Based On Pvm Task Scheduling Strategy

Parallel distributed computing (PDC) has become an extremely effective way to solve many problems in the large-scale scientific computing and engineering field. However, there are many factors to affect the efficiency of parallel distributed computing, such as task gratuity, processors allocating and network topology, in which task-scheduling strategy is a crucial one and thus has become a hot research spot in the field of parallel processing.In this paper, firstly, the general model of scheduling problem is analyzed and stimulatingly the factors affecting the scheduling efficiency are summarized. For these different kinds of factors, some new strategies are promoted and some improved methods are provided: in order to resolve the bottle necklace problems in centralized task scheduling strategy and simultaneously obtain the better load balancing capability, these ideas of two-tier scheduling and active reporting are presented; by introducing the conception of valve and adopting the auto-adaptation and reporting mechanism, the system can timely and accurately get the load information of each workstation; by introducing these ideas of efficiency aggregating and exhaustive releasing, the system get the better scalability; in order to improve the whole system's efficiency, these technologies of multi-thread and message relaying without unfolding are adopted and a few different kinds of task allocating strategy are provided.Resorting to the above achievements and combined with the problems of complex electromechanical coupling system, a visible parallel scheduling program, which possesses many virtues such as easy-to-use and highly modularization, is designed and realized, which is based on the easy-to-transplant parallelrdeveloped environment PVM.Finally, an application instance of rolling mill is taken for an example and the validity of the theory and the practicability of the program are proved.