With recent advances in microelectronics manufacturing, multicore processors have become common. A multicore processor is a single chip that contains multiple cores, or central processing units (CPUs). These cores may communicate among themselves via switches, buses, or other interconnection networks. They may also share caches or memory space. In a nutshell, a multicore processor is somewhat like a traditional multiprocessor system built into a single chip.

Legacy application software typically operates on the assumption that it will run on a single core chip (that is, a single CPU chip). Therefore, even if such a program runs on a multicore chip, it only uses one core. To take full advantage of the power of multicore systems, application software needs to be parallelized to access all parts of the multicore. However, as in traditional multiprocessor systems, when application software is partitioned into several components and components are executed in parallel, scheduling how the components are run becomes complex. There needs to be some guarantee of consistency and correctness in resource sharing. Please note that a component may contain several dependent or independent tasks.

Many algorithms have been developed for traditional multiprocessor systems. The multiprocessors synchronization protocol for real-time open systems (MSOS) provides a resource synchronization protocol that allows one application component to be executed by one processor at a time. This paper extends MSOS to handle the case where a component is assigned to a cluster (a set of processors or cores).

Although the intent is the same for multiprocessor and multicore scheduling, there are subtle differences to be considered. For example, the communication between multicores should be faster and resource sharing should be easier if the resources reside in the same chip. Therefore, although the task scheduling method and resource usage synchronization algorithm for a multiprocessor system could be applied to the multicore system, it may not be the wisest choice.

Each multicore system has its unique characteristics and architecture. To fully utilize the power of multicore systems, a deeper understanding of the multicore structure is needed. There is no cure-all method for such systems.