Scheduling Parallel Computations by Work Stealing: A Survey

摘要

Abstract Work stealing has been proven to be an efficient technique for scheduling parallel computations, and has been gaining popularity as the multiprocessor/multicore-processor load balancing technology of choice in both industry and academia. A review on the work stealing scheduling is provided from the perspective of scheduling algorithms, optimization of algorithm implementation and processor architecture oriented optimization. The future research trends and recommendations driven by theory, emerging applications and motifs, architecture and heterogeneous platforms are also provided.