High speed buffer operation in a multi-processing system

摘要: Described is an interlocking scheme which permits multiprocessing in a shared storage configuration with each central processing unit (CPU) having private high-speed buffer utilizing the store-in-buffer concept. The basic problem solved insuring that all processors access latest copy of common data minimum performance impact. system allows fetch-only copies same block to exist simultaneously storages, but only one store allowed contain currently being stored into. Disclosed, addition normal controls necessary search high speed determine whether or not required by processor buffer, means for interconnecting sharing main storage. interconnection broadcasting address information from control mechanism other purpose invalidating storages obtained most current value. That is, if has been modified another processor, must be returned its form insure receives data.