SPAS: scalable path-sensitive pointer analysis on full-sparse SSA

  • Authors:
  • Yulei Sui;Sen Ye;Jingling Xue;Pen-Chung Yew

  • Affiliations:
  • School of Computer Science and Engineering, UNSW, Australia;School of Computer Science and Engineering, UNSW, Australia;School of Computer Science and Engineering, UNSW, Australia;Department of Computer Science and Engineering, University of Minnesota

  • Venue:
  • APLAS'11 Proceedings of the 9th Asian conference on Programming Languages and Systems
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

We present a new SPAS (Scalable PAth-Sensitive) framework for resolving points-to sets in C programs that exploits recent advances in pointer analysis. SPAS enables intraprocedural path-sensitivity to be obtained in flow-sensitive and context-sensitive (FSCS) techniques scalably, by using BDDs to manipulate program paths and by performing pointer analysis level-by-level on a full-sparse SSA representation similarly as the state-of-the-art LevPA (the FSCS version of SPAS). Compared with LevPA using all 27 C benchmarks in SPEC CPU2000 and CPU2006, SPAS incurs 18.42% increase in analysis time and 10.97% increase in memory usage on average, while guaranteeing that all points-to sets are obtained with non-decreasing precision.