The methodology developed in task 4/5.2 also allows us to perform architecture profiling: having a model for a synchronous program running on a distributed architecture, we can replace each elementary process in this model, by another one, modelling time elapsing. This allows us to produce a program which simulates how much time is needed for the target architecture to execute. Such timing simulator can then be given to packages for subsequent analysis: is this or that particular real-time or timeliness property satisfied?
There are approaches along the lines of timed automata that can be considered. We have been looking instead in a different direction, namely the use of so-called MaxPlus algebra. In particular, MaxPlus automata have been introduced by this school as a way to model time consuming for machines having a finite set of different running modes. This is interesting as it provides us with techniques to evaluate infinite timing behaviours for our system (such as, e.g., steady-state throughput), something not available from competing approaches from temporal logic. And this holds, both for synchronous programs and their desynchronised execution, following the principles and method from tasks 4/5.1 and 4/5.2. Thus there is a need