{"title":"使用签名监控进行处理器错误检测的最佳签名位置","authors":"K. Wilken","doi":"10.1109/FTCS.1991.146681","DOIUrl":null,"url":null,"abstract":"An approach that produces optimal placement of justifying signatures for concurrent processor-error detection using signature monitoring is presented. In this approach, placing justifying signatures on nodes and arcs in a directed program control-flow graph is transformed into placing justifying signatures on edges in an undirected, costed graph. A justifying signature is represented in the costed graph by a deleted edge, and optimal placement is reduced to finding a valid minimum-cost deleted edge set. An equivalent problem is finding this set's maximum-cost complement. For order-independent signature functions, the complement set for optimal placement is shown to be a maximum spanning tree. For cyclic codes, the complement set for optimal placement is a new type of graph, a maximum valuation graph (MVG), which is produced by a new algorithm. Using this algorithm, cyclic codes produce significantly less performance overhead than order-independent functions. Experimental results show that the MVG algorithm yields substantial improvement over previous solutions.<<ETX>>","PeriodicalId":300397,"journal":{"name":"[1991] Digest of Papers. Fault-Tolerant Computing: The Twenty-First International Symposium","volume":"64 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1991-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"11","resultStr":"{\"title\":\"Optimal signature placement for processor-error detection using signature monitoring\",\"authors\":\"K. Wilken\",\"doi\":\"10.1109/FTCS.1991.146681\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"An approach that produces optimal placement of justifying signatures for concurrent processor-error detection using signature monitoring is presented. In this approach, placing justifying signatures on nodes and arcs in a directed program control-flow graph is transformed into placing justifying signatures on edges in an undirected, costed graph. A justifying signature is represented in the costed graph by a deleted edge, and optimal placement is reduced to finding a valid minimum-cost deleted edge set. An equivalent problem is finding this set's maximum-cost complement. For order-independent signature functions, the complement set for optimal placement is shown to be a maximum spanning tree. For cyclic codes, the complement set for optimal placement is a new type of graph, a maximum valuation graph (MVG), which is produced by a new algorithm. Using this algorithm, cyclic codes produce significantly less performance overhead than order-independent functions. Experimental results show that the MVG algorithm yields substantial improvement over previous solutions.<<ETX>>\",\"PeriodicalId\":300397,\"journal\":{\"name\":\"[1991] Digest of Papers. Fault-Tolerant Computing: The Twenty-First International Symposium\",\"volume\":\"64 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"1991-06-25\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"11\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"[1991] Digest of Papers. Fault-Tolerant Computing: The Twenty-First International Symposium\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/FTCS.1991.146681\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"[1991] Digest of Papers. Fault-Tolerant Computing: The Twenty-First International Symposium","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/FTCS.1991.146681","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Optimal signature placement for processor-error detection using signature monitoring
An approach that produces optimal placement of justifying signatures for concurrent processor-error detection using signature monitoring is presented. In this approach, placing justifying signatures on nodes and arcs in a directed program control-flow graph is transformed into placing justifying signatures on edges in an undirected, costed graph. A justifying signature is represented in the costed graph by a deleted edge, and optimal placement is reduced to finding a valid minimum-cost deleted edge set. An equivalent problem is finding this set's maximum-cost complement. For order-independent signature functions, the complement set for optimal placement is shown to be a maximum spanning tree. For cyclic codes, the complement set for optimal placement is a new type of graph, a maximum valuation graph (MVG), which is produced by a new algorithm. Using this algorithm, cyclic codes produce significantly less performance overhead than order-independent functions. Experimental results show that the MVG algorithm yields substantial improvement over previous solutions.<>