ABSTRACT

Some task scheduling algorithms generate the shortest schedule, when its input DAG satisfies a specified condition. Among those scheduling algorithms, TDS algorithm proposed a DAG condition where allocation of two parent tasks of a join task in the same processor cause longer schedule length than allocation in different processors, and it generates the shortest schedule if any input DAG satisfies the condition. In the paper, we propose a post-processing scheduling algorithm that reduces the number of processors while preserving its schedule length. Especially, we propose conditions where two processes can be merged without increasing schedule length. Experimental results show that the number of processor is reduced to 92.3% ~ 98.0% if schedule length is reserved and required computing power is reduced to 84.3% ~ 91.2% if schedule length can be increased.