Please use this identifier to cite or link to this item:
https://doi.org/10.1109/ICPADS.2007.4447753
Title: | Novel critical-path based low-energy scheduling algorithms for heterogeneous multiprocessor real-time embedded systems | Authors: | Liu, Y. Veeravalli, B. Viswanathan, S. |
Issue Date: | 2007 | Citation: | Liu, Y.,Veeravalli, B.,Viswanathan, S. (2007). Novel critical-path based low-energy scheduling algorithms for heterogeneous multiprocessor real-time embedded systems. Proceedings of the International Conference on Parallel and Distributed Systems - ICPADS 1 : -. ScholarBank@NUS Repository. https://doi.org/10.1109/ICPADS.2007.4447753 | Abstract: | In this paper, we propose novel low-energy static and dynamic scheduling algorithms with low computational complexities, for heterogeneous multiprocessor real-time embedded systems. We consider task graphs with deadlines and precedence relationships to satisfy. We propose a novel scheme, referred to as "critical-path information track-and update", based on critical-path analysis to distribute the slack-time over tasks such that energy consumption is minimized, while guaranteeing the precedence and timing constraints. Our dynamic scheduling algorithm applies the static scheduling algorithm during runtime based on the updated average-case execution demands of tasks. Our simulation results show that the proposed static scheduling algorithm consumes only 2% of the computational time with no degradation in energy savings, whereas the dynamic scheduling algorithm delivers up to 25% more energy savings while reducing the computational time overhead by more than 90%, when compared with recent heterogeneous multiprocessor scheduling algorithms. © 2007 IEEE. | Source Title: | Proceedings of the International Conference on Parallel and Distributed Systems - ICPADS | URI: | http://scholarbank.nus.edu.sg/handle/10635/71176 | ISBN: | 9781424418909 | ISSN: | 15219097 | DOI: | 10.1109/ICPADS.2007.4447753 |
Appears in Collections: | Staff Publications |
Show full item record
Files in This Item:
There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.