Incremental compilation of optimized code

  • Authors:
  • Lori L. Pollock;Mary Lou Soffa

  • Affiliations:
  • Department of Computer Science, University of Pittsburgh, Pittsburgh, Penna.;Department of Computer Science, University of Pittsburgh, Pittsburgh, Penna.

  • Venue:
  • POPL '85 Proceedings of the 12th ACM SIGACT-SIGPLAN symposium on Principles of programming languages
  • Year:
  • 1985

Quantified Score

Hi-index 0.00

Visualization

Abstract

Although optimizing compilers have successfully been used to reduce the size and running times of compiled programs, present incremental compilers only support the incremental update of unoptimized code. In this work, we extend the notion of incremental compilation to include optimized code. Techniques to incrementally compile locally optimized code, given intermediate code modifications are developed using a program representation based on flow graphs and dags. A model is designed to represent both unoptimized and optimized code and to maintain an optimizing history. Changes to the optimized code which either destroy optimizations or create conditions for further optimizations are incorporated into the model and the optimized code without recompiling unaffected optimizations.