符号化WCET(worst-case execution time)分析是用符号表达式表示任务的最大执行时间:表达式中包含了参数.通过在运行时刻快速确定表达式值,符号化WCET分析可以更精确地估算WCET.提出了一种针对其分支直接依赖于输入数据的程序的符号化WCET分析方法.首先对Blieberger方法进行扩充,使得WCET符号表达式能够表达依赖输入分支,然后利用程序的控制依赖图对符号表达式进行化简,从而产生带条件的WCET符号表达式,即不同的条件对应不同的符号表达式与已有方法不同,符号化WCET公式直接依赖于输入参数,使得运行时的WCET估算更加简单直接.
Symbolic WCET (worst-case execution time) analysis yields symbolic upper bound expressions for tasks that contain parameters. Quickly evaluated at run-time, such expressions can improve the accuracy of WCET estimate. This paper proposes a symbolic WCET analysis method that is especially for the input-data dependent branches. First, the formulas described by Blieberger is expanded, so that they can express the input-data dependent branches. Simplification of the formulas using control -dependent graph y ields conditional symbolic expressions that have different forms corresponding to different input value range s. Different from the existing methods, the symbolic formulas are directly dependent on input-data, so WCET estimate evaluation at run-time is more simple and straightforward.