Static Worst-Case Execution Time (WCET) analysis is a technique to derive upper bounds for the execution times of programs. Such bounds are crucial when designing and verifying real-time systems. A key component for static derivation of precise WCET estimates is upper bounds on the number of times different loops can be iterated. In this paper we present an approach for deriving upper loop bounds based on a combination of standard program analysis techniques. The idea is to bound the number of different states in the loop which can influence the exit conditions. Given that the loop terminates, this number provides an upper loop bound. An algorithm based on the approach has been implemented in our WCET analysis tool SWEET. We evaluate the algorithm on a number of standard WCET benchmarks, giving evidence that it is capable to derive valid bounds for many types of loops.
Best paper award