Correlation analysis techniques for refining execution timeestimates of real-time applications

摘要

Scheduling techniques based upon worst case execution times, asare commonly used in real-time applications, often result in severeunderutilization of the processor resources since most tasks finish inmuch less time than their anticipated worst-case execution times. Wedescribe techniques for identifying correlation among the executions ofvarious statements within a program. We demonstrate how this informationcan be used to refine the estimate of remaining worst case executiontime of a real-time task as the execution of the task progresses.Refined estimates can be used at run-time to achieve better utilization,of the system and early failure detection and recovery