The Performance Value of Shared Network Caches in Clustered Multiprocessor Workstations

This paper evaluates the bene t of adding a shared cache to the network interface as a means of improving the performance of networked workstations con gured as a distributed shared memory multi processor A cache on the network interface shared by all processors on each cluster o ers the potential bene ts of retaining evicted processor cache lines providing implicit prefetching when network cache lines are longer than processor cache lines and increasing intra cluster sharing Using simulation the performance of eight parallel scienti c applications was evaluated In each case we examined in detail the means by which processor cache misses were satis ed Our results were mixed For the applications studied we found that the network cache o ers substan tial performance bene t when processor caches are too small to hold the application s primary working set or when network contention limits application performance The expected bene ts of implicit prefetching and increased intra cluster sharing did not contribute signi cantly to the performance enhancement of the network cache for most applications Finally the advantage a orded by the network cache diminishes as processor cache size increases and network contention decreases