Young-Sang Lee, Jun-Woo Kang, L. Kim, Seung-Ho Hwang
{"title":"Self-timed shared division and square-root implementation using full redundant signed digit numbers","authors":"Young-Sang Lee, Jun-Woo Kang, L. Kim, Seung-Ho Hwang","doi":"10.1109/ICVC.1999.820995","DOIUrl":null,"url":null,"abstract":"A radix-2 square root implementation for self-timed dividers using redundant signed-digit (RSD) adders is presented. In this method, two self-timed RSD adder stages are used for each result bit selection. A very efficient and simple result bit selection logic compared to the previous designs is implemented by using double self-timed ring stages. The F-term in the RSD format is easily applied to two self-time substages. F-term generation is overlapped with a partial remainder calculation and result-bit selection. This makes the hardware implementation of the F-term generation much easier and less time-constraint. No additional time delay is included in the square-root arithmetic. From the SPICE simulation at 35/spl deg/C and under MOSIS 1.2 /spl mu/m design rule, the speed of this design is estimated to be 124 ns for 54 bit square-root and division calculation.","PeriodicalId":13415,"journal":{"name":"ICVC '99. 6th International Conference on VLSI and CAD (Cat. No.99EX361)","volume":"1 1","pages":"541-544"},"PeriodicalIF":0.0000,"publicationDate":"1999-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"ICVC '99. 6th International Conference on VLSI and CAD (Cat. No.99EX361)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICVC.1999.820995","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1
Abstract
A radix-2 square root implementation for self-timed dividers using redundant signed-digit (RSD) adders is presented. In this method, two self-timed RSD adder stages are used for each result bit selection. A very efficient and simple result bit selection logic compared to the previous designs is implemented by using double self-timed ring stages. The F-term in the RSD format is easily applied to two self-time substages. F-term generation is overlapped with a partial remainder calculation and result-bit selection. This makes the hardware implementation of the F-term generation much easier and less time-constraint. No additional time delay is included in the square-root arithmetic. From the SPICE simulation at 35/spl deg/C and under MOSIS 1.2 /spl mu/m design rule, the speed of this design is estimated to be 124 ns for 54 bit square-root and division calculation.