Date: Tuesday, March 27, 2001
Time: 3:00P.M. (Talk) <========= NOTE!!
Place: Withers 402-A, NCSU Historical Campus (click for
courtesy parking request)
Speaker: Frank Mueller, Humboldt-Universitaet zu Berlin, Gemany
Efficient Analysis of Temporal Properties for Real-Time Systems
Abstract: Schedulability theory for real-time and embedded systems has matured during the past decades to provide a sound base for the verification of a task's deadline. However, this theory is based on the assumption that the worst-case execution time (WCET) of a task be known a priori. This work focuses on analytical methods to determine the WCET statically.
In this work, a formal framework for safe timing analysis is derived, an implementation is provided and results on timing analysis for uniprocessors are presented. It is shown that a component-based approach separating the process into several steps of the analysis based on the modeled entity results in an efficient and practical solution of the problem. Then, an introduction to the overall toolset for static timing analysis is given followed by a more detailed derivation of methods for performing the static cache analysis. For this task, a data-flow framework is developed that provides classifications of cache references suitable for the static prediction of the WCET. A brief overview of the integration of these results into pipeline simulation and path analysis concludes the description of the framework for uniprocessor timing analysis.
Short Bio: Frank
Mueller received his MS and PhD in computer science from from Florida
State University in 1991 and 1994, respectively. From 1995 to 2000,
he held a position as an assistant professor at the Humboldt University
Berlin, Germany. In 2001, he joined the Center for Applied Scientific Computing
at the Lawrence Livermore National Laboratory. His professional interests
include the areas of real-time and embedded systems, compilers, architecture,
concurrent and distributed systems, and programming languages. He is a
member of the ACM and the IEEE Computer Society.
Host: C.
Healey, Computer Science, NCSU