Research on real-time scheduling algorithm of federated learning tasks based on energy optimization on NoC platform
Lin Chen
Hong Zhou
Tao Song
Ziyi Zhou
DOI: https://doi.org/10.59429/esta.v10i4.1586
Keywords: Task mapping; Network on chip; Cloud computing; Federated learning
Abstract
Federal learning technology can realize global data sharing and reduce the risk of privacy disclosure under the premise of ensuring data security. Aiming at the problem of task assignment scheduling in federated learning process, this paper studies the problem of federated learning task scheduling on NoC multi-core platform under cloud computing architecture. Considering the limited computing resources of physical nodes, this paper describes the problem of optimal assignment and execution of tasks on network nodes as a mixed integer nonlinear programming problem. In order to improve the computational effi ciency, the original problem can be equitably converted into a mixed integer linear programming problem. Finally, the scheduling method is verifi ed by real application of task set, and the infl uence of parameter selection on scheduling scheme is studied.
References
[1] D. Li and J. Wu. Minimizing energy consumption for frame-based tasks on heterogeneous multiprocessor platforms [J]. IEEE Transactions on Parallel
Distributed Systems, 2015, 26(3): 810-823.
[2] G. Che n, K. Huang and A. Knoll. Energy optimization for real-time multiprocessor system-on-chip with optimal DVFS and DPM combination [J]. ACM
Transactions on Embedded Computing Systems, 2014, 13(3): 111-130.
[3] P. Arr oba, J. M. Moya, J. L. Ayala and R. Buyya. DVFS-aware consolidation for energy-efficient clouds [C] International Conference on Parallel
Architecture and Compilation, 2015: 494-495.
[4] X. Lin , Y. Wang, Q. Xie and M. Pedram. Task scheduling with dynamic voltage and frequency scaling for energy minimization in the mobile cloud
computing environment [J]. IEEE Transactions on Services Computing, 2015, 8(2): 175-186.
[5] S. Abd Ishak, H. Wu and U. U. Tariq, Energy -aware task scheduling on heterogeneous NoC-based MPSoCs [C] International Conference on Computer
Design, 2017: 165-168.
[6] J. Sin gh, B. Mangipudi, S. Betha and N. Auluck, Restricted duplication-based MILP formulation for scheduling task graphs on unrelated parallel machines
[C] International Symposium on Parallel Architectures, Algorithms and Programming, 2012: 202-209.