Amortized efficiency of list update and paging rules
Communications of the ACM
Scheduling jobs with fixed start and end times
Discrete Applied Mathematics
Competitive paging with locality of reference
STOC '91 Proceedings of the twenty-third annual ACM symposium on Theory of computing
Journal of Algorithms
Strongly competitive algorithms for paging with locality of reference
SODA '92 Proceedings of the third annual ACM-SIAM symposium on Discrete algorithms
A case for two-way skewed-associative caches
ISCA '93 Proceedings of the 20th annual international symposium on computer architecture
Randomized and multipointer paging with locality of reference
STOC '95 Proceedings of the twenty-seventh annual ACM symposium on Theory of computing
Experimental studies of access graph based heuristics: beating the LRU standard?
SODA '97 Proceedings of the eighth annual ACM-SIAM symposium on Discrete algorithms
Active Management of Data Caches by Exploiting Reuse Information
IEEE Transactions on Computers
ISCA '90 Proceedings of the 17th annual international symposium on Computer Architecture
Theoretical Computer Science - Special issue: Online algorithms in memoriam, Steve Seiden
Hi-index | 0.00 |
We study the on-line caching problem in a restricted cache where each memory item can be placed in only a restricted subset of cache locations. Examples of restricted caches in practice include victim caches, assist caches, and skew caches. To the best of our knowledge, all previous on-line caching studies have considered on-line caching in identical or fully-associative caches where every memory item can be placed in any cache location.In this paper, we focus on companion caches, a simple restricted cache that includes victim caches and assist caches as special cases. Our results show that restricted caches are significantly more complex than identical caches. For example, we show that the commonly studied Least Recently Used algorithm is not competitive unless cache reorganization is allowed while the performance of the First In First Out algorithm is competitive but not optimal. We also present two near optimal algorithms for this problem as well as lower bound arguments.