Title :
A comparison of search heuristics for empirical code optimization
Author :
Seymour, Keith ; You, Haihang ; Dongarra, Jack
Author_Institution :
Electr. Eng. & Comput. Sci. Dept., Univ. of Tennessee, Knoxville, TN
fDate :
Sept. 29 2008-Oct. 1 2008
Abstract :
This paper describes the application of various search techniques to the problem of automatic empirical code optimization. The search process is a critical aspect of auto-tuning systems because the large size of the search space and the cost of evaluating the candidate implementations makes it infeasible to find the true optimum point by brute force. We evaluate the effectiveness of Nelder-Mead Simplex, Genetic Algorithms, Simulated Annealing, Particle Swarm Optimization, Orthogonal search, and Random search in terms of the performance of the best candidate found under varying time limits.
Keywords :
genetic algorithms; particle swarm optimisation; query formulation; simulated annealing; autotuning systems; empirical code optimization; genetic algorithms; nelder-mead simplex; orthogonal search; particle swarm optimization; random search; search heuristics; search process; search techniques; simulated annealing; Application software; Cost function; Genetic algorithms; Hardware; Laboratories; Lifting equipment; Linear algebra; Optimizing compilers; Particle swarm optimization; Simulated annealing;
Conference_Titel :
Cluster Computing, 2008 IEEE International Conference on
Conference_Location :
Tsukuba
Print_ISBN :
978-1-4244-2639-3
Electronic_ISBN :
1552-5244
DOI :
10.1109/CLUSTR.2008.4663803