Using Access Patterns to Analyze the Performance of Optimistic Synchronization Algorithms in Simulations of MAS

Lees, Michael; Logan, Brian; Theodoropoulos, Georgios

doi:10.1177/0037549708096691

Cited by 11 publications

(5 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…When independence between the agent updates cannot be guaranteed, the results are discarded. Lees et al studied the effects of access patterns to shared state variables on the performance of optimistic simulation algorithms [29]. Some previous works consider accelerating time-driven agent-based simulations by identifying independent state updates among agents: Scheutz et al [21,41,42] apply a translation function that reflects the furthest possible amount of movement of an agent to determine an event horizon in model time.…”

Section: Exploiting Independent State Updatesmentioning

confidence: 99%

Fidelity and Performance of State Fast-forwarding in Microscopic Traffic Simulations

Andelfinger

Xu²,

Eckhoff

et al. 2020

ACM Trans. Model. Comput. Simul.

View full text Add to dashboard Cite

Common car-following models for microscopic traffic simulation assume a time advancement using fixed-sized time steps. However, a purely time-driven execution is inefficient when the states of some agents are independent of other agents and thus predictable far into the simulated future. We propose a method to accelerate microscopic traffic simulations based on identifying independence among agent state updates. Instead of iteratively updating an agent’s state throughout a sequence of time steps, a computationally inexpensive “fast-forward” function advances the agent’s state to the time of its earliest possible interaction with other agents. We present an algorithm to determine independence intervals in microscopic traffic simulations and derive fast-forward functions for several well-known traffic models. In contrast to existing approaches based on reducing the level of detail, our approach retains the microscopic nature of the simulation. An evaluation is performed for a synthetic scenario and on the road network of Singapore. At low traffic densities, maximum speedup factors of about 2.6 and 1.6 are achieved, while at the highest considered densities, only few opportunities for fast-forwarding exist. We show that the deviation from purely time-driven execution is reduced to a minimum when choosing an adequate numerical integration scheme to execute the time-driven updates. Verification results show that the overall deviation in vehicle travel times is marginal.

show abstract

Section: Exploiting Independent State Updatesmentioning

confidence: 99%

Fidelity and Performance of State Fast-forwarding in Microscopic Traffic Simulations

Andelfinger

Xu²,

Eckhoff

et al. 2020

ACM Trans. Model. Comput. Simul.

View full text Add to dashboard Cite

show abstract

“…As the results were averaged over 3 runs with different pseudo random number generator seeds, the standard deviation for each result in the graphs is depicted using error-bars. The order in which the events, both reads and writes, are processed by PDES-MAS is important in that an ALP commits a rollback if it arrives with a timestamp earlier that its local time (straggler-events), committing the event otherwise [Logan and Theodoropoulos 2001;Lees et al 2008]. A committed rollback also has the potential to regenerate Range-Queries.…”

Section: Rollback Volatilitymentioning

confidence: 99%

Synchronised Range Queries in Distributed Simulations of Multi-agent Systems

Suryanarayanan¹,

Craenen²,

Theodoropoulos³

2010

2010 IEEE/ACM 14th International Symposium on Distributed Simulation and Real Time Applications

View full text Add to dashboard Cite

Peer-reviewedCitation for published item: Suryanarayanan, V. and Theodoropoulos, G. (2013) 'Synchronised range queries in distributed simulations of multi-agent systems.', ACM transactions on modeling and computer simulation., 23 (4). p. 25.Further information on publisher's website:http://dx.doi.org/10.1145/2517449Publisher's copyright statement:Additional information: Use policyThe full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for personal research or study, educational, or not-for-prot purposes provided that:• a full bibliographic reference is made to the original source • a link is made to the metadata record in DRO • the full-text is not changed in any way The full-text must not be sold in any format or medium without the formal permission of the copyright holders.Please consult the full DRO policy for further details. Range-Queries are an increasingly important associative form of data access encountered in different computational environments including Peer to Peer systems, Wireless communications, Database systems, Distributed Virtual Environments and, more recently, distributed simulations. In this paper, we present and evaluate a system for performing logical-time synchronised Range-Queries over data in the context of distributed simulations of Multi-Agent Systems (MAS). The paper presents algorithms performing instantaneous Queries within an optimistic synchronisation framework and in the presence of dynamic migration of the simulation state. A quantitative evaluation of the effectiveness of the proposed algorithms under different conditions is also presented.

