{"title":"Programmable architecture for matrix and signal processing","authors":"B. Hamilton","doi":"10.1109/REG5.1988.15912","DOIUrl":null,"url":null,"abstract":"A matrix signal processor (MSP) is being developed to provide fast and efficient solutions for several different groups of problems. The design balances the speed of a dedicated pipeline with the generality of a reconfigurable architecture. Throughput, flexibility, and efficiency are maximized by incorporating a programmable pipeline. Additional features are incorporated to increase further the throughput. Dual cache memory banks reduce processor idle time to almost zero while performing back-to-back matrix or signal processing algorithms. A third cache memory bank stores both constant coefficients for such algorithms as the FFT (fast Fourier transform) and temporary coefficients for such algorithms as time-invariant filtering.<<ETX>>","PeriodicalId":126733,"journal":{"name":"IEEE Region 5 Conference, 1988: 'Spanning the Peaks of Electrotechnology'","volume":"8 6 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1988-03-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"IEEE Region 5 Conference, 1988: 'Spanning the Peaks of Electrotechnology'","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/REG5.1988.15912","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 matrix signal processor (MSP) is being developed to provide fast and efficient solutions for several different groups of problems. The design balances the speed of a dedicated pipeline with the generality of a reconfigurable architecture. Throughput, flexibility, and efficiency are maximized by incorporating a programmable pipeline. Additional features are incorporated to increase further the throughput. Dual cache memory banks reduce processor idle time to almost zero while performing back-to-back matrix or signal processing algorithms. A third cache memory bank stores both constant coefficients for such algorithms as the FFT (fast Fourier transform) and temporary coefficients for such algorithms as time-invariant filtering.<>