CSR: Small: Collaborative Research: Combining Static Analysis and Dynamic Run-time Optimization for Parallel Discrete Event Simulation in Many-Core Environments

点击翻译按钮获取中文摘要

Proposal Title: CSR:Small:Collaborative Research: Combining Static Analysis andDynamic Run-time Optimization for Parallel Discrete EventSimulation in Many-Core EnvironmentsInstitution: SUNY at BinghamtonAbstract Date: 07/06/09This project investigates how a new processor paradigm (multi-core architectures)changes the way Parallel Discrete Event Simulation (PDES) is done. This topic isimportant given the wide use of simulation and the emergence of multi-corearchitectures. PDES is likely to play an increasingly important role in discrete eventsimulation as Moore?s Law is sharply curtailed and explicit parallelism becomes themajor avenue for improving performance of sequential applications. Improving PDESperformance translates to improved.Discrete Event Simulation (DES) is widely used for performance evaluation in manyapplication domains. The fine grained nature of PDES causes its performance andscalability to be limited by communication latency. The emergence of multi-corearchitectures and their expected evolution into manycore systems offers potential reliefto PDES and other fine grained parallel applications because the cost of communicationwithin a chip is dramatically lower than conventional networked communication. Absentthis dominant effect, PDES performance will be determined by issues such as loadbalancing, synchronization and optimism control, and the choice and configuration ofvarious other algorithms and data structures of the simulator. Operation in a manycoreenvironment introduces new system tradeoffs that must be effectively balanced by thesystem software. Primarily, the pressure on the memory system and resilience to loadfluctuations will emerge as critical issues that we address in the proposed research.Finally, the more predictable nature of communication cost in this environment (due inpart to the more frequent synchronization possible between nearby cores) can beexploited, especially by static analysis, for effective simulation.As multi-cores become the default microprocessor architecture, applications that areperformance constrained must evolve to use parallelism to take advantage of theresources available on the cores. This project?s new PDES can have a significantimpact on a number of applications that rely on discrete event simulations. The PIs planto incorporate the research results into a graduate level course on parallel simulationtechniques and to involve undergraduate students in the project.