This thesis evaluates a source-level approach to estimating the worst-case execution time of software, leveraging recent tools from functional verification. Towards this, we automatically annotate the source code with processor timing models, and apply a combination of Abstract Interpretation and Model Checking to compute a precise estimate that is superior to existing approaches. We further introduce 'Timing Debugging', a novel process which enables the user to inspect the execution time in a debugger.
«
This thesis evaluates a source-level approach to estimating the worst-case execution time of software, leveraging recent tools from functional verification. Towards this, we automatically annotate the source code with processor timing models, and apply a combination of Abstract Interpretation and Model Checking to compute a precise estimate that is superior to existing approaches. We further introduce 'Timing Debugging', a novel process which enables the user to inspect the execution time in a d...
»