SOLUTIONS

Architecture & Systems Engineering

CoMET® — Create Powerful Virtual System Prototypes for Embedded System Modeling and Simulation of System Architecture

Quick sketching, rapid experimentation and visualization of ideas are the metier of system architects. VaST's virtual system prototype technologies enable system architects to rapidly construct and evaluate candidate system architectures. A VaST virtual system prototype simulates fast enough to enable architects to evaluate architectures under real software loads, including real-time operating systems, protocol stacks and large software libraries — no need to compromise by running just a small sample of the software.

The final choice of a system architecture is assisted by the CoMET environment's dynamic visualization of systems, parameters, behavior and interactions.

VaST's technologies and models have a high degree of timing accuracy so that the virtual embedded system simulation reflects the actual behavior of even the most demanding real-time systems. Fast and accurate cycle edge simulation enables device drivers and other code to be run interacting intimately with the hardware. Important parameters such as cache size also can be evaluated and optimized, which is impossible in a less accurate environment where cache hits and misses are not modeled precisely.

Abstract representations such as MATLAB / Simulink and universal modeling language (UML) can be included in a candidate virtual system prototype either as peripherals or as target code. Critical data can be abstracted from the simulation,including detailed performance data such as bus utilization; cache hit and fill rates, code procedures consuming the most time, and platform power dissipation. Designers no longer need to use rough estimation approaches to look at hardware/software tradeoffs, but can run accurate simulations of the various possibilities and then base decisions on real data:

• Microprocessor choice and configuration — including choosing between general purpose or DSP processors.
• Determine system partitioning — quantify, using accurate performance measures, the divisions and interfaces between subsystems created in different technologies.
• Operating system choice — determine which operating system (OS) to use by running the actual OS in the target hardware environment and assessing impact in response time, resource requirements, power consumption and system performance.
• System interaction — quantify the behavior of experimental systems in the context of their projected use; wireless and telecommunications devices in a network system; automotive electronics devices running with mechanical model.

Up to top