Efficient stack sizing for very large software systems

Accurate determination of stack sizes for tasks is essential for software systems whose tasks have predeclared stack sizes (for example, embedded systems). Underestimation of the stack size can lead to stack overflows, a particularly insidious run time error. Overestimation will lead to wastage of precious memory space of these often memory starved systems. In this paper, we describe a tool which efficiently computes the stack sizes for tasks of very large software systems supporting multiple products. This is in contrast to existing tools which are mainly targeted at small software systems supporting single products. The efficiency of the tool is achieved by using two different types of caches.