Visualization tools to explore large execution traces of multicore embedded systems

Affiliated to

Description

Execution traces play an important role to debug real-time aspects in embedded applications, and in particular to check respect of deadlines, delays and other real-time properties. Such aspects are key for a multiprocessor platform, where the interactions between software components deployed on all its processors may have a high impact on the global real-time behavior.

STMicroelectronics has developed a large set of trace management tools, which allow to visualize and analyze the dynamic behavior of such applications. Those tools collect execution traces through different hardware and software mechanisms and make them available to end-users through different visualization tools. Most of those traces are based on operating system trace points and track system activity, but can be extended to application traces.

The PhD student will work on the design and prototyping of new visualization tools to make the exploration of large volumes of traces much easier. Different tracks can be explored: pruning the traces, condensing repetitive sequences, automatic analysis of some properties, better visualization mechanisms, data mining, etc. The intern student should help the local team decide which tracks are the most interesting, depending on his background. Any prototyped tool should be integrated in our trace management framework.