Multiple-Morphs Adaptive Stream Architecture

In modern VLSI technology, hundreds of thousands of arithmetic units fit on a 1cm² chip. The challenge is supplying them with instructions and data. Stream architecture is able to solve the problem well. However, the applications suited for typical stream architecture are limited. This paper presents the definition of regular stream and irregular stream, and then describes MASA (Multiple-morphs Adaptive Stream Architecture) prototype system which supports different execution models according to applications' stream characteristics. This paper first discusses MASA architecture and stream model, and then explores the features and advantages of MASA through mapping stream applications to hardware. Finally MASA is evaluated by ten benchmarks. The result is encouraging. This research was supported by the National Natural Science Foundation of China (Grant No.60473080). Mei Wen is a Ph.D. candidate in computer science at National University of Defense Technology. Her research interests include computer architecture and scientific computing. Nan Wu is a Ph.D. candidate in computer science at National University of Defense Technology. His research interests include computer architecture, stream computing and compiler design. Hai-Yan Li is a Ph.D. candidate in computer science at National University of Defense Technology. Her research interests include computer architecture and media processing. Chun-Yuan Zhang is a professor of computer science at National University of Defense Technology. His research interests include multicomputer architecture and scientific computing. He received the Ph.D. degree in computer science from National University of Defense Technology.