Channel interface and protocols for cache coherency in a scalable symmetric multiprocessor system

作者: Daniel Fu , Carl G. Amdahl , Yuanlong Wang , Earl T. Cohen , Brian R. Biard

DOI:

关键词:

摘要: A preferred embodiment of a symmetric multiprocessor system includes switched fabric (switch matrix) for data transfers that provides multiple concurrent buses enable greatly increased bandwidth between processors and shared memory. high-speed point-to-point Channel couples command initiators memory with the switch matrix I/O subsystems. Each end channel is connected to Interface Block (CIB). The CIB presents logical interface Channel, providing communication path from in another IC. logic similar core-logic transceivers. transport protocol implemented reliably transfer one chip face errors limited buffering.

参考文章(53)
John Michael Kaiser, Jerry Don Lewis, John Steven Dodson, Ravi Kumar Arimilli, Shared intervention protocol for SMP bus using caches, snooping, tags and prioritizing ,(1997)
Raghunath Rao, Jeffrey Niehaus, Zheng Luo, James Divine, Miroslav Dokic, Interprocessor communication circuitry and methods ,(1997)
Ravi Kumar Arimilli, John Michael Kaiser, Jerry Don Lewis, John Steven Dodson, Cache intervention from only one of many cache lines sharing an unmodified value ,(1997)
Margaret R. Herubin, Mark Bluhm, David A. Courtright, Douglas E. Duschatko, Jeffrey S. Byrne, Raul A. Garibay, Marvin W. Martinez, Coherency for write-back cache in a system designed for write-through cache including write-back latency control ,(1994)