Code generators are notoriously difficult to debug, yet their correctness is crucial. This paper demonstrates that concept assignment can be performed in an entirely syntax-directed manner for code generators and other highly structured program modules. Anomalies in this concept assignment indicate the possible existence of bugs. These insights enable the formulation of a new debugging technique for code generators. This paper describes the procedure, a practical implementation, and results from the application of this debugging technique to an experimental code generator.

The Value State Dependence Graph is a sparse data-dependence IR that has been investigated as a midpoint in compilers producing sequential code for uniprocessors. However, we argue that parallelism is an important feature that has been largely overlooked. We conjecture that the VSDG is exploitable for parallelism through partitioning, and investigate the feasibility of generating streaming languages via a VSDG implementation, possibly bridging the gap between streaming languages and von Neumann-style languages such as C. D.3.4 [Programming Lan-