Title :
Analytical modeling and optimization for affinity based thread scheduling on multicore systems
Author :
Song, Fengguang ; Moore, Shirley ; Dongarra, Jack
Author_Institution :
Dept. of Electr. Eng. & Comput. Sci., Univ. of Tennessee, Knoxville, TN, USA
fDate :
Aug. 31 2009-Sept. 4 2009
Abstract :
This paper proposes an analytical model to estimate the cost of running an affinity-based thread schedule on multicore systems. The model consists of three submodels to evaluate the cost of executing a thread schedule: an affinity-graph submodel, a memory hierarchy submodel, and a cost submodel that characterize programs, machines, and costs respectively. We applied the analytical model to both synthetic and real-world applications. The estimated cost accurately predicts which schedule will provide better performance. Due to the NP-hardness of the scheduling problem, we designed an approximation algorithm to compute near-optimal solutions. We have extended the algorithm to support threads with data dependences. We conducted experiments with a computational fluid dynamics (CFD) kernel and Cholesky factorization on both UMA SMP and NUMA DSM machines. The results show that using the optimized thread schedule can improve the program performance by 25% to 400%, demonstrating that our method for determining an optimized thread schedule for multicore systems is efficient and practical.
Keywords :
approximation theory; graph theory; microprocessor chips; optimisation; processor scheduling; Cholesky factorization; NP-hardness; NUMA DSM machines; UMA SMP machine; affinity based thread scheduling; affinity-graph submodel; analytical modeling; approximation algorithm; computational fluid dynamics kernel; cost submodel; memory hierarchy submodel; multicore systems; Algorithm design and analysis; Analytical models; Approximation algorithms; Computational fluid dynamics; Costs; Multicore processing; Optimization methods; Processor scheduling; Scheduling algorithm; Yarn;
Conference_Titel :
Cluster Computing and Workshops, 2009. CLUSTER '09. IEEE International Conference on
Conference_Location :
New Orleans, LA
Print_ISBN :
978-1-4244-5011-4
Electronic_ISBN :
1552-5244
DOI :
10.1109/CLUSTR.2009.5289173