A Scenario-Aware Dataflow Programming Model

Abstract

Abstract-The FSM-SADF model of computation allows to find a tight bound on the throughput of firm real-time applications by capturing dynamic variations in scenarios. We explore an FSM-SADF programming model, and propose three different alternatives for scenario switching. The best candidate for our CompSOC platform was implemented, and experiments confirm that the tight throughput bound results in a reduced resource budget. This comes at the cost of a predictable overhead at runtime as well as increased communication and memory budgets. We show that design choices offer interesting trade-offs between run-time cost and resource budgets.

On the implementation and execution of adaptive streaming applications modeled as MADF

“…In [3], the FSM-SADF model is presented as another analysis model for adaptive streaming applications. To implement an application modeled and analyzed with FSM-SADF, two programming models have been proposed in [9] and [10]. In [9], the programming model is constructed by merging the SDF graphs of all scenarios into a single graph which may be larger than the FSM-SADF analysis graph.…”

“…To implement an application modeled and analyzed with FSM-SADF, two programming models have been proposed in [9] and [10]. In [9], the programming model is constructed by merging the SDF graphs of all scenarios into a single graph which may be larger than the FSM-SADF analysis graph. Then, to enable switching to a new scenario, all actors in all scenarios are constantly kept active while only those actors belonging to the identified new scenario by a detecting actor(s) will be executed after switching.…”

Programming and analysing scenario-aware dataflow on a multi-processor platform

“…An existing FSM-SADF programming model splits off oneactor detector scenarios in a similar way [7]. There all actors fire but execute the encapsulated functions conditionally based on a scenario identifier token.…”

“…before it fires. The behaviour during runtime can be matched to the analysis steps listed in Subsection III-C as follows: (a) the output port of Sw is connected to the FIFO towards df ; (b) Sw fires and consumes all of its tokens (4 in III-C); (c) Sw produces tokens into the connected FIFO (5); (d) execution of S full continues as usual (6,7,8,1).…”

Multi-criteria resource allocation in modal hard real-time systems

“…However, during mode change, the contexts of runnables that are executed on different cores in two subsequent modes need to be migrated from one core to another and the time of this migration shall be bounded, as hard real-time systems are considered in this paper. Therefore, the worst case switching time has to be assumed to provide the timing guarantees [8]. To migrate all involved runnables' contexts during a required interval, some additional requirements for the available communication bandwidth can be imposed.…”