The ROme OpTimistic Simulator: A Tutorial

Proceedings of the 3rd ACM SIGSIM Conference on Principles of Advanced Discrete Simulation

2015

Self Cite

The rollback operation is a fundamental building block to support the correct execution of a speculative Time Warp-based Parallel Discrete Event Simulation. In the literature, several solutions to reduce the execution cost of this operation have been proposed, either based on the creation of a checkpoint of previous simulation state images, or on the execution of negative copies of simulation events which are able to undo the updates on the state. In this paper, we explore the practical design and implementation of a state recoverability technique which allows to restore a previous simulation state either relying on checkpointing or on the reverse execution of the state updates occurred while processing events in forward mode. Differently from other proposals, we address the issue of executing backward updates in a fully-transparent and event granularity-independent way, by relying on static software instrumentation (targeting the x86 architecture and Linux systems) to generate at runtime reverse update code blocks (not to be confused with reverse events, proper of the reverse computing approach). These are able to undo the effects of a forward execution while minimizing the cost of the undo operation. We also present experimental results related to our implementation, which is released as free software and fully integrated into the open source ROOT-Sim (ROme OpTimistic Simulator) package. The experimental data support the viability and effectiveness of our proposal

Section: Test-bed Environmentmentioning

confidence: 99%

“…We again remark that our proposal is fully application transparent, and has been released as free software and integrated into the open source ROOT-Sim package [18,20].…”

Section: Introductionmentioning

confidence: 98%

Transparently Mixing Undo Logs and Software Reversibility for State Recovery in Optimistic PDES

Cingolani

Proceedings of the 3rd ACM SIGSIM Conference on Principles of Advanced Discrete Simulation

2015

Self Cite

“…As hinted, we have integrated all our solutions into the ROme OpTimistic Simulator (ROOT-Sim) [24,22,17], an optimistic simulation platform based on the Time Warp protocol [18] and tailored for UNIX-like systems. ROOTSim is designed to be a general-purpose platform, that can support any kind of discrete event model developed by adhering to a very simple and intuitive programming model (in fact, the platform transparently handles all the mechanisms associated with parallelization and synchronization).…”

Section: Experimental Studymentioning

confidence: 99%

NUMA Time Warp

Proceedings of the 3rd ACM SIGSIM Conference on Principles of Advanced Discrete Simulation

2015

Self Cite

It is well known that Time Warp may suffer from large usage of memory, which may hamper the efficiency of the memory hierarchy. To cope with this issue, several approaches have been devised, mostly based on the reduction of the amount of used virtual memory, e.g., by the avoidance of checkpointing and the exploitation of reverse computing. In this article we present an orthogonal solution aimed at optimizing the latency for memory access operations when running Time Warp systems on Non-Uniform Memory Access (NUMA) multi-processor/multi-core computing systems. More in detail, we provide an innovative Linux-based architecture allowing per simulation-object management of memory segments made up by disjoint sets of pages, and supporting both static and dynamic binding of the memory pages reserved for an individual object to the different NUMA nodes, depending on what worker thread is in charge of running that simulation object along a given wall-clock-time window. Our proposal not only manages the virtual pages used for the live state image of the simulation object, rather, it also copes with memory pages destined to keep the simulation object's event buffers and any recoverability data. Further, the architecture allows memory access optimization for data (messages) exchanged across the different simulation objects running on the NUMA machine. Our proposal is fully transparent to the application code, thus operating in a seamless manner. Also, a free software release of our NUMA memory manager for Time Warp has been made available within the open source ROOT-Sim simulation platform. Experimental data for an assessment of our innovative proposal are also provided in this article

“…We integrated ECS within the open source ROOT-Sim package [26], particularly the symmetric multi-threaded version presented in [4]. A few relevant modifications to this simulation platform have been made for integration purposes.…”

Section: Test-bed Platformmentioning

confidence: 99%

Cross-state events: A new approach to parallel discrete event simulation and its speculative runtime support

Journal of Parallel and Distributed Computing

2019

Self Cite

We present a new approach to Parallel Discrete Event Simulation (PDES), where we enable the execution of so-called cross-state events. During their processing, the state of multiple concurrent simulation objects can be accessed in read/write mode, as opposed to classical partitioned accesses. This is done with no pre-declaration of this type of access by the programmer, hence also coping with non-determinism. In our proposal, cross-state events are supported by a speculative runtime environment fully transparently to the application code. This is done through an ad-hoc memory management architecture and an extension of the classical Time Warp synchronization protocol. This extension, named Event and Cross-State (ECS) synchronization, ensures causally-consistent speculative parallel execution of discrete event applications by allowing all events to observe the snapshot of the model execution trajectory that would have been observed in a timestamp-ordered execution of the same model. An experimental assessment of our proposal shows how it can significantly reduce the application development complexity, while also providing advantages in terms of performance.