show abstract

“…The only option is to go massively parallel due to availability of cheap parallel local hardware with many cores, or cloud services like Amazon EC. This trend has been already recognised in the field of ABS as a research challenge for Large-scale ABMS (Macal 2016) was called out and as a substantial body of research for parallel ABS shows Logan and Theodoropoulos 2001;Lees et al 2008;Riley et al 2003;Gasser and Kakugawa 2002;Himmelspach and Uhrmacher 2007;Minson and Theodoropoulos 2008;Gorur et al 2016;Hay and Wilsey 2015;Abar et al 2017;Cicirelli et al 2015).…”

Section: Introductionmentioning

confidence: 99%

“…In this body of work it has been established that parallelisation of autonomous agents, situated in some spacial, metric environment can be particularly challenging. The reason for this is that the environment constitutes a key medium for the agents interactions, represented as a passive data structure, recording attributes of the environment and the agents (Lees et al 2008). Thus, the problem of parallelising ABS boils down to the problem of how to synchronise access to shared state without violating the causality principle and resource constraints (Logan and Theodoropoulos 2001;).…”

Section: Introductionmentioning

confidence: 99%

“…Adopting PDES to ABS is challenging as agents are autonomous, which means that the topology can change in every step, making it hard to predict the topology of logical processes in advance (Lees et al 2008), posing a difficult problem for parallelisation in general (Cicirelli et al 2015). The work discusses this challenge by giving a detailed and in-depth discussion of the internals and implementation of their powerful and highly complex PDES-MAS system.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A tale of lock-free agents: towards Software Transactional Memory in parallel Agent-Based Simulation

Thaler

Siebers

2019

Complex Adapt Syst Model

View full text Add to dashboard Cite

With the decline of Moore's law and the ever increasing availability of cheap massively parallel hardware, it becomes more and more important to embrace parallel programming methods to implement Agent-Based Simulations (ABS). This has been acknowledged in the field a while ago and numerous research on distributed parallel ABS exists, focusing primarily on Parallel Discrete Event Simulation as the underlying mechanism. However, these concepts and tools are inherently difficult to master and apply and often an excess in case implementers simply want to parallelise their own, custom agent-based model implementation. However, with the established programming languages in the field, Python, Java and C++, it is not easy to address the complexities of parallel programming due to unrestricted side effects and the intricacies of low-level locking semantics. Therefore, in this paper we propose the use of a lock-free approach to parallel ABS using Software Transactional Memory (STM) in conjunction with the pure functional programming language Haskell, which in combination, removes some of the problems and complexities of parallel implementations in imperative approaches. We present two case studies, in which we compare the performance of lock-based and lock-free STM implementations in two different well known Agent-Based Models, where we investigate both the scaling performance under increasing number of CPU cores and the scaling performance under increasing number of agents. We show that the lock-free STM implementations consistently outperform the lock-based ones and scale much better to increasing number of CPU cores both on local hardware and on Amazon EC. Further, by utilizing the pure functional language Haskell we gain the benefits of immutable data and lack of unrestricted side effects guaranteed at compile-time, making validation easier and leading to increased confidence in the correctness of an implementation, something of fundamental importance and benefit in parallel programming in general and scientific computing like ABS in particular.

show abstract

Using Access Patterns to Analyze the Performance of Optimistic Synchronization Algorithms in Simulations of MAS

Cited by 11 publications

References 27 publications

Fidelity and Performance of State Fast-forwarding in Microscopic Traffic Simulations

Fidelity and Performance of State Fast-forwarding in Microscopic Traffic Simulations

Synchronised Range Queries in Distributed Simulations of Multi-agent Systems

A tale of lock-free agents: towards Software Transactional Memory in parallel Agent-Based Simulation

Contact Info

Product

Resources

About