Affinity-driven modeling and scheduling for makespan optimization in heterogeneous multiprocessor systems
Department of Electrical and Computer Engineering
With the advent of heterogeneous multiprocessor architectures, efﬁcient scheduling for high performance has been of signiﬁcant importance. However, joint considerations of reliability, temperature, and stochastic characteristics of precedence-constrained tasks for performance optimization make task scheduling particularly challenging. In this paper, we tackle this challenge by using an afﬁnity (i.e., probability)-driven task allocation and scheduling approach that decouples schedule lengths and thermal proﬁles of processors. Speciﬁcally, we sep-arately model the afﬁnity of a task for processors with respect to schedule lengths and the afﬁnity of a task for processors with regard to chip thermal proﬁles considering task reliability and stochastic characteristics of task execution time and intertask communication time. Subsequently, we combine the two types of afﬁnities, and design a scheduling heuristic that assigns a task to the processor with the highest joint afﬁnity. Extensive simu-lations based on randomly generated stochastic and real-world applications are performed to validate the effectiveness of the proposed approach. Experiment results show that the proposed scheme can reduce the system makespan by up to 30.1% without violating the temperature and reliability constraints compared to benchmarking methods.
IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
Hu, X. S.
Affinity-driven modeling and scheduling for makespan optimization in heterogeneous multiprocessor systems.
IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems,
Retrieved from: https://digitalcommons.mtu.edu/michigantech-p/572