Power-proportional modelling fidelity

  1. Lookup NU author(s)
  2. Dr Ashur Rafiev
  3. Dr Fei Xia
  4. Dr Alexei Iliasov
  5. Rem Gensh
  6. Ali Aalsaud
  7. Professor Alexander Romanovsky
  8. Professor Alex Yakovlev
Author(s)Rafiev A, Xia F, Iliasov A, Gensh R, Aalsaud A, Romanovsky A, Yakovlev A
Publication type Report
Series TitleSchool of Computing Science Technical Report Series
Report Number1443
Full text is available for this publication:
Traditional hierarchical modelling methods tend to have layers of abstraction corresponding to naturally existing layers of concern in multilevel systems. Although convenient, this is not always optimal for analysis and design. For instance, parts of a system which are in the same layer may not contribute to the same degree on some metric, e.g. system power consumption. To moderate the modelling, analysis and design effort, and potentially runtime control overhead for models used at runtime, less significant parts of the system should be studied at higher levels of abstraction and more significant ones with more detail. Concentrating on system power consumption, this paper presents Order Graphs (OGs), which have a clear hierarchical structure, but provide straightforward vertical zooming across multiple layers (orders) of model fidelity, resulting in the discovery of power-proportional cuts that run through different orders to be analysed together in a flat manner. Stochastic Activity Networks (SANs), a good flat modelling method, is suggested as an example of studying technique for cuts discovered with OGs. A series of experiments on an Odroid development system consisting of an ARM big.LITTLE multi-core structure provides initial validation for the approach.
InstitutionSchool of Computing Science, University of Newcastle upon Tyne
Place PublishedNewcastle upon Tyne
ActionsLink to this publication