithreaded and Multiprocessor Architectures Nagi N. Mekhiel Department of Electrical and Computer Engineering Ryerson University, Toronto, Ontario M5B 2K3

More information

warped: A Time Warp Simulation Kernel for Analysis and Application Development Dale E. Martin, Timothy J. McBrayer, and Philip A.

warped: A Time Warp Simulation Kernel for Analysis and Application Development Dale E. Martin, Timothy J. McBrayer, and Philip A. Published in the Proceedings of the Hawaiian International Conference on System Sciences, HICSS-1996. c 1996, IEEE. Personal use of this material is permitted. However permission to reprint or republish

More information

Verification and Validation

Verification and Validation Chapter 5 Verification and Validation Chapter Revision History Revision 0 Revision 1 Revision 2 Revision 3 Revision 4 original 94/03/23 by Fred Popowich modified 94/11/09 by Fred Popowich reorganization

More information

Implementation of Process Networks in Java

Implementation of Process Networks in Java Implementation of Process Networks in Java Richard S, Stevens 1, Marlene Wan, Peggy Laramie, Thomas M. Parks, Edward A. Lee DRAFT: 10 July 1997 Abstract A process network, as described by G. Kahn, is a

More information

ASSEMBLY LANGUAGE MACHINE ORGANIZATION

ASSEMBLY LANGUAGE MACHINE ORGANIZATION ASSEMBLY LANGUAGE MACHINE ORGANIZATION CHAPTER 3 1 Sub-topics The topic will cover: Microprocessor architecture CPU processing methods Pipelining Superscalar RISC Multiprocessing Instruction Cycle Instruction

More information

A Study of the Performance Tradeoffs of a Tape Archive

A Study of the Performance Tradeoffs of a Tape Archive A Study of the Performance Tradeoffs of a Tape Archive Jason Xie (jasonxie@cs.wisc.edu) Naveen Prakash (naveen@cs.wisc.edu) Vishal Kathuria (vishal@cs.wisc.edu) Computer Sciences Department University

More information

MULTIPROCESSORS AND THREAD LEVEL PARALLELISM

MULTIPROCESSORS AND THREAD LEVEL PARALLELISM UNIT III MULTIPROCESSORS AND THREAD LEVEL PARALLELISM 1. Symmetric Shared Memory Architectures: The Symmetric Shared Memory Architecture consists of several processors with a single physical memory shared

More information

6.895 Final Project: Serial and Parallel execution of Funnel Sort

6.895 Final Project: Serial and Parallel execution of Funnel Sort 6.895 Final Project: Serial and Parallel execution of Funnel Sort Paul Youn December 17, 2003 Abstract The speed of a sorting algorithm is often measured based on the sheer number of calculations required

More information

740: Computer Architecture Memory Consistency. Prof. Onur Mutlu Carnegie Mellon University

740: Computer Architecture Memory Consistency. Prof. Onur Mutlu Carnegie Mellon University 740: Computer Architecture Memory Consistency Prof. Onur Mutlu Carnegie Mellon University Readings: Memory Consistency Required Lamport, How to Make a Multiprocessor Computer That Correctly Executes Multiprocess

More information

TRANSACTION-ORIENTED SIMULATION IN AD HOC GRIDS: DESIGN AND EXPERIENCE

TRANSACTION-ORIENTED SIMULATION IN AD HOC GRIDS: DESIGN AND EXPERIENCE TRANSACTION-ORIENTED SIMULATION IN AD HOC GRIDS: DESIGN AND EXPERIENCE Gerald Krafft and Vladimir Getov Harrow School of Computer Science University of Westminster Watford Rd, Northwick Park, Harrow HA1

More information

Distributed Database Management System UNIT-2. Concurrency Control. Transaction ACID rules. MCA 325, Distributed DBMS And Object Oriented Databases

Distributed Database Management System UNIT-2. Concurrency Control. Transaction ACID rules. MCA 325, Distributed DBMS And Object Oriented Databases Distributed Database Management System UNIT-2 Bharati Vidyapeeth s Institute of Computer Applications and Management, New Delhi-63,By Shivendra Goel. U2.1 Concurrency Control Concurrency control is a method

More information

CS 571 Operating Systems. Midterm Review. Angelos Stavrou, George Mason University

CS 571 Operating Systems. Midterm Review. Angelos Stavrou, George Mason University CS 571 Operating Systems Midterm Review Angelos Stavrou, George Mason University Class Midterm: Grading 2 Grading Midterm: 25% Theory Part 60% (1h 30m) Programming Part 40% (1h) Theory Part (Closed Books):

