Instrumentation-driven model detection for dataflow graphs

Dataflow modeling offers a myriad of tools to improve optimization and analysis of signal processing applications, and is often used by designers to help design, implement, and maintain systems on chip for signal processing. However, maintaining and upgrading legacy systems that were not originally designed using dataflow modeling can be challenging. To facilitate maintenance, designers often convert legacy code to dataflow graphs, a process that can be difficult and time consuming. We propose a method to facilitate this conversion process by automatically detecting the dataflow models of the core functions. The contribution of this work is twofold. First, we introduce a generic method for instrumenting dataflow graphs that can be used to measure various statistics and extract run-time information. Second, we use this instrumentation technique to demonstrate a method that facilitates the conversion of legacy code to dataflow-based implementations. This method operates by automatically detecting the dataflow model of the core functions being converted.