Ross Smith, K. Fant, D. Parker, Rick Stephani, Ching-Yi Wang
{"title":"An asynchronous 2-D discrete cosine transform chip","authors":"Ross Smith, K. Fant, D. Parker, Rick Stephani, Ching-Yi Wang","doi":"10.1109/ASYNC.1998.666508","DOIUrl":null,"url":null,"abstract":"This paper describes a fully asynchronous two-dimensional discrete cosine transform chip. The chip has a fixed block size of 8/spl times/8 pixels and uses bit-serial arithmetic. The chip was fabricated through MOSIS using a 0.8 /spl mu/ double-metal CMOS process. The 49.5 mm/sup 2/ core uses /spl sim/162,000 transistors. The chip operates from 0.65 V to 7.0 V, but its pixel rate at 5.0 V, 17 MHz, is significantly below the 27 MHz simulated because none of the signal's capacitances were backextracted. In order to design a completely asynchronous chip, a FIFO-based transposition memory was used, even though it used more area than RAM-based memory. The most interesting aspects of the design are presented here: the memory control structure, the pipelining structures, the use of Xilinx FPGAs and a Quickturn emulation system for emulation, and a comparison with other synchronous and asynchronous designs.","PeriodicalId":425072,"journal":{"name":"Proceedings Fourth International Symposium on Advanced Research in Asynchronous Circuits and Systems","volume":"26 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1998-03-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"9","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings Fourth International Symposium on Advanced Research in Asynchronous Circuits and Systems","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ASYNC.1998.666508","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 9
Abstract
This paper describes a fully asynchronous two-dimensional discrete cosine transform chip. The chip has a fixed block size of 8/spl times/8 pixels and uses bit-serial arithmetic. The chip was fabricated through MOSIS using a 0.8 /spl mu/ double-metal CMOS process. The 49.5 mm/sup 2/ core uses /spl sim/162,000 transistors. The chip operates from 0.65 V to 7.0 V, but its pixel rate at 5.0 V, 17 MHz, is significantly below the 27 MHz simulated because none of the signal's capacitances were backextracted. In order to design a completely asynchronous chip, a FIFO-based transposition memory was used, even though it used more area than RAM-based memory. The most interesting aspects of the design are presented here: the memory control structure, the pipelining structures, the use of Xilinx FPGAs and a Quickturn emulation system for emulation, and a comparison with other synchronous and asynchronous designs.