Model-driven design-space exploration

Model-driven design-space exploration

Predict the past, explore the future

One well-established approach to product development in the high-tech industry is to have separate ‘design-build-integrate-test’ stages. In this approach, major unexpected problems often arise at the start of the ‘integrate-test’ stage. A typical example is a disappointing overall system performance post-integration. Resolving issues at this stage involves costly redesign.

Model-driven design space exploration is an excellent way to avoid such problems early on. This approach constructs executable models that separate the embedded system application’s logic from the execution platform on which it is mapped. The models can be calibrated with available measurements or approximations in order to validate and improve the model’s predictive power.

Design space exploration models bridge the gap between back-of-the-envelope calculations and physical prototypes. They provide system-wide insights into timing bottlenecks and allow for the rapid exploration of design alternatives, such as for the optimization of timing performance. Ultimately, this approach leads to shorter, predictable development times and a much more controlled product quality.

Method description
  1. Predict the past

    • Model the system by decomposing it into an application, platform and mapping view

    • Calibrate the model with available measurements or ‘guesstimates’ and validate its predictive power

  2. Explore the future: run what-if scenarios

    Optimize application functionality, platform dimensioning and mapping

Additional information