We report on a case study in the application of partial evaluation, initiated<br />by the desire to speed up a constraint-based algorithm for control-flow<br /> analysis. We designed and implemented a dedicated partial evaluator,<br />able to specialize the analysis wrt. a given constraint graph and thus <br />remove the interpretive overhead, and measured it with Feeley's Scheme<br />benchmarks. Even though the gain turned out to be rather limited, our<br />investigation yielded valuable feed back in that it provided a better understanding<br />of the analysis, leading us to (re)invent an incremental version.<br />We believe this phenomenon to be a quite frequent spinoff from using <br />partial evaluation, since the removal of interpretive overhead makes the flow<br />of control more explicit and hence pinpoints sources of inefficiency. <br /> Finally, we observed that partial evaluation in our case yields such regular,<br />low-level specialized programs that it begs for run-time code generation.
Partial Evaluation for Constraint-Based Program Analyses