Timing Analysis for Data and Wrap-Around Fill Caches

  • Authors:

  • Randall T. White;Frank Mueller;Chris Healy;David Whalley;Marion Harmon

  • Affiliations:

  • Florida State University, Computer Science Department, Tallahassee, FL 32306-4530;Humboldt-Universität zu Berlin, Institut für Informatik, 10099 Berlin (Germany);Florida State University, Computer Science Department, Tallahassee, FL 32306-4530;Florida State University, Computer Science Department, Tallahassee, FL 32306-4530;Florida A&M University, Computer & Information Systems Department, Tallahassee, FL 32307-3101

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

Quantified Score

Hi-index 0.00

Visualization

Abstract

Thecontributions of this paper are twofold. First, an automatictool-based approach is described to bound worst-case data cacheperformance. The approach works on fully optimized code, performsthe analysis over the entire control flow of a program, detectsand exploits both spatial and temporal locality within data references,and produces results typically within a few seconds. Resultsobtained by running the system on representative programs arepresented and indicate that timing analysis of data cache behaviorusually results in significantly tighter worst-case performancepredictions. Second, a method to deal with realistic cache fillingapproaches, namely wrap-around-filling for cache misses, is presentedas an extension to pipeline analysis. Results indicate that worst-casetiming predictions become significantly tighter when wrap-around-fillanalysis is performed. Overall, the contribution of this paperis a comprehensive report on methods and results of worst-casetiming analysis for data caches and wrap-around caches. The approachtaken is unique and provides a considerable step toward realisticworst-case execution time prediction of contemporary architecturesand its use in schedulability analysis for hard real-time systems.