The evaluation of nonfunctional properties, specially power consumption, at early design phases is very important for making better design choices in the development of an embedded system. This extended abstract presents: a) MAESTRO [4], a system-level methodology for the modeling and evaluation of interdependent nonfunctional properties together with embedded firmware, OS services, and application functionality in a single SystemC-based actor-oriented language, and b) its appliance in the evaluation of power consumption of seven use cases. MAESTRO provides a systematic and extensible approach to incorporate nonfunctional properties evaluators to a system model by formally modeling the complex interdependencies of all nonfunctional properties of an HW architecture component into a so called interaction model. The interaction model defines a formal specification for each nonfunctional property evaluator and determines their Modelof-Computation (MoC). Furthermore, the integration of performance models at heterogeneous levels of abstraction is facilitated by analysis of the interaction model. With MAESTRO, it is possible to investigate trade-offs between modeling effort and accuracy of nonfunctional properties evaluators, OS services and firmware, such as power management strategies and schedulers. MAESTRO builds upon the existing methodology SystemCoDesigner [2], and thus benefits from its established design flow, including automatic Design Space Exploration (DSE) and virtual prototype synthesis.
[1]
Christian Haubelt,et al.
SystemCoDesigner—an automatic ESL synthesis approach by design space exploration and behavioral synthesis for streaming applications
,
2009,
TODE.
[2]
Rafael Rosales,et al.
Scenario-based energy estimation of heterogeneous integrated systems at system level
,
2013,
2013 European Microwave Conference.
[3]
Michael Glaß,et al.
Mahler: Sketch-Based Model-Driven Virtual Prototyping
,
2014,
ARCS.
[4]
Rafael Rosales,et al.
MAESTRO— Holistic Actor-Oriented Modeling of Nonfunctional Properties and Firmware Behavior for MPSoCs
,
2014,
TODE.
[5]
Michael Glaß,et al.
End-to-end power estimation for heterogeneous cellular LTE SoCs in early design phases
,
2014,
2014 24th International Workshop on Power and Timing Modeling, Optimization and Simulation (PATMOS).