.png){kind=icon}
.png){kind=icon}
Return
Energy aware frame based fair scheduling
.png){kind=icon}
PRE
AI
DOI：10.1016/j.suscom.2018.03.003.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Energy has now-a-days become a critical resource in all battery operated devices. Reduction of energy consumption is essential to prolong the battery life in these systems. A significant class of such systems often execute a mix of independent real-time applications with guaranteed Quality of Service (QoS) requirements. Proportional fair schedulers form a very effective mechanism for handling such application sets, with high resource utilisation on multiprocessor systems. In this paper, we propose a DVFS enabled proportional fair strategy called Energy Aware Frame Based Fair Scheduling (EAFBFS) for executing a set of real-time QoS aware tasks on multicores. In addition to providing appreciable energy savings, EAFBFS is able to deliver high and tunable fairness accuracy, low scheduling related overheads along with resource utilisation optimality, by combining and leveraging the benefits of two prominent state-of-theart schedulers ERFair and DPFair. While the proposed scheduler ensures atmost m(-1) task migrations over stipulated intervals called time frames similar to DPFair, EAFBFS exhibits far improved fairness properties. Experimental results show that EAFBFS performs almost at par with energy aware DPFair in terms of energy savings and migrations incurred while achieving much higher proportional fairness accuracy (10-15 times on average over DPFair). (C) 2018 Elsevier Inc. All rights reserved.
Keywords:
EAFBFS
Energy aware
ERFair
DPFair
Frame
Real-time and scheduler
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
S
IF:
5.7
Papers:
966
Citations:
2.9K