More information

Multiprocessor and Real-Time Scheduling. Chapter 10

Multiprocessor and Real-Time Scheduling. Chapter 10 Multiprocessor and Real-Time Scheduling Chapter 10 1 Roadmap Multiprocessor Scheduling Real-Time Scheduling Linux Scheduling Unix SVR4 Scheduling Windows Scheduling Classifications of Multiprocessor Systems

More information

Automatic Counterflow Pipeline Synthesis

Automatic Counterflow Pipeline Synthesis Automatic Counterflow Pipeline Synthesis Bruce R. Childers, Jack W. Davidson Computer Science Department University of Virginia Charlottesville, Virginia 22901 {brc2m, jwd}@cs.virginia.edu Abstract The

More information

CS6303 Computer Architecture Regulation 2013 BE-Computer Science and Engineering III semester 2 MARKS

CS6303 Computer Architecture Regulation 2013 BE-Computer Science and Engineering III semester 2 MARKS CS6303 Computer Architecture Regulation 2013 BE-Computer Science and Engineering III semester 2 MARKS UNIT-I OVERVIEW & INSTRUCTIONS 1. What are the eight great ideas in computer architecture? The eight

More information

Operating Systems. Deadlock. Lecturer: William Fornaciari Politecnico di Milano Homel.deib.polimi.

Operating Systems. Deadlock. Lecturer: William Fornaciari Politecnico di Milano Homel.deib.polimi. Politecnico di Milano Operating Systems Lecturer: William Fornaciari Politecnico di Milano william.fornaciari@elet.polimi.it Homel.deib.polimi.it/fornacia Summary What is a resource Conditions for deadlock

More information

NOW Handout Page 1. Memory Consistency Model. Background for Debate on Memory Consistency Models. Multiprogrammed Uniprocessor Mem.

NOW Handout Page 1. Memory Consistency Model. Background for Debate on Memory Consistency Models. Multiprogrammed Uniprocessor Mem. Memory Consistency Model Background for Debate on Memory Consistency Models CS 258, Spring 99 David E. Culler Computer Science Division U.C. Berkeley for a SAS specifies constraints on the order in which

More information

Addresses in the source program are generally symbolic. A compiler will typically bind these symbolic addresses to re-locatable addresses.

Addresses in the source program are generally symbolic. A compiler will typically bind these symbolic addresses to re-locatable addresses. 1 Memory Management Address Binding The normal procedures is to select one of the processes in the input queue and to load that process into memory. As the process executed, it accesses instructions and

More information

IP LOOK-UP WITH TIME OR MEMORY GUARANTEE AND LOW UPDATE TIME 1

IP LOOK-UP WITH TIME OR MEMORY GUARANTEE AND LOW UPDATE TIME 1 2005 IEEE International Symposium on Signal Processing and Information Technology IP LOOK-UP WITH TIME OR MEMORY GUARANTEE AND LOW UPDATE TIME 1 G.T. Kousiouris and D.N. Serpanos Dept. of Electrical and

More information

Web-based Energy-efficient Cache Invalidation in Wireless Mobile Environment

Web-based Energy-efficient Cache Invalidation in Wireless Mobile Environment Web-based Energy-efficient Cache Invalidation in Wireless Mobile Environment Y.-K. Chang, M.-H. Hong, and Y.-W. Ting Dept. of Computer Science & Information Engineering, National Cheng Kung University

More information

Applying March Tests to K-Way Set-Associative Cache Memories

Applying March Tests to K-Way Set-Associative Cache Memories 13th European Test Symposium Applying March Tests to K-Way Set-Associative Cache Memories Simone Alpe, Stefano Di Carlo, Paolo Prinetto, Alessandro Savino Politecnico di Torino, Dep. of Control and Computer

More information

Improving VoD System Efficiency with Multicast and Caching

Improving VoD System Efficiency with Multicast and Caching Improving VoD System Efficiency with Multicast and Caching Jack Yiu-bun Lee Department of Information Engineering The Chinese University of Hong Kong Contents 1. Introduction 2. Previous Works 3. UVoD

More information

Lecture Topics. Announcements. Today: Advanced Scheduling (Stallings, chapter ) Next: Deadlock (Stallings, chapter

Lecture Topics. Announcements. Today: Advanced Scheduling (Stallings, chapter ) Next: Deadlock (Stallings, chapter Lecture Topics Today: Advanced Scheduling (Stallings, chapter 10.1-10.4) Next: Deadlock (Stallings, chapter 6.1-6.6) 1 Announcements Exam #2 returned today Self-Study Exercise #10 Project #8 (due 11/16)

More information

Page 1 FAULT TOLERANT SYSTEMS. Coordinated Checkpointing. Time-Based Synchronization. A Coordinated Checkpointing Algorithm

