Inferring Message Flows from System Communication Traces

This paper proposes a novel method for automatically inferring message flowspecifications from the communication traces of a system-on-chip (SoC) designthat captures messages exchanged among the components during a systemexecution. The inferred message flows characterize the communication andcoordination of components in a system design for realizing various systemfunctions, and they are essential for SoC validation and debugging. Theproposed method relieves the burden of manual development and maintenance ofsuch specifications on human designers. Our method also uses a new accuracymetric, acceptance ratio, to evaluate the quality of the minedspecifications instead of the specification size often used in the previouswork, enabling more accurate specifications to be mined. Furthermore, thispaper introduces the concept of essential causalities to enhance the accuracyof the message flow mining and accelerate the mining process. The effectivenessof the proposed method is evaluated on both synthetic traces and tracesgenerated from executing several system models in GEM5. In both cases, theproposed method achieves superior accuracies compared to a previous approach.Additionally, this paper includes some practical use cases.