COMPARING PROCESSOR ALLOCATION STRATEGIES IN MULTIPROGRAMMED SHARED-MEMORY MULTIPROCESSORS

Small-scale shared-memory multiprocessors are commonly used in a workg roup environment where multiple applications, both parallel and sequen tial, are executed concurrently while sharing the processors and other system resources. To utilize the processors efficiently, an effective allocation strategy is required. In this paper, we use performance da ta obtained from an SGI multiprocessor to evaluate several processor a llocation strategies when running two parallel programs simultaneously . We examine gang scheduling (coscheduling), static space-sharing (spa ce partitioning), and a dynamic allocation scheme called lop-level pro cess control (LLPC) with three different dynamic allocation heuristics . We use regression analysis to quantify the measured data and thereby explore the relationship between the degree of parallelism of the app lication, specific system parameters (such as the size of the system), the processor allocation strategy, and the resulting performance. Thi s study shows that dynamically partitioning the system using LLPC or s imilar heuristics provides better performance for applications with a high degree of parallelism than either gang scheduling or static space -sharing. (C) 1998 Academic Press.