Repast HPC with optimistic time management

High performance computing (HPC) has great potential to speedup agent-based simulations. In parallel and distributed simulation (PADS) community, a well-known fact is that employing an optimistic time management mechanism instead of a conservative time management mechanism may provide remarkable performance enhancement, because optimistic approach avoids redundant synchronization among logical processes (LPs). In this paper, an existing optimistic time management mechanism, namely Time Warp, by Jefferson, is adapted for a distributed agent based simulation tool. We implemented Time Warp on an open source and distributed agent based modeling and simulation (ABMS) tool, namely Repast for High Performance Computing (Repast HPC), from Argonne National Laboratory, Chicago, IL, USA. We incorporated a simple and self-adaptive technique for adjusting checkpoint intervals. Two case studies have been implemented to compare our optimistic approach and existing Repast HPC's conservative approach. The experiments suggest that optimistic approach is more scalable than conservative approach in agent based simulations.

[1]  Joshua M. Epstein,et al.  Modeling civil violence: An agent-based computational approach , 2002, Proceedings of the National Academy of Sciences of the United States of America.

[2]  Carl Tropper,et al.  On Rolling Back and Checkpointing in Time Warp , 2001, IEEE Trans. Parallel Distributed Syst..

[3]  Nazim Fatès,et al.  A Guided Tour of Asynchronous Cellular Automata , 2013, J. Cell. Autom..

[4]  Peter L. Reiher,et al.  A time warp implementation of Sharks World , 1990, 1990 Winter Simulation Conference Proceedings.

[5]  Raj Jain,et al.  Analysis of the Increase and Decrease Algorithms for Congestion Avoidance in Computer Networks , 1989, Comput. Networks.

[6]  Michael J. North,et al.  Tutorial on agent-based modelling and simulation , 2005, Proceedings of the Winter Simulation Conference, 2005..

[7]  Philip A. Wilsey,et al.  An analytical comparison of periodic checkpointing and incremental state saving , 1993, PADS '93.

[8]  Andreas Tolk,et al.  Engineering Principles of Combat Modeling and Distributed Simulation: Tolk/Engineering Principles Combat Modeling , 2012 .

[9]  Christopher D. Carothers,et al.  Scalable Time Warp on Blue Gene Supercomputers , 2009, 2009 ACM/IEEE/SCS 23rd Workshop on Principles of Advanced and Distributed Simulation.

[10]  David R. Jefferson,et al.  Virtual time , 1985, ICPP.

[11]  Francesco Quaglia.  A Cost Model for Selecting Checkpoint Positions in Time Warp Parallel Simulation , 2001, IEEE Trans. Parallel Distributed Syst..

[12]  Richard M. Fujimoto,et al.  Parallel and Distribution Simulation Systems , 1999 .

[13]  Maria Hybinette,et al.  Quantitative assessment of an agent-based simulation on a time warp executive , 2008, 2008 Winter Simulation Conference.

[14]  Ingo J. Timm,et al.  A Hybrid Time Management Approach to Agent-Based Simulation , 2006, KI.

[15]  Yi-Bing Lin,et al.  Selecting the checkpoint interval in time warp simulation , 1993, PADS '93.

[16]  Christopher D. Carothers,et al.  Warp speed: executing time warp on 1,966,080 cores , 2013, SIGSIM-PADS.

[17]  Philip A. Wilsey,et al.  Comparative analysis of periodic state saving techniques in time warp simulators , 1995, PADS.

[18]  Rassul Ayani,et al.  Adaptive checkpointing in Time Warp , 1994, PADS '94.

[19]  Kalyan S. Perumalla.  Scaling time warp-based discrete event execution to 104 processors on a Blue Gene supercomputer , 2007, CF '07.

[20]  G. Amdhal,et al.  Validity of the single processor approach to achieving large scale computing capabilities , 1967, AFIPS '67 (Spring).

[21]  Michael J. North,et al.  Parallel agent-based simulation with Repast for High Performance Computing , 2013, Simul..