{"title":"Fault-tolerant sorting using VLSI processor arrays","authors":"H. Youn, Kyung Ook Lee","doi":"10.1109/DFTVS.1993.595609","DOIUrl":null,"url":null,"abstract":"Parallel sorting is one of the most important computational problem. An efficient scheme for fault tolerant sorting is proposed, which is based on odd-even transposition sort with a linear array of processing elements (PEs). The faults in the array are tolerated as far as no more than a single compare-and-swap (CS) module fault exists in any three consecutive CS modules using the voting approach. The hardware overhead is basically an additional register and a voter per PE. The scheme can also be easily adapted to a two-dimensional processor arrays, using Shear sort. If the proposed approach is employed for only error detection, then multiple faults can be detected in each step of computation using only a simple XOR circuitry in each PE.","PeriodicalId":213798,"journal":{"name":"Proceedings of 1993 IEEE International Workshop on Defect and Fault Tolerance in VLSI Systems","volume":"14 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1993-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of 1993 IEEE International Workshop on Defect and Fault Tolerance in VLSI Systems","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/DFTVS.1993.595609","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
Parallel sorting is one of the most important computational problem. An efficient scheme for fault tolerant sorting is proposed, which is based on odd-even transposition sort with a linear array of processing elements (PEs). The faults in the array are tolerated as far as no more than a single compare-and-swap (CS) module fault exists in any three consecutive CS modules using the voting approach. The hardware overhead is basically an additional register and a voter per PE. The scheme can also be easily adapted to a two-dimensional processor arrays, using Shear sort. If the proposed approach is employed for only error detection, then multiple faults can be detected in each step of computation using only a simple XOR circuitry in each PE.