The local Time Warp approach to parallel simulation

Rajaei, Hassan; Ayani, Rassul; Thorelli, Lars-Erik

doi:10.1145/158459.158474

Cited by 65 publications

(15 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As a result, combined synchronization is widely used in general-purpose parallel simulation frameworks and languages such as Maisie [1], the High Level Architecture [8], and μsik [16]. Although those frameworks support dynamic switching between conservative and optimistic schemes, each scheme is applied to a whole LP or even a group of LPs [22]. In contrast, our approach dynamically decides for each event individually whether conservative or optimistic execution is the most favorable option to maximize performance.…”

Section: Related Workmentioning

confidence: 99%

Know Thy Simulation Model: Analyzing Event Interactions for Probabilistic Synchronization in Parallel Simulations

Kunz¹,

Stoffers²,

Gross³

et al. 2012

Proceedings of the Fifth International Conference on Simulation Tools and Techniques

View full text Add to dashboard Cite

Efficiently scheduling and synchronizing parallel event execution constitutes the fundamental challenge in parallel discrete event simulation. Existing synchronization algorithms typically do not analyze event interactions within the simulation model -mainly to minimize runtime overhead and complexity. However, we argue that disregarding event interactions results in a lack of insight into the behavior of the simulation model, thereby severely limiting synchronization efficiency and thus parallel performance. In this paper, we present a probabilistic synchronization scheme that obtains extensive knowledge of the simulation behavior at runtime to guide event execution. Specifically, we design three heuristics that dynamically derive event dependencies from tracing event interactions and decide whether or not to speculatively execute events. Our evaluation shows that the proposed probabilistic synchronization scheme considerably outperforms traditional conservative and optimistic schemes.

show abstract

Section: Related Workmentioning

confidence: 99%

Know Thy Simulation Model: Analyzing Event Interactions for Probabilistic Synchronization in Parallel Simulations

Kunz¹,

Stoffers²,

Gross³

et al. 2012

Proceedings of the Fifth International Conference on Simulation Tools and Techniques

View full text Add to dashboard Cite

show abstract

“…One of the first approaches in this area was Local Time Warp by Rajaei et al [20]. Local Time Warp joins multiple LPs into a cluster and performs optimistic synchronization inside each cluster, limited by a time window conservatively negotiated among the clusters.…”

Section: Shared-memory and Multi-layer Pdesmentioning

confidence: 99%

Large-Scale Network Simulation: Leveraging the Strengths of Modern SMP-based Compute Clusters

Stoffers

Schmerling

Kunz

et al. 2014

Proceedings of the Seventh International Conference on Simulation Tools and Techniques

View full text Add to dashboard Cite

Parallelization is crucial for efficient execution of large-scale network simulation. Today's computing clusters commonly used for that purpose are built from a large amount of multiprocessor machines. The traditional approach to utilize all CPU cores in such a system is to partition the network and distribute the partitions to the cores. This, however, does not incorporate the presence of shared memory into the design, such that messages between partitions on the same computing node have to be serialized and synchronization becomes more complex. In this paper, we present an approach that combines the shared-memory parallelization scheme Horizon [9] with the standard approach to distributed simulation to leverage the strengths of today's computing clusters. To further reduce the synchronization overhead, we introduce a novel synchronization algorithm that takes domain knowledge into account to reduce the number of synchronization points. In a case study with a UMTS LTE model, we show that both contributions combined enable much higher scalability achieving almost linear speedup when simulating 1,536 LTE cells on 1,536 CPU cores.

show abstract

“…The Local Time Warp (LTW) [28] approach combines two simulation protocols by using the optimistic simulation between LPs belonging to the same cluster and by maintaining a conservative protocol between clusters. LTW minimizes the impact of any rollback to the LPs in a given cluster.…”

Section: Problem Decompositionmentioning

confidence: 99%