Partitioning dataflow analyses using types

Abstract

We present a simple method for partitioning a dataflow analysis problem into a series of related subproblems. We use type information (either declared by the programmer, or computed via nonstandard type inference) to conservatively approximate analysis-time data dependences between program quantities. This dependence information frees us from the need to model all program quantities simultaneously in a single, monolithic analysis. Instead, we can model quantities in any order, or even in parallel, so long as we respect the dependences. Our approach is independent of the means used to solve the subproblems, and enables the wider application of existing sparse approaches previously restricted to 'separable' dataflow problems. Preliminary experiments applying our technique to flow-sensitive points-to-analysis of C programs have achieved storage savings of 1.3-7.2x over existing methods.