Team Members: Murugappan Senthilvelan	

Title: Tool for the Generation and Optimization of Data Flow Graphs from
standard filter Kernals in High Level Language

Abstract:

There have been many algorithms proposed and implemented to optimize
(retiming, unfolding etc) Data Flow Graphs in the past.  We have learnt
some of these algortihms as a part of this course.  All these algorithms
perform the optimizations on a input matrix that represents the delay on
each edge of the dataflow graph.  This matrix is generated by inspection
of the data flow graph which in turn is generated manually from the
filter algorithm or from the kernal code in high level language.  There
have been very few to no attempts on automatically generating these data
flow graphs directly from the Filter Kernal implemented in any high
level programming language.  This project attempts to develop a tool
that would inspect the program in high level language given to it as
input and generate the data flow graph for the filter kernal.  The data
flow graph thus generated will then be converted into the standard
matrix representation so that the optimization algorithms can take advan
tage of it.