An Evolutionary Algorithm to Optimize Log/Restore Operations within Optimistic Simulation Platforms

Abstract

In this work we address state recoverability in advanced optimistic simulation systems by proposing an evolutionary algorithm to optimize at run-time the parameters associated with state log/restore activities. Optimization takes place by adaptively selecting for each simulation object both (i) the best suited log mode (incremental vs non-incremental) and (ii) the corresponding optimal value of the log interval. Our performance optimization approach allows to indirectly cope with hidden effects (e.g., locality) as well as cross-object effects due to the variation of log/restore parameters for different simulation objects (e.g., rollback thrashing). Both of them are not captured by literature solutions based on analytical models of the overhead associated with log/restore tasks. More in detail, our evolutionary algorithm dynamically adjusts the log/restore parameters of distinct simulation objects as a whole, towards a well suited configuration. In such a way, we prevent negative effects on performance due to the biasing of the optimization towards individual simulation objects, which may cause reduced gains (or even decrease) in performance just due to the aforementioned hidden and/or cross-object phenomena. We also present an application-transparent implementation of the evolutionary algorithm within the ROme OpTimistic Simulator (ROOT-Sim), namely an open source, general purpose simulation environment designed according to the optimistic synchronization paradigm

The ROme OpTimistic Simulator: A Tutorial

“…In particular, as mentioned, the simulation state can be scattered across different segments of allocated memory, and log/restore operations can be carried out either in nonincremental [23] or incremental way [24]. Additionally, ROOT-Sim can be configured to switch between these two differentiated modes autonomically and to optimize the checkpointing interval, either by relying on an analytic model [25], or by relying on a genetic algorithm [26].…”

A Study on the Parallelization of Terrain-Covering Ant Robots Simulations

“…It features several AI algorithms for supporting agents' decision making, provides python binding (which is a benefit for inexperienced programmers), and is complemented by Cassandra, a visualization tool created to detect spatial-temporal patterns generated by the simulation for post-analysis. On the other hand, we use ROOT-Sim [8,9], a general-purpose PDES simulation framework, to support agent-based simulation. We do not rely on specific agent-based facilities, rather the simulation model is written in plain ANSI-C, giving the modeler the freedom to implement the logic using the general-purpose libraries.…”

“…In in this article we present a study on the tradeoff between performance and reliability of the simulation outputs when exploiting optimistic PDES techniques, particularly the ROOT-Sim PDES platform [8,9], for the case of simulations of Terrain-Covering Ant Robots (TCAR). More in detail, we study how the frequency of inspection on committed portions of the parallel simulation run impacts both the achievable speedup and the distribution of the estimated simulated-time for fully exploring a target spatial region.…”

Hijacker: Efficient static software instrumentation with applications in high performance computing: Poster paper

“…computing systems which are able to actively react to changes in the execution dynamics due to internal or external solicitations. This is the case, e.g., of [13] or [18], where the coexistence of differently instrumented versions of the same executable is exploited in order to dinamically reselect the best suited execution mode depending on the actual execution dynamics.…”