Computing time as a program variable: a way around infeasible paths

Author Niklas Holsti



PDF
Thumbnail PDF

File

OASIcs.WCET.2008.1660.pdf
  • Filesize: 0.62 MB
  • 11 pages

Document Identifiers

Author Details

Niklas Holsti

Cite As Get BibTex

Niklas Holsti. Computing time as a program variable: a way around infeasible paths. In 8th International Workshop on Worst-Case Execution Time Analysis (WCET'08). Open Access Series in Informatics (OASIcs), Volume 8, pp. 1-11, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2008) https://doi.org/10.4230/OASIcs.WCET.2008.1660

Abstract

Conditional branches connect the values of program variables with the execution paths and thus with the execution times, including the worst-case execution time (WCET). Flow analysis aims to discover this connection and represent it as loop bounds and other path constraints. Usually, a specific analysis of the dependencies between branch conditions and assign­ments to variables creates some representation of the feasible paths, for example as IPET execution-count constraints, from which a WCET bound is calculated. This paper explores another approach that uses a more direct connection between variable values and execution time. The execution time is modeled as a program variable. An analysis of the dependencies between variables, including the execution-time variable, gives a WCET bound that excludes many infeasible paths. Examples show that the approach often works, in principle. It remains to be seen if it is scalable to real programs.

Subject Classification

Keywords
  • WCET
  • flow analysis
  • infeasible paths
  • dependency analysis

Metrics

  • Access Statistics
  • Total Accesses (updated on a weekly basis)
    0
    PDF Downloads
Questions / Remarks / Feedback
X

Feedback for Dagstuhl Publishing


Thanks for your feedback!

Feedback submitted

Could not send message

Please try again later or send an E-mail