Temperature-Aware Scheduling Based on Dynamic Time-Slice Scaling

摘要

As power density increases with high technology, the high temperature has threatened the system performance, reliability and even system safety. In this paper, we propose a temperature-aware task scheduling approach which combines low-overhead Time-Slice Scaling (TSS) with Alternative Scheduling schemes to reduce temperature. Through fine-grained thermal characterization based on task behavior, dynamically determine Time-slice Scaling factor (TSF) for each task on real-time. Besides, combining alternative scheduling scheme based on boosting thermal model. Experimental results demonstrate our proposed policy can reduce the chip average and peak temperature maximums by 3.1°C and 2°C with negligible performance loss.