Throughput of multiprocessors with replicated shared memories
Sigurd L. Lillevik, John L. Easterday · AFIPS conference proceedings · 1984
Multiprocessors with replicated shared memory use a memory structure consisting of a set of memories, one for each processor, with identical contents. This minimizes read interference since each processor simply accesses its own private copy of the shared memory. To ensure shared-memory integrity, write requests transfer data to all copies in parallel. Compared to traditional shared memories, multiprocessors with replicated shared memories may achieve a speed-up which approaches O(N), with N equal to the number of processors. This speed-up occurs for systems with large N, a small number of shared memories, and large shared-memory use and fractions of read requests.