Propagating Information Using SSA

• Published in: SSA-based Compiler Design pp. 95 - 106
• Main Authors: Brandner, Florian, Novillo, Diego
• Format: Book Chapter
• Language: English
• Published: Cham Springer International Publishing 2022
• Subjects: Compact evaluation graph (CEG); Compiler design; Computer Science; Constant propagation; Control dependence; Copy propagation; Data dependence; Data-flow analysis; Forward analysis; Lattice; Liveness analysis; Maximal fixed point (MFP); Meet over all path (MOP); Monotone analysis; Programming Languages; Propagation; Range propagation; Single-entry/single-exit region (SESE); Sparse data-flow analysis; Sparse evaluation graph (SEG); SSA graph; Static analysis; Static single assignment form (SSA form); Transfer function; SSA Graph; Sparse evaluation graph (SEG); Sparce Data-Flow Analysis; Static Single Assignment Form; Copy propagation; Single-entry/single-exit region (SESE); Meet over all path (MOP); Static Analysis; Compact evaluation graph (CEG); Constant Propagation
• ISBN: 303080514X; 9783030805142
• DOI: 10.1007/978-3-030-80515-9_8

This chapter provides a gentle introduction to classical data-flow analysis and explains how SSA form improves both performance and expressiveness. In the first part, classical data-flow analysis is discussed by introducing the fundamentals of monotone frameworks on control-flow graphs. In the following part, the chapter introduces a simple and elegant data-flow engine that exploits SSA form. The engine relies on SSA form in order to efficiently propagate information on the values of variables on a so-called SSA graph, while at the same time taking into account control dependencies that emerge from the classical control-flow graph. The engine’s operation is explained through several examples based on constant and copy propagation, which illustrate the advantages of jointly processing the program’s data and control flow.