Hot Spot Analysis in Large ScaleShared Memory

Scalable multiprocessors that support a shared-memory image to application programmers are typically based on physical memory modules that are distributed. Consequently, the access times for a particular processor to various parts of physical memory di er. In this paper, we explore the implications of this non-uniformity in memory access times. In particular, we study the e ect of hot-spots in hierarchical large scale NUMA multiprocessors. Hot-spot analysis is of interest because coordinated threads of parallel programs lead to hot spots whose impact on performance may be substantial or even dominant. We have developed an analytical model of access latencies and contention for shared resources in the interconnection network that links the processors and memory modules. Our objective is to provide a better understanding of non-uniform memory access times in scalable architectures. We show the extent to which a variable can be shared before it becomes a performance bottleneck, and assess the potential gain from replication of shared data items. We also demonstrate that the backo value (after a memory request rejection) must be chosen carefully to balance memory access time and network utilization. Finally, we show that memory utilization is improved by allowing memory request bu ering.