Allocation and scheduling of SystemJ programs on chip multiprocessors with weighted TDMA scheduling.

The event-driven model of computation provides a good platform for capturing complex real-time applications with low latency requirements since this model is free of problems like deadlocks, livelocks and performance bottlenecks such as locking overheads. Furthermore, there is great potential for extracting parallelism from an event-driven model on a chip multiprocessor (CMP). But, the execution and scheduling of this model on a CMP faces two primary problems that need to be addressed: (a) allocation of asynchronous event handlers (AEHs) on a minimal number of processing elements of a CMP such that the time guarantees are met and (b) scheduling the AEHs taking into consideration the shared resource access contentions. We propose a two-stage Integer Linear Programming approach for static allocation and scheduling of AEHs executing on a time-analyzable multiprocessor architecture with shared resources. The first stage allocates AEHs on a minimal number of processing elements in a CMP with weighted time-division multiplex access (TDMA) to the shared memory while meeting the timing requirements. The second stage then tunes the TDMA schedule such that the worst case reaction time of AEHs can be further shortened. This approach shows 43.75% better utilization of resources compared to traditional fair TDMA scheduling on CMPs.