Resource-Aware Job Scheduling and Management System for Semiconductor Application.

The increased number of transistors and complexity of chips has led the rapid increase of the semiconductor design data size. It is crucial for turn-around time (TAT) of mask data preparation (MDP) procedures which are translated from designed data into a set of polygons for mask-writing equipment. MDP procedures require huge computational resources and software to maintain desired TAT. One of our concerns is how to decide priorities of all jobs and effectively manage target TAT using limited resources with many users. In this paper, we introduce a job scheduling system with a newly developed algorithm to control job priorities considering planned date and predict a number of required resources in advance. Furthermore, we propose a resource management technique to maximally utilize limited resources and dramatically reduced TAT, simultaneously. As a practical example, TAT and resource utilization of advanced technology node were improved by 58% and 30%, respectively.