Dependency aware ahead of time static scheduler for multicore

Multicore processor necessitated the use of parallel programming in order to make multicore processor utilization optimal. Researchers have dealt multicore scheduling problem to optimize some of the parameters like execution time, cache, memory, and dependency information. Dependency based scheduling algorithms invariably use task graphs for optimal scheduler with an aim to minimize dependency. We have developed a scheduling algorithm that considers dependency information at task level, dependency release information within tasks and load balancing for execution time. Tasks can be executed even before its dependent task completes its execution by using accurate task release information. We have improved the dependency information by the use of machine-learning algorithms resulting in improved makespan. We have evaluated our algorithm on standard benchmarks and we report an improvement of makespan up to a maximum of 66% of tasks. The results also suggest that the algorithm is scalable as its efficiency increases when large numbers of tasks are scheduled on many cores.