Optimizing real-time software to ensure that timing requirements are met can be both costly and time consuming. The tutorial examines common misconceptions and pitfalls in approaches to solving timing problems that can consume effort and resources while failing to address the underlying issues. An effective process and techniques are then presented which avoid these pitfalls. This process identifies code that contributes the most to the overall worst-case execution time, asks questions about the outcome of optimization, and targets optimization effort where it will have the maximum benefit for the minimum cost. The tutorial closes with a summary of the results of case study, applying this process to a large Ada project. There will be an opportunity for hands-on work, including a competition and prize for the best optimization.

This tutorial will benefit embedded software developers and managers who need to engineer reliable, embedded software. Today, software timing analysis does not have to be guesswork. The presentation covers two key aspects of real-time systems performance: (1) how to gain a clear, detailed, and accurate understanding of the execution time behavior of embedded software, (2) how to target optimization effort precisely where it will have the maximum benefit in improving system timing behavior (eliminating timing failures and creating headroom for new functionality) for the minimum cost.

Presenter

Dr. Ian Broster is a founder and Director of Rapita Systems Ltd, a company specializing in software timing analysis, worst-case execution time and profiling. He is an experienced, lively lecturer who has given numerous training courses, lectures and presentations on this and other topics. He has been involved with Ada for several years. He earned his PhD at the Real-Time Systems Research Group of University of York.

>Back