Scalable cache coherence for a network on a chip

作者: Drew E. Wingard , Doddaballapur N. Jayashimha

DOI:

关键词: Bus sniffingCache algorithmsParallel computingMESI protocolSnoopy cacheCacheCache pollutionComputer scienceCache invalidationSmart Cache

摘要: Maintaining cache coherence in a System-on-a-Chip with both multiple coherent master IP cores (CCMs) and non-cache (NCMs). A plug-in manager (CM), logic agents, an interconnect are used for the SoC to provide scalable scheme that scales amount of CCMs SoC. The each includes at least one processor operatively coupled through CM stores data CCM. maintains responsive miss line on first caches, then broadcasts request instance stored corresponding cache. Each CCM its own NCM is configured issue communication transactions into non-coherent address spaces.

参考文章(8)
Ian Rudolf Bratt, Anant Agarwal, Matthew Mattina, Caching in multicore and multiprocessor architectures ,(2007)
Ganesh Kumar, Eric Delano, Vincent R. Freytag, Dheemanth Nagaraj, Gregory S. Averill, Memory mirroring and migration at home agent ,(2009)
Ryan C. Kinter, Reduced handling of writeback data ,(2007)
Robert H. Beers, Brannon Baxton, Phanindra K. Mannava, Seungjoon Park, System and method for a 3-hop cache coherency protocol ,(2006)
Guy L. Guthrie, William J. Starke, Jeffrey Stuecheli, Derek E. Williams, Proactive prefetch throttling ,(2009)