Huong Thien Hoang, P. T. Vo, Y. T. Vo, Liem Tan Pham, N. Otsuki, M. Ito, O. Nishii
{"title":"在多核系统中减少中断处理开销的8处理器8640 MIPS SoC的设计和性能评估","authors":"Huong Thien Hoang, P. T. Vo, Y. T. Vo, Liem Tan Pham, N. Otsuki, M. Ito, O. Nishii","doi":"10.1109/ASSCC.2008.4708761","DOIUrl":null,"url":null,"abstract":"We have developed a platform SoC including eight SuperH processor cores for high performance applications. It achieves 8,640 MIPS at 600 MHz for Dhrystone 2.1. The eight processor cores are divided into two clusters. Each cluster has a snoop controller to maintain cache coherency. The main internal system bus, packet-based split transaction, is 64 bits wide and runs at 300 MHz. As increasing number of processor cores in the system, enhancing overall system performance and optimizing power are important aims and design challenges. In this paper, we introduce one scheme to improve the system performance by reducing overhead of interrupt handling in multi-core system. We have added an automatic-rotating interrupt distribution scheme to processor cores to reduce overhead in handling interrupt requests. As a result, the processing time in Linux kernel is improved by 21% when SPLASH-2 is executed.","PeriodicalId":143173,"journal":{"name":"2008 IEEE Asian Solid-State Circuits Conference","volume":"51 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2008-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":"{\"title\":\"Design and performance evaluation of an 8-processor 8,640 MIPS SoC with overhead reduction of interrupt handling in a multi-core system\",\"authors\":\"Huong Thien Hoang, P. T. Vo, Y. T. Vo, Liem Tan Pham, N. Otsuki, M. Ito, O. Nishii\",\"doi\":\"10.1109/ASSCC.2008.4708761\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"We have developed a platform SoC including eight SuperH processor cores for high performance applications. It achieves 8,640 MIPS at 600 MHz for Dhrystone 2.1. The eight processor cores are divided into two clusters. Each cluster has a snoop controller to maintain cache coherency. The main internal system bus, packet-based split transaction, is 64 bits wide and runs at 300 MHz. As increasing number of processor cores in the system, enhancing overall system performance and optimizing power are important aims and design challenges. In this paper, we introduce one scheme to improve the system performance by reducing overhead of interrupt handling in multi-core system. We have added an automatic-rotating interrupt distribution scheme to processor cores to reduce overhead in handling interrupt requests. As a result, the processing time in Linux kernel is improved by 21% when SPLASH-2 is executed.\",\"PeriodicalId\":143173,\"journal\":{\"name\":\"2008 IEEE Asian Solid-State Circuits Conference\",\"volume\":\"51 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2008-12-12\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"1\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2008 IEEE Asian Solid-State Circuits Conference\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/ASSCC.2008.4708761\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2008 IEEE Asian Solid-State Circuits Conference","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ASSCC.2008.4708761","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Design and performance evaluation of an 8-processor 8,640 MIPS SoC with overhead reduction of interrupt handling in a multi-core system
We have developed a platform SoC including eight SuperH processor cores for high performance applications. It achieves 8,640 MIPS at 600 MHz for Dhrystone 2.1. The eight processor cores are divided into two clusters. Each cluster has a snoop controller to maintain cache coherency. The main internal system bus, packet-based split transaction, is 64 bits wide and runs at 300 MHz. As increasing number of processor cores in the system, enhancing overall system performance and optimizing power are important aims and design challenges. In this paper, we introduce one scheme to improve the system performance by reducing overhead of interrupt handling in multi-core system. We have added an automatic-rotating interrupt distribution scheme to processor cores to reduce overhead in handling interrupt requests. As a result, the processing time in Linux kernel is improved by 21% when SPLASH-2 is executed.