Page 1 FAULT TOLERANT SYSTEMS. Coordinated Checkpointing. Time-Based Synchronization. A Coordinated Checkpointing Algorithm FAULT TOLERANT SYSTEMS Coordinated http://www.ecs.umass.edu/ece/koren/faulttolerantsystems Chapter 6 II Uncoordinated checkpointing may lead to domino effect or to livelock Example: l P wants to take a

More information

Fall 2012 Parallel Computer Architecture Lecture 16: Speculation II. Prof. Onur Mutlu Carnegie Mellon University 10/12/2012

Fall 2012 Parallel Computer Architecture Lecture 16: Speculation II. Prof. Onur Mutlu Carnegie Mellon University 10/12/2012 18-742 Fall 2012 Parallel Computer Architecture Lecture 16: Speculation II Prof. Onur Mutlu Carnegie Mellon University 10/12/2012 Past Due: Review Assignments Was Due: Tuesday, October 9, 11:59pm. Sohi

More information

MANAGING EVENT TRACES FOR A WEB FRONT-END TO A PARALLEL SIMULATION

MANAGING EVENT TRACES FOR A WEB FRONT-END TO A PARALLEL SIMULATION Proceedings of the 2001 Winter Simulation Conference B. A. Peters, J. S. Smith, D. J. Medeiros, and M. W. Rohrer, eds. MANAGING EVENT TRACES FOR A WEB FRONT-END TO A PARALLEL SIMULATION Boon Ping Gan Li

More information

Distributed Scheduling for the Sombrero Single Address Space Distributed Operating System

Distributed Scheduling for the Sombrero Single Address Space Distributed Operating System Distributed Scheduling for the Sombrero Single Address Space Distributed Operating System Donald S. Miller Department of Computer Science and Engineering Arizona State University Tempe, AZ, USA Alan C.

More information

Performance Predictions for Speculative, Synchronous, VLSI Logic Simulation. Bradley L. Noble J. Cris Wade Roger D. Chamberlain

Performance Predictions for Speculative, Synchronous, VLSI Logic Simulation. Bradley L. Noble J. Cris Wade Roger D. Chamberlain Performance Predictions for Speculative, Synchronous, VLSI Logic Simulation Bradley L. Noble J. Cris Wade Roger D. Chamberlain Bradley L. Noble, J. Cris Wade, and Roger D. Chamberlain, Performance Predictions

More information

Principles of Parallel Algorithm Design: Concurrency and Mapping

Principles of Parallel Algorithm Design: Concurrency and Mapping Principles of Parallel Algorithm Design: Concurrency and Mapping John Mellor-Crummey Department of Computer Science Rice University johnmc@rice.edu COMP 422/534 Lecture 3 17 January 2017 Last Thursday

More information

Proceedings of the 2015 Winter Simulation Conference L. Yilmaz, W. K. V. Chan, I. Moon, T. M. K. Roeder, C. Macal, and M. D. Rossetti, eds.

Proceedings of the 2015 Winter Simulation Conference L. Yilmaz, W. K. V. Chan, I. Moon, T. M. K. Roeder, C. Macal, and M. D. Rossetti, eds. Proceedings of the 2015 Winter Simulation Conference L. Yilmaz, W. K. V. Chan, I. Moon, T. M. K. Roeder, C. Macal, and M. D. Rossetti, eds. PARALLEL AND DISTRIBUTED SIMULATION Richard Fujimoto School of

More information

Portland State University ECE 588/688. Dataflow Architectures

Portland State University ECE 588/688. Dataflow Architectures Portland State University ECE 588/688 Dataflow Architectures Copyright by Alaa Alameldeen and Haitham Akkary 2018 Hazards in von Neumann Architectures Pipeline hazards limit performance Structural hazards

More information

ON THE SCALABILITY AND DYNAMIC LOAD BALANCING OF PARALLEL VERILOG SIMULATIONS. Sina Meraji Wei Zhang Carl Tropper

ON THE SCALABILITY AND DYNAMIC LOAD BALANCING OF PARALLEL VERILOG SIMULATIONS. Sina Meraji Wei Zhang Carl Tropper Proceedings of the 2009 Winter Simulation Conference M. D. Rossetti, R. R. Hill, B. Johansson, A. Dunkin, and R. G. Ingalls, eds. ON THE SCALABILITY AND DYNAMIC LOAD BALANCING OF PARALLEL VERILOG SIMULATIONS

More information

Distributed simulation with MPI in ns-3. Josh Pelkey Dr. George Riley

Distributed simulation with MPI in ns-3. Josh Pelkey Dr. George Riley Distributed simulation with MPI in ns-3 Josh Pelkey Dr. George Riley Overview Parallel and distributed discrete event simulation [1] Allows single simulation program to run on multiple interconnected processors

More information