Efficient and Precise Cache Behavior Prediction for Real-TimeSystems

  • Authors:
  • Christian Ferdinand;Reinhard Wilhelm

  • Affiliations:
  • Fachbereich Informatik / Universität des Saarlandes, Postfach 15 11 50 / D-66041 Saarbrücken;Fachbereich Informatik / Universität des Saarlandes, Postfach 15 11 50 / D-66041 Saarbrücken

  • Venue:
  • Real-Time Systems
  • Year:
  • 1999

Quantified Score

Hi-index 0.02

Visualization

Abstract

Abstractinterpretation is a technique for the static detection of dynamicproperties of programs. It is semantics based, that is, it computesapproximative properties of the semantics of programs. On thisbasis, it supports correctness proofs of analyses. It replacescommonly used ad hoc techniques by systematic, provable ones,and it allows for the automatic generation of analyzers fromspecifications by existing tools. In this work, abstract interpretationis applied to the problem of predicting the cache behavior ofprograms. Abstract semantics of machine programs are definedwhich determine the contents of caches. For interprocedural analysis,existing methods are examined and a new approach that is especiallytailored for the cache analysis is presented. This allows fora static classification of the cache behavior of memory referencesof programs. The calculated information can be used to improveworst case execution time estimations. It is possible to analyzeinstruction, data, and combined instruction/data caches for common(re)placement and write strategies. Experimental results arepresented that demonstrate the applicability of the analyses.