{"title":"A Precise Interrupts Mechanism Based on Micro-Operation Tracing of Instruction Boundary for Embedded Processor","authors":"Jun Zhang, Xiaoming Fan, Song-He Liu","doi":"10.1109/NAS.2008.25","DOIUrl":null,"url":null,"abstract":"Precise interrupts is a key technique of embedded processor, for assurance of properly executing and state resuming of the whole system. As interrupt occurs on the instruction boundary, pipeline flushing, interrupt transfer micro-program and interrupt handling routine are executed after the committing soon instruction finishes. This process spends lots of cycles to pre-fetch and decode instructions that will never be executed, and reduces the real time performance. Combined with the CISC processor execution characteristic of micro-operation, this paper proposes a precise interrupts mechanism based on Instruction Boundary Micro-operation Tracing, called IBMT. This technique inspects instruction boundary and interrupt window every clock cycle, and starts up the pipeline flushing, pre-fetch, interrupt transfer micro-program and interrupt handling routine in advance. As a result, 39.34% of the clock cycles can be saved at every interrupt acknowledgement.","PeriodicalId":153238,"journal":{"name":"2008 International Conference on Networking, Architecture, and Storage","volume":"33 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2008-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2008 International Conference on Networking, Architecture, and Storage","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/NAS.2008.25","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 2
Abstract
Precise interrupts is a key technique of embedded processor, for assurance of properly executing and state resuming of the whole system. As interrupt occurs on the instruction boundary, pipeline flushing, interrupt transfer micro-program and interrupt handling routine are executed after the committing soon instruction finishes. This process spends lots of cycles to pre-fetch and decode instructions that will never be executed, and reduces the real time performance. Combined with the CISC processor execution characteristic of micro-operation, this paper proposes a precise interrupts mechanism based on Instruction Boundary Micro-operation Tracing, called IBMT. This technique inspects instruction boundary and interrupt window every clock cycle, and starts up the pipeline flushing, pre-fetch, interrupt transfer micro-program and interrupt handling routine in advance. As a result, 39.34% of the clock cycles can be saved at every interrupt acknowledgement.