Extended Synchronous Dataflow for Efficient DSP System Prototyping

Abstract

Though synchronous dataflow (SDF) graph has been a successful input specification language for digital signal processing (DSP) applications, lack of support for global states makes it unsuitable for multimedia signal processing applications that need global states for efficient implementation. In this paper, we propose synchronous piggybacked dataflow (SPDF), an extension of SDF model to accommodate global states without side effects. Global states are accessed by a special block that piggybacks the global state update request on data samples. Such an extension enlarges the domain of application where dataflow representation can be used for rapid system prototyping. The only penalty it incurs is scheduling complexity since the scheduler now considers control dependency as well as data dependency. We present the static analysis of the SPDF model and an implementation technique for memory efficient code synthesis. Finally, we show experimental results with a real life example, MPEG-audio decoder, to present the novelty and usefulness of our approach.