Job Scheduling in Multiprogrammed Parallel Systems

Scheduling in the context of parallel systems is often thought of in terms of assigning tasks in a program to processors, so as to minimize the makespan. This formulation assumes that the processors are dedicated to the program in question. But when the parallel system is shared by a number of users, this is not necessarily the case. In the context of multiprogrammed parallel machines, scheduling refers to the execution of threads from competing programs. This is an operating system issue, involved with resource allocation, not a program development issue. Scheduling schemes for multiprogrammed parallel systems can be classiied as one or two leveled. Single-level scheduling combines the allocation of processing power with the decision of which thread will use it. Two level scheduling decouples the two issues: rst, processors are allocated to the job, and then the job's threads are scheduled using this pool of processors. The processors of a parallel system can be shared in two basic ways, which are relevant for both one-level and two-level scheduling. One approach is to use time slicing, e.g. when all the processors in the system (or all the processors in the pool) service a global queue of ready threads. The other approach is to use space slicing, and partition the processors statically or dynamically among the diierent jobs. As these approaches are orthogonal to each other, it is also possible to combine them in various ways; for example, this is often done in gang scheduling. Systems using the various approaches are described, and the implications of the diierent mechanisms are discussed. The goals of this survey are to describe the many diierent approaches within a uniied framework based on the mechanisms used to achieve multiprogramming, and at the same time document commercial systems that have not been described in the open literature.