Using proxies to reduce controller contention in large shared-memory multiprocessors

Abstract. Some shared-memory applications have execution times linear in the number of processors due to unfortunate allocation of the home and ownership of cache lines. We present a modi ed coherency protocol which avoids this e ect. Read requests are routed via \proxies", randomly-selected intermediate nodes. We present results from executiondriven simulations of a cc-numa architecture which show that proxying can yield a large speedup in cases where read contention is extreme, while only causing small slowdowns in other benchmarks. We investigate how many proxies should be used and what e ect the scheme has on tra c levels and queuing of requests at node controllers.