Abstract
The inclusion property is essential in reducing the cache coherence complexity for multiprocessors with multilevel cache hierarchies. We give some necessary and sufficient conditions for imposing the inclusion property for fully- and set-associative caches which allow different block sizes at different levels of the hierarchy. Three multiprocessor structures with a two-level cache hierarchy (single cache extension, multiport second-level cache, bus-based) are examined. The feasibility of imposing the inclusion property in these structures is discussed. This leads us to propose a new inclusion-coherence mechanism for two-level bus-based architectures.
- 1 Archibald,J. and J.-L. Baer. An economical solution to the cache coherence problem. In Proc. 11th Symposium on Computer Architecture, pages 355-362, 1984. Google ScholarDigital Library
- 2 Archibald,J. and J.-L. Baer. Cache coherence protocols: Evaluation using a multiprocessor simulation model. ACll4 TOCS, 4(4):273-298, November 1986 Google ScholarDigital Library
- 3 Baer, J.-L. and W.-H. Wang. Architectural choices for multi-level cache hierarchies. In Proc. 16th International Conference on Parallel Processing, pages 258-261, 1987.Google Scholar
- 4 Baer,J.-L. and W.-H.Wang. Architectural choices for multi-level cache hierarchies. Technical Report TR 87- 01-04, University of Washington, January 1987.Google Scholar
- 5 Censier,M. and P.Feautrier. A new solution to coherence .problems in multicache systems. IEEE TC, C- 27(12):1112-1118, December 1978.Google ScholarDigital Library
- 6 Chang, J.H., H. Chao and K. So. Cache design of a submicron cmos system/370. In Proc. 14th Symposium on Computer Architecture, pages 208-213, 1987. Google ScholarDigital Library
- 7 Enslow Jr., P.H. Multiprocessor organizations - a survey. Computing Surveys, 9(1):103-129, March 1977. Google ScholarDigital Library
- 8 Goodman, J. Coherency for multiprocessor virtual address caches. In Proc. Architectuml Support for Programming Languages and Operating Systems(ASPOLS-II), pages 72- 81, 1987. Google ScholarCross Ref
- 9 Hattori,A., Koshino,M. and S.Kamimoto. Three-level hierarchical storage system for FACOM M-380/382. In Proc. Information Processing IFIP, pages 693-697, 1983.Google Scholar
- 10 Katz, R., Eggers, S., Wood, D., Perkins, C. and R.G. Sheldon. Implementing a cache coherence protocol. In Proc. 18th Symposium on Computer Architecture, pages 276- 283, 1985. Google ScholarDigital Library
- 11 Lam,C-Y. and S.Mudnick. Properties of storage hierarchy systems with multiple page sizes and redundant data. ACM TODS, 4(3):345-367, September 1979. Google ScholarDigital Library
- 12 Liptay, J. S. Structural aspects of the System/360 model 85 part II - the cache. IBM System Journal, 7(1):15-21, 1968,Google ScholarDigital Library
- 13 Short, R. T. A Study of Multilevel Cache Memories. Master's Thesis, University of Washington, 1987.Google Scholar
- 14 Sweazey, P. and A.J. Smith. A class of compatible cache consistency protocols and their support by the IEEE futurebus. In Proc. 13th Symposium on Computer Amhitecture, pages 414-423, 1986. Google ScholarDigital Library
- 15 Wilson Jr., A.W. Hierarchical cache/bus architecture for shared memory multiprocessors. In Proc. 14th Symposium on Computer Architecture, pages 244-252, 1987. Google ScholarDigital Library
- 16 Winsor, D.C. and T.N. Mudge. Crosspoint cache architectures. In Proc. 16th International Conference on Parallel Processing, pages 266-269, 1987.Google Scholar
Index Terms
- On the inclusion properties for multi-level cache hierarchies
Recommendations
Characteristics of performance-optimal multi-level cache hierarchies
Special Issue: Proceedings of the 16th annual international symposium on Computer ArchitectureThe increasing speed of new generation processors will exacerbate the already large difference between CPU cycle times and main memory access times. As this difference grows, it will be increasingly difficult to build single-level caches that are both ...
Characteristics of performance-optimal multi-level cache hierarchies
ISCA '89: Proceedings of the 16th annual international symposium on Computer architectureThe increasing speed of new generation processors will exacerbate the already large difference between CPU cycle times and main memory access times. As this difference grows, it will be increasingly difficult to build single-level caches that are both ...
Comments