Tuning the WCET of Embedded Applications

  • Authors:
  • Affiliations:
  • Venue:
  • RTAS '04 Proceedings of the 10th IEEE Real-Time and Embedded Technology and Applications Symposium
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

It is advantageous to not only calculate the WCET of anapplication, but to also perform transformations to reducethe WCET since an application with a lower WCET willbe less likely to violate its timing constraints. In thispaper we describe an environment consisting of an interactivecompilation system and a timing analyzer, where auser can interactively tune the WCET of an application.After each optimization phase is applied, the timing analyzeris automatically invoked to calculate the WCET ofthe function being tuned. Thus, a user can easily gaugethe progress of reducing the WCET. In addition, the usercan apply a genetic algorithm to search for an effectiveoptimization sequence that best reduces the WCET. Usingthe genetic algorithm, we show that the WCET for a numberof applications can be reduced by 7% on average ascompared to the default batch optimization sequence.