WCET

WCET, or worst-case execution time, is the maximum time required for a software component to execute on a given hardware platform under defined conditions. It is used in real-time systems to establish temporal guarantees and to support schedulability analysis, verification, and certification.

Determining WCET involves accounting for hardware characteristics such as caches, pipelines, branch predictors, memory hierarchy, interrupts, and context switches, as well as software factors such as control-flow paths, loop bounds, and input data. Because these factors interact in complex ways, WCET is typically an upper bound rather than an exact value. It may vary with compiler options, task priorities, and other runtime conditions.

Estimation methods include measurement-based approaches, where actual execution times are collected during testing to identify a worst-case run; static timing analysis (STA), which builds a model of the software and hardware to prove an upper bound; and hybrid methods that combine analysis with empirical data. Path-based analyses and loop-bound techniques are common within STA. Some approaches also model cache effects or pipeline timing to improve realism, at the risk of increased conservatism.

WCET is sensitive to hardware configurations and software changes; a bound valid for one platform or compiler may not hold for another. It is used to determine whether deadlines can be met in scheduling and to support safety certifications in domains such as avionics and automotive systems. It is common to report WCET alongside an average-case execution time and to include margins to absorb uncertainty.