Performance scalability of decoupled software pipelining
Performance scalability of decoupled software pipelining
复制标题
解耦软件流水线的性能可扩展性
DOI：
10.1145/1400112.1400113
复制
发表时间：
2008
期刊：
影响因子：
--
通讯作者：
David I. August
中科院分区：
文献类型：
--
作者：
Ram Rangan;Neil Vachharajani;Guilherme Ottoni;David I. August
任何成功的解决方案用于使用多核心处理器来扩展通用计划的性能，都必须与核心间通信成本上升，同时暴露粗粒的并行性。最近提出的管道多线程（PMT）技术已被证明具有通用的适用性，并且还能够通过管道的线程交流通信有效地耐受核心间潜伏期。这些理想的属性使PMT技术在当前和未来的多项处理器上进行计划并行化的强大候选者，并且了解其性能特征对于他们的部署至关重要。为此，本文评估了称为解耦的软件管道（DSWP）的通用PMT技术的性能可伸缩性，并对必须克服的通信瓶颈进行了彻底的分析，以获得最佳的DSWP可伸缩性。
Any successful solution to using multicore processors to scale general-purpose program performance will have to contend with rising intercore communication costs while exposing coarse-grained parallelism. Recently proposed pipelined multithreading (PMT) techniques have been demonstrated to have general-purpose applicability and are also able to effectively tolerate inter-core latencies through pipelined interthread communication. These desirable properties make PMT techniques strong candidates for program parallelization on current and future multicore processors and understanding their performance characteristics is critical to their deployment. To that end, this paper evaluates the performance scalability of a general-purpose PMT technique called decoupled software pipelining (DSWP) and presents a thorough analysis of the communication bottlenecks that must be overcome for optimal DSWP scalability.