The Community for Technology Leaders
High Performance Computing and Grid in Asia Pacific Region, International Conference on (1997)
Seoul, Korea
Apr. 28, 1997 to May 2, 1997
ISBN: 0-8186-7901-8
pp: 7
Won-Kee Hong , Dept. of Comput. Sci., Yonsei Univ., Seoul, South Korea
Tack-Don Han , Dept. of Comput. Sci., Yonsei Univ., Seoul, South Korea
Shin-Dug Kim , Dept. of Comput. Sci., Yonsei Univ., Seoul, South Korea
Sung-Bong Yang , Dept. of Comput. Sci., Yonsei Univ., Seoul, South Korea
ABSTRACT
In multiprocessor systems, the cache misses due to coherence transactions make up many of the total cache misses. However this type of cache miss is strongly dependent on the type of data sharing among processors, especially false sharing. Until now the small cache block size has been used to avoid false sharing mainly in multiprocessor systems, but the smaller the cache block size, the lower the prefetching effect. Moreover it is shown that high spatial locality appears in many parallel programs. The paper presents two advanced full-map directory schemes which provide a low cache miss ratio and communication traffic by avoiding false sharing and taking advantage of the spatial locality existing in many parallel programs. The performance was evaluated by the event-driven simulator and the empirical results show that the proposed scheme can provide about a 6/spl sim/77% decrease in the cache miss ratio and a 46/spl sim/96% decrease in the communication traffic.
INDEX TERMS
cache storage; sectored caches; full-map directory scheme; multiprocessor systems; cache miss; coherence transactions; data sharing; false sharing; cache block size; prefetching; spatial locality; parallel programs; low cache miss ratio; low communication traffic; event-driven simulator
CITATION
Won-Kee Hong, Tack-Don Han, Shin-Dug Kim, Sung-Bong Yang, "An effective full-map directory scheme for the sectored caches", High Performance Computing and Grid in Asia Pacific Region, International Conference on, vol. 00, no. , pp. 7, 1997, doi:10.1109/HPC.1997.592113
79 ms
(Ver 3.3 (11022016))