ETD

• flight control software
• mixed-integer linear programming
• model-based design
• multicore scheduling
• NUMA-aware systems
• real-time systems
• schedule optimization

Nowadays, many industries whose products rely on time-critical control software need to migrate from monocore to multicore platforms. In order to take the best advantage of the benefits of parallelism (reduced clock frequency, reduced consumption, etc.) a re-scheduling of software tasks, originally conceived to execute sequentially, is advisable. The goal of this dissertation is to present a placement-and-scheduling engine that can be used for two main purposes. First, to re-allocate tasks to the available computation units, preserving their original temporal and functional constraints and optimizing, at the same time, a customizable performance metric. Second, to generate a bare-metal static, contention-free run-time environment suited to execute the schedule on a typical multicore NUMA industrial platform.