.png){kind=icon}
返回
Performance Enhancement Using a Dynamic Scoring-Based Task Scheduling Algorithm for Heterogeneous Multicore Systems
.png){kind=icon}
PRE
AI
DOI：10.1109/OJCS.2025.3648227.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Heterogeneous multicore systems, such as ARM's big.LITTLE architecture requires efficient task scheduling to balance deadline misses, energy consumption, and throughput. Traditional methods often assign tasks based on static classifications of computational or memory intensity, which do not fully utilize the dynamic nature of these systems. This article proposes a scheduling algorithm called the Dynamic Scoring-Based Task Scheduling (DSTS) algorithm, which dynamically assigns tasks to cores by considering real-time processor states and task requirements. The algorithm uses task and processor scores, which account for computational needs, memory demands, and urgency, to ensure efficient task processor mapping and core utilization. The proposed approach significantly reduces energy consumption and improves system performance. The results show that DSTS reduces average energy consumption by 30.95 percent compared to four other leading schedulers, cuts missed deadlines by 26.92 percent, and increases throughput by 3.11 percent. These results demonstrate that DSTS effectively reduces energy use and improves deadline adherence, making it suitable for energy-critical applications such as mobile devices, cloud services, and embedded systems.
Keyword:
Energy efficiency
heterogeneous
multi-cores
scheduling