AOW

Key Features

• Large scale architecture optimization in terms of input model
• Extreme scale scheduling problem support by specialized set of decomposition algorithms and optimization pipelines
• Extendable design

Users

•  System engineers
• Software and hardware engineers

Benefits for the User

•  Possibility to perform large- and extreme- scale DSE and find set of optimal designs with respect to user-defined objectives and constraints
• Do not require knowledge in optimization / mathematical programming

Inputs

•  SysML concise modelling of the architecture problem including Rhapsody, Excel models and KPI metrics templates.
• JSON formatted models for scheduling
• JSON formatted models for dataflow mapping

Outputs

• Pareto-frontier of optimal solutions in terms of input model (optimal architectures, optimal scheduling, optimal dataflow mappings)

Role in the CERBERO Toolchain

• HW/SW optimal scheduling (integrated with PREESM and DynAA)
• Dataflow optimization (integrated with MDC)
• Black-box optimization (adaptive camera in Ocean Monitoring use-case)

Tool Highlight

Return to Toolchain