Computation of WCET using Program Slicing and Real-Time Model-Checking

We address the problem of computing accurate Worst-Case Execution Time (WCET) on pipelined architectures with caches. We propose a fully automatic and modular methodology based on program slicing and real-time modelchecking. We have implemented our methodology and applied it to standard benchmarks. To further validate the approach, we also compare our results to the real execution times of the programs measured on a real board.