{"title":"Uniformization of linear recurrence equations: a step toward the automatic synthesis of systolic arrays","authors":"V. van Dongen, P. Quinton","doi":"10.1109/ARRAYS.1988.18084","DOIUrl":"https://doi.org/10.1109/ARRAYS.1988.18084","url":null,"abstract":"Systems of linear recurrences are a very powerful means for specifying algorithms that need to be implemented on a systolic array or, more generally, on a parallel architecture. In particular, the class of linear recurrences called uniform recurrences can be mapped automatically on systolic arrays. However, it is often difficult to express the algorithm directly using strictly uniform recurrences, especially when the algorithm involves recursive computations, as in the transitive closure problem. Novel tools for the uniformization of linear recurrences are provided. Their use is illustrated by solving the so-called routing problem, which is a particular uniformization problem. Two routing strategies are presented, namely, hierarchical routing and split routing.<<ETX>>","PeriodicalId":339807,"journal":{"name":"[1988] Proceedings. International Conference on Systolic Arrays","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1988-05-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128739517","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Systolic communication","authors":"H. Kung","doi":"10.1109/ARRAYS.1988.18106","DOIUrl":"https://doi.org/10.1109/ARRAYS.1988.18106","url":null,"abstract":"The author introduces the notion of systolic communication, a general architecture technique for supporting efficient implementation of systolic algorithms. This notion is described and compared to the usual memory-to-memory communication. The Warp systolic array machine is shown to support the systolic communication. A processor array such as Warp that supports the systolic communication will be most effective in executing systolic algorithms, for which data move during computation. This is illustrated by a simple matrix multiplication example. Advantages of the systolic communication are summarized.<<ETX>>","PeriodicalId":339807,"journal":{"name":"[1988] Proceedings. International Conference on Systolic Arrays","volume":"112 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128052604","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}