提高SMT芯片中主线程性能的两种方法

2007 IFIP International Conference on Network and Parallel Computing Workshops (NPC 2007) Pub Date : 2007-09-18 DOI:10.1109/NPC.2007.45

Ma Pengyong, Chen Shuming, Hu Xiao

{"title":"提高SMT芯片中主线程性能的两种方法","authors":"Ma Pengyong, Chen Shuming, Hu Xiao","doi":"10.1109/NPC.2007.45","DOIUrl":null,"url":null,"abstract":"In SMT processor, every thread's executing time will be enlarged due to sharing resource by several threads. It will result in task lost in real time system. In this paper, we present two methods to alleviate the cache conflict in multithread chip, one is protecting the last instructions of master thread, and the other is locking loop. Both methods only need several additional registers. Simulations show that not only the performance of master thread is improved by 14%, but also the IPC of all threads is improved about 6%.","PeriodicalId":278518,"journal":{"name":"2007 IFIP International Conference on Network and Parallel Computing Workshops (NPC 2007)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2007-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Two methods to enhance the master thread's performance in SMT Chip\",\"authors\":\"Ma Pengyong, Chen Shuming, Hu Xiao\",\"doi\":\"10.1109/NPC.2007.45\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"In SMT processor, every thread's executing time will be enlarged due to sharing resource by several threads. It will result in task lost in real time system. In this paper, we present two methods to alleviate the cache conflict in multithread chip, one is protecting the last instructions of master thread, and the other is locking loop. Both methods only need several additional registers. Simulations show that not only the performance of master thread is improved by 14%, but also the IPC of all threads is improved about 6%.\",\"PeriodicalId\":278518,\"journal\":{\"name\":\"2007 IFIP International Conference on Network and Parallel Computing Workshops (NPC 2007)\",\"volume\":\"6 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2007-09-18\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2007 IFIP International Conference on Network and Parallel Computing Workshops (NPC 2007)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/NPC.2007.45\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2007 IFIP International Conference on Network and Parallel Computing Workshops (NPC 2007)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/NPC.2007.45","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}

引用次数: 0

摘要

在SMT处理器中，由于多个线程共享资源，每个线程的执行时间都会延长。这将导致实时系统中的任务丢失。本文提出了两种缓解多线程芯片缓存冲突的方法，一种是保护主线程的最后一条指令，另一种是锁定循环。这两种方法只需要几个额外的寄存器。仿真结果表明，该方法不仅使主线程的性能提高了14%，而且使所有线程的IPC提高了约6%。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

查看原文本刊更多论文

Two methods to enhance the master thread's performance in SMT Chip

In SMT processor, every thread's executing time will be enlarged due to sharing resource by several threads. It will result in task lost in real time system. In this paper, we present two methods to alleviate the cache conflict in multithread chip, one is protecting the last instructions of master thread, and the other is locking loop. Both methods only need several additional registers. Simulations show that not only the performance of master thread is improved by 14%, but also the IPC of all threads is improved about 6%.

求助全文

通过发布文献求助，成功后即可免费获取论文全文。去求助

来源期刊

2007 IFIP International Conference on Network and Parallel Computing Workshops (NPC 2007)

自引率

0.00%

发文量