Design and Implementation of a Multidimensional Synchronous Dataflow Environment *
Michael J. Chen and Edward A. Lee1995 Proc. IEEE Asilomar Conf. on Signals, Systems, and Computers
| Prepublished version |
| Published version |
ABSTRACT
Multidimensional Synchronous Dataflow is a model of computation where functional blocks produce and consume data on a multidimensional index space. The model can express multidimensional signal processing systems and algorithms elegantly, and is able to reveal the data parallelism of such systems in a way that raises possibilities for better multiprocessor scheduling. We have formalized definitions and specifications for a two-dimensional model and have created a single-processor simulation domain in Ptolemy[1]. Our paper discusses the design issues and methodologies we used to efficiently handle the large amounts of two-dimensional data that the systems operate upon. The primary structures used were matrices to act as buffers and submatrices to access subsets of those buffers. is intended to be a summary and not an in-depth discussion.*This research is part of the Ptolemy project, which is supported by the Advanced Research Projects Agency and the U.S. Air Force (under the RASSP program, contract F33615-93-C-1317), Semiconductor Research Corporation (project 94-DC-008), National Science Foundation (MIP-9201605), Office of Naval Technology (via Naval Research Laboratories), Bell Northern Research, Cadence, Dolby, Hitachi, Mentor Graphics, Mitsubishi, NEC, Pacific Bell, Philips, Rockwell, Sony, and Synopsis.