Use-definition chains with applications

  • Authors:
  • Ken Kennedy

  • Affiliations:
  • Department of Mathematical Sciences, William Marsh Rice University, Houston, TX 77001, U.S.A.

  • Venue:
  • Computer Languages
  • Year:
  • 1978

Quantified Score

Hi-index 0.00

Visualization

Abstract

Methods of compile-time program analysis for automatic code optimization usually include control flow analysis, in which possible execution flow paths are modeled, and data flow analysis, in which data relatiohsips are modeled. One representation of data relations in a program, via use-definition chains, is examined in the light of a particular example problem-the global elimination of ''useless'' computation. Two elimination algorithms which use differently organized use-definition chains are presented and compared for space complexity on two pathologically different families of flow graphs. Algorithms to compute chains of both varieties are also developed.