Experimental Comparison of Scheduling Policies on Coarse-Grain Multiprocessors
Christopher Connelly · 1995
Scalable shared memory multiprocessors commonly employ replication and the associated coherency maintenance of memory blocks, but differ in the granularity from fine-grain (e.g. cache-coherent multiprocessors) to coarse-grain (e.g. page-based distributed shared memory systems). Regardless of the size of coherency blocks, attaining good performance may depend on the number of copies of cached data staying small, due to the inability of current architectures to support widely shared, writable data. Previous workload characterization studies of fine-grain systems have shown sharing patterns that tend toward small worker sets. Coarse-grain architectures offer scalability advantages unless sharing patterns for larger blocks are considerably less favorable. In the first portion of this dissertation, we investigate data sharing patterns in coarse-grain multiprocessors. Our results are encouraging in that, despite the coarse granularity, worker sets scale reasonably. This has implications for ...