A scheduling framework for minimizing memory requirements of multirate DSP systems represented as dataflow graphs

Numerous design environments for signal processing use specification languages with semantics closely related to synchronous dataflow (SDF), a restricted form of dataflow that has proven efficient for describing and compiling multirate signal processing algorithms. An SDF representation allows the compiler freedom to explore different ways to sequence the computations in a program, and to evaluate the associated tradeoffs, such as those involving throughput, code size, and buffering requirements. To guide the scheduling process, compilers may apply some form of clustering, in which multiple computations are grouped together according to different criteria. The authors develop clustering techniques to synthesize minimum code size implementations of SDF programs, and describe techniques to incorporate arbitrary clustering strategies into a minimum code size scheduler